Peter Speelman, MD, PhD

Hospital and Patient Inclusion

This study was approved by the medical ethics committee of the Academic Medical Center, Amsterdam, the Netherlands. After approval, we composed a representative sample out of the total of 93 Dutch hospitals, by including hospitals through a blind drawing. Hospitals were divided into 3 categories: (1) university hospitals, (2) nonuniversity teaching hospitals and (3) nonteaching hospitals. The teaching status of nonuniversity hospitals was based on the (non) presence of an internal or surgical medicine residency training program. After the drawing, each group contained 30% of the total number of Dutch hospitals in its category (source: RIVM, Nationale Atlas Volksgezondheid, 2007).

Subsequently, splenectomized patients were included retrospectively using the Dutch Pathology Registry, since spleens are routinely sent to pathology after removal. In this Registry, a search query *milt* (spleen) was performed, after which all splenectomies performed from 1997 to 2008 were selected and nonrelevant hits such as partial splenectomies or spleen biopsies were removed.

Data Collection

After hospitals and patients were identified, the medical file and all discharge correspondence were assessed on site. All data were collected separately for each hospital by the same 2 investigators (DV, JL) using a standardized survey form. To investigate discharge correspondence, discharge letters as well as all other correspondence up to at least 1 year after splenectomy were included, for example from follow‐up out‐patient visits.

After hospital category was documented, we registered for each hospital the size of the surgical staff at the time of inclusion, the availability of any form of protocol of the surgical department reflecting hospital post‐splenectomy policy, and the practice of systematically registering (surgical) complications by the department of surgery. Patient data included demographics, documentation of vaccine administration and documentation of the prescription of antibiotics. Furthermore, discharge correspondence was checked for mentioning of each of the following: performed splenectomy, vaccination status, the need for revaccination, prescribed prophylactic antibiotics, the need of urgent use of antibiotics in case of suspected infection, and the advice for annual flu‐vaccination.

Data Analysis

When computing vaccination rates, we included only those patients who survived the first 2 weeks after surgery, since correct vaccination is considered by the British Committee to be given 2 weeks prior to or at least 2 or more weeks after surgery. Pneumococcal vaccination was defined as immunization with either the 23‐valent pneumococcal polysaccharides vaccine (PPV‐23, Pneumovax), the 7‐valent pneumococcal conjugate vaccine (PCV‐7, Prevenar), or both.

Prophylactic antibiotics were defined as a prescription of antibiotics for the first 2 years after splenectomy. On demand antibiotics were defined as a prescription to be given to the patient at discharge, to use in case of suspected infection. When investigating prescription rates of prophylactic antibiotics, we excluded patients deceased in the first 2 weeks after surgery, regarding their death as a complication of surgery. In case of on demand antibiotics, patients who died in the hospital before their discharge were excluded as well. When investigating discharge information to the general practitioner (GP), only those patients alive at time of discharge were excluded.

Statistical Analysis

First, we have described the study sample using standard descriptive statistics. Second, to explore differences in performance and calculate P values, we used a chi‐square test between the 3 categories of hospitals (Table 2), between presence or absence of a protocol (Table 3) and complication registry. The influence of surgical staff size (divided into 1‐8 surgeons or >8 surgeons) was calculated using multivariate logistic regression analysis, where surgical staff size and hospital teaching status were used as covariates in the analysis. All statistical analysis of data was performed in SPSS 16.0.

Differences Between University and Nonteaching Hospitals

Hospital performance of Dutch university, nonuniversity teaching, and nonteaching hospitals is shown in Table 2. Admission to a university hospital is associated with better guideline adherence: 22 of 33 of patients (66.7%) in university hospitals were immunized with H. influenzae B as compared to 108 of 268 patients (40.3%) in nonuniversity teaching and 66 of 197 (33.5%) in nonteaching hospitals. Vaccination with N. meningitidis C occurred in 21 of 33 patients (63.6%) as compared to 82 of 268 patients (30.6%) in nonuniversity teaching and 58 of 197 (29.4%) in nonteaching hospitals. In 53.1% of patients no antibiotics were prescribed in university hospitals, as compared to 72.6% in nonuniversity teaching and 82.5% in nonteaching hospitals. Differences between nonuniversity teaching hospitals and nonteaching hospitals were small.

Presence of a Post‐Splenectomy Protocol

The availability of a protocol at the surgical department was not associated with higher vaccination rates (Table 3). It did however show a positive relation on the prescription of prophylactic antibiotics. The effect of a protocol on the quality of discharge information to the GP was minimal.

Size of Surgical Staff

Performance in relation to the size of surgical staff was determined (data not shown). There were no differences in vaccination rates or quality of discharge information between the groups of different sizes (less or more than 8 surgeons). Larger surgical groups seemed to perform better in prescribing antibiotics, however when adjusting for hospital category in multivariate analysis these differences were not significant.

Complication Registry

Complications were systematically registered by all but 2 surgical departments in nonteaching hospitals, composing a cohort of 27 patients.

Although numbers are low, it demonstrates that in the absence of a registry, the guideline adherence for this group of patients was similar, and only prophylactic antibiotics were significantly less prescribed: 62 of 473 patients (13.1%) in the presence of a registry, as compared to 0 of 27 patients in absence of a registry (P value = 0.044) (data not shown). The precise role of the registry herein remains unclear, since both hospitals also lacked a hospital post‐splenectomy protocol.

Main Findings

The aim of the present study was to investigate quality of care for splenectomized patients in Dutch hospitals with different teaching status. In general, beneficial effects of teaching status only extended to university hospitals in the Netherlands. Other teaching hospitals performed similarly to nonteaching hospitals in the Netherlands. Hospitals in which the surgical department developed a local protocol with recommendations for managing patients after splenectomy did not achieve higher vaccination rates. There was, however, an improvement in prescription of antibiotics and in the quality of discharge correspondence from the hospital to the GP. Surgical staff size was not related to hospital performance.

Explanation of Results

In the Netherlands, all categories of hospitals provided over 80% of their post‐splenectomy patients with pneumococcal immunization, reflecting that Dutch physicians in general are aware of the need for pneumococcal protection after splenectomy. However, university hospitals had better performance results regarding immunizing patients with all 3 recommended vaccines, as well as prescribing prophylactic antibiotics in combination with a prescription for on‐demand antibiotics. Collectively, university hospitals offered their patients a more complete post‐splenectomy treatment.

It has been described elsewhere that minor teaching and nonteaching hospitals show small differences, and that nonteaching hospitals even perform better at certain indicators than minor teaching hospitals.13 We indeed found small differences between nonuniversity teaching hospitals and nonteaching hospitals, where nonuniversity teaching hospitals performed better at prescribing antibiotics, and nonteaching hospitals did better at giving recommendations to the GP on booster immunization and use of on‐demand antibiotics.

Hospital characteristics have been shown to have important effects on hospital outcomes.10, 14, 15 We hypothesized this would also be the case regarding the adherence to post‐splenectomy management recommendations. In particular, we were expecting to find that the availability of a protocol at the department of surgery would be associated with better compliance with all key recommendations in the British Standards, however, vaccination rates did not differ form departments without a protocol. The items that were generally most eligible for improvement seemed to benefit most from the presence of a protocol.

Neither the presence of a protocol nor the size of the surgical staff were related to better performance in university hospitals. We can therefore only speculate about the explanation for the differences found between university and other hospitals. Organizational differences may not be disregarded; it has been described elsewhere that better quality and processes of care are delivered in major teaching hospitals.16, 9, 12 Most prior studies have reported a lower risk‐adjusted mortality in major teaching hospitals as compared with minor teaching or nonteaching hospitals.9, 12 It is also possible that residency and fellowship programs contribute to better compliance of guidelines and have a favorable impact on the delivery of patient care in teaching hospitals.12

Limitations

In absence of a Dutch guideline we chose to investigate adherence to the recommendations by the British Committee for Standards in Haematology, assuming that Dutch professionals have some knowledge of these recommendations. Although these recommendations are internationally considered to reflect current best practice and patients should therefore be managed according to at least comparable standards, the extent of familiarity and use of the British standards by Dutch physicians remains to be investigated in the future. Furthermore, we investigated the availability of a locally designed protocol on the management of post‐splenectomy patients by the surgical department. Checking the contents of each of these local protocols was not part of our study and thus we can not exclude that these protocols are lacking certain recommendations. It also remains unclear how hospitals have implemented their protocols.

Implications for Future Research and Policy

In the Netherlands, hospitals could offer better quality of care for hyposplenic and asplenic patients in the prevention of infections by increasing immunization rates. Furthermore, although academic centers performed better than the other hospital categories, only a minority of patients were given or advised to receive on demand antibiotics. Here lies a tremendous opportunity to improve patient care in the prevention of severe infections.

Potential barriers that exist for delivering optimal care to these patients remain to be investigated. Furthermore, although teaching status is related to performance, the explanation for this difference remains unclear. The results of this study suggest that there is a relation between characteristics of practice organization and performance, but these characteristics should be further elucidated.

Hospital and Patient Inclusion

This study was approved by the medical ethics committee of the Academic Medical Center, Amsterdam, the Netherlands. After approval, we composed a representative sample out of the total of 93 Dutch hospitals, by including hospitals through a blind drawing. Hospitals were divided into 3 categories: (1) university hospitals, (2) nonuniversity teaching hospitals and (3) nonteaching hospitals. The teaching status of nonuniversity hospitals was based on the (non) presence of an internal or surgical medicine residency training program. After the drawing, each group contained 30% of the total number of Dutch hospitals in its category (source: RIVM, Nationale Atlas Volksgezondheid, 2007).

Subsequently, splenectomized patients were included retrospectively using the Dutch Pathology Registry, since spleens are routinely sent to pathology after removal. In this Registry, a search query *milt* (spleen) was performed, after which all splenectomies performed from 1997 to 2008 were selected and nonrelevant hits such as partial splenectomies or spleen biopsies were removed.

Data Collection

After hospitals and patients were identified, the medical file and all discharge correspondence were assessed on site. All data were collected separately for each hospital by the same 2 investigators (DV, JL) using a standardized survey form. To investigate discharge correspondence, discharge letters as well as all other correspondence up to at least 1 year after splenectomy were included, for example from follow‐up out‐patient visits.

After hospital category was documented, we registered for each hospital the size of the surgical staff at the time of inclusion, the availability of any form of protocol of the surgical department reflecting hospital post‐splenectomy policy, and the practice of systematically registering (surgical) complications by the department of surgery. Patient data included demographics, documentation of vaccine administration and documentation of the prescription of antibiotics. Furthermore, discharge correspondence was checked for mentioning of each of the following: performed splenectomy, vaccination status, the need for revaccination, prescribed prophylactic antibiotics, the need of urgent use of antibiotics in case of suspected infection, and the advice for annual flu‐vaccination.

Data Analysis

When computing vaccination rates, we included only those patients who survived the first 2 weeks after surgery, since correct vaccination is considered by the British Committee to be given 2 weeks prior to or at least 2 or more weeks after surgery. Pneumococcal vaccination was defined as immunization with either the 23‐valent pneumococcal polysaccharides vaccine (PPV‐23, Pneumovax), the 7‐valent pneumococcal conjugate vaccine (PCV‐7, Prevenar), or both.

Prophylactic antibiotics were defined as a prescription of antibiotics for the first 2 years after splenectomy. On demand antibiotics were defined as a prescription to be given to the patient at discharge, to use in case of suspected infection. When investigating prescription rates of prophylactic antibiotics, we excluded patients deceased in the first 2 weeks after surgery, regarding their death as a complication of surgery. In case of on demand antibiotics, patients who died in the hospital before their discharge were excluded as well. When investigating discharge information to the general practitioner (GP), only those patients alive at time of discharge were excluded.

Statistical Analysis

First, we have described the study sample using standard descriptive statistics. Second, to explore differences in performance and calculate P values, we used a chi‐square test between the 3 categories of hospitals (Table 2), between presence or absence of a protocol (Table 3) and complication registry. The influence of surgical staff size (divided into 1‐8 surgeons or >8 surgeons) was calculated using multivariate logistic regression analysis, where surgical staff size and hospital teaching status were used as covariates in the analysis. All statistical analysis of data was performed in SPSS 16.0.

Differences Between University and Nonteaching Hospitals

Hospital performance of Dutch university, nonuniversity teaching, and nonteaching hospitals is shown in Table 2. Admission to a university hospital is associated with better guideline adherence: 22 of 33 of patients (66.7%) in university hospitals were immunized with H. influenzae B as compared to 108 of 268 patients (40.3%) in nonuniversity teaching and 66 of 197 (33.5%) in nonteaching hospitals. Vaccination with N. meningitidis C occurred in 21 of 33 patients (63.6%) as compared to 82 of 268 patients (30.6%) in nonuniversity teaching and 58 of 197 (29.4%) in nonteaching hospitals. In 53.1% of patients no antibiotics were prescribed in university hospitals, as compared to 72.6% in nonuniversity teaching and 82.5% in nonteaching hospitals. Differences between nonuniversity teaching hospitals and nonteaching hospitals were small.

Presence of a Post‐Splenectomy Protocol

The availability of a protocol at the surgical department was not associated with higher vaccination rates (Table 3). It did however show a positive relation on the prescription of prophylactic antibiotics. The effect of a protocol on the quality of discharge information to the GP was minimal.

Size of Surgical Staff

Performance in relation to the size of surgical staff was determined (data not shown). There were no differences in vaccination rates or quality of discharge information between the groups of different sizes (less or more than 8 surgeons). Larger surgical groups seemed to perform better in prescribing antibiotics, however when adjusting for hospital category in multivariate analysis these differences were not significant.

Complication Registry

Complications were systematically registered by all but 2 surgical departments in nonteaching hospitals, composing a cohort of 27 patients.

Although numbers are low, it demonstrates that in the absence of a registry, the guideline adherence for this group of patients was similar, and only prophylactic antibiotics were significantly less prescribed: 62 of 473 patients (13.1%) in the presence of a registry, as compared to 0 of 27 patients in absence of a registry (P value = 0.044) (data not shown). The precise role of the registry herein remains unclear, since both hospitals also lacked a hospital post‐splenectomy protocol.

Main Findings

The aim of the present study was to investigate quality of care for splenectomized patients in Dutch hospitals with different teaching status. In general, beneficial effects of teaching status only extended to university hospitals in the Netherlands. Other teaching hospitals performed similarly to nonteaching hospitals in the Netherlands. Hospitals in which the surgical department developed a local protocol with recommendations for managing patients after splenectomy did not achieve higher vaccination rates. There was, however, an improvement in prescription of antibiotics and in the quality of discharge correspondence from the hospital to the GP. Surgical staff size was not related to hospital performance.

Explanation of Results

In the Netherlands, all categories of hospitals provided over 80% of their post‐splenectomy patients with pneumococcal immunization, reflecting that Dutch physicians in general are aware of the need for pneumococcal protection after splenectomy. However, university hospitals had better performance results regarding immunizing patients with all 3 recommended vaccines, as well as prescribing prophylactic antibiotics in combination with a prescription for on‐demand antibiotics. Collectively, university hospitals offered their patients a more complete post‐splenectomy treatment.

It has been described elsewhere that minor teaching and nonteaching hospitals show small differences, and that nonteaching hospitals even perform better at certain indicators than minor teaching hospitals.13 We indeed found small differences between nonuniversity teaching hospitals and nonteaching hospitals, where nonuniversity teaching hospitals performed better at prescribing antibiotics, and nonteaching hospitals did better at giving recommendations to the GP on booster immunization and use of on‐demand antibiotics.

Hospital characteristics have been shown to have important effects on hospital outcomes.10, 14, 15 We hypothesized this would also be the case regarding the adherence to post‐splenectomy management recommendations. In particular, we were expecting to find that the availability of a protocol at the department of surgery would be associated with better compliance with all key recommendations in the British Standards, however, vaccination rates did not differ form departments without a protocol. The items that were generally most eligible for improvement seemed to benefit most from the presence of a protocol.

Neither the presence of a protocol nor the size of the surgical staff were related to better performance in university hospitals. We can therefore only speculate about the explanation for the differences found between university and other hospitals. Organizational differences may not be disregarded; it has been described elsewhere that better quality and processes of care are delivered in major teaching hospitals.16, 9, 12 Most prior studies have reported a lower risk‐adjusted mortality in major teaching hospitals as compared with minor teaching or nonteaching hospitals.9, 12 It is also possible that residency and fellowship programs contribute to better compliance of guidelines and have a favorable impact on the delivery of patient care in teaching hospitals.12

Limitations

In absence of a Dutch guideline we chose to investigate adherence to the recommendations by the British Committee for Standards in Haematology, assuming that Dutch professionals have some knowledge of these recommendations. Although these recommendations are internationally considered to reflect current best practice and patients should therefore be managed according to at least comparable standards, the extent of familiarity and use of the British standards by Dutch physicians remains to be investigated in the future. Furthermore, we investigated the availability of a locally designed protocol on the management of post‐splenectomy patients by the surgical department. Checking the contents of each of these local protocols was not part of our study and thus we can not exclude that these protocols are lacking certain recommendations. It also remains unclear how hospitals have implemented their protocols.

Implications for Future Research and Policy

In the Netherlands, hospitals could offer better quality of care for hyposplenic and asplenic patients in the prevention of infections by increasing immunization rates. Furthermore, although academic centers performed better than the other hospital categories, only a minority of patients were given or advised to receive on demand antibiotics. Here lies a tremendous opportunity to improve patient care in the prevention of severe infections.

Potential barriers that exist for delivering optimal care to these patients remain to be investigated. Furthermore, although teaching status is related to performance, the explanation for this difference remains unclear. The results of this study suggest that there is a relation between characteristics of practice organization and performance, but these characteristics should be further elucidated.

Hospital and Patient Inclusion

This study was approved by the medical ethics committee of the Academic Medical Center, Amsterdam, the Netherlands. After approval, we composed a representative sample out of the total of 93 Dutch hospitals, by including hospitals through a blind drawing. Hospitals were divided into 3 categories: (1) university hospitals, (2) nonuniversity teaching hospitals and (3) nonteaching hospitals. The teaching status of nonuniversity hospitals was based on the (non) presence of an internal or surgical medicine residency training program. After the drawing, each group contained 30% of the total number of Dutch hospitals in its category (source: RIVM, Nationale Atlas Volksgezondheid, 2007).

Subsequently, splenectomized patients were included retrospectively using the Dutch Pathology Registry, since spleens are routinely sent to pathology after removal. In this Registry, a search query *milt* (spleen) was performed, after which all splenectomies performed from 1997 to 2008 were selected and nonrelevant hits such as partial splenectomies or spleen biopsies were removed.

Data Collection

After hospitals and patients were identified, the medical file and all discharge correspondence were assessed on site. All data were collected separately for each hospital by the same 2 investigators (DV, JL) using a standardized survey form. To investigate discharge correspondence, discharge letters as well as all other correspondence up to at least 1 year after splenectomy were included, for example from follow‐up out‐patient visits.

After hospital category was documented, we registered for each hospital the size of the surgical staff at the time of inclusion, the availability of any form of protocol of the surgical department reflecting hospital post‐splenectomy policy, and the practice of systematically registering (surgical) complications by the department of surgery. Patient data included demographics, documentation of vaccine administration and documentation of the prescription of antibiotics. Furthermore, discharge correspondence was checked for mentioning of each of the following: performed splenectomy, vaccination status, the need for revaccination, prescribed prophylactic antibiotics, the need of urgent use of antibiotics in case of suspected infection, and the advice for annual flu‐vaccination.

Data Analysis

When computing vaccination rates, we included only those patients who survived the first 2 weeks after surgery, since correct vaccination is considered by the British Committee to be given 2 weeks prior to or at least 2 or more weeks after surgery. Pneumococcal vaccination was defined as immunization with either the 23‐valent pneumococcal polysaccharides vaccine (PPV‐23, Pneumovax), the 7‐valent pneumococcal conjugate vaccine (PCV‐7, Prevenar), or both.

Prophylactic antibiotics were defined as a prescription of antibiotics for the first 2 years after splenectomy. On demand antibiotics were defined as a prescription to be given to the patient at discharge, to use in case of suspected infection. When investigating prescription rates of prophylactic antibiotics, we excluded patients deceased in the first 2 weeks after surgery, regarding their death as a complication of surgery. In case of on demand antibiotics, patients who died in the hospital before their discharge were excluded as well. When investigating discharge information to the general practitioner (GP), only those patients alive at time of discharge were excluded.

Statistical Analysis

First, we have described the study sample using standard descriptive statistics. Second, to explore differences in performance and calculate P values, we used a chi‐square test between the 3 categories of hospitals (Table 2), between presence or absence of a protocol (Table 3) and complication registry. The influence of surgical staff size (divided into 1‐8 surgeons or >8 surgeons) was calculated using multivariate logistic regression analysis, where surgical staff size and hospital teaching status were used as covariates in the analysis. All statistical analysis of data was performed in SPSS 16.0.

Differences Between University and Nonteaching Hospitals

Hospital performance of Dutch university, nonuniversity teaching, and nonteaching hospitals is shown in Table 2. Admission to a university hospital is associated with better guideline adherence: 22 of 33 of patients (66.7%) in university hospitals were immunized with H. influenzae B as compared to 108 of 268 patients (40.3%) in nonuniversity teaching and 66 of 197 (33.5%) in nonteaching hospitals. Vaccination with N. meningitidis C occurred in 21 of 33 patients (63.6%) as compared to 82 of 268 patients (30.6%) in nonuniversity teaching and 58 of 197 (29.4%) in nonteaching hospitals. In 53.1% of patients no antibiotics were prescribed in university hospitals, as compared to 72.6% in nonuniversity teaching and 82.5% in nonteaching hospitals. Differences between nonuniversity teaching hospitals and nonteaching hospitals were small.

Presence of a Post‐Splenectomy Protocol

The availability of a protocol at the surgical department was not associated with higher vaccination rates (Table 3). It did however show a positive relation on the prescription of prophylactic antibiotics. The effect of a protocol on the quality of discharge information to the GP was minimal.

Size of Surgical Staff

Performance in relation to the size of surgical staff was determined (data not shown). There were no differences in vaccination rates or quality of discharge information between the groups of different sizes (less or more than 8 surgeons). Larger surgical groups seemed to perform better in prescribing antibiotics, however when adjusting for hospital category in multivariate analysis these differences were not significant.

Complication Registry

Complications were systematically registered by all but 2 surgical departments in nonteaching hospitals, composing a cohort of 27 patients.

Although numbers are low, it demonstrates that in the absence of a registry, the guideline adherence for this group of patients was similar, and only prophylactic antibiotics were significantly less prescribed: 62 of 473 patients (13.1%) in the presence of a registry, as compared to 0 of 27 patients in absence of a registry (P value = 0.044) (data not shown). The precise role of the registry herein remains unclear, since both hospitals also lacked a hospital post‐splenectomy protocol.

Main Findings

The aim of the present study was to investigate quality of care for splenectomized patients in Dutch hospitals with different teaching status. In general, beneficial effects of teaching status only extended to university hospitals in the Netherlands. Other teaching hospitals performed similarly to nonteaching hospitals in the Netherlands. Hospitals in which the surgical department developed a local protocol with recommendations for managing patients after splenectomy did not achieve higher vaccination rates. There was, however, an improvement in prescription of antibiotics and in the quality of discharge correspondence from the hospital to the GP. Surgical staff size was not related to hospital performance.

Explanation of Results

In the Netherlands, all categories of hospitals provided over 80% of their post‐splenectomy patients with pneumococcal immunization, reflecting that Dutch physicians in general are aware of the need for pneumococcal protection after splenectomy. However, university hospitals had better performance results regarding immunizing patients with all 3 recommended vaccines, as well as prescribing prophylactic antibiotics in combination with a prescription for on‐demand antibiotics. Collectively, university hospitals offered their patients a more complete post‐splenectomy treatment.

It has been described elsewhere that minor teaching and nonteaching hospitals show small differences, and that nonteaching hospitals even perform better at certain indicators than minor teaching hospitals.13 We indeed found small differences between nonuniversity teaching hospitals and nonteaching hospitals, where nonuniversity teaching hospitals performed better at prescribing antibiotics, and nonteaching hospitals did better at giving recommendations to the GP on booster immunization and use of on‐demand antibiotics.

Hospital characteristics have been shown to have important effects on hospital outcomes.10, 14, 15 We hypothesized this would also be the case regarding the adherence to post‐splenectomy management recommendations. In particular, we were expecting to find that the availability of a protocol at the department of surgery would be associated with better compliance with all key recommendations in the British Standards, however, vaccination rates did not differ form departments without a protocol. The items that were generally most eligible for improvement seemed to benefit most from the presence of a protocol.

Neither the presence of a protocol nor the size of the surgical staff were related to better performance in university hospitals. We can therefore only speculate about the explanation for the differences found between university and other hospitals. Organizational differences may not be disregarded; it has been described elsewhere that better quality and processes of care are delivered in major teaching hospitals.16, 9, 12 Most prior studies have reported a lower risk‐adjusted mortality in major teaching hospitals as compared with minor teaching or nonteaching hospitals.9, 12 It is also possible that residency and fellowship programs contribute to better compliance of guidelines and have a favorable impact on the delivery of patient care in teaching hospitals.12

Limitations

In absence of a Dutch guideline we chose to investigate adherence to the recommendations by the British Committee for Standards in Haematology, assuming that Dutch professionals have some knowledge of these recommendations. Although these recommendations are internationally considered to reflect current best practice and patients should therefore be managed according to at least comparable standards, the extent of familiarity and use of the British standards by Dutch physicians remains to be investigated in the future. Furthermore, we investigated the availability of a locally designed protocol on the management of post‐splenectomy patients by the surgical department. Checking the contents of each of these local protocols was not part of our study and thus we can not exclude that these protocols are lacking certain recommendations. It also remains unclear how hospitals have implemented their protocols.

Implications for Future Research and Policy

In the Netherlands, hospitals could offer better quality of care for hyposplenic and asplenic patients in the prevention of infections by increasing immunization rates. Furthermore, although academic centers performed better than the other hospital categories, only a minority of patients were given or advised to receive on demand antibiotics. Here lies a tremendous opportunity to improve patient care in the prevention of severe infections.

Potential barriers that exist for delivering optimal care to these patients remain to be investigated. Furthermore, although teaching status is related to performance, the explanation for this difference remains unclear. The results of this study suggest that there is a relation between characteristics of practice organization and performance, but these characteristics should be further elucidated.