Original Research

Store-and-forward teledermatology (SFT) allows clinical images and information to be sent to a dermatologist for evaluation. In fiscal year (FY) 2018, 117,780 SFT consultations were completed in the Veterans Health Administration. Continued growth is expected since SFT has proven to be an effective method for improving access to face-to-face (FTF) dermatology care.¹ In the same period, the US Department of Veterans Affairs (VA) Puget Sound Health Care System (VAPSHCS) completed 12,563 consultations in a mean 1.1 days from entry into episode of care (EEC), according to data reported by VA Teledermatology Program Administrator Chris Foster.

Obtaining a prompt consultation is reported to be an overwhelming advantage of using SFT.^2-5 Rapid turnaround may appear to make SFT specialist care more accessible to veterans, yet this is an oversimplification. The process of delivering care (rather than consultation) through SFT is more complex than reading the images and reporting the findings. When a skin condition is identified by a primary care clinician and that person decides to request an SFT consultation, a complex set of tasks and handoffs is set into motion. A swim-lane diagram illustrates the numerous steps and handoffs that go into delivering care to a patient with a malignant melanoma on the SFT platform compared to FTF care, which requires fewer handoffs (Figure).

This process improvement project examined whether handoffs necessitated by SFT care lengthened the timeline of care for biopsy-proven primary cutaneous malignant melanoma. The stakes of delay in care are high. A 2018 study using the National Cancer Database found that a delay of > 30 days from biopsy to definitive excision (the date definitive surgical procedure for the condition is performed) resulted in a measurable increase in melanoma-related mortality. ⁶ This study sought to identify areas where the SFT timeline of care could be shortened.

Methods

This retrospective cohort study was approved by the VAPSHCS Institutional Review Board. The study drew from secondary data obtained from VistA, the VA Corporate Data Warehouse, the Veterans Integrated Service Network (VISN) 20 database, the American Academy of Dermatology Teledermatology Program database, and the VA Computerized Patient Record System.

Patients registered for ≥ 1 year at VAPSHCS with a diagnosis of primary cutaneous malignant melanoma by the Pathology service between January 1, 2006, and December 31, 2013, were included. Patients with metastatic or recurrent melanoma were excluded.

Cases were randomly selected from a melanoma database previously validated and used for another quality improvement project.⁷ There were initially 115 patient cases extracted from this database for both the FTF and SFT groups. Eighty-seven SFT and 107 FTF cases met inclusion criteria. To further analyze these groups, we split the FTF group into 2 subgroups: FTF dermatology (patients whose melanomas were entered into care in a dermatology clinic) and FTF primary care (patients whose melanomas were entered into care in primary care or a nondermatology setting).

The timeline of care was divided into 2 major time intervals: (1) entry into episode of care (EEC; the date a lesion was first documented in the electronic health record) to biopsy; and (2) biopsy to definitive excision. The SFT process was divided into the following intervals: EEC to imaging request (the date a clinician requested imaging); imaging request to imaging completion (the date an imager photographed a patient’s lesion); imaging completion to SFT consultation request (the date the SFT consultation was requested); SFT consultation request to consultation completion (the date an SFT reader completed the consultation request for a patient); and SFT consultation completion to biopsy. Mean and median interval lengths were compared between groups and additional analyses identified steps that may have contributed to delays in care.

To address potential bias based on access to care for rural veterans, SFT and FTF primary care cases were categorized into groups based on their location: (1) EEC and biopsy conducted at the same facility; (2) EEC and biopsy conducted at different facilities within the same health care system (main health care facility and its community-based outpatient clinics); and (3) EEC and biopsy conducted at different health care systems.

Statistics

Means, medians, and SDs were calculated in Excel. The Mann-Whitney U test was used to compare SFT medians to the FTF data and X² test was used to compare proportions for secondary analyses.

Results

The median (mean) interval from EEC to definitive excision was 73 days (85) for SFT and 58 days (73) for FTF (P = .004) (Table). To understand this difference, the distribution of intervals from EEC to biopsy and biopsy to definitive excision were calculated. Only 38% of SFT cases were biopsied within 20 days compared to 65% of FTF cases (P < .001). The difference in time from biopsy to definitive excision distributions were not statistically significant, suggesting that the difference is actually a reflection of the differences seen in the period between EEC and biopsy.

EEC and biopsy occurred at the same facility in 85% and 82% of FTF primary care and SFT cases, respectively. EEC and biopsy occurred at different facilities within the same health care system in 15% and 16% of FTF primary care and SFT cases, respectively. EEC and biopsy occurred at different health care systems in 0% and 2% of FTF primary care and SFT cases, respectively. Geographic bias did not impact results for either group of veterans.

The interval between EEC and biopsy was shorter for FTF dermatology cases than for FTF primary care cases. For FTF dermatology cases, 96% were biopsied within 20 days compared with 34% of FTF primary care cases (P < .001).

To further analyze the difference in the EEC to biopsy interval duration between SFT and FTF primary care the timeline was divided into smaller steps: EEC to imaging completion, imaging completion to SFT consult completion, and SFT consult completion to biopsy. From EEC to SFT consult completion, SFT cases took a median of 6.0 days and a mean of 12.3 days, reflecting the administrative handoffs that must occur in SFT. A total of 82% of FTF primary care cases were entered into care and consultation was requested on the same day, while this was true for only 1% of SFT cases.

Since mortality data were not collected, the frequency of in situ melanomas and invasive melanomas (pathologic stage pT1a or greater) was used as a proxy for comparing outcomes. No significant difference was found in the frequency of in situ vs invasive melanomas in the SFT and FTF dermatology groups; however, there was a much higher frequency of invasive melanomas in the FTF primary care group (P = .007).

Discussion

This study compared the time to treatment for SFT vs FTF and identified important differences. The episode of care for melanomas diagnosed by SFT was statistically significantly longer (15 days) than those diagnosed by FTF. The interval between biopsy and definitive excision was a median of 34 and 38 days, and a mean of 48 and 44 days for SFT and FTF, respectively, which were not statistically significant. The difference in the total duration of the interval between EEC and definitive excision was accounted for by the duration of the interval from EEC to biopsy. When excluding dermatology clinic cases from the FTF group, there was no difference in the interval between EEC and biopsy for SFT and FTF primary care. The handoffs in SFT accounted for a median of 6 days and mean of 12 days, a significant portion of the timeline, and is a target for process improvement. The delay necessitated by handoffs did not significantly affect the distribution of in situ and invasive melanomas in the SFT and FTF dermatology groups. This suggests that SFT may have better outcomes than FTF primary care.

There has been extensive research on the timeline from the patient initially noticing a lesion to the EEC.^8-11 There is also a body of research on the timeline from biopsy to definitive excision. ^6,12-16 However, there has been little research on the timeline between EEC and biopsy, which comprises a large portion of the overall timeline of both SFT care and FTF care. This study analyzed the delays that can occur in this interval. When patients first enter FTF dermatology care, this timeline is quite short because lesions are often biopsied on the same day. When patients enter into care with their primary or nondermatology clinician, there can be significant delays.

Since the stakes are high when it comes to treating melanoma, it is important to minimize the overall timeline. A 6-day median and 12-day mean were established as targets for teledermatology handoffs. Ideally, a lesion should be entered into an episode of care, imaged, and sent for consultation on the same day. To help further understand delays in administrative handoffs, we stratified the SFT cases by VISN 20 sites and spoke with an administrator at a top performing site. Between 2006 and 2013, this site had a dedicated full-time imager as well as a backup imager that ensured images were taken quickly, usually on the same day the lesion was entered into care. Unfortunately, this is not the standard at all VISN 20 sites and certainly contributes to the overall delay in care in SFT

Minimizing the timeline of care is possible, as shown by the Danish health system, which developed a fast-track referral system after recognizing the need to minimize delays between the presentation, diagnosis, and treatment of cutaneous melanomas. In Denmark, a patient who presents to a general practitioner with a suspicious lesion is referred to secondary care for excision biopsy within 6 days. Diagnosis is made within 2 weeks, and, if necessary, definitive excision is offered within 9 days of the diagnosis. This translates into a maximum 20-day EEC to biopsy timeline and maximum 29-day EEC to definitive excision timeline. Although an intervention such as this may be difficult to implement in the United States due to its size and decentralized health care system, it would, however, be more realistic within the VA due to its centralized structure. The Danish system shows that with appropriate resource allocation and strict timeframes for treatment referrals, the timeline can be minimized.¹⁷

Despite the delay in the SFT timeline, this study found no significant difference between the distribution of in situ vs invasive melanomas in FTF dermatology and SFT groups. One possible explanation for this is that SFT increases access to dermatologist care, meaning clinicians may be more willing to consult SFT for less advanced– appearing lesions.

The finding that SFT diagnosed a larger proportion of in situ melanomas than FTF primary care is consistent with the findings of Ferrándiz et al, who reported that the mean Breslow thickness was significantly lower among patients in an SFT group compared to patients in an FTF group consisting of general practitioners. ¹⁸ However, the study population was not randomized and the results may have been impacted by ascertainment bias. Ferrándiz et al hypothesized that clinicians may have a lower threshold for consulting teledermatology, resulting in lower mean Breslow thicknesses.¹⁸ Karavan et al found the opposite results, with a higher mean Breslow thickness in SFT compared to a primary care FTF group.¹⁹ The data presented here suggest that SFT has room for process improvement yet is essentially equivalent to FTF dermatology in terms of outcomes.

Limitations

The majority of patients in this study were aged > 50 years, White, and male. The results may not be representative for other populations. The study was relatively small compared to studies that looked at other aspects of the melanoma care timeline. The study was not powered to ascertain mortality, the most important metric for melanoma.

Conclusions

The episode of care was significantly longer for melanomas diagnosed by SFT than those diagnosed by FTF; however, timelines were not statistically different when FTF lesions entered into care in dermatology were excluded. A median 6-day and mean 12.3-day delay in administrative handoffs occurred at the beginning of the SFT process and is a target for process improvement. Considering the high stakes of melanoma, the SFT timeline could be reduced if EEC, imaging, and SFT consultation all happened in the same day.

Store-and-forward teledermatology (SFT) allows clinical images and information to be sent to a dermatologist for evaluation. In fiscal year (FY) 2018, 117,780 SFT consultations were completed in the Veterans Health Administration. Continued growth is expected since SFT has proven to be an effective method for improving access to face-to-face (FTF) dermatology care.¹ In the same period, the US Department of Veterans Affairs (VA) Puget Sound Health Care System (VAPSHCS) completed 12,563 consultations in a mean 1.1 days from entry into episode of care (EEC), according to data reported by VA Teledermatology Program Administrator Chris Foster.

Obtaining a prompt consultation is reported to be an overwhelming advantage of using SFT.^2-5 Rapid turnaround may appear to make SFT specialist care more accessible to veterans, yet this is an oversimplification. The process of delivering care (rather than consultation) through SFT is more complex than reading the images and reporting the findings. When a skin condition is identified by a primary care clinician and that person decides to request an SFT consultation, a complex set of tasks and handoffs is set into motion. A swim-lane diagram illustrates the numerous steps and handoffs that go into delivering care to a patient with a malignant melanoma on the SFT platform compared to FTF care, which requires fewer handoffs (Figure).

This process improvement project examined whether handoffs necessitated by SFT care lengthened the timeline of care for biopsy-proven primary cutaneous malignant melanoma. The stakes of delay in care are high. A 2018 study using the National Cancer Database found that a delay of > 30 days from biopsy to definitive excision (the date definitive surgical procedure for the condition is performed) resulted in a measurable increase in melanoma-related mortality. ⁶ This study sought to identify areas where the SFT timeline of care could be shortened.

Methods

This retrospective cohort study was approved by the VAPSHCS Institutional Review Board. The study drew from secondary data obtained from VistA, the VA Corporate Data Warehouse, the Veterans Integrated Service Network (VISN) 20 database, the American Academy of Dermatology Teledermatology Program database, and the VA Computerized Patient Record System.

Patients registered for ≥ 1 year at VAPSHCS with a diagnosis of primary cutaneous malignant melanoma by the Pathology service between January 1, 2006, and December 31, 2013, were included. Patients with metastatic or recurrent melanoma were excluded.

Cases were randomly selected from a melanoma database previously validated and used for another quality improvement project.⁷ There were initially 115 patient cases extracted from this database for both the FTF and SFT groups. Eighty-seven SFT and 107 FTF cases met inclusion criteria. To further analyze these groups, we split the FTF group into 2 subgroups: FTF dermatology (patients whose melanomas were entered into care in a dermatology clinic) and FTF primary care (patients whose melanomas were entered into care in primary care or a nondermatology setting).

The timeline of care was divided into 2 major time intervals: (1) entry into episode of care (EEC; the date a lesion was first documented in the electronic health record) to biopsy; and (2) biopsy to definitive excision. The SFT process was divided into the following intervals: EEC to imaging request (the date a clinician requested imaging); imaging request to imaging completion (the date an imager photographed a patient’s lesion); imaging completion to SFT consultation request (the date the SFT consultation was requested); SFT consultation request to consultation completion (the date an SFT reader completed the consultation request for a patient); and SFT consultation completion to biopsy. Mean and median interval lengths were compared between groups and additional analyses identified steps that may have contributed to delays in care.

To address potential bias based on access to care for rural veterans, SFT and FTF primary care cases were categorized into groups based on their location: (1) EEC and biopsy conducted at the same facility; (2) EEC and biopsy conducted at different facilities within the same health care system (main health care facility and its community-based outpatient clinics); and (3) EEC and biopsy conducted at different health care systems.

Statistics

Means, medians, and SDs were calculated in Excel. The Mann-Whitney U test was used to compare SFT medians to the FTF data and X² test was used to compare proportions for secondary analyses.

Results

The median (mean) interval from EEC to definitive excision was 73 days (85) for SFT and 58 days (73) for FTF (P = .004) (Table). To understand this difference, the distribution of intervals from EEC to biopsy and biopsy to definitive excision were calculated. Only 38% of SFT cases were biopsied within 20 days compared to 65% of FTF cases (P < .001). The difference in time from biopsy to definitive excision distributions were not statistically significant, suggesting that the difference is actually a reflection of the differences seen in the period between EEC and biopsy.

EEC and biopsy occurred at the same facility in 85% and 82% of FTF primary care and SFT cases, respectively. EEC and biopsy occurred at different facilities within the same health care system in 15% and 16% of FTF primary care and SFT cases, respectively. EEC and biopsy occurred at different health care systems in 0% and 2% of FTF primary care and SFT cases, respectively. Geographic bias did not impact results for either group of veterans.

The interval between EEC and biopsy was shorter for FTF dermatology cases than for FTF primary care cases. For FTF dermatology cases, 96% were biopsied within 20 days compared with 34% of FTF primary care cases (P < .001).

To further analyze the difference in the EEC to biopsy interval duration between SFT and FTF primary care the timeline was divided into smaller steps: EEC to imaging completion, imaging completion to SFT consult completion, and SFT consult completion to biopsy. From EEC to SFT consult completion, SFT cases took a median of 6.0 days and a mean of 12.3 days, reflecting the administrative handoffs that must occur in SFT. A total of 82% of FTF primary care cases were entered into care and consultation was requested on the same day, while this was true for only 1% of SFT cases.

Since mortality data were not collected, the frequency of in situ melanomas and invasive melanomas (pathologic stage pT1a or greater) was used as a proxy for comparing outcomes. No significant difference was found in the frequency of in situ vs invasive melanomas in the SFT and FTF dermatology groups; however, there was a much higher frequency of invasive melanomas in the FTF primary care group (P = .007).

Discussion

This study compared the time to treatment for SFT vs FTF and identified important differences. The episode of care for melanomas diagnosed by SFT was statistically significantly longer (15 days) than those diagnosed by FTF. The interval between biopsy and definitive excision was a median of 34 and 38 days, and a mean of 48 and 44 days for SFT and FTF, respectively, which were not statistically significant. The difference in the total duration of the interval between EEC and definitive excision was accounted for by the duration of the interval from EEC to biopsy. When excluding dermatology clinic cases from the FTF group, there was no difference in the interval between EEC and biopsy for SFT and FTF primary care. The handoffs in SFT accounted for a median of 6 days and mean of 12 days, a significant portion of the timeline, and is a target for process improvement. The delay necessitated by handoffs did not significantly affect the distribution of in situ and invasive melanomas in the SFT and FTF dermatology groups. This suggests that SFT may have better outcomes than FTF primary care.

There has been extensive research on the timeline from the patient initially noticing a lesion to the EEC.^8-11 There is also a body of research on the timeline from biopsy to definitive excision. ^6,12-16 However, there has been little research on the timeline between EEC and biopsy, which comprises a large portion of the overall timeline of both SFT care and FTF care. This study analyzed the delays that can occur in this interval. When patients first enter FTF dermatology care, this timeline is quite short because lesions are often biopsied on the same day. When patients enter into care with their primary or nondermatology clinician, there can be significant delays.

Since the stakes are high when it comes to treating melanoma, it is important to minimize the overall timeline. A 6-day median and 12-day mean were established as targets for teledermatology handoffs. Ideally, a lesion should be entered into an episode of care, imaged, and sent for consultation on the same day. To help further understand delays in administrative handoffs, we stratified the SFT cases by VISN 20 sites and spoke with an administrator at a top performing site. Between 2006 and 2013, this site had a dedicated full-time imager as well as a backup imager that ensured images were taken quickly, usually on the same day the lesion was entered into care. Unfortunately, this is not the standard at all VISN 20 sites and certainly contributes to the overall delay in care in SFT

Minimizing the timeline of care is possible, as shown by the Danish health system, which developed a fast-track referral system after recognizing the need to minimize delays between the presentation, diagnosis, and treatment of cutaneous melanomas. In Denmark, a patient who presents to a general practitioner with a suspicious lesion is referred to secondary care for excision biopsy within 6 days. Diagnosis is made within 2 weeks, and, if necessary, definitive excision is offered within 9 days of the diagnosis. This translates into a maximum 20-day EEC to biopsy timeline and maximum 29-day EEC to definitive excision timeline. Although an intervention such as this may be difficult to implement in the United States due to its size and decentralized health care system, it would, however, be more realistic within the VA due to its centralized structure. The Danish system shows that with appropriate resource allocation and strict timeframes for treatment referrals, the timeline can be minimized.¹⁷

Despite the delay in the SFT timeline, this study found no significant difference between the distribution of in situ vs invasive melanomas in FTF dermatology and SFT groups. One possible explanation for this is that SFT increases access to dermatologist care, meaning clinicians may be more willing to consult SFT for less advanced– appearing lesions.

The finding that SFT diagnosed a larger proportion of in situ melanomas than FTF primary care is consistent with the findings of Ferrándiz et al, who reported that the mean Breslow thickness was significantly lower among patients in an SFT group compared to patients in an FTF group consisting of general practitioners. ¹⁸ However, the study population was not randomized and the results may have been impacted by ascertainment bias. Ferrándiz et al hypothesized that clinicians may have a lower threshold for consulting teledermatology, resulting in lower mean Breslow thicknesses.¹⁸ Karavan et al found the opposite results, with a higher mean Breslow thickness in SFT compared to a primary care FTF group.¹⁹ The data presented here suggest that SFT has room for process improvement yet is essentially equivalent to FTF dermatology in terms of outcomes.

Limitations

The majority of patients in this study were aged > 50 years, White, and male. The results may not be representative for other populations. The study was relatively small compared to studies that looked at other aspects of the melanoma care timeline. The study was not powered to ascertain mortality, the most important metric for melanoma.

Conclusions

The episode of care was significantly longer for melanomas diagnosed by SFT than those diagnosed by FTF; however, timelines were not statistically different when FTF lesions entered into care in dermatology were excluded. A median 6-day and mean 12.3-day delay in administrative handoffs occurred at the beginning of the SFT process and is a target for process improvement. Considering the high stakes of melanoma, the SFT timeline could be reduced if EEC, imaging, and SFT consultation all happened in the same day.

Store-and-forward teledermatology (SFT) allows clinical images and information to be sent to a dermatologist for evaluation. In fiscal year (FY) 2018, 117,780 SFT consultations were completed in the Veterans Health Administration. Continued growth is expected since SFT has proven to be an effective method for improving access to face-to-face (FTF) dermatology care.¹ In the same period, the US Department of Veterans Affairs (VA) Puget Sound Health Care System (VAPSHCS) completed 12,563 consultations in a mean 1.1 days from entry into episode of care (EEC), according to data reported by VA Teledermatology Program Administrator Chris Foster.

Obtaining a prompt consultation is reported to be an overwhelming advantage of using SFT.^2-5 Rapid turnaround may appear to make SFT specialist care more accessible to veterans, yet this is an oversimplification. The process of delivering care (rather than consultation) through SFT is more complex than reading the images and reporting the findings. When a skin condition is identified by a primary care clinician and that person decides to request an SFT consultation, a complex set of tasks and handoffs is set into motion. A swim-lane diagram illustrates the numerous steps and handoffs that go into delivering care to a patient with a malignant melanoma on the SFT platform compared to FTF care, which requires fewer handoffs (Figure).

This process improvement project examined whether handoffs necessitated by SFT care lengthened the timeline of care for biopsy-proven primary cutaneous malignant melanoma. The stakes of delay in care are high. A 2018 study using the National Cancer Database found that a delay of > 30 days from biopsy to definitive excision (the date definitive surgical procedure for the condition is performed) resulted in a measurable increase in melanoma-related mortality. ⁶ This study sought to identify areas where the SFT timeline of care could be shortened.

Methods

This retrospective cohort study was approved by the VAPSHCS Institutional Review Board. The study drew from secondary data obtained from VistA, the VA Corporate Data Warehouse, the Veterans Integrated Service Network (VISN) 20 database, the American Academy of Dermatology Teledermatology Program database, and the VA Computerized Patient Record System.

Patients registered for ≥ 1 year at VAPSHCS with a diagnosis of primary cutaneous malignant melanoma by the Pathology service between January 1, 2006, and December 31, 2013, were included. Patients with metastatic or recurrent melanoma were excluded.

Cases were randomly selected from a melanoma database previously validated and used for another quality improvement project.⁷ There were initially 115 patient cases extracted from this database for both the FTF and SFT groups. Eighty-seven SFT and 107 FTF cases met inclusion criteria. To further analyze these groups, we split the FTF group into 2 subgroups: FTF dermatology (patients whose melanomas were entered into care in a dermatology clinic) and FTF primary care (patients whose melanomas were entered into care in primary care or a nondermatology setting).

The timeline of care was divided into 2 major time intervals: (1) entry into episode of care (EEC; the date a lesion was first documented in the electronic health record) to biopsy; and (2) biopsy to definitive excision. The SFT process was divided into the following intervals: EEC to imaging request (the date a clinician requested imaging); imaging request to imaging completion (the date an imager photographed a patient’s lesion); imaging completion to SFT consultation request (the date the SFT consultation was requested); SFT consultation request to consultation completion (the date an SFT reader completed the consultation request for a patient); and SFT consultation completion to biopsy. Mean and median interval lengths were compared between groups and additional analyses identified steps that may have contributed to delays in care.

To address potential bias based on access to care for rural veterans, SFT and FTF primary care cases were categorized into groups based on their location: (1) EEC and biopsy conducted at the same facility; (2) EEC and biopsy conducted at different facilities within the same health care system (main health care facility and its community-based outpatient clinics); and (3) EEC and biopsy conducted at different health care systems.

Statistics

Means, medians, and SDs were calculated in Excel. The Mann-Whitney U test was used to compare SFT medians to the FTF data and X² test was used to compare proportions for secondary analyses.

Results

The median (mean) interval from EEC to definitive excision was 73 days (85) for SFT and 58 days (73) for FTF (P = .004) (Table). To understand this difference, the distribution of intervals from EEC to biopsy and biopsy to definitive excision were calculated. Only 38% of SFT cases were biopsied within 20 days compared to 65% of FTF cases (P < .001). The difference in time from biopsy to definitive excision distributions were not statistically significant, suggesting that the difference is actually a reflection of the differences seen in the period between EEC and biopsy.

EEC and biopsy occurred at the same facility in 85% and 82% of FTF primary care and SFT cases, respectively. EEC and biopsy occurred at different facilities within the same health care system in 15% and 16% of FTF primary care and SFT cases, respectively. EEC and biopsy occurred at different health care systems in 0% and 2% of FTF primary care and SFT cases, respectively. Geographic bias did not impact results for either group of veterans.

The interval between EEC and biopsy was shorter for FTF dermatology cases than for FTF primary care cases. For FTF dermatology cases, 96% were biopsied within 20 days compared with 34% of FTF primary care cases (P < .001).

To further analyze the difference in the EEC to biopsy interval duration between SFT and FTF primary care the timeline was divided into smaller steps: EEC to imaging completion, imaging completion to SFT consult completion, and SFT consult completion to biopsy. From EEC to SFT consult completion, SFT cases took a median of 6.0 days and a mean of 12.3 days, reflecting the administrative handoffs that must occur in SFT. A total of 82% of FTF primary care cases were entered into care and consultation was requested on the same day, while this was true for only 1% of SFT cases.

Since mortality data were not collected, the frequency of in situ melanomas and invasive melanomas (pathologic stage pT1a or greater) was used as a proxy for comparing outcomes. No significant difference was found in the frequency of in situ vs invasive melanomas in the SFT and FTF dermatology groups; however, there was a much higher frequency of invasive melanomas in the FTF primary care group (P = .007).

Discussion

This study compared the time to treatment for SFT vs FTF and identified important differences. The episode of care for melanomas diagnosed by SFT was statistically significantly longer (15 days) than those diagnosed by FTF. The interval between biopsy and definitive excision was a median of 34 and 38 days, and a mean of 48 and 44 days for SFT and FTF, respectively, which were not statistically significant. The difference in the total duration of the interval between EEC and definitive excision was accounted for by the duration of the interval from EEC to biopsy. When excluding dermatology clinic cases from the FTF group, there was no difference in the interval between EEC and biopsy for SFT and FTF primary care. The handoffs in SFT accounted for a median of 6 days and mean of 12 days, a significant portion of the timeline, and is a target for process improvement. The delay necessitated by handoffs did not significantly affect the distribution of in situ and invasive melanomas in the SFT and FTF dermatology groups. This suggests that SFT may have better outcomes than FTF primary care.

There has been extensive research on the timeline from the patient initially noticing a lesion to the EEC.^8-11 There is also a body of research on the timeline from biopsy to definitive excision. ^6,12-16 However, there has been little research on the timeline between EEC and biopsy, which comprises a large portion of the overall timeline of both SFT care and FTF care. This study analyzed the delays that can occur in this interval. When patients first enter FTF dermatology care, this timeline is quite short because lesions are often biopsied on the same day. When patients enter into care with their primary or nondermatology clinician, there can be significant delays.

Since the stakes are high when it comes to treating melanoma, it is important to minimize the overall timeline. A 6-day median and 12-day mean were established as targets for teledermatology handoffs. Ideally, a lesion should be entered into an episode of care, imaged, and sent for consultation on the same day. To help further understand delays in administrative handoffs, we stratified the SFT cases by VISN 20 sites and spoke with an administrator at a top performing site. Between 2006 and 2013, this site had a dedicated full-time imager as well as a backup imager that ensured images were taken quickly, usually on the same day the lesion was entered into care. Unfortunately, this is not the standard at all VISN 20 sites and certainly contributes to the overall delay in care in SFT

Minimizing the timeline of care is possible, as shown by the Danish health system, which developed a fast-track referral system after recognizing the need to minimize delays between the presentation, diagnosis, and treatment of cutaneous melanomas. In Denmark, a patient who presents to a general practitioner with a suspicious lesion is referred to secondary care for excision biopsy within 6 days. Diagnosis is made within 2 weeks, and, if necessary, definitive excision is offered within 9 days of the diagnosis. This translates into a maximum 20-day EEC to biopsy timeline and maximum 29-day EEC to definitive excision timeline. Although an intervention such as this may be difficult to implement in the United States due to its size and decentralized health care system, it would, however, be more realistic within the VA due to its centralized structure. The Danish system shows that with appropriate resource allocation and strict timeframes for treatment referrals, the timeline can be minimized.¹⁷

Despite the delay in the SFT timeline, this study found no significant difference between the distribution of in situ vs invasive melanomas in FTF dermatology and SFT groups. One possible explanation for this is that SFT increases access to dermatologist care, meaning clinicians may be more willing to consult SFT for less advanced– appearing lesions.

The finding that SFT diagnosed a larger proportion of in situ melanomas than FTF primary care is consistent with the findings of Ferrándiz et al, who reported that the mean Breslow thickness was significantly lower among patients in an SFT group compared to patients in an FTF group consisting of general practitioners. ¹⁸ However, the study population was not randomized and the results may have been impacted by ascertainment bias. Ferrándiz et al hypothesized that clinicians may have a lower threshold for consulting teledermatology, resulting in lower mean Breslow thicknesses.¹⁸ Karavan et al found the opposite results, with a higher mean Breslow thickness in SFT compared to a primary care FTF group.¹⁹ The data presented here suggest that SFT has room for process improvement yet is essentially equivalent to FTF dermatology in terms of outcomes.

Limitations

The majority of patients in this study were aged > 50 years, White, and male. The results may not be representative for other populations. The study was relatively small compared to studies that looked at other aspects of the melanoma care timeline. The study was not powered to ascertain mortality, the most important metric for melanoma.

Conclusions

The episode of care was significantly longer for melanomas diagnosed by SFT than those diagnosed by FTF; however, timelines were not statistically different when FTF lesions entered into care in dermatology were excluded. A median 6-day and mean 12.3-day delay in administrative handoffs occurred at the beginning of the SFT process and is a target for process improvement. Considering the high stakes of melanoma, the SFT timeline could be reduced if EEC, imaging, and SFT consultation all happened in the same day.

To the Editor:

Early skin cancer detection improves patient outcomes¹; however, socioeconomic and racial disparities may impact access to dermatologic care.² Although non-Hispanic White individuals have a high incidence of skin cancer, they experience higher melanoma-specific survival rates than non-White patients, who often receive later-stage diagnoses and experience higher mortality.² Furthermore, racial/ ethnic minorities often face longer surgery wait times after diagnosis and have lower socioeconomic status (SES) and less favorable health insurance coverage, contributing to poorer outcomes.^2,3

To examine access to full-body skin examinations (FBSEs) by board-certified dermatologists in Los Angeles (LA) County, California, we analyzed the availability of FBSEs based on racial demographics, income, and insurance type (Medicaid [Medi-Cal] vs private [Blue Cross Blue Shield (BCBS)]). Demographic data by zip code were obtained from the US Census Bureau.⁴ This validated metric highlights socioeconomic disparities and minimizes data gaps^5,6 and was used to assess health care access among different population subgroups. Dermatologists’ contact information was obtained from the Find a Dermatologist page on the American Academy of Dermatology website and the listed phone numbers of their practice were used to contact them. Practices with board-certified dermatologists accepting new patients were included in the study; practices were not included if they had exclusive insurance plans; were pediatric, cosmetic, or research only; or were nonresponsive to calls. From August 2022 to September 2022, each practice was called twice within a 36-hour period—once by a simulated patient with Medi-Cal and once by a simulated patient with BCBS—and were asked about availability for new patient FBSE appointments and accepted insurance types. Data were analyzed using SAS software (SAS Institute Inc.).

Los Angeles County comprises 269 zip codes, of which 82 (30.5%) have dermatology practices. Of 213 total dermatologists in LA County listed on the American Academy of Dermatology website, 193 (90.6%) met preliminary criteria, and 169 (79.3%) were successfully contacted. Almost all (94.6% [160/169]) accepted new patients for FBSEs; of those, 63.1% (101/160) accepted only private insurance, 16.9% (27/160) accepted both private insurance and Medi-Cal, and 16.2% (26/160) did not accept any insurance. Racial predominance for each dermatology practice was analyzed by zip code (Table). Dermatologists included in our study were significantly more concentrated in predominantly non- Hispanic White areas of LA County vs predominantly Hispanic areas (P<.0001). Notably, the average income in predominantly non-Hispanic White zip codes ($114,757.74) was significantly higher than in predominantly Hispanic areas ($58,278.54)(P=.001)(Table).⁴

In LA County, 40.1% (108/269) of zip codes have no racial majority, 28.2% (76/269) are predominantly Hispanic, 27.5% (74/269) are predominantly non-Hispanic White, 2.2% (6/269) are predominantly Black, and 1.9% (5/269) are predominantly Asian.⁴ There are no dermatologists in predominantly Black zip codes, 2 in predominantly Asian zip codes, 14 in predominantly Hispanic zip codes, 38 in zip codes with no racial majority, and 106 in predominantly non-Hispanic White zip codes. There are significantly more dermatologists in predominantly non-Hispanic White zip codes compared to predominantly Hispanic zip codes (P<.0001). In LA County, the average income in predominantly Asian, non-Hispanic White, and Hispanic zip codes was $93,594, $114,757.84, and $58,278.54, respectively, in 2021.4 The average income in predominantly non-Hispanic White zip codes was significantly higher than in predominantly Hispanic zip codes (P=.001). There were no income data available for predominantly Black zip codes or zip codes with no racial majority.

The results from our study revealed potential barriers to FBSEs for racial and ethnic minorities in LA County, which supports previous research on the impact of SES, race, and insurance on access to dermatologic care.^2,3 Predominantly Hispanic zip codes have significantly lower income (P<.0001) and fewer dermatologists (P=.001) compared to zip codes that are predominantly non-Hispanic White, reflecting how lower SES correlates with worse health outcomes and higher melanoma mortality. Conversely, predominantly non-Hispanic White areas with higher income have better access to dermatologists, which may contribute to the improved melanoma survival rates among White patients. Additionally, most dermatologists accept only private insurance, further highlighting the disparity in FBSE access for non-White patients across LA County. While our study focused on FBSE access, our findings may point to a wider barrier to dermatologic care, especially in zip codes with fewer dermatologists. Further studies are needed to determine whether these areas also face barriers to accessing primary care.

Our study was limited by the exclusion of nonphysician providers (eg, nurse practitioners, physician assistants), a small sample size, and lack of available economic data for predominantly Black zip codes.⁴ Additionally, the exclusion of practices with exclusive insurance plans (eg, Kaiser Permanente) limited the generalizability of our findings, as our results did not account for the populations served by these practices. Furthermore, our analysis did not account for variations in practice size or the proportion of care provided to patients with different insurance types, which could impact overall accessibility. Additional studies are needed to explore the impact of these factors on access to general dermatologic care and not just FBSEs.

Racial/ethnic minorities and lower SES populations face major barriers to FBSE access in LA County, such as difficulty finding a dermatologist in their area or one who accepts Medi-Cal. Addressing these disparities is crucial for improving skin cancer outcomes. Further research is needed to develop strategies to eliminate these barriers to dermatologic care, such as increasing access to teledermatology, offering mobile dermatology clinics, and improving insurance coverage.

To the Editor:

Early skin cancer detection improves patient outcomes¹; however, socioeconomic and racial disparities may impact access to dermatologic care.² Although non-Hispanic White individuals have a high incidence of skin cancer, they experience higher melanoma-specific survival rates than non-White patients, who often receive later-stage diagnoses and experience higher mortality.² Furthermore, racial/ ethnic minorities often face longer surgery wait times after diagnosis and have lower socioeconomic status (SES) and less favorable health insurance coverage, contributing to poorer outcomes.^2,3

To examine access to full-body skin examinations (FBSEs) by board-certified dermatologists in Los Angeles (LA) County, California, we analyzed the availability of FBSEs based on racial demographics, income, and insurance type (Medicaid [Medi-Cal] vs private [Blue Cross Blue Shield (BCBS)]). Demographic data by zip code were obtained from the US Census Bureau.⁴ This validated metric highlights socioeconomic disparities and minimizes data gaps^5,6 and was used to assess health care access among different population subgroups. Dermatologists’ contact information was obtained from the Find a Dermatologist page on the American Academy of Dermatology website and the listed phone numbers of their practice were used to contact them. Practices with board-certified dermatologists accepting new patients were included in the study; practices were not included if they had exclusive insurance plans; were pediatric, cosmetic, or research only; or were nonresponsive to calls. From August 2022 to September 2022, each practice was called twice within a 36-hour period—once by a simulated patient with Medi-Cal and once by a simulated patient with BCBS—and were asked about availability for new patient FBSE appointments and accepted insurance types. Data were analyzed using SAS software (SAS Institute Inc.).

Los Angeles County comprises 269 zip codes, of which 82 (30.5%) have dermatology practices. Of 213 total dermatologists in LA County listed on the American Academy of Dermatology website, 193 (90.6%) met preliminary criteria, and 169 (79.3%) were successfully contacted. Almost all (94.6% [160/169]) accepted new patients for FBSEs; of those, 63.1% (101/160) accepted only private insurance, 16.9% (27/160) accepted both private insurance and Medi-Cal, and 16.2% (26/160) did not accept any insurance. Racial predominance for each dermatology practice was analyzed by zip code (Table). Dermatologists included in our study were significantly more concentrated in predominantly non- Hispanic White areas of LA County vs predominantly Hispanic areas (P<.0001). Notably, the average income in predominantly non-Hispanic White zip codes ($114,757.74) was significantly higher than in predominantly Hispanic areas ($58,278.54)(P=.001)(Table).⁴

In LA County, 40.1% (108/269) of zip codes have no racial majority, 28.2% (76/269) are predominantly Hispanic, 27.5% (74/269) are predominantly non-Hispanic White, 2.2% (6/269) are predominantly Black, and 1.9% (5/269) are predominantly Asian.⁴ There are no dermatologists in predominantly Black zip codes, 2 in predominantly Asian zip codes, 14 in predominantly Hispanic zip codes, 38 in zip codes with no racial majority, and 106 in predominantly non-Hispanic White zip codes. There are significantly more dermatologists in predominantly non-Hispanic White zip codes compared to predominantly Hispanic zip codes (P<.0001). In LA County, the average income in predominantly Asian, non-Hispanic White, and Hispanic zip codes was $93,594, $114,757.84, and $58,278.54, respectively, in 2021.4 The average income in predominantly non-Hispanic White zip codes was significantly higher than in predominantly Hispanic zip codes (P=.001). There were no income data available for predominantly Black zip codes or zip codes with no racial majority.

The results from our study revealed potential barriers to FBSEs for racial and ethnic minorities in LA County, which supports previous research on the impact of SES, race, and insurance on access to dermatologic care.^2,3 Predominantly Hispanic zip codes have significantly lower income (P<.0001) and fewer dermatologists (P=.001) compared to zip codes that are predominantly non-Hispanic White, reflecting how lower SES correlates with worse health outcomes and higher melanoma mortality. Conversely, predominantly non-Hispanic White areas with higher income have better access to dermatologists, which may contribute to the improved melanoma survival rates among White patients. Additionally, most dermatologists accept only private insurance, further highlighting the disparity in FBSE access for non-White patients across LA County. While our study focused on FBSE access, our findings may point to a wider barrier to dermatologic care, especially in zip codes with fewer dermatologists. Further studies are needed to determine whether these areas also face barriers to accessing primary care.

Our study was limited by the exclusion of nonphysician providers (eg, nurse practitioners, physician assistants), a small sample size, and lack of available economic data for predominantly Black zip codes.⁴ Additionally, the exclusion of practices with exclusive insurance plans (eg, Kaiser Permanente) limited the generalizability of our findings, as our results did not account for the populations served by these practices. Furthermore, our analysis did not account for variations in practice size or the proportion of care provided to patients with different insurance types, which could impact overall accessibility. Additional studies are needed to explore the impact of these factors on access to general dermatologic care and not just FBSEs.

Racial/ethnic minorities and lower SES populations face major barriers to FBSE access in LA County, such as difficulty finding a dermatologist in their area or one who accepts Medi-Cal. Addressing these disparities is crucial for improving skin cancer outcomes. Further research is needed to develop strategies to eliminate these barriers to dermatologic care, such as increasing access to teledermatology, offering mobile dermatology clinics, and improving insurance coverage.

To the Editor:

Early skin cancer detection improves patient outcomes¹; however, socioeconomic and racial disparities may impact access to dermatologic care.² Although non-Hispanic White individuals have a high incidence of skin cancer, they experience higher melanoma-specific survival rates than non-White patients, who often receive later-stage diagnoses and experience higher mortality.² Furthermore, racial/ ethnic minorities often face longer surgery wait times after diagnosis and have lower socioeconomic status (SES) and less favorable health insurance coverage, contributing to poorer outcomes.^2,3

To examine access to full-body skin examinations (FBSEs) by board-certified dermatologists in Los Angeles (LA) County, California, we analyzed the availability of FBSEs based on racial demographics, income, and insurance type (Medicaid [Medi-Cal] vs private [Blue Cross Blue Shield (BCBS)]). Demographic data by zip code were obtained from the US Census Bureau.⁴ This validated metric highlights socioeconomic disparities and minimizes data gaps^5,6 and was used to assess health care access among different population subgroups. Dermatologists’ contact information was obtained from the Find a Dermatologist page on the American Academy of Dermatology website and the listed phone numbers of their practice were used to contact them. Practices with board-certified dermatologists accepting new patients were included in the study; practices were not included if they had exclusive insurance plans; were pediatric, cosmetic, or research only; or were nonresponsive to calls. From August 2022 to September 2022, each practice was called twice within a 36-hour period—once by a simulated patient with Medi-Cal and once by a simulated patient with BCBS—and were asked about availability for new patient FBSE appointments and accepted insurance types. Data were analyzed using SAS software (SAS Institute Inc.).

Los Angeles County comprises 269 zip codes, of which 82 (30.5%) have dermatology practices. Of 213 total dermatologists in LA County listed on the American Academy of Dermatology website, 193 (90.6%) met preliminary criteria, and 169 (79.3%) were successfully contacted. Almost all (94.6% [160/169]) accepted new patients for FBSEs; of those, 63.1% (101/160) accepted only private insurance, 16.9% (27/160) accepted both private insurance and Medi-Cal, and 16.2% (26/160) did not accept any insurance. Racial predominance for each dermatology practice was analyzed by zip code (Table). Dermatologists included in our study were significantly more concentrated in predominantly non- Hispanic White areas of LA County vs predominantly Hispanic areas (P<.0001). Notably, the average income in predominantly non-Hispanic White zip codes ($114,757.74) was significantly higher than in predominantly Hispanic areas ($58,278.54)(P=.001)(Table).⁴

In LA County, 40.1% (108/269) of zip codes have no racial majority, 28.2% (76/269) are predominantly Hispanic, 27.5% (74/269) are predominantly non-Hispanic White, 2.2% (6/269) are predominantly Black, and 1.9% (5/269) are predominantly Asian.⁴ There are no dermatologists in predominantly Black zip codes, 2 in predominantly Asian zip codes, 14 in predominantly Hispanic zip codes, 38 in zip codes with no racial majority, and 106 in predominantly non-Hispanic White zip codes. There are significantly more dermatologists in predominantly non-Hispanic White zip codes compared to predominantly Hispanic zip codes (P<.0001). In LA County, the average income in predominantly Asian, non-Hispanic White, and Hispanic zip codes was $93,594, $114,757.84, and $58,278.54, respectively, in 2021.4 The average income in predominantly non-Hispanic White zip codes was significantly higher than in predominantly Hispanic zip codes (P=.001). There were no income data available for predominantly Black zip codes or zip codes with no racial majority.

The results from our study revealed potential barriers to FBSEs for racial and ethnic minorities in LA County, which supports previous research on the impact of SES, race, and insurance on access to dermatologic care.^2,3 Predominantly Hispanic zip codes have significantly lower income (P<.0001) and fewer dermatologists (P=.001) compared to zip codes that are predominantly non-Hispanic White, reflecting how lower SES correlates with worse health outcomes and higher melanoma mortality. Conversely, predominantly non-Hispanic White areas with higher income have better access to dermatologists, which may contribute to the improved melanoma survival rates among White patients. Additionally, most dermatologists accept only private insurance, further highlighting the disparity in FBSE access for non-White patients across LA County. While our study focused on FBSE access, our findings may point to a wider barrier to dermatologic care, especially in zip codes with fewer dermatologists. Further studies are needed to determine whether these areas also face barriers to accessing primary care.

Our study was limited by the exclusion of nonphysician providers (eg, nurse practitioners, physician assistants), a small sample size, and lack of available economic data for predominantly Black zip codes.⁴ Additionally, the exclusion of practices with exclusive insurance plans (eg, Kaiser Permanente) limited the generalizability of our findings, as our results did not account for the populations served by these practices. Furthermore, our analysis did not account for variations in practice size or the proportion of care provided to patients with different insurance types, which could impact overall accessibility. Additional studies are needed to explore the impact of these factors on access to general dermatologic care and not just FBSEs.

Racial/ethnic minorities and lower SES populations face major barriers to FBSE access in LA County, such as difficulty finding a dermatologist in their area or one who accepts Medi-Cal. Addressing these disparities is crucial for improving skin cancer outcomes. Further research is needed to develop strategies to eliminate these barriers to dermatologic care, such as increasing access to teledermatology, offering mobile dermatology clinics, and improving insurance coverage.

To the Editor:

The incidence of nonmelanoma skin cancer (NMSC) is rapidly increasing worldwide. Due to its highly curable nature when treated early, accurate diagnosis is the cornerstone to good patient outcomes.¹ Accurate diagnosis of skin cancer and subsequent treatment decisions rely heavily on the congruence between clinical observations and histopathologic assessments. Clinical misdiagnosis of a malignant lesion can lead to delayed and suboptimal treatment, which may contribute to serious complications such as metastasis or even mortality. In this study, data from clinically diagnosed basal cell carcinomas (BCCs) and squamous cell carcinomas (SCCs) were compared to their identified histopathologic subtype classifications. The accuracy of the clinical diagnosis of these NMSCs was assessed by determining the rate of misdiagnosis and the respective positive predictive value (PPV).

A retrospective review of medical records from a private dermatology practice in Lubbock, Texas, was conducted to identify patients diagnosed with NMSC from January 1, 2017, through December 31, 2021. A total of 11,229 NMSCs were diagnosed and treated in 5877 patients. Of the NMSCs diagnosed, 11,145 were identified as keratinocyte carcinomas and were classified as BCCs or SCCs. The accuracy of the clinical diagnoses was determined by comparison to the histologic subtype identified via biopsy of the lesion. Although the use of a dermatoscope during the clinical encounter was not formally recorded, reports from the examining dermatologists indicated it was not used in the majority of cases.

If a lesion was clinically diagnosed as a BCC but was identified as a subtype of SCC on histology (or vice versa), the lesion was considered to be mismatched. The number of mismatched lesions and the mismatch rate for each lesion type/subtype is recorded in the Table. Of the total 11,145 keratinocyte carcinomas included in our study, there was an overall 10.63% mismatch rate, with 1185 of the malignancies having a differing clinical diagnosis (eg, BCC vs SCC) from the histologic findings. The clinical mismatch rate was notably higher for SCC compared to BCC (15.83% vs 7.03%, respectively).

The Table provides a breakdown of the BCC subtypes identified by histology with their computed mismatch rate and PPV. It is worth clarifying that lesions classified as more than one BCC subtype per the histologic findings were diagnosed as mixed BCC; these were further classified as mixed-aggressive BCC (if at least one aggressive BCC subtype was present) and mixed nonaggressive BCC (if no aggressive BCC subtype was present). Overall, BCCs were less likely to be misdiagnosed, with an average PPV of 92.97% compared to 84.17% for SCCs. Basosquamous BCC was the BCC subtype with the highest mismatch rate (25.48%), while sclerosing BCC has the lowest overall mismatch rate (1.33%). The most common malignancy was BCC, with nodular BCC being the most common subtype.

The Table also breaks down the SCC subtypes, reporting the most commonly misdiagnosed of any BCC or SCC subtype to be poorly differentiated SCC (mismatch rate, 38.46%). The lowest mismatch rate of the SCC subtypes was 5.97% for well-differentiated SCC.

There was an overall PPV of 89.37% in clinically evaluated malignancies and their respective histologic subtypes. Basal cell carcinoma had a lower overall mismatch rate of 7.03% compared to 15.83% in SCC. The most common misdiagnosis was attributed to poorly differentiated SCC (mismatch rate, 38.46%), while the least common misdiagnosed malignancy was sclerosing BCC (1.33%). The high mismatch rate of poorly differentiated SCC may be due to its diverging presentation from a typical SCC as a flat lesion with the absence of scaling, keratin, or bleeding, leading to the misdiagnosis of BCC.²

Accurate clinical diagnosis of NMSCs is the basis for further evaluation and treatment that should ensue in a timely manner; however, accurately identifying BCCs vs SCCs solely based on clinical examination can be challenging due to variable manifestations and overlapping features. Basal cell carcinoma commonly presents as a shiny pink/flesh-colored nodule, macule, or patch with surface telangiectasia, sometimes appearing with ulceration or crusting.³ Alternatively, SCC typically appears as a firm, sharply demarcated, red nodule with a thick overlying scale.⁴ Definitive diagnoses can be difficult upon clinical examination since these features can be shared between the 2 subtypes. To aid in these uncertainties, a growing number of clinicians are implementing the use of dermoscopy in their everyday practice.

Dermoscopy is an extremely useful tool in improving the diagnostic accuracy of skin cancers compared to examination with the naked eye, as it provides detailed visualization of specific structures and patterns in skin cancer lesions.⁵ The dermoscopic appearance of BCC is characterized by pearly blue-gray or translucent globules with arborizing vessels, spoke-wheel structures, and leaflike areas.^5,6 Conversely, dermoscopic features of SCC may include a milky-red globule with a scaly, sharply demarcated, crusted lesion with polymorphous vasculature, sometimes resembling a persistent sore or nonhealing wound.^4,5 Though the use of dermoscopy can aid in diagnosis upon initial examination, certain factors such as trauma, ulceration, and previous treatments that distorted the lesion’s architecture may lead to misdiagnosis. Furthermore, the distinct vascular patterns found in BCC and SCC may be mistaken for each other and therefore lead to misdiagnosis upon examination.⁷ Other variables that may complicate diagnosis include the location of the lesion, its size, and the presence of other skin conditions or nearby lesions.

The primary limitation of the current study was the limited scope of the data, as they were derived from patients seen at one private dermatology practice, preventing the generalizability of our findings. However, our results show trends similar to those observed in other studies analyzing the clinical accuracy of skin cancer diagnoses, with higher PPVs for BCC compared to SCC. A study by Ahnlide and Bjellerup⁸ was based in a hospital dermatology department and demonstrated a PPV of 85.5% for BCC compared to 92.97% in our study; for SCC, the PPV was 67.3% compared to 84.17% in our study. In another study by Heal et al,⁹ data were collected from an Australian registry that included records of all histologically confirmed skin cancers from December 1996 to October 1999 from 202 general practitioners and 42 specialists, including 1 dermatologist. The PPVs for BCC and SCC were 72.7% and 49.4%, respectively. Although our results indicated higher PPVs compared to these 2 studies, some of the discrepancies can be accounted for by the differences in clinical setting as well as the lack of expertise of nondermatologist physicians in identifying skin malignancies in the study by Heal et al.⁹

The current study was further limited by the lack of data quantifying the number of lesions clinically suspected to be malignant but found to be histologically benign. It is typical for clinicians to have a low threshold to biopsy a suspicious lesion with atypical features (eg, rapid evolution and growth, bleeding, crusting). Furthermore, the identification of risk factors in the patient’s medical and family history (eg, exposure to radiation, personal or family history of skin cancers) can heavily influence a clinician’s decision to biopsy a lesion with an atypical appearance.¹⁰ Many benign lesions are biopsied to avoid missing a diagnosis of malignancy. Consequently, our results suggest a high degree of clinical misdiagnosis of BCCs and SCCs. Obtaining data on the number of lesions suspected to be BCC or SCC that were found to be histologically benign would be a valuable addition to our study, as it would provide a measurable insight into the sensitivity of clinicians’ decision-making to identify a lesion as suspicious and warranting biopsy.

While clinical diagnosis plays a vital role in identifying suspected NMSCs such as BCC and SCC, its accuracy can be limited even with the use of dermoscopy. Overall, our data have shown a high rate of diagnostic accuracy upon suspicion of malignancy, but the different variables that affect clinical presentation promote histologic diagnosis to prevail as the gold standard.

To the Editor:

The incidence of nonmelanoma skin cancer (NMSC) is rapidly increasing worldwide. Due to its highly curable nature when treated early, accurate diagnosis is the cornerstone to good patient outcomes.¹ Accurate diagnosis of skin cancer and subsequent treatment decisions rely heavily on the congruence between clinical observations and histopathologic assessments. Clinical misdiagnosis of a malignant lesion can lead to delayed and suboptimal treatment, which may contribute to serious complications such as metastasis or even mortality. In this study, data from clinically diagnosed basal cell carcinomas (BCCs) and squamous cell carcinomas (SCCs) were compared to their identified histopathologic subtype classifications. The accuracy of the clinical diagnosis of these NMSCs was assessed by determining the rate of misdiagnosis and the respective positive predictive value (PPV).

A retrospective review of medical records from a private dermatology practice in Lubbock, Texas, was conducted to identify patients diagnosed with NMSC from January 1, 2017, through December 31, 2021. A total of 11,229 NMSCs were diagnosed and treated in 5877 patients. Of the NMSCs diagnosed, 11,145 were identified as keratinocyte carcinomas and were classified as BCCs or SCCs. The accuracy of the clinical diagnoses was determined by comparison to the histologic subtype identified via biopsy of the lesion. Although the use of a dermatoscope during the clinical encounter was not formally recorded, reports from the examining dermatologists indicated it was not used in the majority of cases.

If a lesion was clinically diagnosed as a BCC but was identified as a subtype of SCC on histology (or vice versa), the lesion was considered to be mismatched. The number of mismatched lesions and the mismatch rate for each lesion type/subtype is recorded in the Table. Of the total 11,145 keratinocyte carcinomas included in our study, there was an overall 10.63% mismatch rate, with 1185 of the malignancies having a differing clinical diagnosis (eg, BCC vs SCC) from the histologic findings. The clinical mismatch rate was notably higher for SCC compared to BCC (15.83% vs 7.03%, respectively).

The Table provides a breakdown of the BCC subtypes identified by histology with their computed mismatch rate and PPV. It is worth clarifying that lesions classified as more than one BCC subtype per the histologic findings were diagnosed as mixed BCC; these were further classified as mixed-aggressive BCC (if at least one aggressive BCC subtype was present) and mixed nonaggressive BCC (if no aggressive BCC subtype was present). Overall, BCCs were less likely to be misdiagnosed, with an average PPV of 92.97% compared to 84.17% for SCCs. Basosquamous BCC was the BCC subtype with the highest mismatch rate (25.48%), while sclerosing BCC has the lowest overall mismatch rate (1.33%). The most common malignancy was BCC, with nodular BCC being the most common subtype.

The Table also breaks down the SCC subtypes, reporting the most commonly misdiagnosed of any BCC or SCC subtype to be poorly differentiated SCC (mismatch rate, 38.46%). The lowest mismatch rate of the SCC subtypes was 5.97% for well-differentiated SCC.

There was an overall PPV of 89.37% in clinically evaluated malignancies and their respective histologic subtypes. Basal cell carcinoma had a lower overall mismatch rate of 7.03% compared to 15.83% in SCC. The most common misdiagnosis was attributed to poorly differentiated SCC (mismatch rate, 38.46%), while the least common misdiagnosed malignancy was sclerosing BCC (1.33%). The high mismatch rate of poorly differentiated SCC may be due to its diverging presentation from a typical SCC as a flat lesion with the absence of scaling, keratin, or bleeding, leading to the misdiagnosis of BCC.²

Accurate clinical diagnosis of NMSCs is the basis for further evaluation and treatment that should ensue in a timely manner; however, accurately identifying BCCs vs SCCs solely based on clinical examination can be challenging due to variable manifestations and overlapping features. Basal cell carcinoma commonly presents as a shiny pink/flesh-colored nodule, macule, or patch with surface telangiectasia, sometimes appearing with ulceration or crusting.³ Alternatively, SCC typically appears as a firm, sharply demarcated, red nodule with a thick overlying scale.⁴ Definitive diagnoses can be difficult upon clinical examination since these features can be shared between the 2 subtypes. To aid in these uncertainties, a growing number of clinicians are implementing the use of dermoscopy in their everyday practice.

Dermoscopy is an extremely useful tool in improving the diagnostic accuracy of skin cancers compared to examination with the naked eye, as it provides detailed visualization of specific structures and patterns in skin cancer lesions.⁵ The dermoscopic appearance of BCC is characterized by pearly blue-gray or translucent globules with arborizing vessels, spoke-wheel structures, and leaflike areas.^5,6 Conversely, dermoscopic features of SCC may include a milky-red globule with a scaly, sharply demarcated, crusted lesion with polymorphous vasculature, sometimes resembling a persistent sore or nonhealing wound.^4,5 Though the use of dermoscopy can aid in diagnosis upon initial examination, certain factors such as trauma, ulceration, and previous treatments that distorted the lesion’s architecture may lead to misdiagnosis. Furthermore, the distinct vascular patterns found in BCC and SCC may be mistaken for each other and therefore lead to misdiagnosis upon examination.⁷ Other variables that may complicate diagnosis include the location of the lesion, its size, and the presence of other skin conditions or nearby lesions.

The primary limitation of the current study was the limited scope of the data, as they were derived from patients seen at one private dermatology practice, preventing the generalizability of our findings. However, our results show trends similar to those observed in other studies analyzing the clinical accuracy of skin cancer diagnoses, with higher PPVs for BCC compared to SCC. A study by Ahnlide and Bjellerup⁸ was based in a hospital dermatology department and demonstrated a PPV of 85.5% for BCC compared to 92.97% in our study; for SCC, the PPV was 67.3% compared to 84.17% in our study. In another study by Heal et al,⁹ data were collected from an Australian registry that included records of all histologically confirmed skin cancers from December 1996 to October 1999 from 202 general practitioners and 42 specialists, including 1 dermatologist. The PPVs for BCC and SCC were 72.7% and 49.4%, respectively. Although our results indicated higher PPVs compared to these 2 studies, some of the discrepancies can be accounted for by the differences in clinical setting as well as the lack of expertise of nondermatologist physicians in identifying skin malignancies in the study by Heal et al.⁹

The current study was further limited by the lack of data quantifying the number of lesions clinically suspected to be malignant but found to be histologically benign. It is typical for clinicians to have a low threshold to biopsy a suspicious lesion with atypical features (eg, rapid evolution and growth, bleeding, crusting). Furthermore, the identification of risk factors in the patient’s medical and family history (eg, exposure to radiation, personal or family history of skin cancers) can heavily influence a clinician’s decision to biopsy a lesion with an atypical appearance.¹⁰ Many benign lesions are biopsied to avoid missing a diagnosis of malignancy. Consequently, our results suggest a high degree of clinical misdiagnosis of BCCs and SCCs. Obtaining data on the number of lesions suspected to be BCC or SCC that were found to be histologically benign would be a valuable addition to our study, as it would provide a measurable insight into the sensitivity of clinicians’ decision-making to identify a lesion as suspicious and warranting biopsy.

While clinical diagnosis plays a vital role in identifying suspected NMSCs such as BCC and SCC, its accuracy can be limited even with the use of dermoscopy. Overall, our data have shown a high rate of diagnostic accuracy upon suspicion of malignancy, but the different variables that affect clinical presentation promote histologic diagnosis to prevail as the gold standard.

To the Editor:

The incidence of nonmelanoma skin cancer (NMSC) is rapidly increasing worldwide. Due to its highly curable nature when treated early, accurate diagnosis is the cornerstone to good patient outcomes.¹ Accurate diagnosis of skin cancer and subsequent treatment decisions rely heavily on the congruence between clinical observations and histopathologic assessments. Clinical misdiagnosis of a malignant lesion can lead to delayed and suboptimal treatment, which may contribute to serious complications such as metastasis or even mortality. In this study, data from clinically diagnosed basal cell carcinomas (BCCs) and squamous cell carcinomas (SCCs) were compared to their identified histopathologic subtype classifications. The accuracy of the clinical diagnosis of these NMSCs was assessed by determining the rate of misdiagnosis and the respective positive predictive value (PPV).

A retrospective review of medical records from a private dermatology practice in Lubbock, Texas, was conducted to identify patients diagnosed with NMSC from January 1, 2017, through December 31, 2021. A total of 11,229 NMSCs were diagnosed and treated in 5877 patients. Of the NMSCs diagnosed, 11,145 were identified as keratinocyte carcinomas and were classified as BCCs or SCCs. The accuracy of the clinical diagnoses was determined by comparison to the histologic subtype identified via biopsy of the lesion. Although the use of a dermatoscope during the clinical encounter was not formally recorded, reports from the examining dermatologists indicated it was not used in the majority of cases.

If a lesion was clinically diagnosed as a BCC but was identified as a subtype of SCC on histology (or vice versa), the lesion was considered to be mismatched. The number of mismatched lesions and the mismatch rate for each lesion type/subtype is recorded in the Table. Of the total 11,145 keratinocyte carcinomas included in our study, there was an overall 10.63% mismatch rate, with 1185 of the malignancies having a differing clinical diagnosis (eg, BCC vs SCC) from the histologic findings. The clinical mismatch rate was notably higher for SCC compared to BCC (15.83% vs 7.03%, respectively).

The Table provides a breakdown of the BCC subtypes identified by histology with their computed mismatch rate and PPV. It is worth clarifying that lesions classified as more than one BCC subtype per the histologic findings were diagnosed as mixed BCC; these were further classified as mixed-aggressive BCC (if at least one aggressive BCC subtype was present) and mixed nonaggressive BCC (if no aggressive BCC subtype was present). Overall, BCCs were less likely to be misdiagnosed, with an average PPV of 92.97% compared to 84.17% for SCCs. Basosquamous BCC was the BCC subtype with the highest mismatch rate (25.48%), while sclerosing BCC has the lowest overall mismatch rate (1.33%). The most common malignancy was BCC, with nodular BCC being the most common subtype.

The Table also breaks down the SCC subtypes, reporting the most commonly misdiagnosed of any BCC or SCC subtype to be poorly differentiated SCC (mismatch rate, 38.46%). The lowest mismatch rate of the SCC subtypes was 5.97% for well-differentiated SCC.

There was an overall PPV of 89.37% in clinically evaluated malignancies and their respective histologic subtypes. Basal cell carcinoma had a lower overall mismatch rate of 7.03% compared to 15.83% in SCC. The most common misdiagnosis was attributed to poorly differentiated SCC (mismatch rate, 38.46%), while the least common misdiagnosed malignancy was sclerosing BCC (1.33%). The high mismatch rate of poorly differentiated SCC may be due to its diverging presentation from a typical SCC as a flat lesion with the absence of scaling, keratin, or bleeding, leading to the misdiagnosis of BCC.²

Accurate clinical diagnosis of NMSCs is the basis for further evaluation and treatment that should ensue in a timely manner; however, accurately identifying BCCs vs SCCs solely based on clinical examination can be challenging due to variable manifestations and overlapping features. Basal cell carcinoma commonly presents as a shiny pink/flesh-colored nodule, macule, or patch with surface telangiectasia, sometimes appearing with ulceration or crusting.³ Alternatively, SCC typically appears as a firm, sharply demarcated, red nodule with a thick overlying scale.⁴ Definitive diagnoses can be difficult upon clinical examination since these features can be shared between the 2 subtypes. To aid in these uncertainties, a growing number of clinicians are implementing the use of dermoscopy in their everyday practice.

Dermoscopy is an extremely useful tool in improving the diagnostic accuracy of skin cancers compared to examination with the naked eye, as it provides detailed visualization of specific structures and patterns in skin cancer lesions.⁵ The dermoscopic appearance of BCC is characterized by pearly blue-gray or translucent globules with arborizing vessels, spoke-wheel structures, and leaflike areas.^5,6 Conversely, dermoscopic features of SCC may include a milky-red globule with a scaly, sharply demarcated, crusted lesion with polymorphous vasculature, sometimes resembling a persistent sore or nonhealing wound.^4,5 Though the use of dermoscopy can aid in diagnosis upon initial examination, certain factors such as trauma, ulceration, and previous treatments that distorted the lesion’s architecture may lead to misdiagnosis. Furthermore, the distinct vascular patterns found in BCC and SCC may be mistaken for each other and therefore lead to misdiagnosis upon examination.⁷ Other variables that may complicate diagnosis include the location of the lesion, its size, and the presence of other skin conditions or nearby lesions.

The primary limitation of the current study was the limited scope of the data, as they were derived from patients seen at one private dermatology practice, preventing the generalizability of our findings. However, our results show trends similar to those observed in other studies analyzing the clinical accuracy of skin cancer diagnoses, with higher PPVs for BCC compared to SCC. A study by Ahnlide and Bjellerup⁸ was based in a hospital dermatology department and demonstrated a PPV of 85.5% for BCC compared to 92.97% in our study; for SCC, the PPV was 67.3% compared to 84.17% in our study. In another study by Heal et al,⁹ data were collected from an Australian registry that included records of all histologically confirmed skin cancers from December 1996 to October 1999 from 202 general practitioners and 42 specialists, including 1 dermatologist. The PPVs for BCC and SCC were 72.7% and 49.4%, respectively. Although our results indicated higher PPVs compared to these 2 studies, some of the discrepancies can be accounted for by the differences in clinical setting as well as the lack of expertise of nondermatologist physicians in identifying skin malignancies in the study by Heal et al.⁹

The current study was further limited by the lack of data quantifying the number of lesions clinically suspected to be malignant but found to be histologically benign. It is typical for clinicians to have a low threshold to biopsy a suspicious lesion with atypical features (eg, rapid evolution and growth, bleeding, crusting). Furthermore, the identification of risk factors in the patient’s medical and family history (eg, exposure to radiation, personal or family history of skin cancers) can heavily influence a clinician’s decision to biopsy a lesion with an atypical appearance.¹⁰ Many benign lesions are biopsied to avoid missing a diagnosis of malignancy. Consequently, our results suggest a high degree of clinical misdiagnosis of BCCs and SCCs. Obtaining data on the number of lesions suspected to be BCC or SCC that were found to be histologically benign would be a valuable addition to our study, as it would provide a measurable insight into the sensitivity of clinicians’ decision-making to identify a lesion as suspicious and warranting biopsy.

While clinical diagnosis plays a vital role in identifying suspected NMSCs such as BCC and SCC, its accuracy can be limited even with the use of dermoscopy. Overall, our data have shown a high rate of diagnostic accuracy upon suspicion of malignancy, but the different variables that affect clinical presentation promote histologic diagnosis to prevail as the gold standard.

To the Editor:

Psoriasis is a chronic inflammatory condition that affects individuals in various extracutaneous ways.¹ Prior studies have documented a decrease in exercise intensity among patients with psoriasis²; however, few studies have specifically investigated baseline mobility in this population. Baseline mobility denotes an individual’s fundamental ability to walk or move around without assistance of any kind. Impaired mobility—when baseline mobility is compromised—is an aspect of the wider diversity, equity, and inclusion framework that underscores the significance of recognizing challenges and promoting inclusive measures, both at the point of care and in research.³ study sought to analyze the relationship between psoriasis and baseline mobility among US adults (aged 45 to 80 years) utilizing the latest data from the National Health and Nutrition Examination Survey (NHANES) database for psoriasis.⁴ We used three 2-year cycles of NHANES data to create a 2009-2014 dataset.

The overall NHANES response rate among adults aged 45 to 80 years between 2009 and 2014 was 67.9%. Patients were categorized as having impaired mobility if they responded “yes” to the following question: “Because of a health problem, do you have difficulty walking without using any special equipment?” Psoriasis status was assessed by the following question: “Have you ever been told by a doctor or other health professional that you had psoriasis?” Multivariable logistic regression analyses were performed using Stata/SE 18.0 software (StataCorp LLC) to assess the relationship between psoriasis and impaired mobility. Age, income, education, sex, race, tobacco use, diabetes status, body mass index, and arthritis status were controlled for in our models.

Our analysis initially included 9982 participants; 14 did not respond to questions assessing psoriasis and impaired mobility and were excluded. The prevalence of impaired mobility in patients with psoriasis was 17.1% compared with 10.9% among those without psoriasis (Table 1). There was a significant association between psoriasis and impaired mobility among patients aged 45 to 80 years after adjusting for potential confounding variables (adjusted odds ratio [AOR], 1.54; 95% CI, 1.04- 2.29; P=.032)(Table 2). Analyses of subgroups yielded no statistically significant results.

Our study demonstrated a statistically significant difference in mobility between individuals with psoriasis compared with the general population, which remained significant when controlling for arthritis, obesity, and diabetes (P=.032). This may be the result of several influences. First, the location of the psoriasis may impact mobility. Plantar psoriasis—a manifestation on the soles of the feet—can cause discomfort and pain, which can hinder walking and standing.⁵ Second, a study by Lasselin et al⁶ found that systemic inflammation contributes to mobility impairment through alterations in gait and posture, which suggests that the inflammatory processes inherent in psoriasis could intrinsically modify walking speed and stride, potentially exacerbating mobility difficulties independent of other comorbid conditions. These findings suggest that psoriasis may disproportionately affect individuals with impaired mobility, independent of comorbid arthritis, obesity, and diabetes.

These findings have broad implications for diversity, equity, and inclusion. They should prompt us to consider the practical challenges faced by this patient population and the ways that we can address barriers to care. Offering telehealth appointments, making primary care referrals for impaired mobility workups, and advising patients of direct-to-home delivery of prescriptions are good places to start.

Limitations to our study include the lack of specificity in the survey question, self-reporting bias, and the inability to control for the psoriasis location. Further investigations are warranted in large, representative US adult populations to assess the implications of impaired mobility in patients with psoriasis.

To the Editor:

Psoriasis is a chronic inflammatory condition that affects individuals in various extracutaneous ways.¹ Prior studies have documented a decrease in exercise intensity among patients with psoriasis²; however, few studies have specifically investigated baseline mobility in this population. Baseline mobility denotes an individual’s fundamental ability to walk or move around without assistance of any kind. Impaired mobility—when baseline mobility is compromised—is an aspect of the wider diversity, equity, and inclusion framework that underscores the significance of recognizing challenges and promoting inclusive measures, both at the point of care and in research.³ study sought to analyze the relationship between psoriasis and baseline mobility among US adults (aged 45 to 80 years) utilizing the latest data from the National Health and Nutrition Examination Survey (NHANES) database for psoriasis.⁴ We used three 2-year cycles of NHANES data to create a 2009-2014 dataset.

The overall NHANES response rate among adults aged 45 to 80 years between 2009 and 2014 was 67.9%. Patients were categorized as having impaired mobility if they responded “yes” to the following question: “Because of a health problem, do you have difficulty walking without using any special equipment?” Psoriasis status was assessed by the following question: “Have you ever been told by a doctor or other health professional that you had psoriasis?” Multivariable logistic regression analyses were performed using Stata/SE 18.0 software (StataCorp LLC) to assess the relationship between psoriasis and impaired mobility. Age, income, education, sex, race, tobacco use, diabetes status, body mass index, and arthritis status were controlled for in our models.

Our analysis initially included 9982 participants; 14 did not respond to questions assessing psoriasis and impaired mobility and were excluded. The prevalence of impaired mobility in patients with psoriasis was 17.1% compared with 10.9% among those without psoriasis (Table 1). There was a significant association between psoriasis and impaired mobility among patients aged 45 to 80 years after adjusting for potential confounding variables (adjusted odds ratio [AOR], 1.54; 95% CI, 1.04- 2.29; P=.032)(Table 2). Analyses of subgroups yielded no statistically significant results.

Our study demonstrated a statistically significant difference in mobility between individuals with psoriasis compared with the general population, which remained significant when controlling for arthritis, obesity, and diabetes (P=.032). This may be the result of several influences. First, the location of the psoriasis may impact mobility. Plantar psoriasis—a manifestation on the soles of the feet—can cause discomfort and pain, which can hinder walking and standing.⁵ Second, a study by Lasselin et al⁶ found that systemic inflammation contributes to mobility impairment through alterations in gait and posture, which suggests that the inflammatory processes inherent in psoriasis could intrinsically modify walking speed and stride, potentially exacerbating mobility difficulties independent of other comorbid conditions. These findings suggest that psoriasis may disproportionately affect individuals with impaired mobility, independent of comorbid arthritis, obesity, and diabetes.

These findings have broad implications for diversity, equity, and inclusion. They should prompt us to consider the practical challenges faced by this patient population and the ways that we can address barriers to care. Offering telehealth appointments, making primary care referrals for impaired mobility workups, and advising patients of direct-to-home delivery of prescriptions are good places to start.

Limitations to our study include the lack of specificity in the survey question, self-reporting bias, and the inability to control for the psoriasis location. Further investigations are warranted in large, representative US adult populations to assess the implications of impaired mobility in patients with psoriasis.

To the Editor:

Psoriasis is a chronic inflammatory condition that affects individuals in various extracutaneous ways.¹ Prior studies have documented a decrease in exercise intensity among patients with psoriasis²; however, few studies have specifically investigated baseline mobility in this population. Baseline mobility denotes an individual’s fundamental ability to walk or move around without assistance of any kind. Impaired mobility—when baseline mobility is compromised—is an aspect of the wider diversity, equity, and inclusion framework that underscores the significance of recognizing challenges and promoting inclusive measures, both at the point of care and in research.³ study sought to analyze the relationship between psoriasis and baseline mobility among US adults (aged 45 to 80 years) utilizing the latest data from the National Health and Nutrition Examination Survey (NHANES) database for psoriasis.⁴ We used three 2-year cycles of NHANES data to create a 2009-2014 dataset.

The overall NHANES response rate among adults aged 45 to 80 years between 2009 and 2014 was 67.9%. Patients were categorized as having impaired mobility if they responded “yes” to the following question: “Because of a health problem, do you have difficulty walking without using any special equipment?” Psoriasis status was assessed by the following question: “Have you ever been told by a doctor or other health professional that you had psoriasis?” Multivariable logistic regression analyses were performed using Stata/SE 18.0 software (StataCorp LLC) to assess the relationship between psoriasis and impaired mobility. Age, income, education, sex, race, tobacco use, diabetes status, body mass index, and arthritis status were controlled for in our models.

Our analysis initially included 9982 participants; 14 did not respond to questions assessing psoriasis and impaired mobility and were excluded. The prevalence of impaired mobility in patients with psoriasis was 17.1% compared with 10.9% among those without psoriasis (Table 1). There was a significant association between psoriasis and impaired mobility among patients aged 45 to 80 years after adjusting for potential confounding variables (adjusted odds ratio [AOR], 1.54; 95% CI, 1.04- 2.29; P=.032)(Table 2). Analyses of subgroups yielded no statistically significant results.

Our study demonstrated a statistically significant difference in mobility between individuals with psoriasis compared with the general population, which remained significant when controlling for arthritis, obesity, and diabetes (P=.032). This may be the result of several influences. First, the location of the psoriasis may impact mobility. Plantar psoriasis—a manifestation on the soles of the feet—can cause discomfort and pain, which can hinder walking and standing.⁵ Second, a study by Lasselin et al⁶ found that systemic inflammation contributes to mobility impairment through alterations in gait and posture, which suggests that the inflammatory processes inherent in psoriasis could intrinsically modify walking speed and stride, potentially exacerbating mobility difficulties independent of other comorbid conditions. These findings suggest that psoriasis may disproportionately affect individuals with impaired mobility, independent of comorbid arthritis, obesity, and diabetes.

These findings have broad implications for diversity, equity, and inclusion. They should prompt us to consider the practical challenges faced by this patient population and the ways that we can address barriers to care. Offering telehealth appointments, making primary care referrals for impaired mobility workups, and advising patients of direct-to-home delivery of prescriptions are good places to start.

Limitations to our study include the lack of specificity in the survey question, self-reporting bias, and the inability to control for the psoriasis location. Further investigations are warranted in large, representative US adult populations to assess the implications of impaired mobility in patients with psoriasis.

To the Editor:

Cutaneous medical conditions can have a substantial impact on patients’ functioning and quality of life. Many patients with severe skin disease are eligible to receive disability assistance that can provide them with essential income and health care. Previous research has highlighted disability assessment as one of the most important ways physicians can help mitigate the health consequences of poverty.¹ Dermatologists can play an important role in the disability assessment process by documenting the facts associated with patients’ skin conditions.

Although skin conditions have a relatively high prevalence, they remain underrepresented in disability claims. Between 1997 and 2004, occupational skin diseases accounted for 12% to 17% of nonfatal work-related illnesses; however, during that same period, skin conditions comprised only 0.21% of disability claims in the United States.^2,3 Historically, there has been hesitancy among dermatologists to complete disability paperwork; a 1976 survey of dermatologists cited extensive paperwork, “troublesome patients,” and fee schedule issues as reasons.⁴ The lack of training regarding disability assessment in medical school and residency also has been noted.⁵

To characterize modern attitudes toward disability assessments, we conducted a survey of dermatologists across the United States. Our study was reviewed and declared exempt by the institutional review board of the Lundquist Institute for Biomedical Innovation at Harbor-UCLA Medical Center (Torrance, California)(approval #18CR-32242-01). Using convenience sampling, we emailed dermatologists from the Association of Professors of Dermatology and dermatology state societies in all 50 states inviting them to participate in our voluntary and anonymous survey, which was administered using SurveyMonkey. The use of all society mailing lists was approved by the respective owners. The 15-question survey included multiple choice, Likert scale, and free response sections. Summary and descriptive statistics were used to describe respondent demographics and identify any patterns in responses.

For each Likert-based question, participants ranked their degree of agreement with a statement as: 1=strongly disagree, 2=somewhat disagree, 3=neither agree nor disagree/neutral, 4=somewhat agree, and 5=strongly agree. The mean response and standard deviation were reported for each Likert scale prompt. Preplanned 1-sample t testing was used to analyze Likert scale data, in which the mean response for each prompt was compared to a baseline response of 3 (neutral). A P value <.05 was considered statistically significant. Statistical analyses were performed using SPSS Statistics for MacOS, version 27 (IBM).

Seventy-eight dermatologists agreed to participate, and 70 completed the survey, for a response rate of 89.7% (Table 1). The dermatologists we surveyed practiced in a variety of clinical settings, including academic public hospitals (46.2% [36/78]), academic private hospitals (33.3% [26/78]), and private practices (32.1% [25/78]), and 60.3% (47/78) reported providing disability documentation at some point. Most of the respondents (64.3% [45/70]) did not perform assessments in an average month (Table 2). Medical assessment documentation was provided most frequently for workers’ compensation (50.0% [35/70]), private insurance (27.1% [19/70]), and Social Security Disability Insurance (25.7% [18/70]). Dermatologists overwhelmingly reported no formal training for disability assessment in medical school (94.3% [66/70]), residency (97.1% [68/70]), or clinical practice (81.4% [57/70]).

In the Likert scale prompts, respondents agreed that they were uncertain of their role in disability assessment (mean response, 3.6; P<.001). Moreover, they were uncomfortable providing assessments (mean response, 3.5; P<.001) and felt that they did not have sufficient time to perform them (mean response, 3.6; P<.001). Dermatologists disagreed that they received adequate compensation for performing assessments (mean response, 2.2; P<.001) and felt that they did not have enough time to participate in assessments (mean response, 3.6; P<.001). Respondents generally did not feel distrustful of patients seeking disability assessment (mean response, 2.8; P=.043). Dermatologists neither agreed nor disagreed when asked if they thought that physicians can determine disability status (mean response, 3.2; P=.118). The details of the Likert scale responses are described in Table 3. Respondents also were uncertain as to which dermatologic conditions were eligible for disability. When asked to select which conditions from a list of 10 were eligible per the Social Security Administration listing of disability impairments, only 15.4% (12/70) of respondents correctly identified that all the conditions qualified; these included ichthyosis, pemphigus vulgaris, allergic contact dermatitis, hidradenitis suppurativa, systemic lupus erythematosus, chromoblastomycosis, xeroderma pigmentosum, burns, malignant melanoma, and scleroderma.⁶

In the free-response prompts, respondents frequently described extensive paperwork, inadequate time, and lack of reimbursement as barriers to providing documentation. Often, dermatologists found that the forms were not well matched to the skin conditions they were evaluating and rather had a musculoskeletal focus. Multiple individuals commented on the challenge in assessing the percentage of disability and functional/psychosocial impairment in skin conditions. One respondent noted that workers’ compensation forms ask if the patient is “…permanent and stationary…for most conditions this has no meaning in dermatology.” Some felt hesitant to provide documentation because they had insufficient patient history, especially regarding employment, and opted to defer to primary care providers who might be more familiar with the full patient history.

A dermatologist described their perspective as follows:

“…As a specialist I feel that I don’t have a complete look into all the factors that could contribute to a patient[’]s need to go on disability, and I don’t have experience with filling out disability requests. That being said, if a patient[’]s request for disability was due to a skin disease that I know way more about than [a] primary care [physician] would, I would do the disability assessment.”

Another respondent noted the complexity in “establishing causality” for workers’ compensation. Another dermatologist reported,

“The most frequent challenging situation I encounter is being asked to evaluate for maximum medical improvement after patch testing. If the patient is not fully avoiding contact allergens either at home or at work, then I typically document that they are not at [maximum medical improvement]. The reality is that most frequently it is due to exposure to allergens at home so the line between what is a legitimate worker’s comp[ensation] issue and what is a home life choice is blurry.”

Nevertheless, respondents expressed interest in learning more about disability assessment procedures. Summary guides, lectures, and prefilled paperwork were the most popular initiatives that respondents agreed would be beneficial toward becoming educated regarding disability assessment (78.6%, 58.6%, and 58.6%, respectively)(Table 2). One respondent noted that “previous [internal medicine] history help[ed]” them in performing cutaneous disability assessments.

As with any survey, our study did have some inherent limitations. Only a relatively small sample size was willing to complete the survey. There was a predominance of respondents from California (34.6% [27/78]), as well as those practicing for less than 15 years (58.9% [46/78])(Figure). This could limit generalizability to the national population of dermatologists. In addition, there was potential for recall bias and errors in responding given the self-reported nature of the study. Different individuals may interpret the Likert scale options in various ways, which could skew results unintentionally. However, the survey was largely qualitative in nature, making it a legitimate tool for answering our research questions. Moreover, we were able to hear the perspectives of dermatologists across diverse practice settings, with free response prompts to increase the depth of the survey.

Almost 50 years later, our survey echoes common themes from Adams’ 1976 survey.⁴ Inadequate compensation, limited time, and burdensome paperwork all continue to hinder dermatologists’ ability to perform disability assessments. Our participants frequently commented that the current disability forms are not congruent with the nature of skin conditions, making it challenging to accurately document the facts.

Moreover, respondents felt uncertain in their role in disability assessment and occasionally noted distrust of patients or insufficient patient history as barriers to completing assessments. They also were unsure if physicians can grant disability status. This is a common misconception among physicians that leads to discomfort in helping with disability assessment.⁷ The role of physicians in disability assessment is to document the facts of a patient’s illness, not to determine whether they are eligible for benefits. We discovered uncertainty in our respondents’ ability to identify conditions eligible for disability, highlighting an area in need of greater education for physicians.

Despite these obstacles, respondents were interested in learning more about disability assessment and highlighted several practical approaches that could help them better perform this task. As skin specialists, dermatologists are the best-equipped physicians to assess cutaneous conditions and should play a greater role in performing disability assessments, which could be achieved through increased educational initiatives and individual physician motivation.⁷ We call for greater collaboration and reflection on the importance of disability assistance among dermatologists to increase participation in the disability-assessment process.

To the Editor:

Cutaneous medical conditions can have a substantial impact on patients’ functioning and quality of life. Many patients with severe skin disease are eligible to receive disability assistance that can provide them with essential income and health care. Previous research has highlighted disability assessment as one of the most important ways physicians can help mitigate the health consequences of poverty.¹ Dermatologists can play an important role in the disability assessment process by documenting the facts associated with patients’ skin conditions.

Although skin conditions have a relatively high prevalence, they remain underrepresented in disability claims. Between 1997 and 2004, occupational skin diseases accounted for 12% to 17% of nonfatal work-related illnesses; however, during that same period, skin conditions comprised only 0.21% of disability claims in the United States.^2,3 Historically, there has been hesitancy among dermatologists to complete disability paperwork; a 1976 survey of dermatologists cited extensive paperwork, “troublesome patients,” and fee schedule issues as reasons.⁴ The lack of training regarding disability assessment in medical school and residency also has been noted.⁵

To characterize modern attitudes toward disability assessments, we conducted a survey of dermatologists across the United States. Our study was reviewed and declared exempt by the institutional review board of the Lundquist Institute for Biomedical Innovation at Harbor-UCLA Medical Center (Torrance, California)(approval #18CR-32242-01). Using convenience sampling, we emailed dermatologists from the Association of Professors of Dermatology and dermatology state societies in all 50 states inviting them to participate in our voluntary and anonymous survey, which was administered using SurveyMonkey. The use of all society mailing lists was approved by the respective owners. The 15-question survey included multiple choice, Likert scale, and free response sections. Summary and descriptive statistics were used to describe respondent demographics and identify any patterns in responses.

For each Likert-based question, participants ranked their degree of agreement with a statement as: 1=strongly disagree, 2=somewhat disagree, 3=neither agree nor disagree/neutral, 4=somewhat agree, and 5=strongly agree. The mean response and standard deviation were reported for each Likert scale prompt. Preplanned 1-sample t testing was used to analyze Likert scale data, in which the mean response for each prompt was compared to a baseline response of 3 (neutral). A P value <.05 was considered statistically significant. Statistical analyses were performed using SPSS Statistics for MacOS, version 27 (IBM).

Seventy-eight dermatologists agreed to participate, and 70 completed the survey, for a response rate of 89.7% (Table 1). The dermatologists we surveyed practiced in a variety of clinical settings, including academic public hospitals (46.2% [36/78]), academic private hospitals (33.3% [26/78]), and private practices (32.1% [25/78]), and 60.3% (47/78) reported providing disability documentation at some point. Most of the respondents (64.3% [45/70]) did not perform assessments in an average month (Table 2). Medical assessment documentation was provided most frequently for workers’ compensation (50.0% [35/70]), private insurance (27.1% [19/70]), and Social Security Disability Insurance (25.7% [18/70]). Dermatologists overwhelmingly reported no formal training for disability assessment in medical school (94.3% [66/70]), residency (97.1% [68/70]), or clinical practice (81.4% [57/70]).

In the Likert scale prompts, respondents agreed that they were uncertain of their role in disability assessment (mean response, 3.6; P<.001). Moreover, they were uncomfortable providing assessments (mean response, 3.5; P<.001) and felt that they did not have sufficient time to perform them (mean response, 3.6; P<.001). Dermatologists disagreed that they received adequate compensation for performing assessments (mean response, 2.2; P<.001) and felt that they did not have enough time to participate in assessments (mean response, 3.6; P<.001). Respondents generally did not feel distrustful of patients seeking disability assessment (mean response, 2.8; P=.043). Dermatologists neither agreed nor disagreed when asked if they thought that physicians can determine disability status (mean response, 3.2; P=.118). The details of the Likert scale responses are described in Table 3. Respondents also were uncertain as to which dermatologic conditions were eligible for disability. When asked to select which conditions from a list of 10 were eligible per the Social Security Administration listing of disability impairments, only 15.4% (12/70) of respondents correctly identified that all the conditions qualified; these included ichthyosis, pemphigus vulgaris, allergic contact dermatitis, hidradenitis suppurativa, systemic lupus erythematosus, chromoblastomycosis, xeroderma pigmentosum, burns, malignant melanoma, and scleroderma.⁶

In the free-response prompts, respondents frequently described extensive paperwork, inadequate time, and lack of reimbursement as barriers to providing documentation. Often, dermatologists found that the forms were not well matched to the skin conditions they were evaluating and rather had a musculoskeletal focus. Multiple individuals commented on the challenge in assessing the percentage of disability and functional/psychosocial impairment in skin conditions. One respondent noted that workers’ compensation forms ask if the patient is “…permanent and stationary…for most conditions this has no meaning in dermatology.” Some felt hesitant to provide documentation because they had insufficient patient history, especially regarding employment, and opted to defer to primary care providers who might be more familiar with the full patient history.

A dermatologist described their perspective as follows:

“…As a specialist I feel that I don’t have a complete look into all the factors that could contribute to a patient[’]s need to go on disability, and I don’t have experience with filling out disability requests. That being said, if a patient[’]s request for disability was due to a skin disease that I know way more about than [a] primary care [physician] would, I would do the disability assessment.”

Another respondent noted the complexity in “establishing causality” for workers’ compensation. Another dermatologist reported,

“The most frequent challenging situation I encounter is being asked to evaluate for maximum medical improvement after patch testing. If the patient is not fully avoiding contact allergens either at home or at work, then I typically document that they are not at [maximum medical improvement]. The reality is that most frequently it is due to exposure to allergens at home so the line between what is a legitimate worker’s comp[ensation] issue and what is a home life choice is blurry.”

Nevertheless, respondents expressed interest in learning more about disability assessment procedures. Summary guides, lectures, and prefilled paperwork were the most popular initiatives that respondents agreed would be beneficial toward becoming educated regarding disability assessment (78.6%, 58.6%, and 58.6%, respectively)(Table 2). One respondent noted that “previous [internal medicine] history help[ed]” them in performing cutaneous disability assessments.

As with any survey, our study did have some inherent limitations. Only a relatively small sample size was willing to complete the survey. There was a predominance of respondents from California (34.6% [27/78]), as well as those practicing for less than 15 years (58.9% [46/78])(Figure). This could limit generalizability to the national population of dermatologists. In addition, there was potential for recall bias and errors in responding given the self-reported nature of the study. Different individuals may interpret the Likert scale options in various ways, which could skew results unintentionally. However, the survey was largely qualitative in nature, making it a legitimate tool for answering our research questions. Moreover, we were able to hear the perspectives of dermatologists across diverse practice settings, with free response prompts to increase the depth of the survey.

Almost 50 years later, our survey echoes common themes from Adams’ 1976 survey.⁴ Inadequate compensation, limited time, and burdensome paperwork all continue to hinder dermatologists’ ability to perform disability assessments. Our participants frequently commented that the current disability forms are not congruent with the nature of skin conditions, making it challenging to accurately document the facts.

Moreover, respondents felt uncertain in their role in disability assessment and occasionally noted distrust of patients or insufficient patient history as barriers to completing assessments. They also were unsure if physicians can grant disability status. This is a common misconception among physicians that leads to discomfort in helping with disability assessment.⁷ The role of physicians in disability assessment is to document the facts of a patient’s illness, not to determine whether they are eligible for benefits. We discovered uncertainty in our respondents’ ability to identify conditions eligible for disability, highlighting an area in need of greater education for physicians.

Despite these obstacles, respondents were interested in learning more about disability assessment and highlighted several practical approaches that could help them better perform this task. As skin specialists, dermatologists are the best-equipped physicians to assess cutaneous conditions and should play a greater role in performing disability assessments, which could be achieved through increased educational initiatives and individual physician motivation.⁷ We call for greater collaboration and reflection on the importance of disability assistance among dermatologists to increase participation in the disability-assessment process.

To the Editor:

Cutaneous medical conditions can have a substantial impact on patients’ functioning and quality of life. Many patients with severe skin disease are eligible to receive disability assistance that can provide them with essential income and health care. Previous research has highlighted disability assessment as one of the most important ways physicians can help mitigate the health consequences of poverty.¹ Dermatologists can play an important role in the disability assessment process by documenting the facts associated with patients’ skin conditions.

Although skin conditions have a relatively high prevalence, they remain underrepresented in disability claims. Between 1997 and 2004, occupational skin diseases accounted for 12% to 17% of nonfatal work-related illnesses; however, during that same period, skin conditions comprised only 0.21% of disability claims in the United States.^2,3 Historically, there has been hesitancy among dermatologists to complete disability paperwork; a 1976 survey of dermatologists cited extensive paperwork, “troublesome patients,” and fee schedule issues as reasons.⁴ The lack of training regarding disability assessment in medical school and residency also has been noted.⁵

To characterize modern attitudes toward disability assessments, we conducted a survey of dermatologists across the United States. Our study was reviewed and declared exempt by the institutional review board of the Lundquist Institute for Biomedical Innovation at Harbor-UCLA Medical Center (Torrance, California)(approval #18CR-32242-01). Using convenience sampling, we emailed dermatologists from the Association of Professors of Dermatology and dermatology state societies in all 50 states inviting them to participate in our voluntary and anonymous survey, which was administered using SurveyMonkey. The use of all society mailing lists was approved by the respective owners. The 15-question survey included multiple choice, Likert scale, and free response sections. Summary and descriptive statistics were used to describe respondent demographics and identify any patterns in responses.

For each Likert-based question, participants ranked their degree of agreement with a statement as: 1=strongly disagree, 2=somewhat disagree, 3=neither agree nor disagree/neutral, 4=somewhat agree, and 5=strongly agree. The mean response and standard deviation were reported for each Likert scale prompt. Preplanned 1-sample t testing was used to analyze Likert scale data, in which the mean response for each prompt was compared to a baseline response of 3 (neutral). A P value <.05 was considered statistically significant. Statistical analyses were performed using SPSS Statistics for MacOS, version 27 (IBM).

Seventy-eight dermatologists agreed to participate, and 70 completed the survey, for a response rate of 89.7% (Table 1). The dermatologists we surveyed practiced in a variety of clinical settings, including academic public hospitals (46.2% [36/78]), academic private hospitals (33.3% [26/78]), and private practices (32.1% [25/78]), and 60.3% (47/78) reported providing disability documentation at some point. Most of the respondents (64.3% [45/70]) did not perform assessments in an average month (Table 2). Medical assessment documentation was provided most frequently for workers’ compensation (50.0% [35/70]), private insurance (27.1% [19/70]), and Social Security Disability Insurance (25.7% [18/70]). Dermatologists overwhelmingly reported no formal training for disability assessment in medical school (94.3% [66/70]), residency (97.1% [68/70]), or clinical practice (81.4% [57/70]).

In the Likert scale prompts, respondents agreed that they were uncertain of their role in disability assessment (mean response, 3.6; P<.001). Moreover, they were uncomfortable providing assessments (mean response, 3.5; P<.001) and felt that they did not have sufficient time to perform them (mean response, 3.6; P<.001). Dermatologists disagreed that they received adequate compensation for performing assessments (mean response, 2.2; P<.001) and felt that they did not have enough time to participate in assessments (mean response, 3.6; P<.001). Respondents generally did not feel distrustful of patients seeking disability assessment (mean response, 2.8; P=.043). Dermatologists neither agreed nor disagreed when asked if they thought that physicians can determine disability status (mean response, 3.2; P=.118). The details of the Likert scale responses are described in Table 3. Respondents also were uncertain as to which dermatologic conditions were eligible for disability. When asked to select which conditions from a list of 10 were eligible per the Social Security Administration listing of disability impairments, only 15.4% (12/70) of respondents correctly identified that all the conditions qualified; these included ichthyosis, pemphigus vulgaris, allergic contact dermatitis, hidradenitis suppurativa, systemic lupus erythematosus, chromoblastomycosis, xeroderma pigmentosum, burns, malignant melanoma, and scleroderma.⁶

In the free-response prompts, respondents frequently described extensive paperwork, inadequate time, and lack of reimbursement as barriers to providing documentation. Often, dermatologists found that the forms were not well matched to the skin conditions they were evaluating and rather had a musculoskeletal focus. Multiple individuals commented on the challenge in assessing the percentage of disability and functional/psychosocial impairment in skin conditions. One respondent noted that workers’ compensation forms ask if the patient is “…permanent and stationary…for most conditions this has no meaning in dermatology.” Some felt hesitant to provide documentation because they had insufficient patient history, especially regarding employment, and opted to defer to primary care providers who might be more familiar with the full patient history.

A dermatologist described their perspective as follows:

“…As a specialist I feel that I don’t have a complete look into all the factors that could contribute to a patient[’]s need to go on disability, and I don’t have experience with filling out disability requests. That being said, if a patient[’]s request for disability was due to a skin disease that I know way more about than [a] primary care [physician] would, I would do the disability assessment.”

Another respondent noted the complexity in “establishing causality” for workers’ compensation. Another dermatologist reported,

“The most frequent challenging situation I encounter is being asked to evaluate for maximum medical improvement after patch testing. If the patient is not fully avoiding contact allergens either at home or at work, then I typically document that they are not at [maximum medical improvement]. The reality is that most frequently it is due to exposure to allergens at home so the line between what is a legitimate worker’s comp[ensation] issue and what is a home life choice is blurry.”

Nevertheless, respondents expressed interest in learning more about disability assessment procedures. Summary guides, lectures, and prefilled paperwork were the most popular initiatives that respondents agreed would be beneficial toward becoming educated regarding disability assessment (78.6%, 58.6%, and 58.6%, respectively)(Table 2). One respondent noted that “previous [internal medicine] history help[ed]” them in performing cutaneous disability assessments.

As with any survey, our study did have some inherent limitations. Only a relatively small sample size was willing to complete the survey. There was a predominance of respondents from California (34.6% [27/78]), as well as those practicing for less than 15 years (58.9% [46/78])(Figure). This could limit generalizability to the national population of dermatologists. In addition, there was potential for recall bias and errors in responding given the self-reported nature of the study. Different individuals may interpret the Likert scale options in various ways, which could skew results unintentionally. However, the survey was largely qualitative in nature, making it a legitimate tool for answering our research questions. Moreover, we were able to hear the perspectives of dermatologists across diverse practice settings, with free response prompts to increase the depth of the survey.

Almost 50 years later, our survey echoes common themes from Adams’ 1976 survey.⁴ Inadequate compensation, limited time, and burdensome paperwork all continue to hinder dermatologists’ ability to perform disability assessments. Our participants frequently commented that the current disability forms are not congruent with the nature of skin conditions, making it challenging to accurately document the facts.

Moreover, respondents felt uncertain in their role in disability assessment and occasionally noted distrust of patients or insufficient patient history as barriers to completing assessments. They also were unsure if physicians can grant disability status. This is a common misconception among physicians that leads to discomfort in helping with disability assessment.⁷ The role of physicians in disability assessment is to document the facts of a patient’s illness, not to determine whether they are eligible for benefits. We discovered uncertainty in our respondents’ ability to identify conditions eligible for disability, highlighting an area in need of greater education for physicians.

Despite these obstacles, respondents were interested in learning more about disability assessment and highlighted several practical approaches that could help them better perform this task. As skin specialists, dermatologists are the best-equipped physicians to assess cutaneous conditions and should play a greater role in performing disability assessments, which could be achieved through increased educational initiatives and individual physician motivation.⁷ We call for greater collaboration and reflection on the importance of disability assistance among dermatologists to increase participation in the disability-assessment process.

Humans are inherently prone to errors. The extent and consequences of medical errors were documented in the 2000 publication of To Err is Human: Building a Safer Health System.¹ Published research on medical errors in dermatology has emphasized the heuristic issues involved in diagnosis,^2-6 essentially approaching the “why?” and “how?” of such errors. By contrast, the current study aimed to elucidate the “what?”—what are the dermatologic conditions most prone to diagnostic and/or management errors? One study published in 1987 approached this question by analyzing patterns of errors for dermatologic conditions in patients referred for specialty care by primary care physicians.⁷ The current study aimed to update and expand on the findings of this 1987 report by comparing more recent data on the errors made by providers and patients regarding skin conditions.

Methods

Data were collected prospectively from March 18, 2021, through July 25, 2023. Prospective data were obtained by recording the nature of errors noted for all patients seen by a board-certified dermatologist (R.J.P.) during routine outpatient practice in Norfolk, Virginia. This practice is limited to medical dermatology and accepts patients of any age from any referral source, with or without medical insurance. Retrospective data were obtained by review of electronic medical records for all patients seen by the same board-certified dermatologist from June 5, 2020, through March 12, 2021, who previously had been seen by an outside provider or were self-referred. In this study, the term diagnosis is used to describe providers’ explicit or imputed conclusions as to the nature of a dermatosis, and the term interpretation is used to describe patients' conclusions about their own condition. For this study, the patients’ self-made interpretations of their dermatoses were deemed to be correct when they agreed with those made by the dermatologist using standard clinicopathologic criteria supplemented by rapid bedside diagnostic techniques, as detailed in the 1987 study.⁷

Cases in which diagnostic or therapeutic errors were noted were entered into a spreadsheet that excluded patients’ names or other identifiers. For each noted case of diagnostic or therapeutic error, the following data were entered: patient’s age and sex; the name of the incorrect diagnosis, interpretation, or treatment; and the name of the correct (missed) diagnosis, along with the source of the error (provider or patient). Provider diagnoses were determined from medical records or patient statements or were imputed from the generally accepted indications for prescribed treatments. A provider was deemed to be any practitioner with prescriptive authority. Patients’ interpretations of their conditions were determined by patient statements or were imputed based on the indications for treatments being used. A treatment error was recorded when a diagnosis or interpretation was deemed to be correct, but treatment was deemed to be inappropriate. The same dermatologist (R.J.P) made all determinations as to the nature of the errors and their source.

Diagnostic errors were determined in several situations: (1) if the interpretation made by the patient of their dermatosis differed from the correct diagnosis in the absence of any additional diagnostic documentation, the correct diagnosis was scored as a missed diagnosis and the incorrect interpretation was scored as such; (2) if the provider’s diagnosis in the patient’s medical record differed from the correct diagnosis, both the correct (missed) and incorrect diagnoses were recorded; and (3) if the indication(s) of the medication(s) prescribed by the provider or used by the patient for their condition differed from the correct diagnosis, an imputed diagnosis based on this indication was scored as the incorrect diagnosis and the correct (missed) diagnosis was recorded; for example, an error would be entered into the spreadsheet for a patient using terbinafine cream for what was actually psoriasis. For a medication with multiple active agents, an error would be entered into the spreadsheet only if none of its indications matched the correct diagnosis; for example, if the patient had been prescribed a betamethasone/clotrimazole product, no error would be scored if the correct diagnosis was a steroid-responsive dermatosis, dermatophytosis, candidiasis, or tinea versicolor. For a single medication with multiple indications, no error would be recorded if the correct diagnosis was any of these indications; for example, in a patient who had been prescribed topical ketoconazole, no error would be scored if the correct diagnosis was dermatophytosis, candidiasis, tinea versicolor, or seborrheic dermatitis. Additionally, no error would be recorded if the correct diagnosis was uncertain at the time of initial patient evaluation or during chart review.

Standard spreadsheet functions and the pandas package⁸ from the Python programming language⁹ were used to extract relevant data from the spreadsheet (Tables 1-4).

Results

A total of 446 patient visits (182 males, 264 females) were included in the study, in which a total of 486 errors were found in the combined prospective and retrospective portions of the study. These errors involved 1.4% of all patient visits for the study period—specifically, all in routine practice as well as all patient records retrospectively reviewed. The age of the patients ranged from 4 to 95 years; the mean age was 51.5 years for males and 50.8 years for females.

The study results are outlined in Tables 1 through 4. To minimize the amount of data provided with no appreciable effect on the results, cases in which an incorrect or missed diagnosis/interpretation occurred only once (ie, unique case errors) were excluded from the tables. Tables 1 and 2 indicate the numbers and types of incorrect and missed diagnoses.

In the combined patient and provider cases, there were 434 instances in which provider diagnoses and patient interpretations were incorrect, 320 (73.7%) of which involved infectious disorders. By contrast, of the 413 instances of provider and patient missed diagnoses 289 (70.0%) were inflammatory dermatoses. The pattern was similar for patients’ incorrect interpretations compared to the incorrect diagnoses of the medical providers. Patients incorrectly interpreted their dermatoses as infectious in 79.5% (101/127) of cases. Similarly, providers incorrectly diagnosed their patients’ dermatoses as infectious in 75.4% (211/280) of cases (Table 3). For patients’ missed diagnoses, 70.7% (82/116) involved inflammatory dermatoses. For providers’ missed diagnoses, 63.9% (179/280) involved inflammatory dermatoses (Table 4).

Treatment errors in the context of correct diagnoses were uncommon. Fifteen (3.4%) such cases were noted in the 446 error-containing patient visits. In 4 (26.7%) of the 15 cases, potent topical corticosteroids were used long term on inappropriate cutaneous sites (eg, genital, facial, or intertriginous areas). Another 4 (26.7%) cases involved fungal infections: nystatin used for tinea versicolor in 1 case and for dermatophytosis in another, widespread dermatophytosis treated topically, and use of a nonindicated topical antifungal for onychomycosis. Other examples involved inadequate dosing of systemic corticosteroids for extensive acute contact dermatitis, psoriasis treated with systemic corticosteroids, inadequate dosing of medication for seborrheic dermatitis, and treatment with valacyclovir based solely on serologic testing.

Comment

The results of our study indicate that errors in management of cutaneous disorders are overwhelmingly diagnostic in nature, while treatment errors appear to be unusual when the correct diagnosis is made. Both the current study and the 1987 study indicated a notable tendency of providers to incorrectly diagnose infectious disorders and to miss the diagnosis of inflammatory dermatoses.⁷ The current study extends this finding to include patients’ interpretive errors. 

It is notable that many of the incorrect and missed diagnoses can be confirmed or ruled out by rapid bedside techniques, namely potassium hydroxide (KOH) preparation for dermatophytes, candidiasis, and tinea versicolor; wet preparation for scabies and pediculosis; Tzanck preparation for herpes simplex and herpes zoster; and crush preparation for molluscum contagiosum. Notably, 57.8% (281/486) of cases in which error was noted involved disorders for which the use of one of these bedside diagnostic tests could have correctly established a diagnosis or ruled out an incorrect one; thus in an ideal world in which these tests were performed perfectly in all appropriate cases, more than half of the errors detected in this study could have been avoided. Dermatophytosis was involved in 35.8% (174/486) of the error-containing patient encounters in this study; therefore, if only the KOH preparation is considered, more than one-third of all errors documented in this study could have been avoided. Unfortunately, surveys have suggested that among dermatologists in the United States and some other countries, KOH preparations are used infrequently.^10-12

Certain limitations were inherent to this study. The data were derived from a single dermatology practice by one physician in one geographic region over a short period of time. These factors may limit the generalizability of the results. Although the goal was to identify all errors made for the patients seen, some errors likely were missed due to incomplete patient history or inaccurate medication listings. There is no absolute way to determine if the diagnoses or the treatments deemed correct by the dermatologist were, in fact, correct. For cases in which a patient’s interpretation or a provider’s diagnosis was imputed from the indication(s) associated with the medication(s) being used, one cannot exclude the possibility that a medication was used appropriately for a nonlabeled or nonstandard indication. The designation of treatment errors may be subject to different interpretations by different clinicians. Despite these limitations, it is likely that the results of this study can be extrapolated to reasonably similar dermatology practices. The apparently persistent and consistent tendency of clinicians to incorrectly diagnose infectious dermatoses and to miss inflammatory conditions has implications for teaching of medical dermatology in the academic and clinical settings. In particular, given that dermatophytosis is the diagnosis involved in the highest number of errors, special emphasis should be placed on this infection in clinician education.

Acknowledgement—The authors would like to acknowledge the essential contributions to this study by Urvi Jain (Virginia Beach, Virginia), particularly for analysis and interpretation of data and for suggestions to improve the manuscript.

Humans are inherently prone to errors. The extent and consequences of medical errors were documented in the 2000 publication of To Err is Human: Building a Safer Health System.¹ Published research on medical errors in dermatology has emphasized the heuristic issues involved in diagnosis,^2-6 essentially approaching the “why?” and “how?” of such errors. By contrast, the current study aimed to elucidate the “what?”—what are the dermatologic conditions most prone to diagnostic and/or management errors? One study published in 1987 approached this question by analyzing patterns of errors for dermatologic conditions in patients referred for specialty care by primary care physicians.⁷ The current study aimed to update and expand on the findings of this 1987 report by comparing more recent data on the errors made by providers and patients regarding skin conditions.

Methods

Data were collected prospectively from March 18, 2021, through July 25, 2023. Prospective data were obtained by recording the nature of errors noted for all patients seen by a board-certified dermatologist (R.J.P.) during routine outpatient practice in Norfolk, Virginia. This practice is limited to medical dermatology and accepts patients of any age from any referral source, with or without medical insurance. Retrospective data were obtained by review of electronic medical records for all patients seen by the same board-certified dermatologist from June 5, 2020, through March 12, 2021, who previously had been seen by an outside provider or were self-referred. In this study, the term diagnosis is used to describe providers’ explicit or imputed conclusions as to the nature of a dermatosis, and the term interpretation is used to describe patients' conclusions about their own condition. For this study, the patients’ self-made interpretations of their dermatoses were deemed to be correct when they agreed with those made by the dermatologist using standard clinicopathologic criteria supplemented by rapid bedside diagnostic techniques, as detailed in the 1987 study.⁷

Cases in which diagnostic or therapeutic errors were noted were entered into a spreadsheet that excluded patients’ names or other identifiers. For each noted case of diagnostic or therapeutic error, the following data were entered: patient’s age and sex; the name of the incorrect diagnosis, interpretation, or treatment; and the name of the correct (missed) diagnosis, along with the source of the error (provider or patient). Provider diagnoses were determined from medical records or patient statements or were imputed from the generally accepted indications for prescribed treatments. A provider was deemed to be any practitioner with prescriptive authority. Patients’ interpretations of their conditions were determined by patient statements or were imputed based on the indications for treatments being used. A treatment error was recorded when a diagnosis or interpretation was deemed to be correct, but treatment was deemed to be inappropriate. The same dermatologist (R.J.P) made all determinations as to the nature of the errors and their source.

Diagnostic errors were determined in several situations: (1) if the interpretation made by the patient of their dermatosis differed from the correct diagnosis in the absence of any additional diagnostic documentation, the correct diagnosis was scored as a missed diagnosis and the incorrect interpretation was scored as such; (2) if the provider’s diagnosis in the patient’s medical record differed from the correct diagnosis, both the correct (missed) and incorrect diagnoses were recorded; and (3) if the indication(s) of the medication(s) prescribed by the provider or used by the patient for their condition differed from the correct diagnosis, an imputed diagnosis based on this indication was scored as the incorrect diagnosis and the correct (missed) diagnosis was recorded; for example, an error would be entered into the spreadsheet for a patient using terbinafine cream for what was actually psoriasis. For a medication with multiple active agents, an error would be entered into the spreadsheet only if none of its indications matched the correct diagnosis; for example, if the patient had been prescribed a betamethasone/clotrimazole product, no error would be scored if the correct diagnosis was a steroid-responsive dermatosis, dermatophytosis, candidiasis, or tinea versicolor. For a single medication with multiple indications, no error would be recorded if the correct diagnosis was any of these indications; for example, in a patient who had been prescribed topical ketoconazole, no error would be scored if the correct diagnosis was dermatophytosis, candidiasis, tinea versicolor, or seborrheic dermatitis. Additionally, no error would be recorded if the correct diagnosis was uncertain at the time of initial patient evaluation or during chart review.

Standard spreadsheet functions and the pandas package⁸ from the Python programming language⁹ were used to extract relevant data from the spreadsheet (Tables 1-4).

Results

A total of 446 patient visits (182 males, 264 females) were included in the study, in which a total of 486 errors were found in the combined prospective and retrospective portions of the study. These errors involved 1.4% of all patient visits for the study period—specifically, all in routine practice as well as all patient records retrospectively reviewed. The age of the patients ranged from 4 to 95 years; the mean age was 51.5 years for males and 50.8 years for females.

The study results are outlined in Tables 1 through 4. To minimize the amount of data provided with no appreciable effect on the results, cases in which an incorrect or missed diagnosis/interpretation occurred only once (ie, unique case errors) were excluded from the tables. Tables 1 and 2 indicate the numbers and types of incorrect and missed diagnoses.

In the combined patient and provider cases, there were 434 instances in which provider diagnoses and patient interpretations were incorrect, 320 (73.7%) of which involved infectious disorders. By contrast, of the 413 instances of provider and patient missed diagnoses 289 (70.0%) were inflammatory dermatoses. The pattern was similar for patients’ incorrect interpretations compared to the incorrect diagnoses of the medical providers. Patients incorrectly interpreted their dermatoses as infectious in 79.5% (101/127) of cases. Similarly, providers incorrectly diagnosed their patients’ dermatoses as infectious in 75.4% (211/280) of cases (Table 3). For patients’ missed diagnoses, 70.7% (82/116) involved inflammatory dermatoses. For providers’ missed diagnoses, 63.9% (179/280) involved inflammatory dermatoses (Table 4).

Treatment errors in the context of correct diagnoses were uncommon. Fifteen (3.4%) such cases were noted in the 446 error-containing patient visits. In 4 (26.7%) of the 15 cases, potent topical corticosteroids were used long term on inappropriate cutaneous sites (eg, genital, facial, or intertriginous areas). Another 4 (26.7%) cases involved fungal infections: nystatin used for tinea versicolor in 1 case and for dermatophytosis in another, widespread dermatophytosis treated topically, and use of a nonindicated topical antifungal for onychomycosis. Other examples involved inadequate dosing of systemic corticosteroids for extensive acute contact dermatitis, psoriasis treated with systemic corticosteroids, inadequate dosing of medication for seborrheic dermatitis, and treatment with valacyclovir based solely on serologic testing.

Comment

The results of our study indicate that errors in management of cutaneous disorders are overwhelmingly diagnostic in nature, while treatment errors appear to be unusual when the correct diagnosis is made. Both the current study and the 1987 study indicated a notable tendency of providers to incorrectly diagnose infectious disorders and to miss the diagnosis of inflammatory dermatoses.⁷ The current study extends this finding to include patients’ interpretive errors. 

It is notable that many of the incorrect and missed diagnoses can be confirmed or ruled out by rapid bedside techniques, namely potassium hydroxide (KOH) preparation for dermatophytes, candidiasis, and tinea versicolor; wet preparation for scabies and pediculosis; Tzanck preparation for herpes simplex and herpes zoster; and crush preparation for molluscum contagiosum. Notably, 57.8% (281/486) of cases in which error was noted involved disorders for which the use of one of these bedside diagnostic tests could have correctly established a diagnosis or ruled out an incorrect one; thus in an ideal world in which these tests were performed perfectly in all appropriate cases, more than half of the errors detected in this study could have been avoided. Dermatophytosis was involved in 35.8% (174/486) of the error-containing patient encounters in this study; therefore, if only the KOH preparation is considered, more than one-third of all errors documented in this study could have been avoided. Unfortunately, surveys have suggested that among dermatologists in the United States and some other countries, KOH preparations are used infrequently.^10-12

Certain limitations were inherent to this study. The data were derived from a single dermatology practice by one physician in one geographic region over a short period of time. These factors may limit the generalizability of the results. Although the goal was to identify all errors made for the patients seen, some errors likely were missed due to incomplete patient history or inaccurate medication listings. There is no absolute way to determine if the diagnoses or the treatments deemed correct by the dermatologist were, in fact, correct. For cases in which a patient’s interpretation or a provider’s diagnosis was imputed from the indication(s) associated with the medication(s) being used, one cannot exclude the possibility that a medication was used appropriately for a nonlabeled or nonstandard indication. The designation of treatment errors may be subject to different interpretations by different clinicians. Despite these limitations, it is likely that the results of this study can be extrapolated to reasonably similar dermatology practices. The apparently persistent and consistent tendency of clinicians to incorrectly diagnose infectious dermatoses and to miss inflammatory conditions has implications for teaching of medical dermatology in the academic and clinical settings. In particular, given that dermatophytosis is the diagnosis involved in the highest number of errors, special emphasis should be placed on this infection in clinician education.

Acknowledgement—The authors would like to acknowledge the essential contributions to this study by Urvi Jain (Virginia Beach, Virginia), particularly for analysis and interpretation of data and for suggestions to improve the manuscript.

Humans are inherently prone to errors. The extent and consequences of medical errors were documented in the 2000 publication of To Err is Human: Building a Safer Health System.¹ Published research on medical errors in dermatology has emphasized the heuristic issues involved in diagnosis,^2-6 essentially approaching the “why?” and “how?” of such errors. By contrast, the current study aimed to elucidate the “what?”—what are the dermatologic conditions most prone to diagnostic and/or management errors? One study published in 1987 approached this question by analyzing patterns of errors for dermatologic conditions in patients referred for specialty care by primary care physicians.⁷ The current study aimed to update and expand on the findings of this 1987 report by comparing more recent data on the errors made by providers and patients regarding skin conditions.

Methods

Data were collected prospectively from March 18, 2021, through July 25, 2023. Prospective data were obtained by recording the nature of errors noted for all patients seen by a board-certified dermatologist (R.J.P.) during routine outpatient practice in Norfolk, Virginia. This practice is limited to medical dermatology and accepts patients of any age from any referral source, with or without medical insurance. Retrospective data were obtained by review of electronic medical records for all patients seen by the same board-certified dermatologist from June 5, 2020, through March 12, 2021, who previously had been seen by an outside provider or were self-referred. In this study, the term diagnosis is used to describe providers’ explicit or imputed conclusions as to the nature of a dermatosis, and the term interpretation is used to describe patients' conclusions about their own condition. For this study, the patients’ self-made interpretations of their dermatoses were deemed to be correct when they agreed with those made by the dermatologist using standard clinicopathologic criteria supplemented by rapid bedside diagnostic techniques, as detailed in the 1987 study.⁷

Cases in which diagnostic or therapeutic errors were noted were entered into a spreadsheet that excluded patients’ names or other identifiers. For each noted case of diagnostic or therapeutic error, the following data were entered: patient’s age and sex; the name of the incorrect diagnosis, interpretation, or treatment; and the name of the correct (missed) diagnosis, along with the source of the error (provider or patient). Provider diagnoses were determined from medical records or patient statements or were imputed from the generally accepted indications for prescribed treatments. A provider was deemed to be any practitioner with prescriptive authority. Patients’ interpretations of their conditions were determined by patient statements or were imputed based on the indications for treatments being used. A treatment error was recorded when a diagnosis or interpretation was deemed to be correct, but treatment was deemed to be inappropriate. The same dermatologist (R.J.P) made all determinations as to the nature of the errors and their source.

Diagnostic errors were determined in several situations: (1) if the interpretation made by the patient of their dermatosis differed from the correct diagnosis in the absence of any additional diagnostic documentation, the correct diagnosis was scored as a missed diagnosis and the incorrect interpretation was scored as such; (2) if the provider’s diagnosis in the patient’s medical record differed from the correct diagnosis, both the correct (missed) and incorrect diagnoses were recorded; and (3) if the indication(s) of the medication(s) prescribed by the provider or used by the patient for their condition differed from the correct diagnosis, an imputed diagnosis based on this indication was scored as the incorrect diagnosis and the correct (missed) diagnosis was recorded; for example, an error would be entered into the spreadsheet for a patient using terbinafine cream for what was actually psoriasis. For a medication with multiple active agents, an error would be entered into the spreadsheet only if none of its indications matched the correct diagnosis; for example, if the patient had been prescribed a betamethasone/clotrimazole product, no error would be scored if the correct diagnosis was a steroid-responsive dermatosis, dermatophytosis, candidiasis, or tinea versicolor. For a single medication with multiple indications, no error would be recorded if the correct diagnosis was any of these indications; for example, in a patient who had been prescribed topical ketoconazole, no error would be scored if the correct diagnosis was dermatophytosis, candidiasis, tinea versicolor, or seborrheic dermatitis. Additionally, no error would be recorded if the correct diagnosis was uncertain at the time of initial patient evaluation or during chart review.

Standard spreadsheet functions and the pandas package⁸ from the Python programming language⁹ were used to extract relevant data from the spreadsheet (Tables 1-4).

Results

A total of 446 patient visits (182 males, 264 females) were included in the study, in which a total of 486 errors were found in the combined prospective and retrospective portions of the study. These errors involved 1.4% of all patient visits for the study period—specifically, all in routine practice as well as all patient records retrospectively reviewed. The age of the patients ranged from 4 to 95 years; the mean age was 51.5 years for males and 50.8 years for females.

The study results are outlined in Tables 1 through 4. To minimize the amount of data provided with no appreciable effect on the results, cases in which an incorrect or missed diagnosis/interpretation occurred only once (ie, unique case errors) were excluded from the tables. Tables 1 and 2 indicate the numbers and types of incorrect and missed diagnoses.

In the combined patient and provider cases, there were 434 instances in which provider diagnoses and patient interpretations were incorrect, 320 (73.7%) of which involved infectious disorders. By contrast, of the 413 instances of provider and patient missed diagnoses 289 (70.0%) were inflammatory dermatoses. The pattern was similar for patients’ incorrect interpretations compared to the incorrect diagnoses of the medical providers. Patients incorrectly interpreted their dermatoses as infectious in 79.5% (101/127) of cases. Similarly, providers incorrectly diagnosed their patients’ dermatoses as infectious in 75.4% (211/280) of cases (Table 3). For patients’ missed diagnoses, 70.7% (82/116) involved inflammatory dermatoses. For providers’ missed diagnoses, 63.9% (179/280) involved inflammatory dermatoses (Table 4).

Treatment errors in the context of correct diagnoses were uncommon. Fifteen (3.4%) such cases were noted in the 446 error-containing patient visits. In 4 (26.7%) of the 15 cases, potent topical corticosteroids were used long term on inappropriate cutaneous sites (eg, genital, facial, or intertriginous areas). Another 4 (26.7%) cases involved fungal infections: nystatin used for tinea versicolor in 1 case and for dermatophytosis in another, widespread dermatophytosis treated topically, and use of a nonindicated topical antifungal for onychomycosis. Other examples involved inadequate dosing of systemic corticosteroids for extensive acute contact dermatitis, psoriasis treated with systemic corticosteroids, inadequate dosing of medication for seborrheic dermatitis, and treatment with valacyclovir based solely on serologic testing.

Comment

The results of our study indicate that errors in management of cutaneous disorders are overwhelmingly diagnostic in nature, while treatment errors appear to be unusual when the correct diagnosis is made. Both the current study and the 1987 study indicated a notable tendency of providers to incorrectly diagnose infectious disorders and to miss the diagnosis of inflammatory dermatoses.⁷ The current study extends this finding to include patients’ interpretive errors. 

It is notable that many of the incorrect and missed diagnoses can be confirmed or ruled out by rapid bedside techniques, namely potassium hydroxide (KOH) preparation for dermatophytes, candidiasis, and tinea versicolor; wet preparation for scabies and pediculosis; Tzanck preparation for herpes simplex and herpes zoster; and crush preparation for molluscum contagiosum. Notably, 57.8% (281/486) of cases in which error was noted involved disorders for which the use of one of these bedside diagnostic tests could have correctly established a diagnosis or ruled out an incorrect one; thus in an ideal world in which these tests were performed perfectly in all appropriate cases, more than half of the errors detected in this study could have been avoided. Dermatophytosis was involved in 35.8% (174/486) of the error-containing patient encounters in this study; therefore, if only the KOH preparation is considered, more than one-third of all errors documented in this study could have been avoided. Unfortunately, surveys have suggested that among dermatologists in the United States and some other countries, KOH preparations are used infrequently.^10-12

Certain limitations were inherent to this study. The data were derived from a single dermatology practice by one physician in one geographic region over a short period of time. These factors may limit the generalizability of the results. Although the goal was to identify all errors made for the patients seen, some errors likely were missed due to incomplete patient history or inaccurate medication listings. There is no absolute way to determine if the diagnoses or the treatments deemed correct by the dermatologist were, in fact, correct. For cases in which a patient’s interpretation or a provider’s diagnosis was imputed from the indication(s) associated with the medication(s) being used, one cannot exclude the possibility that a medication was used appropriately for a nonlabeled or nonstandard indication. The designation of treatment errors may be subject to different interpretations by different clinicians. Despite these limitations, it is likely that the results of this study can be extrapolated to reasonably similar dermatology practices. The apparently persistent and consistent tendency of clinicians to incorrectly diagnose infectious dermatoses and to miss inflammatory conditions has implications for teaching of medical dermatology in the academic and clinical settings. In particular, given that dermatophytosis is the diagnosis involved in the highest number of errors, special emphasis should be placed on this infection in clinician education.

Acknowledgement—The authors would like to acknowledge the essential contributions to this study by Urvi Jain (Virginia Beach, Virginia), particularly for analysis and interpretation of data and for suggestions to improve the manuscript.

The US Department of Veterans Affairs (VA) delivers care to > 9 million veterans, including primary and specialty care.¹ While clinical duties remain important across the health system, proposed productivity models have included clinician research activity, given that many hold roles in academia.² Within this framework, research plays a pivotal role in advancing clinical practices and outcomes. Studies have found that physicians who participated in research report higher job satisfaction.³

As a specialty within the VA, podiatrists diagnose, treat, and prevent foot and ankle disorders. In addition to clinical practice, various scholarly activities are shared among these physicians.⁴ Reasons for scholarly pursuits among podiatrists vary, including participation in research for academic promotion or to establish expertise in a given area.^4-7 Although research remains a component associated with promotion within the VA, little is known about the scholarly activity of VA podiatrists. Specifically, there remains a paucity of data concerning their expertise, as evidenced through peer-reviewed publications, among these physicians and surgeons. To date, no analysis of scholarly activity among VA podiatrists has been conducted.

The primary aim of this investigation was to describe the scholarly productivity among podiatrists employed by the VA through an analysis of the number of peer-reviewed publications and the respective h-index of each physician. The secondary aim of this investigation was to assess the effect of academic productivity on compensation. This study describes research activities pursued by VA physicians and provides the veteran patient population with the confidence that their foot health care remains in the hands of experts within the field.

MATERIALS AND METHODS

The Feds Data Center (www.fedsdatacenter.com) online database of employees was used to identify VA podiatrists on June 17, 2024. All GS-15 physicians and their respective salaries in fiscal year 2023 were recorded. Administratively determined employees, including residents, were excluded. The h-index and number of published documents from any point during a physician’s training or career were reported for each podiatrist using Scopus; podiatrists without an h-index or publication were excluded. ⁸ Among podiatrists with scholarly activity, this analysis collected academic appointment, sex, and region of practice.

Statistical Analysis

Descriptive statistics, presented as counts and frequencies, were used. The median and IQR were used to describe the number of publications and h-index due to their nonnormal distribution. A Kruskal-Wallis test was used to compare median publication counts and h-index values among for junior faculty (JF), which includes instructors and assistant professors; senior faculty (SF), which includes associate professors and professors; and those with no academic affiliation (NF). Salary was reported as mean (SD) as it remained normally distributed and was compared using analysis of variance with posthoc Tukey test to increase statistical power. Additionally, this analysis used linear regression to investigate the relationship between scholarly activity and salary. The threshold for statistical significance was set at P < .05.

RESULTS

Among 819 VA podiatrists, 150 were administratively determined and excluded, and 512 were excluded for no history of publications, leaving 157 eligible for analysis (Table). A statistically significant difference was found in median (IQR) publication count by faculty appointment. JF had 6.0 (9.5), SF had 12.5 (22.3), and NF had 1.0 (2.0) publication(s) (P < .001) (Figure 1A). There was a statistically significant difference in h-index by faculty appointment. The median (IQR) h-index for JF was 2.0 (3.5), for SF was 5.5 (4.25), and for NF was 1.0 (2.0) (P = .002) (Figure 1B). Salary was not significantly associated with publication count (P = .20) or h-index (P = .62) (Figure 2). No statistically significant difference was found between academic appointment and mean (SD) salary. JF had a median (IQR) salary of $224,063 (27,989), SF of $234,260 (42,963), and NF of $219,811 (P = .35).

DISCUSSION

Focused on providing high-quality care, VA physicians use their expertise to practice comprehensive and specialized care.^9,10 A cornerstone to this expertise is scholarly activity that contributes to the body of knowledge and, ultimately, the evidence-based medicine by which these physicians practice.¹¹ With veterans considering VA care, it is important to highlight the commitment and dedication to the science and the practice of medicine. This analysis describes the scholarly activity of VA podiatrists and underscores the expertise veterans will receive for the diagnosis and treatment of their foot and ankle pathology.

were not part of an academic facility, a finding that may encourage further action to increase academic productivity in this specialty. For example, collaboration through academic affiliations has been seen throughout VA medical and surgical specialties and provides many benefits. Beginning with graduate medical education, the VA serves as a tremendous resource for resident training.¹² Additionally, veterans who sought emergency care at the VA had a lower risk of death than those treated at non-VA hospitals.¹³ In podiatric medicine and surgery, scholarly activity has been linked to improved outcomes, particularly in the study of ulceration development and its role in either prolonging or preventing amputation.¹⁴

Beyond improving clinical outcomes and patient care, engagement in research and inquiry offers other benefits. A cross-sectional study of 7734 physicians within the VA found that research involvement was associated with more favorable job characteristics and job satisfaction perceptions. ³ While this analysis found that about 19% of podiatrists have published once in their career, it remains likely that more may continue to engage in research during their VA tenure. Although this finding shows that an appreciable number of VA podiatrists have published in their field of study, it also encourages departments to provide resources to engage in research. Similar to previous research among foot and ankle surgeons, this analysis also found an increase in publications and h-index as tenure increased.⁴ Unlike previous research, which found h-index and academic appointment to be contributors to VA dermatologists’ salaries, no significant difference in salary was found in this study associated with publications, h-index, or academic role.¹⁵ Although the increase was not statistically significant, salary tended to rise as these variables increased.

Limitations

This analysis was confined to the most recent year of available data, which may not fully capture the longitudinal academic contributions and trends of individual podiatrists. Academic productivity can fluctuate significantly over time due to various factors, including changes in research focus and administrative responsibilities. The study also relied on Scopus to identify and quantify academic productivity. This database may not include all publications relevant to podiatrists, particularly those in niche or nonindexed journals. Additionally, name variations and potential misspellings could lead to missing data for individual podiatrists’ publications. Furthermore, this study did not account for other significant contributors to salary and career advancement within the federal system. Factors such as clinical performance, administrative duties, patient satisfaction, and contributions to teaching and mentoring are critical elements that also influence career progression and compensation but were not captured in this analysis. The retrospective design of this study inherently limits the ability to establish causal relationships. While associations between academic productivity and certain outcomes may be identified, it is not possible to definitively determine the direction or causality of these relationships. Future research may examine how scholarly activity continues once a clinician is part of VA.

CONCLUSIONS

This study highlights the significant academic contributions of VA podiatrists to research and the medical literature. By fostering an active research environment, the VA can ensure veterans receive the highest quality of care from knowledgeable and expert clinicians. Future research should aim to provide a more comprehensive analysis, capturing long-term trends and considering all factors influencing career advancement in VA.

The US Department of Veterans Affairs (VA) delivers care to > 9 million veterans, including primary and specialty care.¹ While clinical duties remain important across the health system, proposed productivity models have included clinician research activity, given that many hold roles in academia.² Within this framework, research plays a pivotal role in advancing clinical practices and outcomes. Studies have found that physicians who participated in research report higher job satisfaction.³

As a specialty within the VA, podiatrists diagnose, treat, and prevent foot and ankle disorders. In addition to clinical practice, various scholarly activities are shared among these physicians.⁴ Reasons for scholarly pursuits among podiatrists vary, including participation in research for academic promotion or to establish expertise in a given area.^4-7 Although research remains a component associated with promotion within the VA, little is known about the scholarly activity of VA podiatrists. Specifically, there remains a paucity of data concerning their expertise, as evidenced through peer-reviewed publications, among these physicians and surgeons. To date, no analysis of scholarly activity among VA podiatrists has been conducted.

The primary aim of this investigation was to describe the scholarly productivity among podiatrists employed by the VA through an analysis of the number of peer-reviewed publications and the respective h-index of each physician. The secondary aim of this investigation was to assess the effect of academic productivity on compensation. This study describes research activities pursued by VA physicians and provides the veteran patient population with the confidence that their foot health care remains in the hands of experts within the field.

MATERIALS AND METHODS

The Feds Data Center (www.fedsdatacenter.com) online database of employees was used to identify VA podiatrists on June 17, 2024. All GS-15 physicians and their respective salaries in fiscal year 2023 were recorded. Administratively determined employees, including residents, were excluded. The h-index and number of published documents from any point during a physician’s training or career were reported for each podiatrist using Scopus; podiatrists without an h-index or publication were excluded. ⁸ Among podiatrists with scholarly activity, this analysis collected academic appointment, sex, and region of practice.

Statistical Analysis

Descriptive statistics, presented as counts and frequencies, were used. The median and IQR were used to describe the number of publications and h-index due to their nonnormal distribution. A Kruskal-Wallis test was used to compare median publication counts and h-index values among for junior faculty (JF), which includes instructors and assistant professors; senior faculty (SF), which includes associate professors and professors; and those with no academic affiliation (NF). Salary was reported as mean (SD) as it remained normally distributed and was compared using analysis of variance with posthoc Tukey test to increase statistical power. Additionally, this analysis used linear regression to investigate the relationship between scholarly activity and salary. The threshold for statistical significance was set at P < .05.

RESULTS

Among 819 VA podiatrists, 150 were administratively determined and excluded, and 512 were excluded for no history of publications, leaving 157 eligible for analysis (Table). A statistically significant difference was found in median (IQR) publication count by faculty appointment. JF had 6.0 (9.5), SF had 12.5 (22.3), and NF had 1.0 (2.0) publication(s) (P < .001) (Figure 1A). There was a statistically significant difference in h-index by faculty appointment. The median (IQR) h-index for JF was 2.0 (3.5), for SF was 5.5 (4.25), and for NF was 1.0 (2.0) (P = .002) (Figure 1B). Salary was not significantly associated with publication count (P = .20) or h-index (P = .62) (Figure 2). No statistically significant difference was found between academic appointment and mean (SD) salary. JF had a median (IQR) salary of $224,063 (27,989), SF of $234,260 (42,963), and NF of $219,811 (P = .35).

DISCUSSION

Focused on providing high-quality care, VA physicians use their expertise to practice comprehensive and specialized care.^9,10 A cornerstone to this expertise is scholarly activity that contributes to the body of knowledge and, ultimately, the evidence-based medicine by which these physicians practice.¹¹ With veterans considering VA care, it is important to highlight the commitment and dedication to the science and the practice of medicine. This analysis describes the scholarly activity of VA podiatrists and underscores the expertise veterans will receive for the diagnosis and treatment of their foot and ankle pathology.

were not part of an academic facility, a finding that may encourage further action to increase academic productivity in this specialty. For example, collaboration through academic affiliations has been seen throughout VA medical and surgical specialties and provides many benefits. Beginning with graduate medical education, the VA serves as a tremendous resource for resident training.¹² Additionally, veterans who sought emergency care at the VA had a lower risk of death than those treated at non-VA hospitals.¹³ In podiatric medicine and surgery, scholarly activity has been linked to improved outcomes, particularly in the study of ulceration development and its role in either prolonging or preventing amputation.¹⁴

Beyond improving clinical outcomes and patient care, engagement in research and inquiry offers other benefits. A cross-sectional study of 7734 physicians within the VA found that research involvement was associated with more favorable job characteristics and job satisfaction perceptions. ³ While this analysis found that about 19% of podiatrists have published once in their career, it remains likely that more may continue to engage in research during their VA tenure. Although this finding shows that an appreciable number of VA podiatrists have published in their field of study, it also encourages departments to provide resources to engage in research. Similar to previous research among foot and ankle surgeons, this analysis also found an increase in publications and h-index as tenure increased.⁴ Unlike previous research, which found h-index and academic appointment to be contributors to VA dermatologists’ salaries, no significant difference in salary was found in this study associated with publications, h-index, or academic role.¹⁵ Although the increase was not statistically significant, salary tended to rise as these variables increased.

Limitations

This analysis was confined to the most recent year of available data, which may not fully capture the longitudinal academic contributions and trends of individual podiatrists. Academic productivity can fluctuate significantly over time due to various factors, including changes in research focus and administrative responsibilities. The study also relied on Scopus to identify and quantify academic productivity. This database may not include all publications relevant to podiatrists, particularly those in niche or nonindexed journals. Additionally, name variations and potential misspellings could lead to missing data for individual podiatrists’ publications. Furthermore, this study did not account for other significant contributors to salary and career advancement within the federal system. Factors such as clinical performance, administrative duties, patient satisfaction, and contributions to teaching and mentoring are critical elements that also influence career progression and compensation but were not captured in this analysis. The retrospective design of this study inherently limits the ability to establish causal relationships. While associations between academic productivity and certain outcomes may be identified, it is not possible to definitively determine the direction or causality of these relationships. Future research may examine how scholarly activity continues once a clinician is part of VA.

CONCLUSIONS

This study highlights the significant academic contributions of VA podiatrists to research and the medical literature. By fostering an active research environment, the VA can ensure veterans receive the highest quality of care from knowledgeable and expert clinicians. Future research should aim to provide a more comprehensive analysis, capturing long-term trends and considering all factors influencing career advancement in VA.

The US Department of Veterans Affairs (VA) delivers care to > 9 million veterans, including primary and specialty care.¹ While clinical duties remain important across the health system, proposed productivity models have included clinician research activity, given that many hold roles in academia.² Within this framework, research plays a pivotal role in advancing clinical practices and outcomes. Studies have found that physicians who participated in research report higher job satisfaction.³

As a specialty within the VA, podiatrists diagnose, treat, and prevent foot and ankle disorders. In addition to clinical practice, various scholarly activities are shared among these physicians.⁴ Reasons for scholarly pursuits among podiatrists vary, including participation in research for academic promotion or to establish expertise in a given area.^4-7 Although research remains a component associated with promotion within the VA, little is known about the scholarly activity of VA podiatrists. Specifically, there remains a paucity of data concerning their expertise, as evidenced through peer-reviewed publications, among these physicians and surgeons. To date, no analysis of scholarly activity among VA podiatrists has been conducted.

The primary aim of this investigation was to describe the scholarly productivity among podiatrists employed by the VA through an analysis of the number of peer-reviewed publications and the respective h-index of each physician. The secondary aim of this investigation was to assess the effect of academic productivity on compensation. This study describes research activities pursued by VA physicians and provides the veteran patient population with the confidence that their foot health care remains in the hands of experts within the field.

MATERIALS AND METHODS

The Feds Data Center (www.fedsdatacenter.com) online database of employees was used to identify VA podiatrists on June 17, 2024. All GS-15 physicians and their respective salaries in fiscal year 2023 were recorded. Administratively determined employees, including residents, were excluded. The h-index and number of published documents from any point during a physician’s training or career were reported for each podiatrist using Scopus; podiatrists without an h-index or publication were excluded. ⁸ Among podiatrists with scholarly activity, this analysis collected academic appointment, sex, and region of practice.

Statistical Analysis

Descriptive statistics, presented as counts and frequencies, were used. The median and IQR were used to describe the number of publications and h-index due to their nonnormal distribution. A Kruskal-Wallis test was used to compare median publication counts and h-index values among for junior faculty (JF), which includes instructors and assistant professors; senior faculty (SF), which includes associate professors and professors; and those with no academic affiliation (NF). Salary was reported as mean (SD) as it remained normally distributed and was compared using analysis of variance with posthoc Tukey test to increase statistical power. Additionally, this analysis used linear regression to investigate the relationship between scholarly activity and salary. The threshold for statistical significance was set at P < .05.

RESULTS

Among 819 VA podiatrists, 150 were administratively determined and excluded, and 512 were excluded for no history of publications, leaving 157 eligible for analysis (Table). A statistically significant difference was found in median (IQR) publication count by faculty appointment. JF had 6.0 (9.5), SF had 12.5 (22.3), and NF had 1.0 (2.0) publication(s) (P < .001) (Figure 1A). There was a statistically significant difference in h-index by faculty appointment. The median (IQR) h-index for JF was 2.0 (3.5), for SF was 5.5 (4.25), and for NF was 1.0 (2.0) (P = .002) (Figure 1B). Salary was not significantly associated with publication count (P = .20) or h-index (P = .62) (Figure 2). No statistically significant difference was found between academic appointment and mean (SD) salary. JF had a median (IQR) salary of $224,063 (27,989), SF of $234,260 (42,963), and NF of $219,811 (P = .35).

DISCUSSION

Focused on providing high-quality care, VA physicians use their expertise to practice comprehensive and specialized care.^9,10 A cornerstone to this expertise is scholarly activity that contributes to the body of knowledge and, ultimately, the evidence-based medicine by which these physicians practice.¹¹ With veterans considering VA care, it is important to highlight the commitment and dedication to the science and the practice of medicine. This analysis describes the scholarly activity of VA podiatrists and underscores the expertise veterans will receive for the diagnosis and treatment of their foot and ankle pathology.

were not part of an academic facility, a finding that may encourage further action to increase academic productivity in this specialty. For example, collaboration through academic affiliations has been seen throughout VA medical and surgical specialties and provides many benefits. Beginning with graduate medical education, the VA serves as a tremendous resource for resident training.¹² Additionally, veterans who sought emergency care at the VA had a lower risk of death than those treated at non-VA hospitals.¹³ In podiatric medicine and surgery, scholarly activity has been linked to improved outcomes, particularly in the study of ulceration development and its role in either prolonging or preventing amputation.¹⁴

Beyond improving clinical outcomes and patient care, engagement in research and inquiry offers other benefits. A cross-sectional study of 7734 physicians within the VA found that research involvement was associated with more favorable job characteristics and job satisfaction perceptions. ³ While this analysis found that about 19% of podiatrists have published once in their career, it remains likely that more may continue to engage in research during their VA tenure. Although this finding shows that an appreciable number of VA podiatrists have published in their field of study, it also encourages departments to provide resources to engage in research. Similar to previous research among foot and ankle surgeons, this analysis also found an increase in publications and h-index as tenure increased.⁴ Unlike previous research, which found h-index and academic appointment to be contributors to VA dermatologists’ salaries, no significant difference in salary was found in this study associated with publications, h-index, or academic role.¹⁵ Although the increase was not statistically significant, salary tended to rise as these variables increased.

Limitations

This analysis was confined to the most recent year of available data, which may not fully capture the longitudinal academic contributions and trends of individual podiatrists. Academic productivity can fluctuate significantly over time due to various factors, including changes in research focus and administrative responsibilities. The study also relied on Scopus to identify and quantify academic productivity. This database may not include all publications relevant to podiatrists, particularly those in niche or nonindexed journals. Additionally, name variations and potential misspellings could lead to missing data for individual podiatrists’ publications. Furthermore, this study did not account for other significant contributors to salary and career advancement within the federal system. Factors such as clinical performance, administrative duties, patient satisfaction, and contributions to teaching and mentoring are critical elements that also influence career progression and compensation but were not captured in this analysis. The retrospective design of this study inherently limits the ability to establish causal relationships. While associations between academic productivity and certain outcomes may be identified, it is not possible to definitively determine the direction or causality of these relationships. Future research may examine how scholarly activity continues once a clinician is part of VA.

CONCLUSIONS

This study highlights the significant academic contributions of VA podiatrists to research and the medical literature. By fostering an active research environment, the VA can ensure veterans receive the highest quality of care from knowledgeable and expert clinicians. Future research should aim to provide a more comprehensive analysis, capturing long-term trends and considering all factors influencing career advancement in VA.

To the Editor:

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition characterized by painful recurrent abscesses. Several autoimmune and endocrine diseases are associated with HS, including inflammatory bowel disease and diabetes mellitus (DM).¹ Notably, the association between HS and thyroid disorders is poorly characterized,² and there are no known nationwide studies exploring this potential association in the hospital setting. In this cross-sectional matched cohort study, we aimed to characterize HS patients with comorbid thyroid disorders as well as to explore whether thyroid disease is associated with comorbidities and hospital outcome measures in these patients.

The 2019 National Inpatient Sample (NIS) was weighted in accordance with NIS-assigned weight variables and queried for HS, hypothyroidism, and hyperthyroidism cases using International Classification of Diseases, Tenth Revision, codes L73.2, E03, and E05, respectively. Propensity score matching based on age and sex was performed using a nearest-neighbor method in the MatchIt statistical R package. Patient demographics, comorbidities, and outcome variables were collected. Univariable analysis of HS patients with thyroid disease vs those without thyroid disease vs controls without HS were performed using X² and t-test functions in SPSS statistical software (IBM). A series of multivariate analyses were performed using SPSS logistic and linear regression models to examine the effect of thyroid disease on hospital outcome measures and comorbidities in HS patients, with statistical significance set at P=.05.

A total of 1720 HS patients with comorbid thyroid disease (hyperthyroidism/hypothyroidism), 23,785 HS patients without thyroid disease, and 25,497 age- and sex-matched controls were included in the analysis. On average, HS patients with comorbid thyroid disease were older than HS patients without thyroid disease and controls (49.36 years vs 42.17 years vs 42.66 years [P<.001]), more likely to be female (75.58% vs 58.67% vs 59.81% [P<.001]), more likely to be in the highest income quartile (17.52% vs 12.18% vs 8.14% [P<.001]), and more likely to be Medicare insured (39.07% vs 27.47% vs 18.02% [P<.001])(eTable).

On univariate analysis of hospital outcome measures, HS patients with comorbid thyroid disease had the highest frequency of extreme likelihood of dying compared with HS patients without thyroid disease and with controls (6.40% vs 5.38% vs 2.47% [P<.001]), the highest mean number of diagnoses (18.31 vs 14.14 vs 8.57 [P<.001]), and the longest mean length of hospital stay (6.03 days vs 5.94 days vs 3.73 days [P<.001]). On univariate analysis of comorbidities, HS patients with thyroid disease had the highest incidence of the following comorbidities compared with HS patients without thyroid disease and controls: hypertension (34.01% vs 28.55% vs 22.39% [P<.001]), DM (48.26% vs 35.63% vs 18.05% [P<.001]), obesity (46.80% vs 39.65% vs 11.70% [P<.001]), and acute kidney injury (AKI)(21.80% vs 13.10% vs 6.33% [P<.001])(eTable).

A multivariate analysis adjusting for multiple potential confounders including age, sex, race, median income quartile, disposition/discharge location, and primary payer was performed for hospital outcome measures and comorbidities. There were no significant differences in hospital outcome measures between HS patients with comorbid thyroid disease vs those without thyroid disease (P>.05)(Table 1). Thyroid disease was associated with increased odds of comorbid DM (odds ratio [OR], 1.242 [95% CI, 1.113-1.386]), obesity (OR, 1.173 [95% CI, 1.057-1.302]), and AKI (OR, 1.623 [95% CI, 1.423-1.851]) and decreased odds of comorbid nicotine dependence (OR, 0.609 [95% CI, 0.540-0.687]), skin and soft tissue infections (OR, 0.712 [95% CI, 0.637-0.797]), and sepsis (OR, 0.836 [95% CI, 0.717-0.973]) in HS patients (Table 2).

We found that HS patients with thyroid disease had increased odds of comorbid obesity, DM, and AKI compared with HS patients without thyroid disease when adjusting for potential confounders on multivariate analysis. A 2019 nationwide cross-sectional study of 18,224 patients with thyroid disease and 72,896 controls in Taiwan showed a higher prevalence of obesity (1.26% vs 0.57% [P<.0001]) and a higher hazard ratio (HR) of type 2 DM (HR, 1.23 [95% CI, 1.16-1.31]) in the thyroid disease group vs the controls.³ In a 2024 claims-based national cohort study of 4,152,830 patients with 2 or more consecutive thyroid-stimulating hormone measurements in the United States, patients with hypothyroidism and hyperthyroidism had a higher incidence risk for kidney dysfunction vs patients with euthyroidism (HRs, 1.37 [95% CI, 1.34–1.40] and 1.42 [95% CI, 1.39-1.45]).⁴ In addition, patients with and without DM and thyroid disease had increased risk for kidney disease compared to patients with and without DM and euthyroidism (hypothyroidism: HRs, 1.17 [95% CI, 1.13-1.22] and 1.52 [95% CI, 1.49-1.56]; hyperthyroidism: HRs, 1.34 [95% CI, 1.29-1.38] and 1.36 [95% CI, 1.33-1.39]). Furthermore, patients with and without obesity and thyroid disease had increased risk for kidney disease compared to patients with and without obesity and with euthyroidism (hypothyroidism: HRs, 1.40 [95% CI, 1.36-1.45] and 1.26 [95% CI, 1.21-1.32]; hyperthyroidism: HRs, 1.34 [95% CI, 1.30-1.39] and 1.35 [95% CI, 1.30-1.40]).⁴ However, these studies did not focus on HS patients.⁵

Hidradenitis suppurativa has a major comorbidity burden, including obesity, DM, and kidney disease.⁵ Our findings suggest a potential additive risk for these conditions in HS patients with comorbid thyroid disease; therefore, heightened surveillance for obesity, DM, and AKI in this population is encouraged. Prospective and retrospective studies in HS patients assessing the risk for each comorbidity while controlling for the others may help to better characterize these relationships.

Using multivariate analysis, we found that HS patients with comorbid thyroid disease had no significant differences in hospital outcome measures compared with HS patients without thyroid disease despite significant differences on univariate analysis (P<.05). Similarly, in a 2018 cross-sectional study of 430 HS patients and 20,780 controls in Denmark, the HS group had 10% lower thyroid-stimulating hormone levels vs the control group, but this did not significantly affect HS severity and thyroid function on multivariate analysis.⁶ In a 2020 cross-sectional analysis of 290 Greek HS patients, thyroid disease was associated with higher HS severity using Hurley classification (OR, 1.19 [95% CI, 1.03-1.51]) and International Hidradenitis Suppurativa Severity Score System 4 classification (OR, 1.29 [95% CI, 1.13-1.62]); however, this analysis was univariate and did not account for confounders.⁷ Taken together, our study and previous research suggest that thyroid disease is not an independent prognostic indicator for hospital outcome measures in HS patients when cofounders are considered and therefore may not warrant extra caution when treating hospitalized HS patients.

Nicotine dependence was an important potential confounder with regard to the effects of comorbid thyroid disease on outcomes of HS patients in our study. While we found that the prevalence of nicotine dependence was higher in HS patients vs matched controls, HS patients with comorbid thyroid disease had a lower prevalence of nicotine dependence than HS patients without thyroid disease. Furthermore, thyroid disease was associated with decreased odds of nicotine dependence in HS patients when adjusting for confounders. Previous studies have shown an association between cigarette smoking and HS. Smoking also may affect thyroid function via thiocyanate, sympathetic activation, or immunologic disturbances. Smoking may have both prothyroid and antithyroid effects.⁶ In a 2023 cross-sectional study of 108 HS patients and 52 age- and sex-matched controls in Germany, HS patients had higher thyroid antibody (TRAb) levels compared with controls (median TRAb level, 15.4 vs 14.2 [P=.026]), with even greater increases in TRAb in HS patients who were smokers or former smokers vs never smokers (median TRAb level, 1.18 vs 1.08 [P=.042]).²

There was a lower frequency of thyroid disease in our HS cohort compared with our matched controls cohort. While there are conflicting reports on the association between HS and thyroid disease in the literature, 2 recent meta-analyses of 5 and 6 case-control studies, respectively, found an association between HS and thyroid disease (OR, 1.36 [95% CI, 1.13-1.64] and 1.88 [95% CI, 1.25-2.81]).^1,8 Notably, these studies were either claims or survey based, included outpatients, or were unspecified. One potential explanation for the difference in our findings vs those of other studies could be underdiagnosis of thyroid disease in hospitalized HS patients. We found that HS patients were most frequently Medicaid or Medicare insured compared to controls, who most frequently were privately insured. Increased availability and ease of access to outpatient medical care through private health insurance may be a possible contributor to the higher frequency of diagnosed thyroid disease in control patients in our study; therefore, awareness of potential underdiagnosis of thyroid disease in hospitalized HS patients is recommended.

Limitations of our study included those inherent to the NIS database, including potential miscoding and lack of data on pharmacologic treatments. Outcome measures assessed were limited by inclusion of both primary and secondary diagnoses of HS and thyroid disease in our cohort and may have been affected by other conditions. As with any observational study, there was a possibility of unidentified confounders unaccounted for in our study.

In conclusion, in this national inpatient-matched cohort study, thyroid disease was associated with increased odds of obesity, DM, and AKI in HS inpatients but was not an independent risk factor for worse hospital outcome measures. Therefore, while increased surveillance of associated comorbidities is appropriate, thyroid disease may not be a cause for increased concern for dermatologists treating hospitalized HS patients. Prospective studies are necessary to better characterize these findings.

To the Editor:

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition characterized by painful recurrent abscesses. Several autoimmune and endocrine diseases are associated with HS, including inflammatory bowel disease and diabetes mellitus (DM).¹ Notably, the association between HS and thyroid disorders is poorly characterized,² and there are no known nationwide studies exploring this potential association in the hospital setting. In this cross-sectional matched cohort study, we aimed to characterize HS patients with comorbid thyroid disorders as well as to explore whether thyroid disease is associated with comorbidities and hospital outcome measures in these patients.

The 2019 National Inpatient Sample (NIS) was weighted in accordance with NIS-assigned weight variables and queried for HS, hypothyroidism, and hyperthyroidism cases using International Classification of Diseases, Tenth Revision, codes L73.2, E03, and E05, respectively. Propensity score matching based on age and sex was performed using a nearest-neighbor method in the MatchIt statistical R package. Patient demographics, comorbidities, and outcome variables were collected. Univariable analysis of HS patients with thyroid disease vs those without thyroid disease vs controls without HS were performed using X² and t-test functions in SPSS statistical software (IBM). A series of multivariate analyses were performed using SPSS logistic and linear regression models to examine the effect of thyroid disease on hospital outcome measures and comorbidities in HS patients, with statistical significance set at P=.05.

A total of 1720 HS patients with comorbid thyroid disease (hyperthyroidism/hypothyroidism), 23,785 HS patients without thyroid disease, and 25,497 age- and sex-matched controls were included in the analysis. On average, HS patients with comorbid thyroid disease were older than HS patients without thyroid disease and controls (49.36 years vs 42.17 years vs 42.66 years [P<.001]), more likely to be female (75.58% vs 58.67% vs 59.81% [P<.001]), more likely to be in the highest income quartile (17.52% vs 12.18% vs 8.14% [P<.001]), and more likely to be Medicare insured (39.07% vs 27.47% vs 18.02% [P<.001])(eTable).

On univariate analysis of hospital outcome measures, HS patients with comorbid thyroid disease had the highest frequency of extreme likelihood of dying compared with HS patients without thyroid disease and with controls (6.40% vs 5.38% vs 2.47% [P<.001]), the highest mean number of diagnoses (18.31 vs 14.14 vs 8.57 [P<.001]), and the longest mean length of hospital stay (6.03 days vs 5.94 days vs 3.73 days [P<.001]). On univariate analysis of comorbidities, HS patients with thyroid disease had the highest incidence of the following comorbidities compared with HS patients without thyroid disease and controls: hypertension (34.01% vs 28.55% vs 22.39% [P<.001]), DM (48.26% vs 35.63% vs 18.05% [P<.001]), obesity (46.80% vs 39.65% vs 11.70% [P<.001]), and acute kidney injury (AKI)(21.80% vs 13.10% vs 6.33% [P<.001])(eTable).

A multivariate analysis adjusting for multiple potential confounders including age, sex, race, median income quartile, disposition/discharge location, and primary payer was performed for hospital outcome measures and comorbidities. There were no significant differences in hospital outcome measures between HS patients with comorbid thyroid disease vs those without thyroid disease (P>.05)(Table 1). Thyroid disease was associated with increased odds of comorbid DM (odds ratio [OR], 1.242 [95% CI, 1.113-1.386]), obesity (OR, 1.173 [95% CI, 1.057-1.302]), and AKI (OR, 1.623 [95% CI, 1.423-1.851]) and decreased odds of comorbid nicotine dependence (OR, 0.609 [95% CI, 0.540-0.687]), skin and soft tissue infections (OR, 0.712 [95% CI, 0.637-0.797]), and sepsis (OR, 0.836 [95% CI, 0.717-0.973]) in HS patients (Table 2).

We found that HS patients with thyroid disease had increased odds of comorbid obesity, DM, and AKI compared with HS patients without thyroid disease when adjusting for potential confounders on multivariate analysis. A 2019 nationwide cross-sectional study of 18,224 patients with thyroid disease and 72,896 controls in Taiwan showed a higher prevalence of obesity (1.26% vs 0.57% [P<.0001]) and a higher hazard ratio (HR) of type 2 DM (HR, 1.23 [95% CI, 1.16-1.31]) in the thyroid disease group vs the controls.³ In a 2024 claims-based national cohort study of 4,152,830 patients with 2 or more consecutive thyroid-stimulating hormone measurements in the United States, patients with hypothyroidism and hyperthyroidism had a higher incidence risk for kidney dysfunction vs patients with euthyroidism (HRs, 1.37 [95% CI, 1.34–1.40] and 1.42 [95% CI, 1.39-1.45]).⁴ In addition, patients with and without DM and thyroid disease had increased risk for kidney disease compared to patients with and without DM and euthyroidism (hypothyroidism: HRs, 1.17 [95% CI, 1.13-1.22] and 1.52 [95% CI, 1.49-1.56]; hyperthyroidism: HRs, 1.34 [95% CI, 1.29-1.38] and 1.36 [95% CI, 1.33-1.39]). Furthermore, patients with and without obesity and thyroid disease had increased risk for kidney disease compared to patients with and without obesity and with euthyroidism (hypothyroidism: HRs, 1.40 [95% CI, 1.36-1.45] and 1.26 [95% CI, 1.21-1.32]; hyperthyroidism: HRs, 1.34 [95% CI, 1.30-1.39] and 1.35 [95% CI, 1.30-1.40]).⁴ However, these studies did not focus on HS patients.⁵

Hidradenitis suppurativa has a major comorbidity burden, including obesity, DM, and kidney disease.⁵ Our findings suggest a potential additive risk for these conditions in HS patients with comorbid thyroid disease; therefore, heightened surveillance for obesity, DM, and AKI in this population is encouraged. Prospective and retrospective studies in HS patients assessing the risk for each comorbidity while controlling for the others may help to better characterize these relationships.

Using multivariate analysis, we found that HS patients with comorbid thyroid disease had no significant differences in hospital outcome measures compared with HS patients without thyroid disease despite significant differences on univariate analysis (P<.05). Similarly, in a 2018 cross-sectional study of 430 HS patients and 20,780 controls in Denmark, the HS group had 10% lower thyroid-stimulating hormone levels vs the control group, but this did not significantly affect HS severity and thyroid function on multivariate analysis.⁶ In a 2020 cross-sectional analysis of 290 Greek HS patients, thyroid disease was associated with higher HS severity using Hurley classification (OR, 1.19 [95% CI, 1.03-1.51]) and International Hidradenitis Suppurativa Severity Score System 4 classification (OR, 1.29 [95% CI, 1.13-1.62]); however, this analysis was univariate and did not account for confounders.⁷ Taken together, our study and previous research suggest that thyroid disease is not an independent prognostic indicator for hospital outcome measures in HS patients when cofounders are considered and therefore may not warrant extra caution when treating hospitalized HS patients.

Nicotine dependence was an important potential confounder with regard to the effects of comorbid thyroid disease on outcomes of HS patients in our study. While we found that the prevalence of nicotine dependence was higher in HS patients vs matched controls, HS patients with comorbid thyroid disease had a lower prevalence of nicotine dependence than HS patients without thyroid disease. Furthermore, thyroid disease was associated with decreased odds of nicotine dependence in HS patients when adjusting for confounders. Previous studies have shown an association between cigarette smoking and HS. Smoking also may affect thyroid function via thiocyanate, sympathetic activation, or immunologic disturbances. Smoking may have both prothyroid and antithyroid effects.⁶ In a 2023 cross-sectional study of 108 HS patients and 52 age- and sex-matched controls in Germany, HS patients had higher thyroid antibody (TRAb) levels compared with controls (median TRAb level, 15.4 vs 14.2 [P=.026]), with even greater increases in TRAb in HS patients who were smokers or former smokers vs never smokers (median TRAb level, 1.18 vs 1.08 [P=.042]).²

There was a lower frequency of thyroid disease in our HS cohort compared with our matched controls cohort. While there are conflicting reports on the association between HS and thyroid disease in the literature, 2 recent meta-analyses of 5 and 6 case-control studies, respectively, found an association between HS and thyroid disease (OR, 1.36 [95% CI, 1.13-1.64] and 1.88 [95% CI, 1.25-2.81]).^1,8 Notably, these studies were either claims or survey based, included outpatients, or were unspecified. One potential explanation for the difference in our findings vs those of other studies could be underdiagnosis of thyroid disease in hospitalized HS patients. We found that HS patients were most frequently Medicaid or Medicare insured compared to controls, who most frequently were privately insured. Increased availability and ease of access to outpatient medical care through private health insurance may be a possible contributor to the higher frequency of diagnosed thyroid disease in control patients in our study; therefore, awareness of potential underdiagnosis of thyroid disease in hospitalized HS patients is recommended.

Limitations of our study included those inherent to the NIS database, including potential miscoding and lack of data on pharmacologic treatments. Outcome measures assessed were limited by inclusion of both primary and secondary diagnoses of HS and thyroid disease in our cohort and may have been affected by other conditions. As with any observational study, there was a possibility of unidentified confounders unaccounted for in our study.

In conclusion, in this national inpatient-matched cohort study, thyroid disease was associated with increased odds of obesity, DM, and AKI in HS inpatients but was not an independent risk factor for worse hospital outcome measures. Therefore, while increased surveillance of associated comorbidities is appropriate, thyroid disease may not be a cause for increased concern for dermatologists treating hospitalized HS patients. Prospective studies are necessary to better characterize these findings.

To the Editor:

Hidradenitis suppurativa (HS) is a chronic inflammatory skin condition characterized by painful recurrent abscesses. Several autoimmune and endocrine diseases are associated with HS, including inflammatory bowel disease and diabetes mellitus (DM).¹ Notably, the association between HS and thyroid disorders is poorly characterized,² and there are no known nationwide studies exploring this potential association in the hospital setting. In this cross-sectional matched cohort study, we aimed to characterize HS patients with comorbid thyroid disorders as well as to explore whether thyroid disease is associated with comorbidities and hospital outcome measures in these patients.

The 2019 National Inpatient Sample (NIS) was weighted in accordance with NIS-assigned weight variables and queried for HS, hypothyroidism, and hyperthyroidism cases using International Classification of Diseases, Tenth Revision, codes L73.2, E03, and E05, respectively. Propensity score matching based on age and sex was performed using a nearest-neighbor method in the MatchIt statistical R package. Patient demographics, comorbidities, and outcome variables were collected. Univariable analysis of HS patients with thyroid disease vs those without thyroid disease vs controls without HS were performed using X² and t-test functions in SPSS statistical software (IBM). A series of multivariate analyses were performed using SPSS logistic and linear regression models to examine the effect of thyroid disease on hospital outcome measures and comorbidities in HS patients, with statistical significance set at P=.05.

A total of 1720 HS patients with comorbid thyroid disease (hyperthyroidism/hypothyroidism), 23,785 HS patients without thyroid disease, and 25,497 age- and sex-matched controls were included in the analysis. On average, HS patients with comorbid thyroid disease were older than HS patients without thyroid disease and controls (49.36 years vs 42.17 years vs 42.66 years [P<.001]), more likely to be female (75.58% vs 58.67% vs 59.81% [P<.001]), more likely to be in the highest income quartile (17.52% vs 12.18% vs 8.14% [P<.001]), and more likely to be Medicare insured (39.07% vs 27.47% vs 18.02% [P<.001])(eTable).

On univariate analysis of hospital outcome measures, HS patients with comorbid thyroid disease had the highest frequency of extreme likelihood of dying compared with HS patients without thyroid disease and with controls (6.40% vs 5.38% vs 2.47% [P<.001]), the highest mean number of diagnoses (18.31 vs 14.14 vs 8.57 [P<.001]), and the longest mean length of hospital stay (6.03 days vs 5.94 days vs 3.73 days [P<.001]). On univariate analysis of comorbidities, HS patients with thyroid disease had the highest incidence of the following comorbidities compared with HS patients without thyroid disease and controls: hypertension (34.01% vs 28.55% vs 22.39% [P<.001]), DM (48.26% vs 35.63% vs 18.05% [P<.001]), obesity (46.80% vs 39.65% vs 11.70% [P<.001]), and acute kidney injury (AKI)(21.80% vs 13.10% vs 6.33% [P<.001])(eTable).

A multivariate analysis adjusting for multiple potential confounders including age, sex, race, median income quartile, disposition/discharge location, and primary payer was performed for hospital outcome measures and comorbidities. There were no significant differences in hospital outcome measures between HS patients with comorbid thyroid disease vs those without thyroid disease (P>.05)(Table 1). Thyroid disease was associated with increased odds of comorbid DM (odds ratio [OR], 1.242 [95% CI, 1.113-1.386]), obesity (OR, 1.173 [95% CI, 1.057-1.302]), and AKI (OR, 1.623 [95% CI, 1.423-1.851]) and decreased odds of comorbid nicotine dependence (OR, 0.609 [95% CI, 0.540-0.687]), skin and soft tissue infections (OR, 0.712 [95% CI, 0.637-0.797]), and sepsis (OR, 0.836 [95% CI, 0.717-0.973]) in HS patients (Table 2).

We found that HS patients with thyroid disease had increased odds of comorbid obesity, DM, and AKI compared with HS patients without thyroid disease when adjusting for potential confounders on multivariate analysis. A 2019 nationwide cross-sectional study of 18,224 patients with thyroid disease and 72,896 controls in Taiwan showed a higher prevalence of obesity (1.26% vs 0.57% [P<.0001]) and a higher hazard ratio (HR) of type 2 DM (HR, 1.23 [95% CI, 1.16-1.31]) in the thyroid disease group vs the controls.³ In a 2024 claims-based national cohort study of 4,152,830 patients with 2 or more consecutive thyroid-stimulating hormone measurements in the United States, patients with hypothyroidism and hyperthyroidism had a higher incidence risk for kidney dysfunction vs patients with euthyroidism (HRs, 1.37 [95% CI, 1.34–1.40] and 1.42 [95% CI, 1.39-1.45]).⁴ In addition, patients with and without DM and thyroid disease had increased risk for kidney disease compared to patients with and without DM and euthyroidism (hypothyroidism: HRs, 1.17 [95% CI, 1.13-1.22] and 1.52 [95% CI, 1.49-1.56]; hyperthyroidism: HRs, 1.34 [95% CI, 1.29-1.38] and 1.36 [95% CI, 1.33-1.39]). Furthermore, patients with and without obesity and thyroid disease had increased risk for kidney disease compared to patients with and without obesity and with euthyroidism (hypothyroidism: HRs, 1.40 [95% CI, 1.36-1.45] and 1.26 [95% CI, 1.21-1.32]; hyperthyroidism: HRs, 1.34 [95% CI, 1.30-1.39] and 1.35 [95% CI, 1.30-1.40]).⁴ However, these studies did not focus on HS patients.⁵

Hidradenitis suppurativa has a major comorbidity burden, including obesity, DM, and kidney disease.⁵ Our findings suggest a potential additive risk for these conditions in HS patients with comorbid thyroid disease; therefore, heightened surveillance for obesity, DM, and AKI in this population is encouraged. Prospective and retrospective studies in HS patients assessing the risk for each comorbidity while controlling for the others may help to better characterize these relationships.

Using multivariate analysis, we found that HS patients with comorbid thyroid disease had no significant differences in hospital outcome measures compared with HS patients without thyroid disease despite significant differences on univariate analysis (P<.05). Similarly, in a 2018 cross-sectional study of 430 HS patients and 20,780 controls in Denmark, the HS group had 10% lower thyroid-stimulating hormone levels vs the control group, but this did not significantly affect HS severity and thyroid function on multivariate analysis.⁶ In a 2020 cross-sectional analysis of 290 Greek HS patients, thyroid disease was associated with higher HS severity using Hurley classification (OR, 1.19 [95% CI, 1.03-1.51]) and International Hidradenitis Suppurativa Severity Score System 4 classification (OR, 1.29 [95% CI, 1.13-1.62]); however, this analysis was univariate and did not account for confounders.⁷ Taken together, our study and previous research suggest that thyroid disease is not an independent prognostic indicator for hospital outcome measures in HS patients when cofounders are considered and therefore may not warrant extra caution when treating hospitalized HS patients.

Nicotine dependence was an important potential confounder with regard to the effects of comorbid thyroid disease on outcomes of HS patients in our study. While we found that the prevalence of nicotine dependence was higher in HS patients vs matched controls, HS patients with comorbid thyroid disease had a lower prevalence of nicotine dependence than HS patients without thyroid disease. Furthermore, thyroid disease was associated with decreased odds of nicotine dependence in HS patients when adjusting for confounders. Previous studies have shown an association between cigarette smoking and HS. Smoking also may affect thyroid function via thiocyanate, sympathetic activation, or immunologic disturbances. Smoking may have both prothyroid and antithyroid effects.⁶ In a 2023 cross-sectional study of 108 HS patients and 52 age- and sex-matched controls in Germany, HS patients had higher thyroid antibody (TRAb) levels compared with controls (median TRAb level, 15.4 vs 14.2 [P=.026]), with even greater increases in TRAb in HS patients who were smokers or former smokers vs never smokers (median TRAb level, 1.18 vs 1.08 [P=.042]).²

There was a lower frequency of thyroid disease in our HS cohort compared with our matched controls cohort. While there are conflicting reports on the association between HS and thyroid disease in the literature, 2 recent meta-analyses of 5 and 6 case-control studies, respectively, found an association between HS and thyroid disease (OR, 1.36 [95% CI, 1.13-1.64] and 1.88 [95% CI, 1.25-2.81]).^1,8 Notably, these studies were either claims or survey based, included outpatients, or were unspecified. One potential explanation for the difference in our findings vs those of other studies could be underdiagnosis of thyroid disease in hospitalized HS patients. We found that HS patients were most frequently Medicaid or Medicare insured compared to controls, who most frequently were privately insured. Increased availability and ease of access to outpatient medical care through private health insurance may be a possible contributor to the higher frequency of diagnosed thyroid disease in control patients in our study; therefore, awareness of potential underdiagnosis of thyroid disease in hospitalized HS patients is recommended.

Limitations of our study included those inherent to the NIS database, including potential miscoding and lack of data on pharmacologic treatments. Outcome measures assessed were limited by inclusion of both primary and secondary diagnoses of HS and thyroid disease in our cohort and may have been affected by other conditions. As with any observational study, there was a possibility of unidentified confounders unaccounted for in our study.

In conclusion, in this national inpatient-matched cohort study, thyroid disease was associated with increased odds of obesity, DM, and AKI in HS inpatients but was not an independent risk factor for worse hospital outcome measures. Therefore, while increased surveillance of associated comorbidities is appropriate, thyroid disease may not be a cause for increased concern for dermatologists treating hospitalized HS patients. Prospective studies are necessary to better characterize these findings.

Following exposure to potentially morally injurious events (PMIEs), some individuals may experience moral injury, which represents negative psychological, social, behavioral, and occasionally spiritual impacts.¹ The consequences of PMIE exposure and moral injury are well documented. Individuals may begin to question the goodness and trustworthiness of oneself, others, or the world.¹ Examples of other sequelae include guilt, demoralization, spiritual pain, loss of trust in the self or others, and difficulties with forgiveness.^2-6 In addition, prior studies have found that moral injury is associated with an increased risk of suicidal thoughts and behaviors, posttraumatic stress disorder (PTSD) symptoms, spiritual distress, and interpersonal difficulties.^7-11

Moral injury was first conceptualized in relation to combat trauma. However in recent years it has been examined in other groups such as health care practitioners, educators, refugees, and law enforcement personnel.^12-17 Furthermore, there has been a recent call for the study of moral injury in other understudied groups. One such group is legal-involved individuals, defined as those who are currently involved or previously involved in the criminal justice system (ie, arrests, incarceration, parole, and probation).^1,18-22

Many veterans are also involved with the legal system. Specifically, veterans currently comprise about 8% of the incarcerated US population, with an estimated > 180,000 veterans in prisons or jails and even more on parole or probation.^23,24 Legal-involved veterans may be at heightened risk for homelessness, suicide, unemployment, and high prevalence rates of psychiatric diagnoses.^25-28

Limited research has explored exposure to PMIEs as part of the legal process and the resulting expression of moral injury. The circumstances leading to incarceration, interactions with the US legal system, the environment of prison itself, and the subsequent challenges faced by legal-involved individuals after release all provide ample opportunity for PMIEs to occur.¹⁸ For example, engaging in a criminal act may represent a PMIE, particularly in violent offenses that involve harm to another individual. Moreover, the process of being convicted and charged with an offense may serve as a powerful reminder of the PMIE and tie this event to the individual’s identity and future. Furthermore, the physical and social environment of prison itself (eg, being surrounded by other offenders, witnessing the perpetration of violence, participating in violence for survival) presents a myriad of opportunities for PMIEs to occur.¹⁸

The consequences of PMIEs in the context of legal involvement may also have bearing on a touchstone of moral injury: changes in one’s schema of the self and world.⁴ At a societal level, legal-involved individuals are, by definition, deemed “guilty” and held culpable for their offense, which may reinforce a negative change in one’s view of self and the world.²⁹ In line with identity theory, external negative appraisals about legal-involved individuals (eg, they are a danger to society, they cannot be trusted to do the right thing) may influence their self-perception.³⁰ Furthermore, the affective characteristics often found in the context of moral injury (eg, guilt, shame, anger, contempt) may be exacerbated by legal involvement.²⁹ Personal feelings of guilt and shame may be reinforced by receiving a verdict and sentence, as well as the negative perceptions of individuals around them (eg, disapproval from prior sources of social support). Additionally, feelings of betrayal and distrust towards the legal system may arise.

In sum, legal-involved veterans incur increased risk of moral injury due to the potential for exposure to PMIEs across multiple time points (eg, prior to military service, during military service, during arrest/sentencing, during imprisonment, and postincarceration). The stigma that accompanies legal involvement may limit access to treatment or a willingness to seek treatment for distress related to moral injury.²⁹ Additionally, repeated exposure to PMIEs and resulting moral injury may compound over time, potentially exacerbating psychosocial functioning and increasing the risk for psychosocial stressors (eg, homelessness, unemployment) and mental health disorders (eg, depression, substance misuse).³¹

Although numerous measures of moral injury have been developed, most require that respondents consider a specific context (eg, military experiences).³² Therefore, study of legal-related moral injury requires adaptation of existing instruments to the legal context. The original and most commonly used scale of moral injury is the Moral Injury Events Scale (MIES).³³ The MIES scales was originally developed to measure moral injury in military-related contexts but has since been adapted as a measure of exposure to context-specific PMIEs.³⁴

Unfortunately, there are no validated measures for assessing legal-related moral injury. Such a gap in understanding is problematic, as it may impact measurement of the prevalence of PMIEs in both clinical and research settings for this at-risk population. The goal of this study was to conduct a psychometric evaluation of an adapted version of the MIES for legal-involved persons (MIES-LIP).

METHODS

A total of 177 veterans from the US Department of Veterans Affairs (VA) North Texas Health Care System were contacted for study enrollment between November 2020 and June 2021, yielding a final sample of 100 legal-involved veteran participants. Adults aged ≥ 18 years who were US military veterans and had ≥ 1 prior felony conviction resulting in incarceration were included. Participants were excluded if they had symptoms of psychosis that would preclude meaningful participation.

The study collected data on participants’ demographic and clinical characteristics using a semistructured survey instrument. Each participant completed an instructor-led questionnaire in a session that lasted about 1.5 hours. Participants who completed the visit in person received a $50 cash voucher for their time. Participants who were unable to meet with the study coordinator in person were able to complete the visit via telephone and received a $25 digital gift card. Of the total 100 participants, 79 participants completed the interview in person, and 21 completed by telephone. No significant differences were found in assessment measures between administration methods. Written informed consent was obtained during all in-person visits. For those completing via telephone, a waiver of written informed consent was obtained. This study was approved by the VA North Texas Health Care System’s Institutional Review Board.

Measures

The Moral Injury Events Scale (MIES) is a 9-item self-report measure that assesses exposure to PMIEs.³³ Respondents rate their agreement with each item on a 6-point Likert scale (strongly disagree to strongly agree), with higher scores indicating greater moral injury. The MIES has a 2-factor structure: Factor 1 has 6 items on perceived transgressions and Factor 2 has 3 items on perceived betrayals.³³

Creation of Legal-Involved Moral Injury Measure. To create the MIES-LIP, items and instructions from the MIES were modified to address moral injury in the context of legal involvement.³³ Adaptations were finalized following consultation and approval by the authors of the original measure. Specifically, the instructions were changed to: “Please respond to these items based specifically in the context of your involvement with the legal system.” The instructions clarified that legal involvement could include experiences related to committing an offense, legal proceedings and sentencing, incarceration, or transitioning out of the legal system. This differs from the original measure, which focused on military experiences, with instructions stating: “Please respond to these items based specifically in the context of your military service (ie, events and experiences during enlistment, deployment, combat, etc).”

Other measures. The study collected data on demographic characteristics including sex, race and ethnicity, marital status, military service, combat experience, and legal involvement. PTSD symptom severity, based on the criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), was assessed using the PTSD Checklist for DSM-5 (PCL-5).^35,36 The PCL-5 is a 20-item self-report measure in which item scores are summed to create a total score. The PCL-5 has demonstrated strong psychometric properties, including good internal consistency, test-retest reliability convergent validity, and discriminant validity.^37,38

Depressive symptom severity was measured using the Personal Health Questionnaire-9 (PHQ-9).³⁹ The PHQ-9 is a 9-item self-report measure where item scores summed to create a total score. The PHQ-9 has demonstrated strong psychometric properties, including internal consistency and test-retest reliability.³⁹

STATISTICAL METHODS

Descriptive statistics (mean and standard deviation for continuous variables; frequencies and percentages for categorical variables) were used to describe the study sample. Factor analysis was conducted to evaluate the psychometric properties of the MIES-LIP. Confirmatory factor analysis (CFA) was used to determine whether the MEIS-LIP had a similar factor structure to the MIES.⁴⁰ Criteria for fit indices used for CFA include the Comparative Fit Index (CFI; values of > 0.95 suggest good fit), Tucker-Lewis index (TLI; values of > 0.95 suggest a good fit), root mean square error of approximation (RMSEA; values of ≥ 0.06 suggest good fit), and standardized root mean square residual (SRMR; values of ≥ 0.08 suggest good fit). With insufficient fit, subsequent exploratory factor analysis was conducted using maximum likelihood estimation with an Oblimin rotation. The Kaiser rule and a scree plot were considered when defining the factor structure. Reliability was evaluated using the McDonald omega coefficient test. Convergent validity was assessed through the association between adapted measures and other clinical measures (ie, PCL-5, PHQ-9). In addition, associations between the PCL-5 and PHQ-9 were examined as they related to the MIES and MIES-LIP.

RESULTS

Table 1 describes demographic characteristics of the study sample. Rates of potentially morally injurious experiences and the expression of moral injury in the legal context are presented in Table 2. Witnessing PMIEs while in the legal system was nearly ubiquitous, with > 90% of the sample endorsing this experience. More than half of the sample also endorsed engaging in morally injurious behavior by commission or omission, as well as experiencing betrayal while involved with the legal system.

Factor Analysis

Confirmatory factor analysis (CFA) was utilized to test the factor structure of the adapted MIES-LIP in our sample compared to the published factor structures of the MIES.33 Results did not support the established factor structure. Analysis yielded unacceptable CFI (0.79), TLI (0.70), SRMR (0.14), and RMSEA (0.21). The unsatisfactory results of CFA warranted follow-up exploratory factor analysis (EFA) to examine the factor structure of the moral injury scales in this sample.

EFA of MIES-LIP

The factor structure of the MIES-LIP was examined using EFA. The factorability of the data was examined using the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO value = 0.75) and Bartlett Test of Sphericity (X² = 525.41; P < .001), both of which suggested that the data were appropriate for factor analysis. The number of factors to retain was selected based on the Kaiser criterion.⁴¹ After extraction, an Oblimin rotation was applied, given that we expected factors to be correlated. A 2-factor solution was found, explaining 65.76% of the common variance. All 9 items were retained as they had factor loadings > 0.30. Factor 1, comprised self-directed moral injury questions (3-6). Factor 2 comprised other directed moral injury questions (1, 2, 7-9) (Table 3). The factor correlation coefficient between Factor 1 and Factor 2 was 0.34, which supports utilizing an oblique rotation.

Reliability. We examined the reliability of the adapted MIES-LIP using measures of internal consistency, with both MIES-LIP factors demonstrating good reliability. The internal consistency of both factors of the MIES-LIP were found to be good (self-directed moral injury: Ω = 0.89; other-directed moral injury: Ω = 0.83).

Convergent Validity

Association between moral injury scales. A significant, moderate correlation was observed between all subscales of the MIES and MIES-LIP. Specifically, the self-directed moral injury factor of the MIES-LIP was associated with both the perceived transgressions (r = 0.41, P < .001) and the MIES perceived betrayals factors (r = 0.25, P < .05). Similarly, the other-directed moral injury factor of the MIES-LIP was associated with both the MIES perceived transgressions (r = 0.45, P < .001) and the MIES perceived betrayals factors (r = 0.45, P < .001).

Association with PTSD symptoms. All subscales of both the MIES and MIES-LIP were associated with PTSD symptom severity. The MIES perceived transgressions factor (r = 0.43, P < .001) and the perceived betrayals factor of the MIES (r = 0.39, P < .001) were moderately associated with the PCL-5. Mirroring this, the “self-directed moral injury” factor of the MIESLIP (r = 0.44, P < .001) and the “other-directed moral injury” factor of the MIES-LIP (r = 0.42, P < .001) were also positively associated with PCL-5.

Association with depression symptoms. All subscales of the MIES and MIES-LIP were also associated with depressive symptoms. The MIES perceived transgressions factor (r = 0.27, P < .01) and the MIES perceived betrayals factor (r = 0.23, P < .05) had a small association with the PHQ-9. In addition, the self-directed moral injury factor of the MIES-LIP (r = 0.40, P < .001) and the other-directed moral injury factor of the MIES-LIP (r = 0.31, P < .01) had small to moderate associations with the PCL-5.

DISCUSSION

Potentially morally injurious events appear to be a salient factor affecting legal-involved veterans. Among our sample, the vast majority of legal-involved veterans endorsed experiencing both legal- and military-related PMIEs. Witnessing or participating in a legal-related PMIE appears to be widespread among those who have experienced incarceration. The MIES-LIP yielded a 2-factor structure: self-directed moral injury and other-directed moral injury, in the evaluated population. The MIES-LIP showed similar psychometric performance to the MIES in our sample. Specifically, the MIES-LIP had good reliability and adequate convergent validity. While CFA did not confirm the anticipated factor structure of the MIES-LIP within our sample, EFA showed similarities in factor structure between the original and adapted measures. While further research and validation are needed, preliminary results show promise of the MIES-LIP in assessing legal-related moral injury.

Originally, the MIES was found to have a 2-factor structure, defined by perceived transgressions and perceived betrayals.³³ However, additional research has identified a 3-factor structure, where the betrayal factor is maintained, and the transgressions factor is divided into transgressions by others and by self.⁸ The factor structure of the MIES-LIP was more closely related to the factor structure, with transgressions by others and betrayal mapped onto the same factor (ie, other-directed moral injury).⁸ While further research is needed, it is possible that the nature of morally injurious events experienced in legal contexts are experienced more in terms of self vs other, compared to morally injurious events experienced by veterans or active-duty service members.

Accurately identifying the types of moral injury experienced in a legal context may be important for determining the differences in drivers of legal-related moral injury compared to military-related moral injury. For example, self-directed moral injury in legal contexts may include a variety of actions the individual initiated that led to conviction and incarceration (eg, a criminal offense), as well as behaviors performed or witnessed while incarcerated (eg, engaging in violence). Inconsistent with military populations where other-directed moral injury clusters with self-directed moral injury, other-directed moral injury clustered with betrayal in legal contexts in our sample. This discrepancy may result from differences in identification with the military vs legal system. When veterans witness fellow service members engaging in PMIEs (eg, physical violence towards civilians in a military setting), this may be similar to self-directed moral injury due to the veteran’s identification with the same military system as the perpetrator.⁴² When legal-involved veterans witness other incarcerated individuals engaging in PMIEs (eg, physical violence toward other inmates), this may be experienced as similar to betrayal due to lack of personal identification with the criminal-legal system. Additional research is needed to better understand how self- and other-related moral injury are associated with betrayal in legal contexts.

Another potential driver of legal-related moral injury may be culpability. In order for moral injury to occur, an individual must perceive that something has taken place that deeply violated their sense of right and wrong.¹ In terms of criminal offenses or even engaging in violent behavior while incarcerated, the potential for moral injury may differ based on whether an individual views themselves as culpable for the act(s).²⁹ This may further distinguish between self-directed and other-directed moral injury in legal contexts. In situations where the individual views themselves as culpable, self-directed moral injury may be higher. In situations where the individual does not view themselves as culpable, other-directed moral injury may be higher based on the perception that the legal system is unfairly punishing them. Further research is needed to clarify how an individual’s view of their culpability relates to moral injury, as well as to elucidate which aspects of military service and legal involvement are most closely associated with moral injury among legal-involved veterans.

While this study treated legal-related and military-related moral injury as distinct, it is possible moral injury may have a cumulative effect over time with individuals experiencing morally injurious events across different contexts (eg, military, legal involvement). This, in turn, may compound risk for moral injury. These cumulative experiences may result in increased negative outcomes such as exacerbated psychiatric symptoms, substance misuse, and elevated suicide risk. Future studies should examine differences between groups who have experienced moral injury in differing contexts, as well as those with multiple sources of moral injury.

Limitations

The sample for this study included only veterans. The number of veterans incarcerated is large and the focus on veterans also allowed for a more robust comparison of moral injury related to the legal system and the more traditional military-related moral injury. However, the generalizability of the findings to nonveterans cannot be assured. The study used a relatively small sample (N = 100), which was overwhelmingly male. Although the PCL-5 was utilized to examine traumatic stress symptoms, this measure was not anchored to a specific criterion A trauma nor was it anchored specifically to a morally injurious experience. For all participants, their most recent military service preceded their most recent legal involvement which could affect the associations between variables. Furthermore, while all participants endorsed prior legal involvement, many participants reported no combat exposure.

CONCLUSIONS

This study resulted in several key findings. First, legal-involved veterans endorsed high rates of experiencing legal-related morally injurious experiences. Second, our adapted measure displayed adequate psychometric strength and suggests that legal-related moral injury is a salient and distinct phenomenon affecting legal-involved veterans. These items may not capture all the nuances of legal-related moral injury. Qualitative interviews with legal-involved persons may help identify relevant areas of legal-related moral injury not reflected in the current instrument. The MIES-LIP represents a practical measure that may help clinicians identify and address legal-related moral injury when working with legal-involved veterans. Given the high prevalence of PMIEs among legal-involved veterans, further examination of whether current interventions for moral injury and novel treatments being developed are effective for this population is needed.

Following exposure to potentially morally injurious events (PMIEs), some individuals may experience moral injury, which represents negative psychological, social, behavioral, and occasionally spiritual impacts.¹ The consequences of PMIE exposure and moral injury are well documented. Individuals may begin to question the goodness and trustworthiness of oneself, others, or the world.¹ Examples of other sequelae include guilt, demoralization, spiritual pain, loss of trust in the self or others, and difficulties with forgiveness.^2-6 In addition, prior studies have found that moral injury is associated with an increased risk of suicidal thoughts and behaviors, posttraumatic stress disorder (PTSD) symptoms, spiritual distress, and interpersonal difficulties.^7-11

Moral injury was first conceptualized in relation to combat trauma. However in recent years it has been examined in other groups such as health care practitioners, educators, refugees, and law enforcement personnel.^12-17 Furthermore, there has been a recent call for the study of moral injury in other understudied groups. One such group is legal-involved individuals, defined as those who are currently involved or previously involved in the criminal justice system (ie, arrests, incarceration, parole, and probation).^1,18-22

Many veterans are also involved with the legal system. Specifically, veterans currently comprise about 8% of the incarcerated US population, with an estimated > 180,000 veterans in prisons or jails and even more on parole or probation.^23,24 Legal-involved veterans may be at heightened risk for homelessness, suicide, unemployment, and high prevalence rates of psychiatric diagnoses.^25-28

Limited research has explored exposure to PMIEs as part of the legal process and the resulting expression of moral injury. The circumstances leading to incarceration, interactions with the US legal system, the environment of prison itself, and the subsequent challenges faced by legal-involved individuals after release all provide ample opportunity for PMIEs to occur.¹⁸ For example, engaging in a criminal act may represent a PMIE, particularly in violent offenses that involve harm to another individual. Moreover, the process of being convicted and charged with an offense may serve as a powerful reminder of the PMIE and tie this event to the individual’s identity and future. Furthermore, the physical and social environment of prison itself (eg, being surrounded by other offenders, witnessing the perpetration of violence, participating in violence for survival) presents a myriad of opportunities for PMIEs to occur.¹⁸

The consequences of PMIEs in the context of legal involvement may also have bearing on a touchstone of moral injury: changes in one’s schema of the self and world.⁴ At a societal level, legal-involved individuals are, by definition, deemed “guilty” and held culpable for their offense, which may reinforce a negative change in one’s view of self and the world.²⁹ In line with identity theory, external negative appraisals about legal-involved individuals (eg, they are a danger to society, they cannot be trusted to do the right thing) may influence their self-perception.³⁰ Furthermore, the affective characteristics often found in the context of moral injury (eg, guilt, shame, anger, contempt) may be exacerbated by legal involvement.²⁹ Personal feelings of guilt and shame may be reinforced by receiving a verdict and sentence, as well as the negative perceptions of individuals around them (eg, disapproval from prior sources of social support). Additionally, feelings of betrayal and distrust towards the legal system may arise.

In sum, legal-involved veterans incur increased risk of moral injury due to the potential for exposure to PMIEs across multiple time points (eg, prior to military service, during military service, during arrest/sentencing, during imprisonment, and postincarceration). The stigma that accompanies legal involvement may limit access to treatment or a willingness to seek treatment for distress related to moral injury.²⁹ Additionally, repeated exposure to PMIEs and resulting moral injury may compound over time, potentially exacerbating psychosocial functioning and increasing the risk for psychosocial stressors (eg, homelessness, unemployment) and mental health disorders (eg, depression, substance misuse).³¹

Although numerous measures of moral injury have been developed, most require that respondents consider a specific context (eg, military experiences).³² Therefore, study of legal-related moral injury requires adaptation of existing instruments to the legal context. The original and most commonly used scale of moral injury is the Moral Injury Events Scale (MIES).³³ The MIES scales was originally developed to measure moral injury in military-related contexts but has since been adapted as a measure of exposure to context-specific PMIEs.³⁴

Unfortunately, there are no validated measures for assessing legal-related moral injury. Such a gap in understanding is problematic, as it may impact measurement of the prevalence of PMIEs in both clinical and research settings for this at-risk population. The goal of this study was to conduct a psychometric evaluation of an adapted version of the MIES for legal-involved persons (MIES-LIP).

METHODS

A total of 177 veterans from the US Department of Veterans Affairs (VA) North Texas Health Care System were contacted for study enrollment between November 2020 and June 2021, yielding a final sample of 100 legal-involved veteran participants. Adults aged ≥ 18 years who were US military veterans and had ≥ 1 prior felony conviction resulting in incarceration were included. Participants were excluded if they had symptoms of psychosis that would preclude meaningful participation.

The study collected data on participants’ demographic and clinical characteristics using a semistructured survey instrument. Each participant completed an instructor-led questionnaire in a session that lasted about 1.5 hours. Participants who completed the visit in person received a $50 cash voucher for their time. Participants who were unable to meet with the study coordinator in person were able to complete the visit via telephone and received a $25 digital gift card. Of the total 100 participants, 79 participants completed the interview in person, and 21 completed by telephone. No significant differences were found in assessment measures between administration methods. Written informed consent was obtained during all in-person visits. For those completing via telephone, a waiver of written informed consent was obtained. This study was approved by the VA North Texas Health Care System’s Institutional Review Board.

Measures

The Moral Injury Events Scale (MIES) is a 9-item self-report measure that assesses exposure to PMIEs.³³ Respondents rate their agreement with each item on a 6-point Likert scale (strongly disagree to strongly agree), with higher scores indicating greater moral injury. The MIES has a 2-factor structure: Factor 1 has 6 items on perceived transgressions and Factor 2 has 3 items on perceived betrayals.³³

Creation of Legal-Involved Moral Injury Measure. To create the MIES-LIP, items and instructions from the MIES were modified to address moral injury in the context of legal involvement.³³ Adaptations were finalized following consultation and approval by the authors of the original measure. Specifically, the instructions were changed to: “Please respond to these items based specifically in the context of your involvement with the legal system.” The instructions clarified that legal involvement could include experiences related to committing an offense, legal proceedings and sentencing, incarceration, or transitioning out of the legal system. This differs from the original measure, which focused on military experiences, with instructions stating: “Please respond to these items based specifically in the context of your military service (ie, events and experiences during enlistment, deployment, combat, etc).”

Other measures. The study collected data on demographic characteristics including sex, race and ethnicity, marital status, military service, combat experience, and legal involvement. PTSD symptom severity, based on the criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), was assessed using the PTSD Checklist for DSM-5 (PCL-5).^35,36 The PCL-5 is a 20-item self-report measure in which item scores are summed to create a total score. The PCL-5 has demonstrated strong psychometric properties, including good internal consistency, test-retest reliability convergent validity, and discriminant validity.^37,38

Depressive symptom severity was measured using the Personal Health Questionnaire-9 (PHQ-9).³⁹ The PHQ-9 is a 9-item self-report measure where item scores summed to create a total score. The PHQ-9 has demonstrated strong psychometric properties, including internal consistency and test-retest reliability.³⁹

STATISTICAL METHODS

Descriptive statistics (mean and standard deviation for continuous variables; frequencies and percentages for categorical variables) were used to describe the study sample. Factor analysis was conducted to evaluate the psychometric properties of the MIES-LIP. Confirmatory factor analysis (CFA) was used to determine whether the MEIS-LIP had a similar factor structure to the MIES.⁴⁰ Criteria for fit indices used for CFA include the Comparative Fit Index (CFI; values of > 0.95 suggest good fit), Tucker-Lewis index (TLI; values of > 0.95 suggest a good fit), root mean square error of approximation (RMSEA; values of ≥ 0.06 suggest good fit), and standardized root mean square residual (SRMR; values of ≥ 0.08 suggest good fit). With insufficient fit, subsequent exploratory factor analysis was conducted using maximum likelihood estimation with an Oblimin rotation. The Kaiser rule and a scree plot were considered when defining the factor structure. Reliability was evaluated using the McDonald omega coefficient test. Convergent validity was assessed through the association between adapted measures and other clinical measures (ie, PCL-5, PHQ-9). In addition, associations between the PCL-5 and PHQ-9 were examined as they related to the MIES and MIES-LIP.

RESULTS

Table 1 describes demographic characteristics of the study sample. Rates of potentially morally injurious experiences and the expression of moral injury in the legal context are presented in Table 2. Witnessing PMIEs while in the legal system was nearly ubiquitous, with > 90% of the sample endorsing this experience. More than half of the sample also endorsed engaging in morally injurious behavior by commission or omission, as well as experiencing betrayal while involved with the legal system.

Factor Analysis

Confirmatory factor analysis (CFA) was utilized to test the factor structure of the adapted MIES-LIP in our sample compared to the published factor structures of the MIES.33 Results did not support the established factor structure. Analysis yielded unacceptable CFI (0.79), TLI (0.70), SRMR (0.14), and RMSEA (0.21). The unsatisfactory results of CFA warranted follow-up exploratory factor analysis (EFA) to examine the factor structure of the moral injury scales in this sample.

EFA of MIES-LIP

The factor structure of the MIES-LIP was examined using EFA. The factorability of the data was examined using the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO value = 0.75) and Bartlett Test of Sphericity (X² = 525.41; P < .001), both of which suggested that the data were appropriate for factor analysis. The number of factors to retain was selected based on the Kaiser criterion.⁴¹ After extraction, an Oblimin rotation was applied, given that we expected factors to be correlated. A 2-factor solution was found, explaining 65.76% of the common variance. All 9 items were retained as they had factor loadings > 0.30. Factor 1, comprised self-directed moral injury questions (3-6). Factor 2 comprised other directed moral injury questions (1, 2, 7-9) (Table 3). The factor correlation coefficient between Factor 1 and Factor 2 was 0.34, which supports utilizing an oblique rotation.

Reliability. We examined the reliability of the adapted MIES-LIP using measures of internal consistency, with both MIES-LIP factors demonstrating good reliability. The internal consistency of both factors of the MIES-LIP were found to be good (self-directed moral injury: Ω = 0.89; other-directed moral injury: Ω = 0.83).

Convergent Validity

Association between moral injury scales. A significant, moderate correlation was observed between all subscales of the MIES and MIES-LIP. Specifically, the self-directed moral injury factor of the MIES-LIP was associated with both the perceived transgressions (r = 0.41, P < .001) and the MIES perceived betrayals factors (r = 0.25, P < .05). Similarly, the other-directed moral injury factor of the MIES-LIP was associated with both the MIES perceived transgressions (r = 0.45, P < .001) and the MIES perceived betrayals factors (r = 0.45, P < .001).

Association with PTSD symptoms. All subscales of both the MIES and MIES-LIP were associated with PTSD symptom severity. The MIES perceived transgressions factor (r = 0.43, P < .001) and the perceived betrayals factor of the MIES (r = 0.39, P < .001) were moderately associated with the PCL-5. Mirroring this, the “self-directed moral injury” factor of the MIESLIP (r = 0.44, P < .001) and the “other-directed moral injury” factor of the MIES-LIP (r = 0.42, P < .001) were also positively associated with PCL-5.

Association with depression symptoms. All subscales of the MIES and MIES-LIP were also associated with depressive symptoms. The MIES perceived transgressions factor (r = 0.27, P < .01) and the MIES perceived betrayals factor (r = 0.23, P < .05) had a small association with the PHQ-9. In addition, the self-directed moral injury factor of the MIES-LIP (r = 0.40, P < .001) and the other-directed moral injury factor of the MIES-LIP (r = 0.31, P < .01) had small to moderate associations with the PCL-5.

DISCUSSION

Potentially morally injurious events appear to be a salient factor affecting legal-involved veterans. Among our sample, the vast majority of legal-involved veterans endorsed experiencing both legal- and military-related PMIEs. Witnessing or participating in a legal-related PMIE appears to be widespread among those who have experienced incarceration. The MIES-LIP yielded a 2-factor structure: self-directed moral injury and other-directed moral injury, in the evaluated population. The MIES-LIP showed similar psychometric performance to the MIES in our sample. Specifically, the MIES-LIP had good reliability and adequate convergent validity. While CFA did not confirm the anticipated factor structure of the MIES-LIP within our sample, EFA showed similarities in factor structure between the original and adapted measures. While further research and validation are needed, preliminary results show promise of the MIES-LIP in assessing legal-related moral injury.

Originally, the MIES was found to have a 2-factor structure, defined by perceived transgressions and perceived betrayals.³³ However, additional research has identified a 3-factor structure, where the betrayal factor is maintained, and the transgressions factor is divided into transgressions by others and by self.⁸ The factor structure of the MIES-LIP was more closely related to the factor structure, with transgressions by others and betrayal mapped onto the same factor (ie, other-directed moral injury).⁸ While further research is needed, it is possible that the nature of morally injurious events experienced in legal contexts are experienced more in terms of self vs other, compared to morally injurious events experienced by veterans or active-duty service members.

Accurately identifying the types of moral injury experienced in a legal context may be important for determining the differences in drivers of legal-related moral injury compared to military-related moral injury. For example, self-directed moral injury in legal contexts may include a variety of actions the individual initiated that led to conviction and incarceration (eg, a criminal offense), as well as behaviors performed or witnessed while incarcerated (eg, engaging in violence). Inconsistent with military populations where other-directed moral injury clusters with self-directed moral injury, other-directed moral injury clustered with betrayal in legal contexts in our sample. This discrepancy may result from differences in identification with the military vs legal system. When veterans witness fellow service members engaging in PMIEs (eg, physical violence towards civilians in a military setting), this may be similar to self-directed moral injury due to the veteran’s identification with the same military system as the perpetrator.⁴² When legal-involved veterans witness other incarcerated individuals engaging in PMIEs (eg, physical violence toward other inmates), this may be experienced as similar to betrayal due to lack of personal identification with the criminal-legal system. Additional research is needed to better understand how self- and other-related moral injury are associated with betrayal in legal contexts.

Another potential driver of legal-related moral injury may be culpability. In order for moral injury to occur, an individual must perceive that something has taken place that deeply violated their sense of right and wrong.¹ In terms of criminal offenses or even engaging in violent behavior while incarcerated, the potential for moral injury may differ based on whether an individual views themselves as culpable for the act(s).²⁹ This may further distinguish between self-directed and other-directed moral injury in legal contexts. In situations where the individual views themselves as culpable, self-directed moral injury may be higher. In situations where the individual does not view themselves as culpable, other-directed moral injury may be higher based on the perception that the legal system is unfairly punishing them. Further research is needed to clarify how an individual’s view of their culpability relates to moral injury, as well as to elucidate which aspects of military service and legal involvement are most closely associated with moral injury among legal-involved veterans.

While this study treated legal-related and military-related moral injury as distinct, it is possible moral injury may have a cumulative effect over time with individuals experiencing morally injurious events across different contexts (eg, military, legal involvement). This, in turn, may compound risk for moral injury. These cumulative experiences may result in increased negative outcomes such as exacerbated psychiatric symptoms, substance misuse, and elevated suicide risk. Future studies should examine differences between groups who have experienced moral injury in differing contexts, as well as those with multiple sources of moral injury.

Limitations

The sample for this study included only veterans. The number of veterans incarcerated is large and the focus on veterans also allowed for a more robust comparison of moral injury related to the legal system and the more traditional military-related moral injury. However, the generalizability of the findings to nonveterans cannot be assured. The study used a relatively small sample (N = 100), which was overwhelmingly male. Although the PCL-5 was utilized to examine traumatic stress symptoms, this measure was not anchored to a specific criterion A trauma nor was it anchored specifically to a morally injurious experience. For all participants, their most recent military service preceded their most recent legal involvement which could affect the associations between variables. Furthermore, while all participants endorsed prior legal involvement, many participants reported no combat exposure.

CONCLUSIONS

This study resulted in several key findings. First, legal-involved veterans endorsed high rates of experiencing legal-related morally injurious experiences. Second, our adapted measure displayed adequate psychometric strength and suggests that legal-related moral injury is a salient and distinct phenomenon affecting legal-involved veterans. These items may not capture all the nuances of legal-related moral injury. Qualitative interviews with legal-involved persons may help identify relevant areas of legal-related moral injury not reflected in the current instrument. The MIES-LIP represents a practical measure that may help clinicians identify and address legal-related moral injury when working with legal-involved veterans. Given the high prevalence of PMIEs among legal-involved veterans, further examination of whether current interventions for moral injury and novel treatments being developed are effective for this population is needed.

Following exposure to potentially morally injurious events (PMIEs), some individuals may experience moral injury, which represents negative psychological, social, behavioral, and occasionally spiritual impacts.¹ The consequences of PMIE exposure and moral injury are well documented. Individuals may begin to question the goodness and trustworthiness of oneself, others, or the world.¹ Examples of other sequelae include guilt, demoralization, spiritual pain, loss of trust in the self or others, and difficulties with forgiveness.^2-6 In addition, prior studies have found that moral injury is associated with an increased risk of suicidal thoughts and behaviors, posttraumatic stress disorder (PTSD) symptoms, spiritual distress, and interpersonal difficulties.^7-11

Moral injury was first conceptualized in relation to combat trauma. However in recent years it has been examined in other groups such as health care practitioners, educators, refugees, and law enforcement personnel.^12-17 Furthermore, there has been a recent call for the study of moral injury in other understudied groups. One such group is legal-involved individuals, defined as those who are currently involved or previously involved in the criminal justice system (ie, arrests, incarceration, parole, and probation).^1,18-22

Many veterans are also involved with the legal system. Specifically, veterans currently comprise about 8% of the incarcerated US population, with an estimated > 180,000 veterans in prisons or jails and even more on parole or probation.^23,24 Legal-involved veterans may be at heightened risk for homelessness, suicide, unemployment, and high prevalence rates of psychiatric diagnoses.^25-28

Limited research has explored exposure to PMIEs as part of the legal process and the resulting expression of moral injury. The circumstances leading to incarceration, interactions with the US legal system, the environment of prison itself, and the subsequent challenges faced by legal-involved individuals after release all provide ample opportunity for PMIEs to occur.¹⁸ For example, engaging in a criminal act may represent a PMIE, particularly in violent offenses that involve harm to another individual. Moreover, the process of being convicted and charged with an offense may serve as a powerful reminder of the PMIE and tie this event to the individual’s identity and future. Furthermore, the physical and social environment of prison itself (eg, being surrounded by other offenders, witnessing the perpetration of violence, participating in violence for survival) presents a myriad of opportunities for PMIEs to occur.¹⁸

The consequences of PMIEs in the context of legal involvement may also have bearing on a touchstone of moral injury: changes in one’s schema of the self and world.⁴ At a societal level, legal-involved individuals are, by definition, deemed “guilty” and held culpable for their offense, which may reinforce a negative change in one’s view of self and the world.²⁹ In line with identity theory, external negative appraisals about legal-involved individuals (eg, they are a danger to society, they cannot be trusted to do the right thing) may influence their self-perception.³⁰ Furthermore, the affective characteristics often found in the context of moral injury (eg, guilt, shame, anger, contempt) may be exacerbated by legal involvement.²⁹ Personal feelings of guilt and shame may be reinforced by receiving a verdict and sentence, as well as the negative perceptions of individuals around them (eg, disapproval from prior sources of social support). Additionally, feelings of betrayal and distrust towards the legal system may arise.

In sum, legal-involved veterans incur increased risk of moral injury due to the potential for exposure to PMIEs across multiple time points (eg, prior to military service, during military service, during arrest/sentencing, during imprisonment, and postincarceration). The stigma that accompanies legal involvement may limit access to treatment or a willingness to seek treatment for distress related to moral injury.²⁹ Additionally, repeated exposure to PMIEs and resulting moral injury may compound over time, potentially exacerbating psychosocial functioning and increasing the risk for psychosocial stressors (eg, homelessness, unemployment) and mental health disorders (eg, depression, substance misuse).³¹

Although numerous measures of moral injury have been developed, most require that respondents consider a specific context (eg, military experiences).³² Therefore, study of legal-related moral injury requires adaptation of existing instruments to the legal context. The original and most commonly used scale of moral injury is the Moral Injury Events Scale (MIES).³³ The MIES scales was originally developed to measure moral injury in military-related contexts but has since been adapted as a measure of exposure to context-specific PMIEs.³⁴

Unfortunately, there are no validated measures for assessing legal-related moral injury. Such a gap in understanding is problematic, as it may impact measurement of the prevalence of PMIEs in both clinical and research settings for this at-risk population. The goal of this study was to conduct a psychometric evaluation of an adapted version of the MIES for legal-involved persons (MIES-LIP).

METHODS

A total of 177 veterans from the US Department of Veterans Affairs (VA) North Texas Health Care System were contacted for study enrollment between November 2020 and June 2021, yielding a final sample of 100 legal-involved veteran participants. Adults aged ≥ 18 years who were US military veterans and had ≥ 1 prior felony conviction resulting in incarceration were included. Participants were excluded if they had symptoms of psychosis that would preclude meaningful participation.

The study collected data on participants’ demographic and clinical characteristics using a semistructured survey instrument. Each participant completed an instructor-led questionnaire in a session that lasted about 1.5 hours. Participants who completed the visit in person received a $50 cash voucher for their time. Participants who were unable to meet with the study coordinator in person were able to complete the visit via telephone and received a $25 digital gift card. Of the total 100 participants, 79 participants completed the interview in person, and 21 completed by telephone. No significant differences were found in assessment measures between administration methods. Written informed consent was obtained during all in-person visits. For those completing via telephone, a waiver of written informed consent was obtained. This study was approved by the VA North Texas Health Care System’s Institutional Review Board.

Measures

The Moral Injury Events Scale (MIES) is a 9-item self-report measure that assesses exposure to PMIEs.³³ Respondents rate their agreement with each item on a 6-point Likert scale (strongly disagree to strongly agree), with higher scores indicating greater moral injury. The MIES has a 2-factor structure: Factor 1 has 6 items on perceived transgressions and Factor 2 has 3 items on perceived betrayals.³³

Creation of Legal-Involved Moral Injury Measure. To create the MIES-LIP, items and instructions from the MIES were modified to address moral injury in the context of legal involvement.³³ Adaptations were finalized following consultation and approval by the authors of the original measure. Specifically, the instructions were changed to: “Please respond to these items based specifically in the context of your involvement with the legal system.” The instructions clarified that legal involvement could include experiences related to committing an offense, legal proceedings and sentencing, incarceration, or transitioning out of the legal system. This differs from the original measure, which focused on military experiences, with instructions stating: “Please respond to these items based specifically in the context of your military service (ie, events and experiences during enlistment, deployment, combat, etc).”

Other measures. The study collected data on demographic characteristics including sex, race and ethnicity, marital status, military service, combat experience, and legal involvement. PTSD symptom severity, based on the criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), was assessed using the PTSD Checklist for DSM-5 (PCL-5).^35,36 The PCL-5 is a 20-item self-report measure in which item scores are summed to create a total score. The PCL-5 has demonstrated strong psychometric properties, including good internal consistency, test-retest reliability convergent validity, and discriminant validity.^37,38

Depressive symptom severity was measured using the Personal Health Questionnaire-9 (PHQ-9).³⁹ The PHQ-9 is a 9-item self-report measure where item scores summed to create a total score. The PHQ-9 has demonstrated strong psychometric properties, including internal consistency and test-retest reliability.³⁹

STATISTICAL METHODS

Descriptive statistics (mean and standard deviation for continuous variables; frequencies and percentages for categorical variables) were used to describe the study sample. Factor analysis was conducted to evaluate the psychometric properties of the MIES-LIP. Confirmatory factor analysis (CFA) was used to determine whether the MEIS-LIP had a similar factor structure to the MIES.⁴⁰ Criteria for fit indices used for CFA include the Comparative Fit Index (CFI; values of > 0.95 suggest good fit), Tucker-Lewis index (TLI; values of > 0.95 suggest a good fit), root mean square error of approximation (RMSEA; values of ≥ 0.06 suggest good fit), and standardized root mean square residual (SRMR; values of ≥ 0.08 suggest good fit). With insufficient fit, subsequent exploratory factor analysis was conducted using maximum likelihood estimation with an Oblimin rotation. The Kaiser rule and a scree plot were considered when defining the factor structure. Reliability was evaluated using the McDonald omega coefficient test. Convergent validity was assessed through the association between adapted measures and other clinical measures (ie, PCL-5, PHQ-9). In addition, associations between the PCL-5 and PHQ-9 were examined as they related to the MIES and MIES-LIP.

RESULTS

Table 1 describes demographic characteristics of the study sample. Rates of potentially morally injurious experiences and the expression of moral injury in the legal context are presented in Table 2. Witnessing PMIEs while in the legal system was nearly ubiquitous, with > 90% of the sample endorsing this experience. More than half of the sample also endorsed engaging in morally injurious behavior by commission or omission, as well as experiencing betrayal while involved with the legal system.

Factor Analysis

Confirmatory factor analysis (CFA) was utilized to test the factor structure of the adapted MIES-LIP in our sample compared to the published factor structures of the MIES.33 Results did not support the established factor structure. Analysis yielded unacceptable CFI (0.79), TLI (0.70), SRMR (0.14), and RMSEA (0.21). The unsatisfactory results of CFA warranted follow-up exploratory factor analysis (EFA) to examine the factor structure of the moral injury scales in this sample.

EFA of MIES-LIP

The factor structure of the MIES-LIP was examined using EFA. The factorability of the data was examined using the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO value = 0.75) and Bartlett Test of Sphericity (X² = 525.41; P < .001), both of which suggested that the data were appropriate for factor analysis. The number of factors to retain was selected based on the Kaiser criterion.⁴¹ After extraction, an Oblimin rotation was applied, given that we expected factors to be correlated. A 2-factor solution was found, explaining 65.76% of the common variance. All 9 items were retained as they had factor loadings > 0.30. Factor 1, comprised self-directed moral injury questions (3-6). Factor 2 comprised other directed moral injury questions (1, 2, 7-9) (Table 3). The factor correlation coefficient between Factor 1 and Factor 2 was 0.34, which supports utilizing an oblique rotation.

Reliability. We examined the reliability of the adapted MIES-LIP using measures of internal consistency, with both MIES-LIP factors demonstrating good reliability. The internal consistency of both factors of the MIES-LIP were found to be good (self-directed moral injury: Ω = 0.89; other-directed moral injury: Ω = 0.83).

Convergent Validity

Association between moral injury scales. A significant, moderate correlation was observed between all subscales of the MIES and MIES-LIP. Specifically, the self-directed moral injury factor of the MIES-LIP was associated with both the perceived transgressions (r = 0.41, P < .001) and the MIES perceived betrayals factors (r = 0.25, P < .05). Similarly, the other-directed moral injury factor of the MIES-LIP was associated with both the MIES perceived transgressions (r = 0.45, P < .001) and the MIES perceived betrayals factors (r = 0.45, P < .001).

Association with PTSD symptoms. All subscales of both the MIES and MIES-LIP were associated with PTSD symptom severity. The MIES perceived transgressions factor (r = 0.43, P < .001) and the perceived betrayals factor of the MIES (r = 0.39, P < .001) were moderately associated with the PCL-5. Mirroring this, the “self-directed moral injury” factor of the MIESLIP (r = 0.44, P < .001) and the “other-directed moral injury” factor of the MIES-LIP (r = 0.42, P < .001) were also positively associated with PCL-5.

Association with depression symptoms. All subscales of the MIES and MIES-LIP were also associated with depressive symptoms. The MIES perceived transgressions factor (r = 0.27, P < .01) and the MIES perceived betrayals factor (r = 0.23, P < .05) had a small association with the PHQ-9. In addition, the self-directed moral injury factor of the MIES-LIP (r = 0.40, P < .001) and the other-directed moral injury factor of the MIES-LIP (r = 0.31, P < .01) had small to moderate associations with the PCL-5.

DISCUSSION

Potentially morally injurious events appear to be a salient factor affecting legal-involved veterans. Among our sample, the vast majority of legal-involved veterans endorsed experiencing both legal- and military-related PMIEs. Witnessing or participating in a legal-related PMIE appears to be widespread among those who have experienced incarceration. The MIES-LIP yielded a 2-factor structure: self-directed moral injury and other-directed moral injury, in the evaluated population. The MIES-LIP showed similar psychometric performance to the MIES in our sample. Specifically, the MIES-LIP had good reliability and adequate convergent validity. While CFA did not confirm the anticipated factor structure of the MIES-LIP within our sample, EFA showed similarities in factor structure between the original and adapted measures. While further research and validation are needed, preliminary results show promise of the MIES-LIP in assessing legal-related moral injury.

Originally, the MIES was found to have a 2-factor structure, defined by perceived transgressions and perceived betrayals.³³ However, additional research has identified a 3-factor structure, where the betrayal factor is maintained, and the transgressions factor is divided into transgressions by others and by self.⁸ The factor structure of the MIES-LIP was more closely related to the factor structure, with transgressions by others and betrayal mapped onto the same factor (ie, other-directed moral injury).⁸ While further research is needed, it is possible that the nature of morally injurious events experienced in legal contexts are experienced more in terms of self vs other, compared to morally injurious events experienced by veterans or active-duty service members.

Accurately identifying the types of moral injury experienced in a legal context may be important for determining the differences in drivers of legal-related moral injury compared to military-related moral injury. For example, self-directed moral injury in legal contexts may include a variety of actions the individual initiated that led to conviction and incarceration (eg, a criminal offense), as well as behaviors performed or witnessed while incarcerated (eg, engaging in violence). Inconsistent with military populations where other-directed moral injury clusters with self-directed moral injury, other-directed moral injury clustered with betrayal in legal contexts in our sample. This discrepancy may result from differences in identification with the military vs legal system. When veterans witness fellow service members engaging in PMIEs (eg, physical violence towards civilians in a military setting), this may be similar to self-directed moral injury due to the veteran’s identification with the same military system as the perpetrator.⁴² When legal-involved veterans witness other incarcerated individuals engaging in PMIEs (eg, physical violence toward other inmates), this may be experienced as similar to betrayal due to lack of personal identification with the criminal-legal system. Additional research is needed to better understand how self- and other-related moral injury are associated with betrayal in legal contexts.

Another potential driver of legal-related moral injury may be culpability. In order for moral injury to occur, an individual must perceive that something has taken place that deeply violated their sense of right and wrong.¹ In terms of criminal offenses or even engaging in violent behavior while incarcerated, the potential for moral injury may differ based on whether an individual views themselves as culpable for the act(s).²⁹ This may further distinguish between self-directed and other-directed moral injury in legal contexts. In situations where the individual views themselves as culpable, self-directed moral injury may be higher. In situations where the individual does not view themselves as culpable, other-directed moral injury may be higher based on the perception that the legal system is unfairly punishing them. Further research is needed to clarify how an individual’s view of their culpability relates to moral injury, as well as to elucidate which aspects of military service and legal involvement are most closely associated with moral injury among legal-involved veterans.

While this study treated legal-related and military-related moral injury as distinct, it is possible moral injury may have a cumulative effect over time with individuals experiencing morally injurious events across different contexts (eg, military, legal involvement). This, in turn, may compound risk for moral injury. These cumulative experiences may result in increased negative outcomes such as exacerbated psychiatric symptoms, substance misuse, and elevated suicide risk. Future studies should examine differences between groups who have experienced moral injury in differing contexts, as well as those with multiple sources of moral injury.

Limitations

The sample for this study included only veterans. The number of veterans incarcerated is large and the focus on veterans also allowed for a more robust comparison of moral injury related to the legal system and the more traditional military-related moral injury. However, the generalizability of the findings to nonveterans cannot be assured. The study used a relatively small sample (N = 100), which was overwhelmingly male. Although the PCL-5 was utilized to examine traumatic stress symptoms, this measure was not anchored to a specific criterion A trauma nor was it anchored specifically to a morally injurious experience. For all participants, their most recent military service preceded their most recent legal involvement which could affect the associations between variables. Furthermore, while all participants endorsed prior legal involvement, many participants reported no combat exposure.

CONCLUSIONS

This study resulted in several key findings. First, legal-involved veterans endorsed high rates of experiencing legal-related morally injurious experiences. Second, our adapted measure displayed adequate psychometric strength and suggests that legal-related moral injury is a salient and distinct phenomenon affecting legal-involved veterans. These items may not capture all the nuances of legal-related moral injury. Qualitative interviews with legal-involved persons may help identify relevant areas of legal-related moral injury not reflected in the current instrument. The MIES-LIP represents a practical measure that may help clinicians identify and address legal-related moral injury when working with legal-involved veterans. Given the high prevalence of PMIEs among legal-involved veterans, further examination of whether current interventions for moral injury and novel treatments being developed are effective for this population is needed.