MDedge Rheumatology

You know what the happiest animal in the world is? A goldfish. You know why? It’s got a 10-second memory. Be a goldfish. — Ted Lasso

I don’t play much golf. When I do, it’s when my dad is in town. He shoots his age (78). I shoot double mine (52). He was recently here. We played and watched the Masters where he pointed out how I looked a lot like Scottie Scheffler, the now two-time Masters champion. On the 10th hole of his third round, you could see the resemblance. Scheffler’s third shot flew past the hole into the galley. He rifled the fourth past the hole on its way back toward the fairway. It was now a good distance further from the cup than a minute ago. He proceeded to misread his bogey putt, ending his misery with a double bogey. Scheffler went on to bogey the next hole and dropped from first on the leaderboard to fifth. Yes, I looked just like that on my last round. But here is where Scheffler and I differ. After a hole like that, I’d have been apoplectic, seething with self loathing. Scheffler was not. He kept moving. Head up, he sauntered to the next hole as if he had no awareness of what just transpired.

The ability to compartmentalize is useful not only to become the Masters champion, but also to become master of your day. In this way, golf is a nice approximation for life. The best golfers in the world will always have horrible shots and dreadful holes. The winning ones are often those who recover rather than continue in a downward spiral of one bad shot after another.

Dr. Benabio

Kaiser Permanente

Scheffler went on to eagle the 13th hole on that round. He eventually won the 2024 Masters Tournament. Remember that the next time you find yourself in a day that feels like it is spiraling toward disaster. Close the door on the compartment that was the last miserable hole and saunter to the next patient like it never happened.

And maybe close the clubface a bit on address for your next drive. 

Dr. Benabio is director of Healthcare Transformation and chief of dermatology at Kaiser Permanente San Diego. The opinions expressed in this column are his own and do not represent those of Kaiser Permanente. Dr. Benabio is @Dermdoc on X. Write to him at [email protected].

You know what the happiest animal in the world is? A goldfish. You know why? It’s got a 10-second memory. Be a goldfish. — Ted Lasso

I don’t play much golf. When I do, it’s when my dad is in town. He shoots his age (78). I shoot double mine (52). He was recently here. We played and watched the Masters where he pointed out how I looked a lot like Scottie Scheffler, the now two-time Masters champion. On the 10th hole of his third round, you could see the resemblance. Scheffler’s third shot flew past the hole into the galley. He rifled the fourth past the hole on its way back toward the fairway. It was now a good distance further from the cup than a minute ago. He proceeded to misread his bogey putt, ending his misery with a double bogey. Scheffler went on to bogey the next hole and dropped from first on the leaderboard to fifth. Yes, I looked just like that on my last round. But here is where Scheffler and I differ. After a hole like that, I’d have been apoplectic, seething with self loathing. Scheffler was not. He kept moving. Head up, he sauntered to the next hole as if he had no awareness of what just transpired.

The ability to compartmentalize is useful not only to become the Masters champion, but also to become master of your day. In this way, golf is a nice approximation for life. The best golfers in the world will always have horrible shots and dreadful holes. The winning ones are often those who recover rather than continue in a downward spiral of one bad shot after another.

Dr. Benabio

Kaiser Permanente

Scheffler went on to eagle the 13th hole on that round. He eventually won the 2024 Masters Tournament. Remember that the next time you find yourself in a day that feels like it is spiraling toward disaster. Close the door on the compartment that was the last miserable hole and saunter to the next patient like it never happened.

And maybe close the clubface a bit on address for your next drive. 

Dr. Benabio is director of Healthcare Transformation and chief of dermatology at Kaiser Permanente San Diego. The opinions expressed in this column are his own and do not represent those of Kaiser Permanente. Dr. Benabio is @Dermdoc on X. Write to him at [email protected].

You know what the happiest animal in the world is? A goldfish. You know why? It’s got a 10-second memory. Be a goldfish. — Ted Lasso

I don’t play much golf. When I do, it’s when my dad is in town. He shoots his age (78). I shoot double mine (52). He was recently here. We played and watched the Masters where he pointed out how I looked a lot like Scottie Scheffler, the now two-time Masters champion. On the 10th hole of his third round, you could see the resemblance. Scheffler’s third shot flew past the hole into the galley. He rifled the fourth past the hole on its way back toward the fairway. It was now a good distance further from the cup than a minute ago. He proceeded to misread his bogey putt, ending his misery with a double bogey. Scheffler went on to bogey the next hole and dropped from first on the leaderboard to fifth. Yes, I looked just like that on my last round. But here is where Scheffler and I differ. After a hole like that, I’d have been apoplectic, seething with self loathing. Scheffler was not. He kept moving. Head up, he sauntered to the next hole as if he had no awareness of what just transpired.

The ability to compartmentalize is useful not only to become the Masters champion, but also to become master of your day. In this way, golf is a nice approximation for life. The best golfers in the world will always have horrible shots and dreadful holes. The winning ones are often those who recover rather than continue in a downward spiral of one bad shot after another.

Dr. Benabio

Kaiser Permanente

Scheffler went on to eagle the 13th hole on that round. He eventually won the 2024 Masters Tournament. Remember that the next time you find yourself in a day that feels like it is spiraling toward disaster. Close the door on the compartment that was the last miserable hole and saunter to the next patient like it never happened.

And maybe close the clubface a bit on address for your next drive. 

Dr. Benabio is director of Healthcare Transformation and chief of dermatology at Kaiser Permanente San Diego. The opinions expressed in this column are his own and do not represent those of Kaiser Permanente. Dr. Benabio is @Dermdoc on X. Write to him at [email protected].

Patients with granulomatosis with polyangiitis (GPA) who scored high on a sinonasal symptom test are nearly three times as likely to relapse, according to a new study.

These patients reported higher scores months and up to 2 years before a disease flare, despite having low disease activity otherwise.

The study uses a different approach to try to predict relapse, compared with measuring biomarkers in lab tests, said Zachary Wallace, MD, a rheumatologist at Massachusetts General Hospital in Boston, Massachusetts. He was not involved with the study.

“It’s exciting because it might suggest that we could use something as simple as a survey to stratify someone’s risk of relapse,” he told this news organization.

Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a rare disease, with an estimated prevalence of 200-400 cases per million people. It is also heterogeneous, which makes it difficult to predict the risk for relapse for a given patient.

“Investigators have long searched for a reliable prognostic marker that identifies patients at high vs low risk of relapse in AAV but, except for ANCA type, no prognostic biomarker is routinely used to inform treatment decision-making,” wrote lead author Ellen Romich, MD, a rheumatology fellow at the University of Pennsylvania in Philadelphia, Pennsylvania, and colleagues.

Proteinase 3-ANCA (compared with myeloperoxidase-ANCA) has been tied to a higher risk for relapse, as well as gastrointestinal complications, sinonasal disease, and patient global assessment scores. In this new study, Dr. Romich and colleagues evaluated if patient-reported outcomes of sinonasal disease could predict AAV disease activity and relapse.

Researchers used data from a prospective, longitudinal cohort study through the University of Pennsylvania Vasculitis Center from 2016 to 2022. They included 107 patients with GPA, 40 patients with eosinophilic granulomatosis with polyangiitis (EGPA), 21 patients with microscopic polyangiitis (MPA), and 51 healthy controls.

Patients completed a median of four clinic visits during the duration of the study.

During each visit, patients filled out the 22-item SinoNasal Outcome Test (SNOT-22), a validated questionnaire that assesses rhinosinusitis. The tool asks patients to rate a list of symptoms from 0 to 5 in five categories: Rhinologic, extra-nasal, ear and face, psychologic, and sleep. The possible total score ranges from 0 to 110.

Disease activity was measured via the Birmingham Vasculitis Activity Score for Wegener’s Granulomatosis. The results were published online in Arthritis Care & Research.

Patients were, on average, 55 years old with an AAV duration of 3 years. (The mean age of healthy participants was 59.) More than half (58%) of patients were female, and 95% were White. The majority had a history of a flare (54%), and 60% of those flares had sinonasal involvement.

Even in remission, patients with AAV generally had on average higher SNOT-22 scores than healthy comparators (20 vs 5). Higher disease activity also correlated with higher SNOT-22 scores.

In patients with GPA, a high SNOT-22 score (total score of 41 or above) was associated with an increased risk for relapse within 2 years (hazard ratio, 2.7; P = .02). This association was not found for EGPA or MPA. This higher risk remained in a sensitivity analysis that included only patients with no history of sinonasal disease.

“Interestingly, among patients with GPA, SNOT-22 scores are elevated months to years prior to onset of systemic relapse but remain low in patients in sustained remission,” the authors wrote.

While other patient-reported outcomes have been validated for AAV, SNOT-22 may provide more detail on upper airway disease, commented Paul Monach, MD, PhD, an adjunct associate professor of medicine at the Boston University Chobanian & Avedisian School of Medicine and a rheumatologist at the VA Boston Healthcare System, Boston, Massachusetts.

“Upper airway GPA has not been studied as much as systemic GPA and MPA,” he told this news organization. “There are very few clinical trials, and we need better outcome measures like this.”

Dr. Romich noted that this work is still in the “early stages,” and SNOT-22 will need to be further studied and validated in other patient cohorts before its inclusion in clinical practice.

“There’s certainly more work that needs to be done to understand how the SNOT-22 questionnaire works in this patient population,” she said, including its predictive value compared with other known risk factors for relapse. “But I think it’s something that’s promising that we could use as a patient-reported outcome to try to, visit-to-visit, track their sinonasal symptoms.”

This study was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases and an NIH Rheumatology Research Training Grant. Dr. Romich, Dr. Wallace, and Monach reported no other relevant disclosures.

A version of this article appeared on Medscape.com.

Patients with granulomatosis with polyangiitis (GPA) who scored high on a sinonasal symptom test are nearly three times as likely to relapse, according to a new study.

These patients reported higher scores months and up to 2 years before a disease flare, despite having low disease activity otherwise.

The study uses a different approach to try to predict relapse, compared with measuring biomarkers in lab tests, said Zachary Wallace, MD, a rheumatologist at Massachusetts General Hospital in Boston, Massachusetts. He was not involved with the study.

“It’s exciting because it might suggest that we could use something as simple as a survey to stratify someone’s risk of relapse,” he told this news organization.

Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a rare disease, with an estimated prevalence of 200-400 cases per million people. It is also heterogeneous, which makes it difficult to predict the risk for relapse for a given patient.

“Investigators have long searched for a reliable prognostic marker that identifies patients at high vs low risk of relapse in AAV but, except for ANCA type, no prognostic biomarker is routinely used to inform treatment decision-making,” wrote lead author Ellen Romich, MD, a rheumatology fellow at the University of Pennsylvania in Philadelphia, Pennsylvania, and colleagues.

Proteinase 3-ANCA (compared with myeloperoxidase-ANCA) has been tied to a higher risk for relapse, as well as gastrointestinal complications, sinonasal disease, and patient global assessment scores. In this new study, Dr. Romich and colleagues evaluated if patient-reported outcomes of sinonasal disease could predict AAV disease activity and relapse.

Researchers used data from a prospective, longitudinal cohort study through the University of Pennsylvania Vasculitis Center from 2016 to 2022. They included 107 patients with GPA, 40 patients with eosinophilic granulomatosis with polyangiitis (EGPA), 21 patients with microscopic polyangiitis (MPA), and 51 healthy controls.

Patients completed a median of four clinic visits during the duration of the study.

During each visit, patients filled out the 22-item SinoNasal Outcome Test (SNOT-22), a validated questionnaire that assesses rhinosinusitis. The tool asks patients to rate a list of symptoms from 0 to 5 in five categories: Rhinologic, extra-nasal, ear and face, psychologic, and sleep. The possible total score ranges from 0 to 110.

Disease activity was measured via the Birmingham Vasculitis Activity Score for Wegener’s Granulomatosis. The results were published online in Arthritis Care & Research.

Patients were, on average, 55 years old with an AAV duration of 3 years. (The mean age of healthy participants was 59.) More than half (58%) of patients were female, and 95% were White. The majority had a history of a flare (54%), and 60% of those flares had sinonasal involvement.

Even in remission, patients with AAV generally had on average higher SNOT-22 scores than healthy comparators (20 vs 5). Higher disease activity also correlated with higher SNOT-22 scores.

In patients with GPA, a high SNOT-22 score (total score of 41 or above) was associated with an increased risk for relapse within 2 years (hazard ratio, 2.7; P = .02). This association was not found for EGPA or MPA. This higher risk remained in a sensitivity analysis that included only patients with no history of sinonasal disease.

“Interestingly, among patients with GPA, SNOT-22 scores are elevated months to years prior to onset of systemic relapse but remain low in patients in sustained remission,” the authors wrote.

While other patient-reported outcomes have been validated for AAV, SNOT-22 may provide more detail on upper airway disease, commented Paul Monach, MD, PhD, an adjunct associate professor of medicine at the Boston University Chobanian & Avedisian School of Medicine and a rheumatologist at the VA Boston Healthcare System, Boston, Massachusetts.

“Upper airway GPA has not been studied as much as systemic GPA and MPA,” he told this news organization. “There are very few clinical trials, and we need better outcome measures like this.”

Dr. Romich noted that this work is still in the “early stages,” and SNOT-22 will need to be further studied and validated in other patient cohorts before its inclusion in clinical practice.

“There’s certainly more work that needs to be done to understand how the SNOT-22 questionnaire works in this patient population,” she said, including its predictive value compared with other known risk factors for relapse. “But I think it’s something that’s promising that we could use as a patient-reported outcome to try to, visit-to-visit, track their sinonasal symptoms.”

This study was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases and an NIH Rheumatology Research Training Grant. Dr. Romich, Dr. Wallace, and Monach reported no other relevant disclosures.

A version of this article appeared on Medscape.com.

Patients with granulomatosis with polyangiitis (GPA) who scored high on a sinonasal symptom test are nearly three times as likely to relapse, according to a new study.

These patients reported higher scores months and up to 2 years before a disease flare, despite having low disease activity otherwise.

The study uses a different approach to try to predict relapse, compared with measuring biomarkers in lab tests, said Zachary Wallace, MD, a rheumatologist at Massachusetts General Hospital in Boston, Massachusetts. He was not involved with the study.

“It’s exciting because it might suggest that we could use something as simple as a survey to stratify someone’s risk of relapse,” he told this news organization.

Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a rare disease, with an estimated prevalence of 200-400 cases per million people. It is also heterogeneous, which makes it difficult to predict the risk for relapse for a given patient.

“Investigators have long searched for a reliable prognostic marker that identifies patients at high vs low risk of relapse in AAV but, except for ANCA type, no prognostic biomarker is routinely used to inform treatment decision-making,” wrote lead author Ellen Romich, MD, a rheumatology fellow at the University of Pennsylvania in Philadelphia, Pennsylvania, and colleagues.

Proteinase 3-ANCA (compared with myeloperoxidase-ANCA) has been tied to a higher risk for relapse, as well as gastrointestinal complications, sinonasal disease, and patient global assessment scores. In this new study, Dr. Romich and colleagues evaluated if patient-reported outcomes of sinonasal disease could predict AAV disease activity and relapse.

Researchers used data from a prospective, longitudinal cohort study through the University of Pennsylvania Vasculitis Center from 2016 to 2022. They included 107 patients with GPA, 40 patients with eosinophilic granulomatosis with polyangiitis (EGPA), 21 patients with microscopic polyangiitis (MPA), and 51 healthy controls.

Patients completed a median of four clinic visits during the duration of the study.

During each visit, patients filled out the 22-item SinoNasal Outcome Test (SNOT-22), a validated questionnaire that assesses rhinosinusitis. The tool asks patients to rate a list of symptoms from 0 to 5 in five categories: Rhinologic, extra-nasal, ear and face, psychologic, and sleep. The possible total score ranges from 0 to 110.

Disease activity was measured via the Birmingham Vasculitis Activity Score for Wegener’s Granulomatosis. The results were published online in Arthritis Care & Research.

Patients were, on average, 55 years old with an AAV duration of 3 years. (The mean age of healthy participants was 59.) More than half (58%) of patients were female, and 95% were White. The majority had a history of a flare (54%), and 60% of those flares had sinonasal involvement.

Even in remission, patients with AAV generally had on average higher SNOT-22 scores than healthy comparators (20 vs 5). Higher disease activity also correlated with higher SNOT-22 scores.

In patients with GPA, a high SNOT-22 score (total score of 41 or above) was associated with an increased risk for relapse within 2 years (hazard ratio, 2.7; P = .02). This association was not found for EGPA or MPA. This higher risk remained in a sensitivity analysis that included only patients with no history of sinonasal disease.

“Interestingly, among patients with GPA, SNOT-22 scores are elevated months to years prior to onset of systemic relapse but remain low in patients in sustained remission,” the authors wrote.

While other patient-reported outcomes have been validated for AAV, SNOT-22 may provide more detail on upper airway disease, commented Paul Monach, MD, PhD, an adjunct associate professor of medicine at the Boston University Chobanian & Avedisian School of Medicine and a rheumatologist at the VA Boston Healthcare System, Boston, Massachusetts.

“Upper airway GPA has not been studied as much as systemic GPA and MPA,” he told this news organization. “There are very few clinical trials, and we need better outcome measures like this.”

Dr. Romich noted that this work is still in the “early stages,” and SNOT-22 will need to be further studied and validated in other patient cohorts before its inclusion in clinical practice.

“There’s certainly more work that needs to be done to understand how the SNOT-22 questionnaire works in this patient population,” she said, including its predictive value compared with other known risk factors for relapse. “But I think it’s something that’s promising that we could use as a patient-reported outcome to try to, visit-to-visit, track their sinonasal symptoms.”

This study was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases and an NIH Rheumatology Research Training Grant. Dr. Romich, Dr. Wallace, and Monach reported no other relevant disclosures.

A version of this article appeared on Medscape.com.

Picture a healthcare system where physicians aren’t bogged down by excessive charting but are instead fully present with their patients, offering undivided attention and personalized care. In a recent X post, Stuart Blitz, COO and co-founder of Hone Health, sparked a thought-provoking conversation. “The problem with US healthcare is physicians are burned out since they spend way too much time charting, not enough with patients,” he wrote. “If you created a health system that did zero charting, you’d attract the best physicians and all patients would go there. Who is working on this?” 

This resonates with many in the medical community, myself included, because the strain of extensive documentation detracts from patient care. Having worked in both large and small healthcare systems, I know the burden of extensive charting is a palpable challenge, often detracting from the time we can devote to our patients.

The first part of this two-part series examines the overarching benefits of artificial intelligence (AI)–based clinical documentation in modern healthcare, a field witnessing a paradigm shift thanks to advancements in AI.
 

Transformative Evolution of Clinical Documentation

The transition from manual documentation to AI-driven solutions marks a significant shift in the field, with a number of products in development including Nuance, Abridge, Ambience, ScribeAmerica, 3M, and DeepScribe. These tools use ambient clinical intelligence (ACI) to automate documentation, capturing patient conversations and translating them into structured clinical summaries. This innovation aligns with the vision of reducing charting burdens and enhancing patient-physician interactions.

How does it work? ACI refers to a sophisticated form of AI applied in healthcare settings, particularly focusing on enhancing the clinical documentation process without disrupting the natural flow of the consultation. Here’s a technical yet practical breakdown of ACI and the algorithms it typically employs:

Data capture and processing: ACI systems employ various sensors and processing units, typically integrated into clinical settings. These sensors, like microphones and cameras, gather diverse data such as audio from patient-doctor dialogues and visual cues. This information is then processed in real-time or near–real-time.

Natural language processing (NLP): A core component of ACI is advanced NLP algorithms. These algorithms analyze the captured audio data, transcribing spoken words into text. NLP goes beyond mere transcription; it involves understanding context, extracting relevant medical information (like symptoms, diagnoses, and treatment plans), and interpreting the nuances of human language.

Deep learning: Machine learning, particularly deep-learning techniques, are employed to improve the accuracy of ACI systems continually. These algorithms can learn from vast datasets of clinical interactions, enhancing their ability to transcribe and interpret future conversations accurately. As they learn, they become better at understanding different accents, complex medical terms, and variations in speech patterns.

Integration with electronic health records (EHRs): ACI systems are often designed to integrate seamlessly with existing EHR systems. They can automatically populate patient records with information from patient-clinician interactions, reducing manual entry and potential errors.

Customization and personalization: Many ACI systems offer customizable templates or allow clinicians to tailor documentation workflows. This flexibility ensures that the output aligns with the specific needs and preferences of healthcare providers.

Ethical and privacy considerations: ACI systems must navigate significant ethical and privacy concerns, especially related to patient consent and data security. These systems need to comply with healthcare privacy regulations such as HIPAA. They need to securely manage sensitive patient data and restrict access to authorized personnel only.
 

Broad-Spectrum Benefits of AI in Documentation

• Reducing clinician burnout: By automating the documentation process, AI tools like DAX Copilot alleviate a significant contributor to physician burnout, enabling clinicians to focus more on patient care.
• Enhanced patient care: With AI handling documentation, clinicians can engage more with their patients, leading to improved care quality and patient satisfaction.
• Data accuracy and quality: AI-driven documentation captures detailed patient encounters accurately, ensuring high-quality and comprehensive medical records.
• Response to the growing need for efficient healthcare: AI-based documentation is a direct response to the growing call for more efficient healthcare practices, where clinicians spend less time on paperwork and more with patients.

The shift toward AI-based clinical documentation represents a critical step in addressing the inefficiencies in healthcare systems. It’s a move towards a more patient-centered approach, where clinicians can focus more on patient care by reducing the time spent on excessive charting. Hopefully, we can integrate these solutions into our clinics at a large enough scale to make such an impact.

In the next column, we will explore in-depth insights from Kenneth Harper at Nuance on the technical implementation of these tools, with DAX as an example.

I would love to read your comments on AI in clinical trials as well as other AI-related topics. Write me at [email protected] or find me on X @DrBonillaOnc.

Dr. Loaiza-Bonilla is the co-founder and chief medical officer at Massive Bio, a company connecting patients to clinical trials using artificial intelligence. His research and professional interests focus on precision medicine, clinical trial design, digital health, entrepreneurship, and patient advocacy. Dr Loaiza-Bonilla serves as medical director of oncology research at Capital Health in New Jersey, where he maintains a connection to patient care by attending to patients 2 days a week. He has served as a consultant for Verify, PSI CRO, Bayer, AstraZeneca, Cardinal Health, BrightInsight, The Lynx Group, Fresenius, Pfizer, Ipsen, and Guardant; served as a speaker or a member of a speakers bureau for Amgen, Guardant, Eisai, Ipsen, Natera, Merck, Bristol-Myers Squibb, and AstraZeneca. He holds a 5% or greater equity interest in Massive Bio.

A version of this article appeared on Medscape.com.

Picture a healthcare system where physicians aren’t bogged down by excessive charting but are instead fully present with their patients, offering undivided attention and personalized care. In a recent X post, Stuart Blitz, COO and co-founder of Hone Health, sparked a thought-provoking conversation. “The problem with US healthcare is physicians are burned out since they spend way too much time charting, not enough with patients,” he wrote. “If you created a health system that did zero charting, you’d attract the best physicians and all patients would go there. Who is working on this?” 

This resonates with many in the medical community, myself included, because the strain of extensive documentation detracts from patient care. Having worked in both large and small healthcare systems, I know the burden of extensive charting is a palpable challenge, often detracting from the time we can devote to our patients.

The first part of this two-part series examines the overarching benefits of artificial intelligence (AI)–based clinical documentation in modern healthcare, a field witnessing a paradigm shift thanks to advancements in AI.
 

Transformative Evolution of Clinical Documentation

The transition from manual documentation to AI-driven solutions marks a significant shift in the field, with a number of products in development including Nuance, Abridge, Ambience, ScribeAmerica, 3M, and DeepScribe. These tools use ambient clinical intelligence (ACI) to automate documentation, capturing patient conversations and translating them into structured clinical summaries. This innovation aligns with the vision of reducing charting burdens and enhancing patient-physician interactions.

How does it work? ACI refers to a sophisticated form of AI applied in healthcare settings, particularly focusing on enhancing the clinical documentation process without disrupting the natural flow of the consultation. Here’s a technical yet practical breakdown of ACI and the algorithms it typically employs:

Data capture and processing: ACI systems employ various sensors and processing units, typically integrated into clinical settings. These sensors, like microphones and cameras, gather diverse data such as audio from patient-doctor dialogues and visual cues. This information is then processed in real-time or near–real-time.

Natural language processing (NLP): A core component of ACI is advanced NLP algorithms. These algorithms analyze the captured audio data, transcribing spoken words into text. NLP goes beyond mere transcription; it involves understanding context, extracting relevant medical information (like symptoms, diagnoses, and treatment plans), and interpreting the nuances of human language.

Deep learning: Machine learning, particularly deep-learning techniques, are employed to improve the accuracy of ACI systems continually. These algorithms can learn from vast datasets of clinical interactions, enhancing their ability to transcribe and interpret future conversations accurately. As they learn, they become better at understanding different accents, complex medical terms, and variations in speech patterns.

Integration with electronic health records (EHRs): ACI systems are often designed to integrate seamlessly with existing EHR systems. They can automatically populate patient records with information from patient-clinician interactions, reducing manual entry and potential errors.

Customization and personalization: Many ACI systems offer customizable templates or allow clinicians to tailor documentation workflows. This flexibility ensures that the output aligns with the specific needs and preferences of healthcare providers.

Ethical and privacy considerations: ACI systems must navigate significant ethical and privacy concerns, especially related to patient consent and data security. These systems need to comply with healthcare privacy regulations such as HIPAA. They need to securely manage sensitive patient data and restrict access to authorized personnel only.
 

Broad-Spectrum Benefits of AI in Documentation

• Reducing clinician burnout: By automating the documentation process, AI tools like DAX Copilot alleviate a significant contributor to physician burnout, enabling clinicians to focus more on patient care.
• Enhanced patient care: With AI handling documentation, clinicians can engage more with their patients, leading to improved care quality and patient satisfaction.
• Data accuracy and quality: AI-driven documentation captures detailed patient encounters accurately, ensuring high-quality and comprehensive medical records.
• Response to the growing need for efficient healthcare: AI-based documentation is a direct response to the growing call for more efficient healthcare practices, where clinicians spend less time on paperwork and more with patients.

The shift toward AI-based clinical documentation represents a critical step in addressing the inefficiencies in healthcare systems. It’s a move towards a more patient-centered approach, where clinicians can focus more on patient care by reducing the time spent on excessive charting. Hopefully, we can integrate these solutions into our clinics at a large enough scale to make such an impact.

In the next column, we will explore in-depth insights from Kenneth Harper at Nuance on the technical implementation of these tools, with DAX as an example.

I would love to read your comments on AI in clinical trials as well as other AI-related topics. Write me at [email protected] or find me on X @DrBonillaOnc.

Dr. Loaiza-Bonilla is the co-founder and chief medical officer at Massive Bio, a company connecting patients to clinical trials using artificial intelligence. His research and professional interests focus on precision medicine, clinical trial design, digital health, entrepreneurship, and patient advocacy. Dr Loaiza-Bonilla serves as medical director of oncology research at Capital Health in New Jersey, where he maintains a connection to patient care by attending to patients 2 days a week. He has served as a consultant for Verify, PSI CRO, Bayer, AstraZeneca, Cardinal Health, BrightInsight, The Lynx Group, Fresenius, Pfizer, Ipsen, and Guardant; served as a speaker or a member of a speakers bureau for Amgen, Guardant, Eisai, Ipsen, Natera, Merck, Bristol-Myers Squibb, and AstraZeneca. He holds a 5% or greater equity interest in Massive Bio.

A version of this article appeared on Medscape.com.

Picture a healthcare system where physicians aren’t bogged down by excessive charting but are instead fully present with their patients, offering undivided attention and personalized care. In a recent X post, Stuart Blitz, COO and co-founder of Hone Health, sparked a thought-provoking conversation. “The problem with US healthcare is physicians are burned out since they spend way too much time charting, not enough with patients,” he wrote. “If you created a health system that did zero charting, you’d attract the best physicians and all patients would go there. Who is working on this?” 

This resonates with many in the medical community, myself included, because the strain of extensive documentation detracts from patient care. Having worked in both large and small healthcare systems, I know the burden of extensive charting is a palpable challenge, often detracting from the time we can devote to our patients.

The first part of this two-part series examines the overarching benefits of artificial intelligence (AI)–based clinical documentation in modern healthcare, a field witnessing a paradigm shift thanks to advancements in AI.
 

Transformative Evolution of Clinical Documentation

The transition from manual documentation to AI-driven solutions marks a significant shift in the field, with a number of products in development including Nuance, Abridge, Ambience, ScribeAmerica, 3M, and DeepScribe. These tools use ambient clinical intelligence (ACI) to automate documentation, capturing patient conversations and translating them into structured clinical summaries. This innovation aligns with the vision of reducing charting burdens and enhancing patient-physician interactions.

How does it work? ACI refers to a sophisticated form of AI applied in healthcare settings, particularly focusing on enhancing the clinical documentation process without disrupting the natural flow of the consultation. Here’s a technical yet practical breakdown of ACI and the algorithms it typically employs:

Data capture and processing: ACI systems employ various sensors and processing units, typically integrated into clinical settings. These sensors, like microphones and cameras, gather diverse data such as audio from patient-doctor dialogues and visual cues. This information is then processed in real-time or near–real-time.

Natural language processing (NLP): A core component of ACI is advanced NLP algorithms. These algorithms analyze the captured audio data, transcribing spoken words into text. NLP goes beyond mere transcription; it involves understanding context, extracting relevant medical information (like symptoms, diagnoses, and treatment plans), and interpreting the nuances of human language.

Deep learning: Machine learning, particularly deep-learning techniques, are employed to improve the accuracy of ACI systems continually. These algorithms can learn from vast datasets of clinical interactions, enhancing their ability to transcribe and interpret future conversations accurately. As they learn, they become better at understanding different accents, complex medical terms, and variations in speech patterns.

Integration with electronic health records (EHRs): ACI systems are often designed to integrate seamlessly with existing EHR systems. They can automatically populate patient records with information from patient-clinician interactions, reducing manual entry and potential errors.

Customization and personalization: Many ACI systems offer customizable templates or allow clinicians to tailor documentation workflows. This flexibility ensures that the output aligns with the specific needs and preferences of healthcare providers.

Ethical and privacy considerations: ACI systems must navigate significant ethical and privacy concerns, especially related to patient consent and data security. These systems need to comply with healthcare privacy regulations such as HIPAA. They need to securely manage sensitive patient data and restrict access to authorized personnel only.
 

Broad-Spectrum Benefits of AI in Documentation

• Reducing clinician burnout: By automating the documentation process, AI tools like DAX Copilot alleviate a significant contributor to physician burnout, enabling clinicians to focus more on patient care.
• Enhanced patient care: With AI handling documentation, clinicians can engage more with their patients, leading to improved care quality and patient satisfaction.
• Data accuracy and quality: AI-driven documentation captures detailed patient encounters accurately, ensuring high-quality and comprehensive medical records.
• Response to the growing need for efficient healthcare: AI-based documentation is a direct response to the growing call for more efficient healthcare practices, where clinicians spend less time on paperwork and more with patients.

The shift toward AI-based clinical documentation represents a critical step in addressing the inefficiencies in healthcare systems. It’s a move towards a more patient-centered approach, where clinicians can focus more on patient care by reducing the time spent on excessive charting. Hopefully, we can integrate these solutions into our clinics at a large enough scale to make such an impact.

In the next column, we will explore in-depth insights from Kenneth Harper at Nuance on the technical implementation of these tools, with DAX as an example.

I would love to read your comments on AI in clinical trials as well as other AI-related topics. Write me at [email protected] or find me on X @DrBonillaOnc.

Dr. Loaiza-Bonilla is the co-founder and chief medical officer at Massive Bio, a company connecting patients to clinical trials using artificial intelligence. His research and professional interests focus on precision medicine, clinical trial design, digital health, entrepreneurship, and patient advocacy. Dr Loaiza-Bonilla serves as medical director of oncology research at Capital Health in New Jersey, where he maintains a connection to patient care by attending to patients 2 days a week. He has served as a consultant for Verify, PSI CRO, Bayer, AstraZeneca, Cardinal Health, BrightInsight, The Lynx Group, Fresenius, Pfizer, Ipsen, and Guardant; served as a speaker or a member of a speakers bureau for Amgen, Guardant, Eisai, Ipsen, Natera, Merck, Bristol-Myers Squibb, and AstraZeneca. He holds a 5% or greater equity interest in Massive Bio.

A version of this article appeared on Medscape.com.

Patient-reported outcomes (PROs) in rheumatology are not just personal lists of physical complaints or so-called “organ recitals.” In fact, PROs can both guide treatment decisions in daily practice and serve as key endpoints for clinical trials.

That’s the informed opinion of Clifton O. Bingham III, MD, director of the Johns Arthritis Center in Baltimore, Maryland, who discussed clinical and research applications of PROs at the 2024 Rheumatoid Arthritis Research Summit presented by the Arthritis Foundation and the Hospital for Special Surgery in New York City.

Johns Hopkins Medicine, Johns Hopkins Division of Rheumatology

“Integrating PROs into practice settings can enhance the clinician’s ability to understand their patients and monitor disease impact, and they are increasingly available for clinical care and are being qualified for outcome measures for clinical trials,” Dr. Bingham said.

“I posit to you that some of this ability to better characterize things like anxiety and depression levels of patients more precisely may help us to identify those patients who are less likely to respond to therapy and may require different interventions than disease-modifying therapies for their disease,” he told the audience.
 

PRO Examples

The term PRO encompasses a broad range of measures that may include health-related quality of life measures, symptoms and their affects, patient satisfaction, and the patient’s experience with care.

PROs are important for rheumatology care and research because “we now have the capacity to make what we used to think were the subjective experiences of disease more objective. We now have ways that we can put numbers and measurements to the experiences that patients have about their illness and use that information as a way to understand more about the patients who are in front of us and also how their disease changes over time,” Dr. Bingham said.

Patients are the best — or in some cases, the only — judges of many aspects of their health, and they are best suited to report on certain events and outcomes, he said.

PROs that are currently included in core outcome measures used to guide care and in clinical trials in pain scores as reported by visual analog scales; functioning, as measured by the Health Assessment Questionnaire-Disability Index (HAQ-DI); and patient global assessment.

In international qualitative studies in which patients with rheumatoid arthritis (RA) were asked what was most important to them, the usual suspects of pain, function, and fatigue were routinely cited across the studies. But patients in studies from these groups (RAPP-PI, RAID, and OMERACT) also said that other factors important to their well-being included good sleep, enjoyment of life, independence, ability to participate in valued activities, and freedom from emotional distress, Dr. Bingham noted.
 

The Promise of PROMIS

The science of clinical measurement has advanced dramatically during his career, as Dr. Bingham said.

“There have been significant changes in the science behind how you develop and validate outcomes measures. The fields of clinimetrics and psychometrics have evolved substantially. These are now grounded in what we call ‘modern measurement’ approaches, which focus on item-response theory, constructing interval scales of measurement in things that are very precise in their ability to detect change over time,” he said.

One such measurement instrument is the Patient-Reported Outcome Measurement Information System (PROMIS®), developed at the National Institutes of Health using advanced measurement science.

The system, administered through either computer or paper questionnaires, is designed to improve precision of health-related quality of life assessments in multiple domains, including most domains identified by patients with RA. It uses a T-score metric standardized to the US population.

“You can use this in a disease like rheumatoid arthritis, and you can find out how patients are doing in reference to the normative United States populations,” he said.

Dr. Bingham noted that his team has “very good data” to show that PROMIS system significantly outperforms existing instruments such as the HAQ.
 

How It Works

The system uses item banks, each with multiple items. For example, there are approximately 150 items for the physical function assessment portion. All the items are scored along a continuum, “from people who are completely disabled to those who can run marathons,” Dr. Bingham said.

Each item on the scale has a question and response component, ranging from “are you able to get in and out of bed?” to “are you able to walk from one room to another?” to “are you able to run 5 miles?”

To evaluate the PROMIS scale, Dr. Bingham and colleagues looked at the distribution of PROMIS T-scores for 1029 patients with RA at their center. The scales showed that patients with RA have higher levels of pain, fatigue, and sleep disturbances, as well as worse physical function, than population norms.

Dr. Bingham and colleagues also evaluated the performance of the system in patients with active RA who were starting on or switching to a different disease-modifying antirheumatic drug (DMARD). As they reported in 2019, among 106 participants who completed the 12-week study, all PROMIS scores improved after DMARD initiation (P ≤ .05). In addition, except for the depression domain, changes in all assessed PROMIS measures correlated with changes in Clinical Disease Activity Index scores.

To see whether integrating PROs into routine clinics could have an effect on care, Dr. Bingham and colleagues conducted a prospective cohort study, which showed that with the additional patient-reported data, clinicians changed or adjusted RA treatment in 16%-19% of visits, identified new symptoms in 27%-38%, and suggested nonpharmacologic interventions in 4%-11%.

“This is information that’s being used, and it’s going into changing medical decision making,” he said.

Summarizing his work, Dr. Bingham told the audience “I hope that I have convinced you that patients with RA prioritize domains that are impacted by their disease. PROMIS measures are really state-of-the-science methods to evaluate multiple aspects of health-related quality of life, and what I’ll note to you is that these have been translated into multiple languages internationally. There are Spanish-language versions, there are Chinese language versions, there are versions for every country in the [European Union] that have been validated and can be used.”
 

It’s a Start

In the Q & A following the presentation, Daniel H. Solomon, MD, MPH, from Brigham and Women’s Hospital in Boston, commented that “the measurement issues and automating measurements seems like it’s a fundamental practice issue — how to manage the system and how to manage patients better, and I feel like we’re kind of scratching the surface.”

He said that artificial intelligence and PROs in clinic offer some promise for improving care but added that “we can do better than this. We can figure out better systems for measuring PROs: Having patients measure PROs, having patients tell us about their PROs so they don’t have to come in, or coming in only when they need to come in, when they’re really flaring. There are lots of innovative ways of thinking about these tools, and it feels like we’re kind of on the cusp of really taking advantage.”

Dr. Bingham’s work is supported by the Patient-Centered Outcomes Research Institute, National Institutes of Health, Ira T. Fine Discovery Fund, Johns Hopkins Arthritis Center Discovery Fund, Camille J. Morgan Arthritis Research and Education Fund, and Scheer Family Foundation and Joanne and John Rogers. He disclosed consulting for AbbVie, Janssen, Lilly, and Sanofi and serving as a board member of the PROMIS health organization, co-chair of the Omeract Technical Advisory Group, and member of the C-PATH RA PRO working group. Dr. Solomon had no relevant disclosures.

A version of this article appeared on Medscape.com.

Patient-reported outcomes (PROs) in rheumatology are not just personal lists of physical complaints or so-called “organ recitals.” In fact, PROs can both guide treatment decisions in daily practice and serve as key endpoints for clinical trials.

That’s the informed opinion of Clifton O. Bingham III, MD, director of the Johns Arthritis Center in Baltimore, Maryland, who discussed clinical and research applications of PROs at the 2024 Rheumatoid Arthritis Research Summit presented by the Arthritis Foundation and the Hospital for Special Surgery in New York City.

Johns Hopkins Medicine, Johns Hopkins Division of Rheumatology

“Integrating PROs into practice settings can enhance the clinician’s ability to understand their patients and monitor disease impact, and they are increasingly available for clinical care and are being qualified for outcome measures for clinical trials,” Dr. Bingham said.

“I posit to you that some of this ability to better characterize things like anxiety and depression levels of patients more precisely may help us to identify those patients who are less likely to respond to therapy and may require different interventions than disease-modifying therapies for their disease,” he told the audience.
 

PRO Examples

The term PRO encompasses a broad range of measures that may include health-related quality of life measures, symptoms and their affects, patient satisfaction, and the patient’s experience with care.

PROs are important for rheumatology care and research because “we now have the capacity to make what we used to think were the subjective experiences of disease more objective. We now have ways that we can put numbers and measurements to the experiences that patients have about their illness and use that information as a way to understand more about the patients who are in front of us and also how their disease changes over time,” Dr. Bingham said.

Patients are the best — or in some cases, the only — judges of many aspects of their health, and they are best suited to report on certain events and outcomes, he said.

PROs that are currently included in core outcome measures used to guide care and in clinical trials in pain scores as reported by visual analog scales; functioning, as measured by the Health Assessment Questionnaire-Disability Index (HAQ-DI); and patient global assessment.

In international qualitative studies in which patients with rheumatoid arthritis (RA) were asked what was most important to them, the usual suspects of pain, function, and fatigue were routinely cited across the studies. But patients in studies from these groups (RAPP-PI, RAID, and OMERACT) also said that other factors important to their well-being included good sleep, enjoyment of life, independence, ability to participate in valued activities, and freedom from emotional distress, Dr. Bingham noted.
 

The Promise of PROMIS

The science of clinical measurement has advanced dramatically during his career, as Dr. Bingham said.

“There have been significant changes in the science behind how you develop and validate outcomes measures. The fields of clinimetrics and psychometrics have evolved substantially. These are now grounded in what we call ‘modern measurement’ approaches, which focus on item-response theory, constructing interval scales of measurement in things that are very precise in their ability to detect change over time,” he said.

One such measurement instrument is the Patient-Reported Outcome Measurement Information System (PROMIS®), developed at the National Institutes of Health using advanced measurement science.

The system, administered through either computer or paper questionnaires, is designed to improve precision of health-related quality of life assessments in multiple domains, including most domains identified by patients with RA. It uses a T-score metric standardized to the US population.

“You can use this in a disease like rheumatoid arthritis, and you can find out how patients are doing in reference to the normative United States populations,” he said.

Dr. Bingham noted that his team has “very good data” to show that PROMIS system significantly outperforms existing instruments such as the HAQ.
 

How It Works

The system uses item banks, each with multiple items. For example, there are approximately 150 items for the physical function assessment portion. All the items are scored along a continuum, “from people who are completely disabled to those who can run marathons,” Dr. Bingham said.

Each item on the scale has a question and response component, ranging from “are you able to get in and out of bed?” to “are you able to walk from one room to another?” to “are you able to run 5 miles?”

To evaluate the PROMIS scale, Dr. Bingham and colleagues looked at the distribution of PROMIS T-scores for 1029 patients with RA at their center. The scales showed that patients with RA have higher levels of pain, fatigue, and sleep disturbances, as well as worse physical function, than population norms.

Dr. Bingham and colleagues also evaluated the performance of the system in patients with active RA who were starting on or switching to a different disease-modifying antirheumatic drug (DMARD). As they reported in 2019, among 106 participants who completed the 12-week study, all PROMIS scores improved after DMARD initiation (P ≤ .05). In addition, except for the depression domain, changes in all assessed PROMIS measures correlated with changes in Clinical Disease Activity Index scores.

To see whether integrating PROs into routine clinics could have an effect on care, Dr. Bingham and colleagues conducted a prospective cohort study, which showed that with the additional patient-reported data, clinicians changed or adjusted RA treatment in 16%-19% of visits, identified new symptoms in 27%-38%, and suggested nonpharmacologic interventions in 4%-11%.

“This is information that’s being used, and it’s going into changing medical decision making,” he said.

Summarizing his work, Dr. Bingham told the audience “I hope that I have convinced you that patients with RA prioritize domains that are impacted by their disease. PROMIS measures are really state-of-the-science methods to evaluate multiple aspects of health-related quality of life, and what I’ll note to you is that these have been translated into multiple languages internationally. There are Spanish-language versions, there are Chinese language versions, there are versions for every country in the [European Union] that have been validated and can be used.”
 

It’s a Start

In the Q & A following the presentation, Daniel H. Solomon, MD, MPH, from Brigham and Women’s Hospital in Boston, commented that “the measurement issues and automating measurements seems like it’s a fundamental practice issue — how to manage the system and how to manage patients better, and I feel like we’re kind of scratching the surface.”

He said that artificial intelligence and PROs in clinic offer some promise for improving care but added that “we can do better than this. We can figure out better systems for measuring PROs: Having patients measure PROs, having patients tell us about their PROs so they don’t have to come in, or coming in only when they need to come in, when they’re really flaring. There are lots of innovative ways of thinking about these tools, and it feels like we’re kind of on the cusp of really taking advantage.”

Dr. Bingham’s work is supported by the Patient-Centered Outcomes Research Institute, National Institutes of Health, Ira T. Fine Discovery Fund, Johns Hopkins Arthritis Center Discovery Fund, Camille J. Morgan Arthritis Research and Education Fund, and Scheer Family Foundation and Joanne and John Rogers. He disclosed consulting for AbbVie, Janssen, Lilly, and Sanofi and serving as a board member of the PROMIS health organization, co-chair of the Omeract Technical Advisory Group, and member of the C-PATH RA PRO working group. Dr. Solomon had no relevant disclosures.

A version of this article appeared on Medscape.com.

Patient-reported outcomes (PROs) in rheumatology are not just personal lists of physical complaints or so-called “organ recitals.” In fact, PROs can both guide treatment decisions in daily practice and serve as key endpoints for clinical trials.

That’s the informed opinion of Clifton O. Bingham III, MD, director of the Johns Arthritis Center in Baltimore, Maryland, who discussed clinical and research applications of PROs at the 2024 Rheumatoid Arthritis Research Summit presented by the Arthritis Foundation and the Hospital for Special Surgery in New York City.

Johns Hopkins Medicine, Johns Hopkins Division of Rheumatology

“Integrating PROs into practice settings can enhance the clinician’s ability to understand their patients and monitor disease impact, and they are increasingly available for clinical care and are being qualified for outcome measures for clinical trials,” Dr. Bingham said.

“I posit to you that some of this ability to better characterize things like anxiety and depression levels of patients more precisely may help us to identify those patients who are less likely to respond to therapy and may require different interventions than disease-modifying therapies for their disease,” he told the audience.
 

PRO Examples

The term PRO encompasses a broad range of measures that may include health-related quality of life measures, symptoms and their affects, patient satisfaction, and the patient’s experience with care.

PROs are important for rheumatology care and research because “we now have the capacity to make what we used to think were the subjective experiences of disease more objective. We now have ways that we can put numbers and measurements to the experiences that patients have about their illness and use that information as a way to understand more about the patients who are in front of us and also how their disease changes over time,” Dr. Bingham said.

Patients are the best — or in some cases, the only — judges of many aspects of their health, and they are best suited to report on certain events and outcomes, he said.

PROs that are currently included in core outcome measures used to guide care and in clinical trials in pain scores as reported by visual analog scales; functioning, as measured by the Health Assessment Questionnaire-Disability Index (HAQ-DI); and patient global assessment.

In international qualitative studies in which patients with rheumatoid arthritis (RA) were asked what was most important to them, the usual suspects of pain, function, and fatigue were routinely cited across the studies. But patients in studies from these groups (RAPP-PI, RAID, and OMERACT) also said that other factors important to their well-being included good sleep, enjoyment of life, independence, ability to participate in valued activities, and freedom from emotional distress, Dr. Bingham noted.
 

The Promise of PROMIS

The science of clinical measurement has advanced dramatically during his career, as Dr. Bingham said.

“There have been significant changes in the science behind how you develop and validate outcomes measures. The fields of clinimetrics and psychometrics have evolved substantially. These are now grounded in what we call ‘modern measurement’ approaches, which focus on item-response theory, constructing interval scales of measurement in things that are very precise in their ability to detect change over time,” he said.

One such measurement instrument is the Patient-Reported Outcome Measurement Information System (PROMIS®), developed at the National Institutes of Health using advanced measurement science.

The system, administered through either computer or paper questionnaires, is designed to improve precision of health-related quality of life assessments in multiple domains, including most domains identified by patients with RA. It uses a T-score metric standardized to the US population.

“You can use this in a disease like rheumatoid arthritis, and you can find out how patients are doing in reference to the normative United States populations,” he said.

Dr. Bingham noted that his team has “very good data” to show that PROMIS system significantly outperforms existing instruments such as the HAQ.
 

How It Works

The system uses item banks, each with multiple items. For example, there are approximately 150 items for the physical function assessment portion. All the items are scored along a continuum, “from people who are completely disabled to those who can run marathons,” Dr. Bingham said.

Each item on the scale has a question and response component, ranging from “are you able to get in and out of bed?” to “are you able to walk from one room to another?” to “are you able to run 5 miles?”

To evaluate the PROMIS scale, Dr. Bingham and colleagues looked at the distribution of PROMIS T-scores for 1029 patients with RA at their center. The scales showed that patients with RA have higher levels of pain, fatigue, and sleep disturbances, as well as worse physical function, than population norms.

Dr. Bingham and colleagues also evaluated the performance of the system in patients with active RA who were starting on or switching to a different disease-modifying antirheumatic drug (DMARD). As they reported in 2019, among 106 participants who completed the 12-week study, all PROMIS scores improved after DMARD initiation (P ≤ .05). In addition, except for the depression domain, changes in all assessed PROMIS measures correlated with changes in Clinical Disease Activity Index scores.

To see whether integrating PROs into routine clinics could have an effect on care, Dr. Bingham and colleagues conducted a prospective cohort study, which showed that with the additional patient-reported data, clinicians changed or adjusted RA treatment in 16%-19% of visits, identified new symptoms in 27%-38%, and suggested nonpharmacologic interventions in 4%-11%.

“This is information that’s being used, and it’s going into changing medical decision making,” he said.

Summarizing his work, Dr. Bingham told the audience “I hope that I have convinced you that patients with RA prioritize domains that are impacted by their disease. PROMIS measures are really state-of-the-science methods to evaluate multiple aspects of health-related quality of life, and what I’ll note to you is that these have been translated into multiple languages internationally. There are Spanish-language versions, there are Chinese language versions, there are versions for every country in the [European Union] that have been validated and can be used.”
 

It’s a Start

In the Q & A following the presentation, Daniel H. Solomon, MD, MPH, from Brigham and Women’s Hospital in Boston, commented that “the measurement issues and automating measurements seems like it’s a fundamental practice issue — how to manage the system and how to manage patients better, and I feel like we’re kind of scratching the surface.”

He said that artificial intelligence and PROs in clinic offer some promise for improving care but added that “we can do better than this. We can figure out better systems for measuring PROs: Having patients measure PROs, having patients tell us about their PROs so they don’t have to come in, or coming in only when they need to come in, when they’re really flaring. There are lots of innovative ways of thinking about these tools, and it feels like we’re kind of on the cusp of really taking advantage.”

Dr. Bingham’s work is supported by the Patient-Centered Outcomes Research Institute, National Institutes of Health, Ira T. Fine Discovery Fund, Johns Hopkins Arthritis Center Discovery Fund, Camille J. Morgan Arthritis Research and Education Fund, and Scheer Family Foundation and Joanne and John Rogers. He disclosed consulting for AbbVie, Janssen, Lilly, and Sanofi and serving as a board member of the PROMIS health organization, co-chair of the Omeract Technical Advisory Group, and member of the C-PATH RA PRO working group. Dr. Solomon had no relevant disclosures.

A version of this article appeared on Medscape.com.

Pregnant physicians may receive more workplace accommodations and protection against discrimination thanks to an updated rule from the US Equal Employment Opportunity Commission (EEOC). The guidelines could prevent women from losing critical career momentum. 

The Pregnant Workers Fairness Act (PWFA) aims to help workers balance professional demands with healthy pregnancies. It requires employers to provide reasonable accommodations for a “worker’s known limitations,” including physical or mental conditions associated with “pregnancy, childbirth, or related medical conditions.”

Reasonable accommodations vary but may involve time off to attend healthcare appointments or recover from childbirth, extra breaks during a shift, shorter work hours, or the ability to sit instead of stand. Private and public sector employers, including state and local governments, federal agencies, and employment agencies, must abide by the new guidelines unless they can provide evidence that doing so will cause undue hardship. 

Female doctors have historically encountered significant barriers to family planning. Years of training cause them to delay having children, often leading to higher rates of infertility, miscarriage, and pregnancy complications than in the general population. 

Some specialties, like surgeons, are particularly at risk, with 42% reporting at least one pregnancy loss. Most surgeons work their regular schedules until delivery despite desiring workload reductions, commonly citing unsupportive workplaces as a reason for not seeking accommodations. 

Trauma surgeon Qaali Hussein, MD, became pregnant with her first child during her intern year in 2008. She told this news organization that her residency program didn’t even have a maternity policy at the time, and her male supervisor was certain that motherhood would end her surgical career. 

She shared how “women usually waited until the end of their training to get pregnant. No one had ever gotten pregnant during the program and returned from maternity leave. I was the first to do so, so there wasn’t a policy or any program support to say, ‘What can we do to help?’ ”

Dr. Hussein used her vacation and sick time, returning to work 4 weeks after delivery. She had five more children, including twins her chief year and another baby during fellowship training in 2014. 

Each subsequent pregnancy was met with the same response from program leadership, she recalled. “They’d say, ‘This is it. You may have been able to do the first and second child, but this one will be impossible.’ ”

After the PWFA regulations first became enforceable in June, the EEOC accepted public feedback. The guidelines received nearly 100,000 comments, spurred mainly by the inclusion of abortion care as a qualifying condition for which an employee could receive accommodations. About 54,000 comments called for abortion to be excluded from the final rule, and 40,000 supported keeping the clause. 

The EEOC issued the final rule on April 15. It includes abortion care. However, the updated rule “does not require any employee to have — or not to have — an abortion, does not require taxpayers to pay for any abortions, and does not compel health care providers to provide any abortions,” the unpublished version of the final rule said. It is scheduled to take effect 60 days after its publication in the Federal Register on April 19.
 

Increasing Support for Doctor-Moms

The PWFA supplements other EEOC protections, such as pregnancy discrimination under Title VII of the Civil Rights Act of 1964 and access to reasonable accommodations under the Americans with Disabilities Act. In addition, it builds upon Department of Labor regulations, like the PUMP Act for breastfeeding employees and the Family and Medical Leave Act, which provides 12 weeks of unpaid, job-protected leave for the arrival of a child or certain medical conditions.

FMLA applies only to employees who have worked full-time for at least 12 months for an employer with 50 or more employees. Meanwhile, the unpaid, job-protected leave under the PWFA has no waiting period, lowers the required number of employees to 15, and permits accommodations for up to 40 weeks. 

Employers are encouraged to honor “common and simple” requests, like using a closer parking space or pumping or nursing at work, without requiring a doctor’s note, the rule said. 

Efforts to improve family leave policies for physicians and residents have been gaining traction. In 2021, the American Board of Medical Specialties began requiring its member boards with training programs lasting 2 or more years to allow at least 6 weeks off for parental, caregiver, and medical leave. This time can be taken without exhausting vacation or sick leave or requiring an extension in training. Over half of the 24 member boards permit leave beyond 6 weeks, including the American Boards of Allergy and Immunology, Emergency Medicine, Family Medicine, Radiology, and Surgery. 

Estefania Oliveros, MD, MSc, cardiologist and assistant professor at the Lewis Katz School of Medicine at Temple University, Philadelphia, told this news organization that the Accreditation Council for Graduate Medical Education also requires that residents and fellows receive 6 weeks of paid leave. 

“We add to that vacation time, so it gives them at least 8 weeks,” she said. The school has created spaces for nursing mothers — something neither she nor Dr. Hussein had access to when breastfeeding — and encourages the attendings to be proactive in excusing pregnant fellows for appointments. 

This differs significantly from her fellowship training experience 6 years ago at another institution, where she worked without accommodations until the day before her cesarean delivery. Dr. Oliveros had to use all her vacation time for recovery, returning to the program after 4 weeks instead of the recommended 6. 

“And that’s the story you hear all the time. Not because people are ill-intended; I just don’t think the system is designed to accommodate women, so we lose a lot of talent that way,” said Dr. Oliveros, whose 2019 survey in the Journal of the American College of Cardiology called for more support and protections for pregnant doctors. 

Both doctors believe the PWFA will be beneficial but only if leadership in the field takes up the cause. 

“The cultures of these institutions determine whether women feel safe or even confident enough to have children in medical school or residency,” said Dr. Hussein. 
 

A version of this article appeared on Medscape.com.

Pregnant physicians may receive more workplace accommodations and protection against discrimination thanks to an updated rule from the US Equal Employment Opportunity Commission (EEOC). The guidelines could prevent women from losing critical career momentum. 

The Pregnant Workers Fairness Act (PWFA) aims to help workers balance professional demands with healthy pregnancies. It requires employers to provide reasonable accommodations for a “worker’s known limitations,” including physical or mental conditions associated with “pregnancy, childbirth, or related medical conditions.”

Reasonable accommodations vary but may involve time off to attend healthcare appointments or recover from childbirth, extra breaks during a shift, shorter work hours, or the ability to sit instead of stand. Private and public sector employers, including state and local governments, federal agencies, and employment agencies, must abide by the new guidelines unless they can provide evidence that doing so will cause undue hardship. 

Female doctors have historically encountered significant barriers to family planning. Years of training cause them to delay having children, often leading to higher rates of infertility, miscarriage, and pregnancy complications than in the general population. 

Some specialties, like surgeons, are particularly at risk, with 42% reporting at least one pregnancy loss. Most surgeons work their regular schedules until delivery despite desiring workload reductions, commonly citing unsupportive workplaces as a reason for not seeking accommodations. 

Trauma surgeon Qaali Hussein, MD, became pregnant with her first child during her intern year in 2008. She told this news organization that her residency program didn’t even have a maternity policy at the time, and her male supervisor was certain that motherhood would end her surgical career. 

She shared how “women usually waited until the end of their training to get pregnant. No one had ever gotten pregnant during the program and returned from maternity leave. I was the first to do so, so there wasn’t a policy or any program support to say, ‘What can we do to help?’ ”

Dr. Hussein used her vacation and sick time, returning to work 4 weeks after delivery. She had five more children, including twins her chief year and another baby during fellowship training in 2014. 

Each subsequent pregnancy was met with the same response from program leadership, she recalled. “They’d say, ‘This is it. You may have been able to do the first and second child, but this one will be impossible.’ ”

After the PWFA regulations first became enforceable in June, the EEOC accepted public feedback. The guidelines received nearly 100,000 comments, spurred mainly by the inclusion of abortion care as a qualifying condition for which an employee could receive accommodations. About 54,000 comments called for abortion to be excluded from the final rule, and 40,000 supported keeping the clause. 

The EEOC issued the final rule on April 15. It includes abortion care. However, the updated rule “does not require any employee to have — or not to have — an abortion, does not require taxpayers to pay for any abortions, and does not compel health care providers to provide any abortions,” the unpublished version of the final rule said. It is scheduled to take effect 60 days after its publication in the Federal Register on April 19.
 

Increasing Support for Doctor-Moms

The PWFA supplements other EEOC protections, such as pregnancy discrimination under Title VII of the Civil Rights Act of 1964 and access to reasonable accommodations under the Americans with Disabilities Act. In addition, it builds upon Department of Labor regulations, like the PUMP Act for breastfeeding employees and the Family and Medical Leave Act, which provides 12 weeks of unpaid, job-protected leave for the arrival of a child or certain medical conditions.

FMLA applies only to employees who have worked full-time for at least 12 months for an employer with 50 or more employees. Meanwhile, the unpaid, job-protected leave under the PWFA has no waiting period, lowers the required number of employees to 15, and permits accommodations for up to 40 weeks. 

Employers are encouraged to honor “common and simple” requests, like using a closer parking space or pumping or nursing at work, without requiring a doctor’s note, the rule said. 

Efforts to improve family leave policies for physicians and residents have been gaining traction. In 2021, the American Board of Medical Specialties began requiring its member boards with training programs lasting 2 or more years to allow at least 6 weeks off for parental, caregiver, and medical leave. This time can be taken without exhausting vacation or sick leave or requiring an extension in training. Over half of the 24 member boards permit leave beyond 6 weeks, including the American Boards of Allergy and Immunology, Emergency Medicine, Family Medicine, Radiology, and Surgery. 

Estefania Oliveros, MD, MSc, cardiologist and assistant professor at the Lewis Katz School of Medicine at Temple University, Philadelphia, told this news organization that the Accreditation Council for Graduate Medical Education also requires that residents and fellows receive 6 weeks of paid leave. 

“We add to that vacation time, so it gives them at least 8 weeks,” she said. The school has created spaces for nursing mothers — something neither she nor Dr. Hussein had access to when breastfeeding — and encourages the attendings to be proactive in excusing pregnant fellows for appointments. 

This differs significantly from her fellowship training experience 6 years ago at another institution, where she worked without accommodations until the day before her cesarean delivery. Dr. Oliveros had to use all her vacation time for recovery, returning to the program after 4 weeks instead of the recommended 6. 

“And that’s the story you hear all the time. Not because people are ill-intended; I just don’t think the system is designed to accommodate women, so we lose a lot of talent that way,” said Dr. Oliveros, whose 2019 survey in the Journal of the American College of Cardiology called for more support and protections for pregnant doctors. 

Both doctors believe the PWFA will be beneficial but only if leadership in the field takes up the cause. 

“The cultures of these institutions determine whether women feel safe or even confident enough to have children in medical school or residency,” said Dr. Hussein. 
 

A version of this article appeared on Medscape.com.

Pregnant physicians may receive more workplace accommodations and protection against discrimination thanks to an updated rule from the US Equal Employment Opportunity Commission (EEOC). The guidelines could prevent women from losing critical career momentum. 

The Pregnant Workers Fairness Act (PWFA) aims to help workers balance professional demands with healthy pregnancies. It requires employers to provide reasonable accommodations for a “worker’s known limitations,” including physical or mental conditions associated with “pregnancy, childbirth, or related medical conditions.”

Reasonable accommodations vary but may involve time off to attend healthcare appointments or recover from childbirth, extra breaks during a shift, shorter work hours, or the ability to sit instead of stand. Private and public sector employers, including state and local governments, federal agencies, and employment agencies, must abide by the new guidelines unless they can provide evidence that doing so will cause undue hardship. 

Female doctors have historically encountered significant barriers to family planning. Years of training cause them to delay having children, often leading to higher rates of infertility, miscarriage, and pregnancy complications than in the general population. 

Some specialties, like surgeons, are particularly at risk, with 42% reporting at least one pregnancy loss. Most surgeons work their regular schedules until delivery despite desiring workload reductions, commonly citing unsupportive workplaces as a reason for not seeking accommodations. 

Trauma surgeon Qaali Hussein, MD, became pregnant with her first child during her intern year in 2008. She told this news organization that her residency program didn’t even have a maternity policy at the time, and her male supervisor was certain that motherhood would end her surgical career. 

She shared how “women usually waited until the end of their training to get pregnant. No one had ever gotten pregnant during the program and returned from maternity leave. I was the first to do so, so there wasn’t a policy or any program support to say, ‘What can we do to help?’ ”

Dr. Hussein used her vacation and sick time, returning to work 4 weeks after delivery. She had five more children, including twins her chief year and another baby during fellowship training in 2014. 

Each subsequent pregnancy was met with the same response from program leadership, she recalled. “They’d say, ‘This is it. You may have been able to do the first and second child, but this one will be impossible.’ ”

After the PWFA regulations first became enforceable in June, the EEOC accepted public feedback. The guidelines received nearly 100,000 comments, spurred mainly by the inclusion of abortion care as a qualifying condition for which an employee could receive accommodations. About 54,000 comments called for abortion to be excluded from the final rule, and 40,000 supported keeping the clause. 

The EEOC issued the final rule on April 15. It includes abortion care. However, the updated rule “does not require any employee to have — or not to have — an abortion, does not require taxpayers to pay for any abortions, and does not compel health care providers to provide any abortions,” the unpublished version of the final rule said. It is scheduled to take effect 60 days after its publication in the Federal Register on April 19.
 

Increasing Support for Doctor-Moms

The PWFA supplements other EEOC protections, such as pregnancy discrimination under Title VII of the Civil Rights Act of 1964 and access to reasonable accommodations under the Americans with Disabilities Act. In addition, it builds upon Department of Labor regulations, like the PUMP Act for breastfeeding employees and the Family and Medical Leave Act, which provides 12 weeks of unpaid, job-protected leave for the arrival of a child or certain medical conditions.

FMLA applies only to employees who have worked full-time for at least 12 months for an employer with 50 or more employees. Meanwhile, the unpaid, job-protected leave under the PWFA has no waiting period, lowers the required number of employees to 15, and permits accommodations for up to 40 weeks. 

Employers are encouraged to honor “common and simple” requests, like using a closer parking space or pumping or nursing at work, without requiring a doctor’s note, the rule said. 

Efforts to improve family leave policies for physicians and residents have been gaining traction. In 2021, the American Board of Medical Specialties began requiring its member boards with training programs lasting 2 or more years to allow at least 6 weeks off for parental, caregiver, and medical leave. This time can be taken without exhausting vacation or sick leave or requiring an extension in training. Over half of the 24 member boards permit leave beyond 6 weeks, including the American Boards of Allergy and Immunology, Emergency Medicine, Family Medicine, Radiology, and Surgery. 

Estefania Oliveros, MD, MSc, cardiologist and assistant professor at the Lewis Katz School of Medicine at Temple University, Philadelphia, told this news organization that the Accreditation Council for Graduate Medical Education also requires that residents and fellows receive 6 weeks of paid leave. 

“We add to that vacation time, so it gives them at least 8 weeks,” she said. The school has created spaces for nursing mothers — something neither she nor Dr. Hussein had access to when breastfeeding — and encourages the attendings to be proactive in excusing pregnant fellows for appointments. 

This differs significantly from her fellowship training experience 6 years ago at another institution, where she worked without accommodations until the day before her cesarean delivery. Dr. Oliveros had to use all her vacation time for recovery, returning to the program after 4 weeks instead of the recommended 6. 

“And that’s the story you hear all the time. Not because people are ill-intended; I just don’t think the system is designed to accommodate women, so we lose a lot of talent that way,” said Dr. Oliveros, whose 2019 survey in the Journal of the American College of Cardiology called for more support and protections for pregnant doctors. 

Both doctors believe the PWFA will be beneficial but only if leadership in the field takes up the cause. 

“The cultures of these institutions determine whether women feel safe or even confident enough to have children in medical school or residency,” said Dr. Hussein. 
 

A version of this article appeared on Medscape.com.

TOPLINE: 

A study comparing the clinical reasoning of an artificial intelligence (AI) model with that of physicians found the AI outperformed residents and attending physicians in simulated cases. The AI had more instances of incorrect reasoning than the doctors did but scored better overall.

METHODOLOGY:

• The study involved 39 physicians from two academic medical centers in Boston and the generative AI model GPT-4.
• Participants were presented with 20 simulated clinical cases involving common problems such as pharyngitis, headache, abdominal pain, cough, and chest pain. Each case included sections describing the triage presentation, review of systems, physical examination, and diagnostic testing.
• The primary outcome was the Revised-IDEA (R-IDEA) score, a 10-point scale evaluating clinical reasoning documentation across four domains: Interpretive summary, differential diagnosis, explanation of the lead diagnosis, and alternative diagnoses.

TAKEAWAY: 

• AI achieved a median R-IDEA score of 10, higher than attending physicians (median score, 9) and residents (8).
• The chatbot had a significantly higher estimated probability of achieving a high R-IDEA score of 8-10 (0.99) compared with attendings (0.76) and residents (0.56).
• AI provided more responses that contained instances of incorrect clinical reasoning (13.8%) than residents (2.8%) and attending physicians (12.5%). It performed similarly to physicians in diagnostic accuracy and inclusion of cannot-miss diagnoses.

IN PRACTICE:

“Future research should assess clinical reasoning of the LLM-physician interaction, as LLMs will more likely augment, not replace, the human reasoning process,” the authors of the study wrote. 

SOURCE:

Adam Rodman, MD, MPH, with Beth Israel Deaconess Medical Center, Boston, was the corresponding author on the paper. The research was published online in JAMA Internal Medicine. 

LIMITATIONS: 

Simulated clinical cases may not replicate performance in real-world scenarios. Further training could enhance the performance of the AI, so the study may underestimate its capabilities, the researchers noted. 

DISCLOSURES:

The study was supported by the Harvard Clinical and Translational Science Center and Harvard University. Authors disclosed financial ties to publishing companies and Solera Health. Dr. Rodman received funding from the Gordon and Betty Moore Foundation.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication. A version of this article appeared on Medscape.com.

TOPLINE: 

A study comparing the clinical reasoning of an artificial intelligence (AI) model with that of physicians found the AI outperformed residents and attending physicians in simulated cases. The AI had more instances of incorrect reasoning than the doctors did but scored better overall.

METHODOLOGY:

• The study involved 39 physicians from two academic medical centers in Boston and the generative AI model GPT-4.
• Participants were presented with 20 simulated clinical cases involving common problems such as pharyngitis, headache, abdominal pain, cough, and chest pain. Each case included sections describing the triage presentation, review of systems, physical examination, and diagnostic testing.
• The primary outcome was the Revised-IDEA (R-IDEA) score, a 10-point scale evaluating clinical reasoning documentation across four domains: Interpretive summary, differential diagnosis, explanation of the lead diagnosis, and alternative diagnoses.

TAKEAWAY: 

• AI achieved a median R-IDEA score of 10, higher than attending physicians (median score, 9) and residents (8).
• The chatbot had a significantly higher estimated probability of achieving a high R-IDEA score of 8-10 (0.99) compared with attendings (0.76) and residents (0.56).
• AI provided more responses that contained instances of incorrect clinical reasoning (13.8%) than residents (2.8%) and attending physicians (12.5%). It performed similarly to physicians in diagnostic accuracy and inclusion of cannot-miss diagnoses.

IN PRACTICE:

“Future research should assess clinical reasoning of the LLM-physician interaction, as LLMs will more likely augment, not replace, the human reasoning process,” the authors of the study wrote. 

SOURCE:

Adam Rodman, MD, MPH, with Beth Israel Deaconess Medical Center, Boston, was the corresponding author on the paper. The research was published online in JAMA Internal Medicine. 

LIMITATIONS: 

Simulated clinical cases may not replicate performance in real-world scenarios. Further training could enhance the performance of the AI, so the study may underestimate its capabilities, the researchers noted. 

DISCLOSURES:

The study was supported by the Harvard Clinical and Translational Science Center and Harvard University. Authors disclosed financial ties to publishing companies and Solera Health. Dr. Rodman received funding from the Gordon and Betty Moore Foundation.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication. A version of this article appeared on Medscape.com.

TOPLINE: 

A study comparing the clinical reasoning of an artificial intelligence (AI) model with that of physicians found the AI outperformed residents and attending physicians in simulated cases. The AI had more instances of incorrect reasoning than the doctors did but scored better overall.

METHODOLOGY:

• The study involved 39 physicians from two academic medical centers in Boston and the generative AI model GPT-4.
• Participants were presented with 20 simulated clinical cases involving common problems such as pharyngitis, headache, abdominal pain, cough, and chest pain. Each case included sections describing the triage presentation, review of systems, physical examination, and diagnostic testing.
• The primary outcome was the Revised-IDEA (R-IDEA) score, a 10-point scale evaluating clinical reasoning documentation across four domains: Interpretive summary, differential diagnosis, explanation of the lead diagnosis, and alternative diagnoses.

TAKEAWAY: 

• AI achieved a median R-IDEA score of 10, higher than attending physicians (median score, 9) and residents (8).
• The chatbot had a significantly higher estimated probability of achieving a high R-IDEA score of 8-10 (0.99) compared with attendings (0.76) and residents (0.56).
• AI provided more responses that contained instances of incorrect clinical reasoning (13.8%) than residents (2.8%) and attending physicians (12.5%). It performed similarly to physicians in diagnostic accuracy and inclusion of cannot-miss diagnoses.

IN PRACTICE:

“Future research should assess clinical reasoning of the LLM-physician interaction, as LLMs will more likely augment, not replace, the human reasoning process,” the authors of the study wrote. 

SOURCE:

Adam Rodman, MD, MPH, with Beth Israel Deaconess Medical Center, Boston, was the corresponding author on the paper. The research was published online in JAMA Internal Medicine. 

LIMITATIONS: 

Simulated clinical cases may not replicate performance in real-world scenarios. Further training could enhance the performance of the AI, so the study may underestimate its capabilities, the researchers noted. 

DISCLOSURES:

The study was supported by the Harvard Clinical and Translational Science Center and Harvard University. Authors disclosed financial ties to publishing companies and Solera Health. Dr. Rodman received funding from the Gordon and Betty Moore Foundation.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication. A version of this article appeared on Medscape.com.

TOPLINE:

The time interval between onset of interstitial lung disease (ILD) and diagnosis of anti–melanoma differentiation-associated gene 5 (MDA5) antibody-positive dermatomyositis (DM) “has not been well described,” the authors say.

METHODOLOGY:

• MDA5 antibody-positive DM is a rare DM subtype associated with ILD, which is categorized into rapidly progressive ILD (RPILD) and chronic ILD, with the former having a particularly high mortality rate.
• In this retrospective cohort study using electronic medical records, researchers evaluated 774 patients with DM between 2008 and 2023 to learn more about the time interval between ILD and the time of an MDA5 antibody-positive DM diagnosis, which has not been well described.
• The primary outcome was ILD diagnosis and time in days between documented ILD and MDA5 antibody-positive DM diagnoses.

TAKEAWAY:

• Overall, 14 patients with DM (1.8%) were diagnosed with MDA5 antibody-positive DM in dermatology, rheumatology, or pulmonology departments (nine women and five men; age, 24-77 years; 79% were White and 7% were Black).
• ILD was diagnosed in 9 of the 14 patients (64%); 6 of the 14 (43%) met the criteria for RPILD. Two cases were diagnosed concurrently and two prior to MDA5 antibody-positive DM diagnosis.
• The median time between ILD and MDA5 antibody-positive DM diagnoses was 163 days.
• Gottron papules/sign and midfacial erythema were the most common dermatologic findings, and no association was seen between cutaneous signs and type of ILD.

IN PRACTICE:

“Establishing an accurate timeline between MDA5 antibody-positive DM and ILD can promote urgency among dermatologists to evaluate extracutaneous manifestations in their management of patients with DM for more accurate risk stratification and appropriate treatment,” the authors wrote.

SOURCE:

This study, led by Rachel R. Lin, from the University of Miami, Miami, Florida, was published online as a research letter in JAMA Dermatology.

LIMITATIONS:

Study limitations were the study’s retrospective design and small sample size.

DISCLOSURES:

No information on study funding was provided. One author reported personal fees from argenX outside this submitted work. Other authors did not disclose any competing interests.

A version of this article appeared on Medscape.com.

TOPLINE:

The time interval between onset of interstitial lung disease (ILD) and diagnosis of anti–melanoma differentiation-associated gene 5 (MDA5) antibody-positive dermatomyositis (DM) “has not been well described,” the authors say.

METHODOLOGY:

• MDA5 antibody-positive DM is a rare DM subtype associated with ILD, which is categorized into rapidly progressive ILD (RPILD) and chronic ILD, with the former having a particularly high mortality rate.
• In this retrospective cohort study using electronic medical records, researchers evaluated 774 patients with DM between 2008 and 2023 to learn more about the time interval between ILD and the time of an MDA5 antibody-positive DM diagnosis, which has not been well described.
• The primary outcome was ILD diagnosis and time in days between documented ILD and MDA5 antibody-positive DM diagnoses.

TAKEAWAY:

• Overall, 14 patients with DM (1.8%) were diagnosed with MDA5 antibody-positive DM in dermatology, rheumatology, or pulmonology departments (nine women and five men; age, 24-77 years; 79% were White and 7% were Black).
• ILD was diagnosed in 9 of the 14 patients (64%); 6 of the 14 (43%) met the criteria for RPILD. Two cases were diagnosed concurrently and two prior to MDA5 antibody-positive DM diagnosis.
• The median time between ILD and MDA5 antibody-positive DM diagnoses was 163 days.
• Gottron papules/sign and midfacial erythema were the most common dermatologic findings, and no association was seen between cutaneous signs and type of ILD.

IN PRACTICE:

“Establishing an accurate timeline between MDA5 antibody-positive DM and ILD can promote urgency among dermatologists to evaluate extracutaneous manifestations in their management of patients with DM for more accurate risk stratification and appropriate treatment,” the authors wrote.

SOURCE:

This study, led by Rachel R. Lin, from the University of Miami, Miami, Florida, was published online as a research letter in JAMA Dermatology.

LIMITATIONS:

Study limitations were the study’s retrospective design and small sample size.

DISCLOSURES:

No information on study funding was provided. One author reported personal fees from argenX outside this submitted work. Other authors did not disclose any competing interests.

A version of this article appeared on Medscape.com.

TOPLINE:

The time interval between onset of interstitial lung disease (ILD) and diagnosis of anti–melanoma differentiation-associated gene 5 (MDA5) antibody-positive dermatomyositis (DM) “has not been well described,” the authors say.

METHODOLOGY:

• MDA5 antibody-positive DM is a rare DM subtype associated with ILD, which is categorized into rapidly progressive ILD (RPILD) and chronic ILD, with the former having a particularly high mortality rate.
• In this retrospective cohort study using electronic medical records, researchers evaluated 774 patients with DM between 2008 and 2023 to learn more about the time interval between ILD and the time of an MDA5 antibody-positive DM diagnosis, which has not been well described.
• The primary outcome was ILD diagnosis and time in days between documented ILD and MDA5 antibody-positive DM diagnoses.

TAKEAWAY:

• Overall, 14 patients with DM (1.8%) were diagnosed with MDA5 antibody-positive DM in dermatology, rheumatology, or pulmonology departments (nine women and five men; age, 24-77 years; 79% were White and 7% were Black).
• ILD was diagnosed in 9 of the 14 patients (64%); 6 of the 14 (43%) met the criteria for RPILD. Two cases were diagnosed concurrently and two prior to MDA5 antibody-positive DM diagnosis.
• The median time between ILD and MDA5 antibody-positive DM diagnoses was 163 days.
• Gottron papules/sign and midfacial erythema were the most common dermatologic findings, and no association was seen between cutaneous signs and type of ILD.

IN PRACTICE:

“Establishing an accurate timeline between MDA5 antibody-positive DM and ILD can promote urgency among dermatologists to evaluate extracutaneous manifestations in their management of patients with DM for more accurate risk stratification and appropriate treatment,” the authors wrote.

SOURCE:

This study, led by Rachel R. Lin, from the University of Miami, Miami, Florida, was published online as a research letter in JAMA Dermatology.

LIMITATIONS:

Study limitations were the study’s retrospective design and small sample size.

DISCLOSURES:

No information on study funding was provided. One author reported personal fees from argenX outside this submitted work. Other authors did not disclose any competing interests.

A version of this article appeared on Medscape.com.

TOPLINE:

Combined pediatric dermatology-rheumatology clinics can improve patient care and patient satisfaction, a survey of dermatologists suggested.

METHODOLOGY:

• Combined pediatric dermatology-rheumatology clinics can improve patient outcomes and experiences, particularly for pediatric autoimmune conditions presenting with both cutaneous and systemic manifestations.
• The researchers surveyed 208 pediatric dermatologists working in combined pediatric dermatology-rheumatology clinics.
• A total of 13 member responses were recorded from three countries: 10 from the United States, two from Mexico, and one from Canada.

TAKEAWAY:

• Perceived benefits of combined clinics were improved patient care through coordinated treatment decisions and timely communication between providers.
• Patient satisfaction was favorable, and patients and families endorsed the combined clinic approach.
• Barriers to clinic establishment included differences in the pace between dermatology and rheumatology clinic flow, the need to generate more relative value units, resistance from colleagues, and limited time.
• Areas that needed improvement included more time for patient visits, dedicated research assistants, new patient referrals, additional patient rooms, resources for research, and patient care infrastructure.

IN PRACTICE:

The insights from this survey “will hopefully inspire further development of these combined clinics,” the authors wrote.

SOURCE:

The investigation, led by Olga S. Cherepakhin, BS, University of Washington, Seattle, Washington, was published in Pediatric Dermatology.

LIMITATIONS:

Limitations included the subjective nature, lack of some information, selection bias, and small number of respondents, and the survey reflected the perspective of the pediatric dermatologists only.

DISCLOSURES:

The study was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health. One author reported full-time employment at Janssen R&D, and the other authors had no disclosures.

A version of this article appeared on Medscape.com.

TOPLINE:

Combined pediatric dermatology-rheumatology clinics can improve patient care and patient satisfaction, a survey of dermatologists suggested.

METHODOLOGY:

• Combined pediatric dermatology-rheumatology clinics can improve patient outcomes and experiences, particularly for pediatric autoimmune conditions presenting with both cutaneous and systemic manifestations.
• The researchers surveyed 208 pediatric dermatologists working in combined pediatric dermatology-rheumatology clinics.
• A total of 13 member responses were recorded from three countries: 10 from the United States, two from Mexico, and one from Canada.

TAKEAWAY:

• Perceived benefits of combined clinics were improved patient care through coordinated treatment decisions and timely communication between providers.
• Patient satisfaction was favorable, and patients and families endorsed the combined clinic approach.
• Barriers to clinic establishment included differences in the pace between dermatology and rheumatology clinic flow, the need to generate more relative value units, resistance from colleagues, and limited time.
• Areas that needed improvement included more time for patient visits, dedicated research assistants, new patient referrals, additional patient rooms, resources for research, and patient care infrastructure.

IN PRACTICE:

The insights from this survey “will hopefully inspire further development of these combined clinics,” the authors wrote.

SOURCE:

The investigation, led by Olga S. Cherepakhin, BS, University of Washington, Seattle, Washington, was published in Pediatric Dermatology.

LIMITATIONS:

Limitations included the subjective nature, lack of some information, selection bias, and small number of respondents, and the survey reflected the perspective of the pediatric dermatologists only.

DISCLOSURES:

The study was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health. One author reported full-time employment at Janssen R&D, and the other authors had no disclosures.

A version of this article appeared on Medscape.com.

TOPLINE:

Combined pediatric dermatology-rheumatology clinics can improve patient care and patient satisfaction, a survey of dermatologists suggested.

METHODOLOGY:

• Combined pediatric dermatology-rheumatology clinics can improve patient outcomes and experiences, particularly for pediatric autoimmune conditions presenting with both cutaneous and systemic manifestations.
• The researchers surveyed 208 pediatric dermatologists working in combined pediatric dermatology-rheumatology clinics.
• A total of 13 member responses were recorded from three countries: 10 from the United States, two from Mexico, and one from Canada.

TAKEAWAY:

• Perceived benefits of combined clinics were improved patient care through coordinated treatment decisions and timely communication between providers.
• Patient satisfaction was favorable, and patients and families endorsed the combined clinic approach.
• Barriers to clinic establishment included differences in the pace between dermatology and rheumatology clinic flow, the need to generate more relative value units, resistance from colleagues, and limited time.
• Areas that needed improvement included more time for patient visits, dedicated research assistants, new patient referrals, additional patient rooms, resources for research, and patient care infrastructure.

IN PRACTICE:

The insights from this survey “will hopefully inspire further development of these combined clinics,” the authors wrote.

SOURCE:

The investigation, led by Olga S. Cherepakhin, BS, University of Washington, Seattle, Washington, was published in Pediatric Dermatology.

LIMITATIONS:

Limitations included the subjective nature, lack of some information, selection bias, and small number of respondents, and the survey reflected the perspective of the pediatric dermatologists only.

DISCLOSURES:

The study was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health. One author reported full-time employment at Janssen R&D, and the other authors had no disclosures.

A version of this article appeared on Medscape.com.

In last month’s column, I discussed employees who are “clock watchers” and how to address this issue in your practice if it exists. Here’s another scenario you may encounter from the Office Politics Forum at the recent American Academy of Dermatology annual meeting:

A 40-year-old dermatologist has practiced in the same office since residency and is loved by patients and staff. He remained with the practice through its takeover by a local hospital three years previously. Recently, over a 3-month period, everyone in the office notices a change in this dermatologist’s behavior. He no longer appears happy, is argumentative with staff and patients alike, often dismisses patients’ concerns, and calls in sick during the practice’s busiest days.

It is not difficult to recognize these changes as hallmarks of burnout, which continues to be pervasive across all practice settings and specialties. According to the American Medical Association’s National Burnout Benchmarking report, over 50% of physicians report some characteristics of burnout, which include emotional exhaustion, depersonalization, and a feeling of decreased personal achievement.

olm26250/Thinkstock

Addressing physician burnout requires a systems-based approach that targets these root causes at all levels, from individual coping strategies to organizational and systemic changes in the healthcare industry. Here are some strategies that have worked for me and others:

Optimize Practice Efficiency: This is the consistent theme of this column over several decades: Streamline office processes to enhance the quality of care while reducing unnecessary workload. This can involve adopting efficient patient scheduling systems, improving clinic flow, and utilizing technology like patient portals judiciously to avoid increasing the task load without compensation.

Promote Work-Life Balance: Encourage a culture that values work-life balance. This can include flexible scheduling, respecting off-duty hours by limiting non-emergency work communications, and using your vacation time. Remember Eastern’s First Law: Your last words will NOT be, “I wish I had spent more time in the office.”

Implement Medical Scribes: I’ve written frequently about this, including a recent column on the new artificial intelligence (AI) scribes, such as DeepCura, DeepScribe, Nuance, Suki, Augmedix, Tali AI, Iodine Software, ScribeLink, and Amazon Web Services’ new HealthScribe product. Utilizing medical scribes to handle documentation can significantly reduce the administrative burden, allowing physicians to focus more on patient care rather than paperwork, potentially improving both physician and patient satisfaction. (As always, I have no financial interest in any product or service mentioned in this column.)

Provide Professional Development Opportunities: Offer opportunities for professional growth and development. This can include attending conferences, participating in research, or providing time and resources for continuing education. Such opportunities can reinvigorate a physician’s passion for medicine and improve job satisfaction.

Foster a Supportive Work Environment: Create a supportive work culture where staff and physicians feel comfortable discussing challenges and seeking support. Regular meetings or check-ins can help identify early signs of burnout and address them proactively.

Evaluate and Adjust Workloads: Regularly assess physician workloads to ensure they are manageable. Adjusting patient loads, redistributing tasks among team members, or hiring additional staff can help prevent burnout.

Leadership Training and Support: Provide training for leaders within the practice on recognizing signs of burnout and effective management strategies. Supportive leadership is crucial in creating an environment where physicians feel valued and heard.

Peer Support and Mentorship Programs: Establish peer support or mentorship programs where physicians can share experiences, offer advice, and provide emotional support to each other.

Feedback and Continuous Improvement: Managers should regularly solicit feedback from physicians regarding their workload, job satisfaction, and suggestions for improvements. Actively work on implementing feasible changes to address concerns.

Dr. Eastern practices dermatology and dermatologic surgery in Belleville, N.J. He is the author of numerous articles and textbook chapters, and is a longtime monthly columnist for Dermatology News. Write to him at [email protected].

In last month’s column, I discussed employees who are “clock watchers” and how to address this issue in your practice if it exists. Here’s another scenario you may encounter from the Office Politics Forum at the recent American Academy of Dermatology annual meeting:

A 40-year-old dermatologist has practiced in the same office since residency and is loved by patients and staff. He remained with the practice through its takeover by a local hospital three years previously. Recently, over a 3-month period, everyone in the office notices a change in this dermatologist’s behavior. He no longer appears happy, is argumentative with staff and patients alike, often dismisses patients’ concerns, and calls in sick during the practice’s busiest days.

It is not difficult to recognize these changes as hallmarks of burnout, which continues to be pervasive across all practice settings and specialties. According to the American Medical Association’s National Burnout Benchmarking report, over 50% of physicians report some characteristics of burnout, which include emotional exhaustion, depersonalization, and a feeling of decreased personal achievement.

olm26250/Thinkstock

Addressing physician burnout requires a systems-based approach that targets these root causes at all levels, from individual coping strategies to organizational and systemic changes in the healthcare industry. Here are some strategies that have worked for me and others:

Optimize Practice Efficiency: This is the consistent theme of this column over several decades: Streamline office processes to enhance the quality of care while reducing unnecessary workload. This can involve adopting efficient patient scheduling systems, improving clinic flow, and utilizing technology like patient portals judiciously to avoid increasing the task load without compensation.

Promote Work-Life Balance: Encourage a culture that values work-life balance. This can include flexible scheduling, respecting off-duty hours by limiting non-emergency work communications, and using your vacation time. Remember Eastern’s First Law: Your last words will NOT be, “I wish I had spent more time in the office.”

Implement Medical Scribes: I’ve written frequently about this, including a recent column on the new artificial intelligence (AI) scribes, such as DeepCura, DeepScribe, Nuance, Suki, Augmedix, Tali AI, Iodine Software, ScribeLink, and Amazon Web Services’ new HealthScribe product. Utilizing medical scribes to handle documentation can significantly reduce the administrative burden, allowing physicians to focus more on patient care rather than paperwork, potentially improving both physician and patient satisfaction. (As always, I have no financial interest in any product or service mentioned in this column.)

Provide Professional Development Opportunities: Offer opportunities for professional growth and development. This can include attending conferences, participating in research, or providing time and resources for continuing education. Such opportunities can reinvigorate a physician’s passion for medicine and improve job satisfaction.

Foster a Supportive Work Environment: Create a supportive work culture where staff and physicians feel comfortable discussing challenges and seeking support. Regular meetings or check-ins can help identify early signs of burnout and address them proactively.

Evaluate and Adjust Workloads: Regularly assess physician workloads to ensure they are manageable. Adjusting patient loads, redistributing tasks among team members, or hiring additional staff can help prevent burnout.

Leadership Training and Support: Provide training for leaders within the practice on recognizing signs of burnout and effective management strategies. Supportive leadership is crucial in creating an environment where physicians feel valued and heard.

Peer Support and Mentorship Programs: Establish peer support or mentorship programs where physicians can share experiences, offer advice, and provide emotional support to each other.

Feedback and Continuous Improvement: Managers should regularly solicit feedback from physicians regarding their workload, job satisfaction, and suggestions for improvements. Actively work on implementing feasible changes to address concerns.

Dr. Eastern practices dermatology and dermatologic surgery in Belleville, N.J. He is the author of numerous articles and textbook chapters, and is a longtime monthly columnist for Dermatology News. Write to him at [email protected].

In last month’s column, I discussed employees who are “clock watchers” and how to address this issue in your practice if it exists. Here’s another scenario you may encounter from the Office Politics Forum at the recent American Academy of Dermatology annual meeting:

A 40-year-old dermatologist has practiced in the same office since residency and is loved by patients and staff. He remained with the practice through its takeover by a local hospital three years previously. Recently, over a 3-month period, everyone in the office notices a change in this dermatologist’s behavior. He no longer appears happy, is argumentative with staff and patients alike, often dismisses patients’ concerns, and calls in sick during the practice’s busiest days.

It is not difficult to recognize these changes as hallmarks of burnout, which continues to be pervasive across all practice settings and specialties. According to the American Medical Association’s National Burnout Benchmarking report, over 50% of physicians report some characteristics of burnout, which include emotional exhaustion, depersonalization, and a feeling of decreased personal achievement.

olm26250/Thinkstock

Addressing physician burnout requires a systems-based approach that targets these root causes at all levels, from individual coping strategies to organizational and systemic changes in the healthcare industry. Here are some strategies that have worked for me and others:

Optimize Practice Efficiency: This is the consistent theme of this column over several decades: Streamline office processes to enhance the quality of care while reducing unnecessary workload. This can involve adopting efficient patient scheduling systems, improving clinic flow, and utilizing technology like patient portals judiciously to avoid increasing the task load without compensation.

Promote Work-Life Balance: Encourage a culture that values work-life balance. This can include flexible scheduling, respecting off-duty hours by limiting non-emergency work communications, and using your vacation time. Remember Eastern’s First Law: Your last words will NOT be, “I wish I had spent more time in the office.”

Implement Medical Scribes: I’ve written frequently about this, including a recent column on the new artificial intelligence (AI) scribes, such as DeepCura, DeepScribe, Nuance, Suki, Augmedix, Tali AI, Iodine Software, ScribeLink, and Amazon Web Services’ new HealthScribe product. Utilizing medical scribes to handle documentation can significantly reduce the administrative burden, allowing physicians to focus more on patient care rather than paperwork, potentially improving both physician and patient satisfaction. (As always, I have no financial interest in any product or service mentioned in this column.)

Provide Professional Development Opportunities: Offer opportunities for professional growth and development. This can include attending conferences, participating in research, or providing time and resources for continuing education. Such opportunities can reinvigorate a physician’s passion for medicine and improve job satisfaction.

Foster a Supportive Work Environment: Create a supportive work culture where staff and physicians feel comfortable discussing challenges and seeking support. Regular meetings or check-ins can help identify early signs of burnout and address them proactively.

Evaluate and Adjust Workloads: Regularly assess physician workloads to ensure they are manageable. Adjusting patient loads, redistributing tasks among team members, or hiring additional staff can help prevent burnout.

Leadership Training and Support: Provide training for leaders within the practice on recognizing signs of burnout and effective management strategies. Supportive leadership is crucial in creating an environment where physicians feel valued and heard.

Peer Support and Mentorship Programs: Establish peer support or mentorship programs where physicians can share experiences, offer advice, and provide emotional support to each other.

Feedback and Continuous Improvement: Managers should regularly solicit feedback from physicians regarding their workload, job satisfaction, and suggestions for improvements. Actively work on implementing feasible changes to address concerns.

Dr. Eastern practices dermatology and dermatologic surgery in Belleville, N.J. He is the author of numerous articles and textbook chapters, and is a longtime monthly columnist for Dermatology News. Write to him at [email protected].

Rheumatologists and their staff have been dutifully recording disease activity and patient-reported outcomes for decades, and now, all that drudgery is beginning to pay off with the introduction of artificial intelligence (AI) and natural language processing systems that can mine electronic health records (EHRs) for nuggets of research gold and accurately predict short-term rheumatoid arthritis (RA) outcomes.

“I think we have learned from our very early experiments that longitudinal deep learning models can forecast rheumatoid arthritis [RA] outcomes with actually surprising efficiency, with fewer patients than we assumed would be needed,” said Jinoos Yazdany, MD, MPH, chief of rheumatology at Zuckerberg San Francisco General Hospital and Trauma Center, and codirector of the University of California San Francisco (UCSF) Quality and Informatics Lab.

At the 2024 Rheumatoid Arthritis Research Summit (RA Summit 2024), presented by the Arthritis Foundation and the Hospital for Special Surgery in New York City, Dr. Yazdany discussed why rheumatologists are well positioned to take advantage of predictive analytics and how natural language processing systems can be used to extract previously hard-to-find data from EHRs, which can then be applied to RA prognostics and research.
 

Data Galore

EHR data can be particularly useful for RA research because of the large volume of information, clinical data such as notes and imaging, less selection bias compared with other data sources such as cohorts or randomized controlled trials, real-time access, and the fact that many records contain longitudinal data (follow-ups, etc.).

However, EHR data may have gaps or inaccurate coding, and data such as text and images may require significant data processing and scrubbing before it can be used to advance research. In addition, EHR data are subject to patient privacy and security concerns, can be plagued by incompatibility across different systems, and may not represent patients who have less access to care, Dr. Yazdany said.

She noted that most rheumatologists record some measure of RA disease activity and patient physical function, and that patient-reported outcomes have been routinely incorporated into clinical records, especially since the 1980 introduction of the Health Assessment Questionnaire.

“In rheumatology, by achieving consensus and building a national quality measurement program, we have a cohesive national RA outcome measure selection strategy. RA outcomes are available for a majority of patients seen by rheumatologists, and that’s a critical strength of EHR data,” she said.
 

Spinning Text Into Analytics

The challenge for investigators who want to use this treasure trove of RA data is that more than 80% of the data are in the form of text, which raises questions about how to best extract outcomes data and drug dosing information from the written record.

As described in an article published online in Arthritis Care & Research February 14, 2023, Dr. Yazdany and colleagues at UCSF and Stanford University developed a natural language processing “pipeline” designed to extract RA outcomes from clinical notes on all patients included in the American College of Rheumatology’s Rheumatology Informatics System for Effectiveness (RISE) registry.

The model used expert-curated terms and a text processing tool to identify patterns and numerical scores linked to outcome measures in the records.

“This was an enormously difficult and ambitious project because we had many, many sites, the data was very messy, we had very complicated [independent review board] procedures, and we actually had to go through de-identification procedures because we were using this data for research, so we learned a lot,” Dr. Yazdany said.

The model processed 34 million notes on 854,628 patients across 158 practices and 24 different EHR systems.

In internal validation studies, the models had 95% sensitivity, 87% positive predictive value (PPV), and an F1 score (a measure of predictive performance) of 91%. Applying the model to an EHR from a large, non-RISE health system for external validation, the natural language processing pipeline had a 92% sensitivity, 69% PPV, and an F1 score of 79%.

The investigators also looked at the use of OpenAI large language models, including GPT 3.5 and 4 to interpret complex prescription orders and found that after training with 100 examples, GPT 4 was able to correctly interpret 95.6% of orders. But this experiment came at a high computational and financial cost, with one experiment running north of $3000, Dr. Yazdany cautioned.
 

Predicting Outcomes

Experiments to see whether an AI system can forecast RA disease activity at the next clinic visit are in their early stages.

Dr. Yazdany and colleagues used EHR data from UCSF and Zuckerberg San Francisco General Hospital on patients with two RA diagnostic codes at 30 days apart, who had at least one disease-modifying antirheumatic drug prescription and two Clinical Disease Activity Index (CDAI) scores 30 days apart.

One model, designed to predict CDAI at the next visit by “playing the odds” based on clinical experience, showed that about 60% of patients at UCSF achieved treat-to-target goals, while the remaining 40% did not.

This model performed barely better than pure chance, with an area under the receiver operating characteristic curve (AUC) of 0.54.

A second model that included the patient’s last CDAI score also fared little better than a roll of the dice, with an AUC of 0.55.

However, a neural network or “deep learning” model designed to process data akin to the way that the human brain works performed much better at predicting outcomes at the second visit, with an AUC of 0.91.

Applying the UCSF-trained neural network model to the Zuckerberg San Francisco General Hospital population, with different patient characteristics from those of UCSF, the AUC was 0.74. Although this result was not as good as that seen when applied to UCSF patients, it demonstrated that the model retains some predictive capability across different hospital systems, Dr. Yazdany said.

The next steps, she said, are to build more robust models based on vast and varied patient data pools that will allow the predictive models to be generalized across various healthcare settings.
 

The Here and Now

In the Q & A following the presentation, an audience member said that the study was “very cool stuff.”

“Is there a way to sort of get ahead and think of the technology that we’re starting to pilot? Hospitals are already using AI scribes, for example, to collect the data that is going to make it much easier to feed it to the predictive analytics that we’re going to use,” she said.

Dr. Yazdany replied that “over the last couple of years, one of the projects that we’ve worked on is to interview rheumatologists who are participating in the RISE registry about the ways that they are collecting [patient-reported outcomes], and it has been fascinating: A vast majority of people are still using paper forms.”

“The challenge is that our patient populations are very diverse. Technology, and especially filling out forms via online platforms, doesn’t work for everybody, and in some ways, filling out the paper forms when you go to the doctor’s office is a great equalizer. So, I think that we have some real challenges, and the solutions have to be embedded in the real world,” she added.

Dr. Yazdany’s research was supported by grants from the Agency for Healthcare Research & Quality and the National Institutes of Health. She disclosed consulting fees and/or research support from AstraZeneca, Aurinia, Bristol Myers Squibb, Gilead, and Pfizer.

A version of this article appeared on Medscape.com.

Rheumatologists and their staff have been dutifully recording disease activity and patient-reported outcomes for decades, and now, all that drudgery is beginning to pay off with the introduction of artificial intelligence (AI) and natural language processing systems that can mine electronic health records (EHRs) for nuggets of research gold and accurately predict short-term rheumatoid arthritis (RA) outcomes.

“I think we have learned from our very early experiments that longitudinal deep learning models can forecast rheumatoid arthritis [RA] outcomes with actually surprising efficiency, with fewer patients than we assumed would be needed,” said Jinoos Yazdany, MD, MPH, chief of rheumatology at Zuckerberg San Francisco General Hospital and Trauma Center, and codirector of the University of California San Francisco (UCSF) Quality and Informatics Lab.

At the 2024 Rheumatoid Arthritis Research Summit (RA Summit 2024), presented by the Arthritis Foundation and the Hospital for Special Surgery in New York City, Dr. Yazdany discussed why rheumatologists are well positioned to take advantage of predictive analytics and how natural language processing systems can be used to extract previously hard-to-find data from EHRs, which can then be applied to RA prognostics and research.
 

Data Galore

EHR data can be particularly useful for RA research because of the large volume of information, clinical data such as notes and imaging, less selection bias compared with other data sources such as cohorts or randomized controlled trials, real-time access, and the fact that many records contain longitudinal data (follow-ups, etc.).

However, EHR data may have gaps or inaccurate coding, and data such as text and images may require significant data processing and scrubbing before it can be used to advance research. In addition, EHR data are subject to patient privacy and security concerns, can be plagued by incompatibility across different systems, and may not represent patients who have less access to care, Dr. Yazdany said.

She noted that most rheumatologists record some measure of RA disease activity and patient physical function, and that patient-reported outcomes have been routinely incorporated into clinical records, especially since the 1980 introduction of the Health Assessment Questionnaire.

“In rheumatology, by achieving consensus and building a national quality measurement program, we have a cohesive national RA outcome measure selection strategy. RA outcomes are available for a majority of patients seen by rheumatologists, and that’s a critical strength of EHR data,” she said.
 

Spinning Text Into Analytics

The challenge for investigators who want to use this treasure trove of RA data is that more than 80% of the data are in the form of text, which raises questions about how to best extract outcomes data and drug dosing information from the written record.

As described in an article published online in Arthritis Care & Research February 14, 2023, Dr. Yazdany and colleagues at UCSF and Stanford University developed a natural language processing “pipeline” designed to extract RA outcomes from clinical notes on all patients included in the American College of Rheumatology’s Rheumatology Informatics System for Effectiveness (RISE) registry.

The model used expert-curated terms and a text processing tool to identify patterns and numerical scores linked to outcome measures in the records.

“This was an enormously difficult and ambitious project because we had many, many sites, the data was very messy, we had very complicated [independent review board] procedures, and we actually had to go through de-identification procedures because we were using this data for research, so we learned a lot,” Dr. Yazdany said.

The model processed 34 million notes on 854,628 patients across 158 practices and 24 different EHR systems.

In internal validation studies, the models had 95% sensitivity, 87% positive predictive value (PPV), and an F1 score (a measure of predictive performance) of 91%. Applying the model to an EHR from a large, non-RISE health system for external validation, the natural language processing pipeline had a 92% sensitivity, 69% PPV, and an F1 score of 79%.

The investigators also looked at the use of OpenAI large language models, including GPT 3.5 and 4 to interpret complex prescription orders and found that after training with 100 examples, GPT 4 was able to correctly interpret 95.6% of orders. But this experiment came at a high computational and financial cost, with one experiment running north of $3000, Dr. Yazdany cautioned.
 

Predicting Outcomes

Experiments to see whether an AI system can forecast RA disease activity at the next clinic visit are in their early stages.

Dr. Yazdany and colleagues used EHR data from UCSF and Zuckerberg San Francisco General Hospital on patients with two RA diagnostic codes at 30 days apart, who had at least one disease-modifying antirheumatic drug prescription and two Clinical Disease Activity Index (CDAI) scores 30 days apart.

One model, designed to predict CDAI at the next visit by “playing the odds” based on clinical experience, showed that about 60% of patients at UCSF achieved treat-to-target goals, while the remaining 40% did not.

This model performed barely better than pure chance, with an area under the receiver operating characteristic curve (AUC) of 0.54.

A second model that included the patient’s last CDAI score also fared little better than a roll of the dice, with an AUC of 0.55.

However, a neural network or “deep learning” model designed to process data akin to the way that the human brain works performed much better at predicting outcomes at the second visit, with an AUC of 0.91.

Applying the UCSF-trained neural network model to the Zuckerberg San Francisco General Hospital population, with different patient characteristics from those of UCSF, the AUC was 0.74. Although this result was not as good as that seen when applied to UCSF patients, it demonstrated that the model retains some predictive capability across different hospital systems, Dr. Yazdany said.

The next steps, she said, are to build more robust models based on vast and varied patient data pools that will allow the predictive models to be generalized across various healthcare settings.
 

The Here and Now

In the Q & A following the presentation, an audience member said that the study was “very cool stuff.”

“Is there a way to sort of get ahead and think of the technology that we’re starting to pilot? Hospitals are already using AI scribes, for example, to collect the data that is going to make it much easier to feed it to the predictive analytics that we’re going to use,” she said.

Dr. Yazdany replied that “over the last couple of years, one of the projects that we’ve worked on is to interview rheumatologists who are participating in the RISE registry about the ways that they are collecting [patient-reported outcomes], and it has been fascinating: A vast majority of people are still using paper forms.”

“The challenge is that our patient populations are very diverse. Technology, and especially filling out forms via online platforms, doesn’t work for everybody, and in some ways, filling out the paper forms when you go to the doctor’s office is a great equalizer. So, I think that we have some real challenges, and the solutions have to be embedded in the real world,” she added.

Dr. Yazdany’s research was supported by grants from the Agency for Healthcare Research & Quality and the National Institutes of Health. She disclosed consulting fees and/or research support from AstraZeneca, Aurinia, Bristol Myers Squibb, Gilead, and Pfizer.

A version of this article appeared on Medscape.com.

Rheumatologists and their staff have been dutifully recording disease activity and patient-reported outcomes for decades, and now, all that drudgery is beginning to pay off with the introduction of artificial intelligence (AI) and natural language processing systems that can mine electronic health records (EHRs) for nuggets of research gold and accurately predict short-term rheumatoid arthritis (RA) outcomes.

“I think we have learned from our very early experiments that longitudinal deep learning models can forecast rheumatoid arthritis [RA] outcomes with actually surprising efficiency, with fewer patients than we assumed would be needed,” said Jinoos Yazdany, MD, MPH, chief of rheumatology at Zuckerberg San Francisco General Hospital and Trauma Center, and codirector of the University of California San Francisco (UCSF) Quality and Informatics Lab.

At the 2024 Rheumatoid Arthritis Research Summit (RA Summit 2024), presented by the Arthritis Foundation and the Hospital for Special Surgery in New York City, Dr. Yazdany discussed why rheumatologists are well positioned to take advantage of predictive analytics and how natural language processing systems can be used to extract previously hard-to-find data from EHRs, which can then be applied to RA prognostics and research.
 

Data Galore

EHR data can be particularly useful for RA research because of the large volume of information, clinical data such as notes and imaging, less selection bias compared with other data sources such as cohorts or randomized controlled trials, real-time access, and the fact that many records contain longitudinal data (follow-ups, etc.).

However, EHR data may have gaps or inaccurate coding, and data such as text and images may require significant data processing and scrubbing before it can be used to advance research. In addition, EHR data are subject to patient privacy and security concerns, can be plagued by incompatibility across different systems, and may not represent patients who have less access to care, Dr. Yazdany said.

She noted that most rheumatologists record some measure of RA disease activity and patient physical function, and that patient-reported outcomes have been routinely incorporated into clinical records, especially since the 1980 introduction of the Health Assessment Questionnaire.

“In rheumatology, by achieving consensus and building a national quality measurement program, we have a cohesive national RA outcome measure selection strategy. RA outcomes are available for a majority of patients seen by rheumatologists, and that’s a critical strength of EHR data,” she said.
 

Spinning Text Into Analytics

The challenge for investigators who want to use this treasure trove of RA data is that more than 80% of the data are in the form of text, which raises questions about how to best extract outcomes data and drug dosing information from the written record.

As described in an article published online in Arthritis Care & Research February 14, 2023, Dr. Yazdany and colleagues at UCSF and Stanford University developed a natural language processing “pipeline” designed to extract RA outcomes from clinical notes on all patients included in the American College of Rheumatology’s Rheumatology Informatics System for Effectiveness (RISE) registry.

The model used expert-curated terms and a text processing tool to identify patterns and numerical scores linked to outcome measures in the records.

“This was an enormously difficult and ambitious project because we had many, many sites, the data was very messy, we had very complicated [independent review board] procedures, and we actually had to go through de-identification procedures because we were using this data for research, so we learned a lot,” Dr. Yazdany said.

The model processed 34 million notes on 854,628 patients across 158 practices and 24 different EHR systems.

In internal validation studies, the models had 95% sensitivity, 87% positive predictive value (PPV), and an F1 score (a measure of predictive performance) of 91%. Applying the model to an EHR from a large, non-RISE health system for external validation, the natural language processing pipeline had a 92% sensitivity, 69% PPV, and an F1 score of 79%.

The investigators also looked at the use of OpenAI large language models, including GPT 3.5 and 4 to interpret complex prescription orders and found that after training with 100 examples, GPT 4 was able to correctly interpret 95.6% of orders. But this experiment came at a high computational and financial cost, with one experiment running north of $3000, Dr. Yazdany cautioned.
 

Predicting Outcomes

Experiments to see whether an AI system can forecast RA disease activity at the next clinic visit are in their early stages.

Dr. Yazdany and colleagues used EHR data from UCSF and Zuckerberg San Francisco General Hospital on patients with two RA diagnostic codes at 30 days apart, who had at least one disease-modifying antirheumatic drug prescription and two Clinical Disease Activity Index (CDAI) scores 30 days apart.

One model, designed to predict CDAI at the next visit by “playing the odds” based on clinical experience, showed that about 60% of patients at UCSF achieved treat-to-target goals, while the remaining 40% did not.

This model performed barely better than pure chance, with an area under the receiver operating characteristic curve (AUC) of 0.54.

A second model that included the patient’s last CDAI score also fared little better than a roll of the dice, with an AUC of 0.55.

However, a neural network or “deep learning” model designed to process data akin to the way that the human brain works performed much better at predicting outcomes at the second visit, with an AUC of 0.91.

Applying the UCSF-trained neural network model to the Zuckerberg San Francisco General Hospital population, with different patient characteristics from those of UCSF, the AUC was 0.74. Although this result was not as good as that seen when applied to UCSF patients, it demonstrated that the model retains some predictive capability across different hospital systems, Dr. Yazdany said.

The next steps, she said, are to build more robust models based on vast and varied patient data pools that will allow the predictive models to be generalized across various healthcare settings.
 

The Here and Now

In the Q & A following the presentation, an audience member said that the study was “very cool stuff.”

“Is there a way to sort of get ahead and think of the technology that we’re starting to pilot? Hospitals are already using AI scribes, for example, to collect the data that is going to make it much easier to feed it to the predictive analytics that we’re going to use,” she said.

Dr. Yazdany replied that “over the last couple of years, one of the projects that we’ve worked on is to interview rheumatologists who are participating in the RISE registry about the ways that they are collecting [patient-reported outcomes], and it has been fascinating: A vast majority of people are still using paper forms.”

“The challenge is that our patient populations are very diverse. Technology, and especially filling out forms via online platforms, doesn’t work for everybody, and in some ways, filling out the paper forms when you go to the doctor’s office is a great equalizer. So, I think that we have some real challenges, and the solutions have to be embedded in the real world,” she added.

Dr. Yazdany’s research was supported by grants from the Agency for Healthcare Research & Quality and the National Institutes of Health. She disclosed consulting fees and/or research support from AstraZeneca, Aurinia, Bristol Myers Squibb, Gilead, and Pfizer.

A version of this article appeared on Medscape.com.