Artificial intelligence (AI) is revolutionising medicine. The aim of this study was to detail its use, opinions, knowledge, and concerns in rheumatology and paediatric rheumatology. A web-based survey open to all professionals working in the field was developed by the Emerging EULAR Network (EMEUNET) and disseminated between March and July 2025 in collaboration with other international rheumatology societies (AFLAR, ArLAR, CARRA, PAFLAR, PANLAR). The survey was divided into 4 sections: (i) participants' characteristics, (ii) AI use and applications, (iii) opinions and knowledge, and (iv) concerns, needs, and expectations. Overall, 461 responses were collected from 59 countries. Respondents were mostly physicians who completed their training (316, 68.7%) and were based in Europe (170, 36.9%). Most participants (397, 86.7%) used AI for medical purposes, especially large language models (385, 83.7%) for grammar correction and brainstorming. Although there was broad optimism about its use (366, 79.6%), self-reported practical skills were predominantly basic or still in development (346, 75.1%), and knowledge was rarely defined as strong or expert-level (63, 13.7%). Concerns focused on ethics (314, 69%), lack of trust (316, 69.5%), and insufficient training (270, 59.3%). Disparities emerged across geographic regions in use, knowledge, and practical skills. AI is widely used and positively perceived in rheumatology, despite limited knowledge and practical skills, and regional disparities. Addressing gaps in ethics, transparency, and insufficient training through targeted education and implementation strategies will be essential to ensure an equitable and effective integration into clinical and research practice.
We aimed to develop expert consensus recommendations for musculoskeletal ultrasound (MSUS) education in Canadian rheumatology residency programs. This three-stage consensus study used an international working group of 13 MSUS experts and 5 non-expert stakeholders and incorporated cognitive load theory to ensure educational feasibility. In stage one, MSUS experts participated in a modified nominal group technique (NGT) to generate and prioritize items for stage two. In stage two, the working group completed three survey rounds using a modified Delphi technique, with consensus defined as agreement >70%. In stage three, experts reviewed items from stage two in a structured focus group, rating each for clinical utility, learnability, and feasibility of national implementation. The NGT generated 30 prioritized anatomical areas and pathologies. The Delphi working group represented 7/12 (58%) Canadian rheumatology programs (English stream) and two U.S. experts. The response rate across all Delphi rounds was 100%. Round one included 83 items, round two included 52, and round three included nine. Consensus was reached on all but five items, yielding 42 core and 38 optional items. In the structured focus group, consensus was reached that all core items are clinically valuable, learnable for rheumatology residents, and feasible for national implementation. Core items were then organized into 33 MSUS competencies, encompassing general ultrasound skills, features of inflammatory arthritis in the hands, wrists, and MTP joints, double contour sign in the MTP joints, and effusion detection in the knee and ankle joints. These recommendations provide a foundation for standardized MSUS curricula in rheumatology residency programs.
Real-world data (RWD) from nonrandomized clinical settings, such as registries and observational cohorts, can be used to address research questions for which randomized controlled trials (RCTs) are not feasible or appropriate. However, these data are susceptible to important sources of bias, including confounding by indication and immortal time bias. Target trial emulation (TTE) has emerged as a framework to improve the quality, credibility, and reproducibility of observational causal inference by aligning the design and analysis of observational studies with the structure of a hypothetical RCT. Appropriate application of the TTE framework requires careful specification of key design elements. Recent methodological guidance documents and reporting guidelines now provide practical support for researchers seeking to specify, conduct, and report TTE studies. Although TTE is increasingly used in rheumatology, its application in pediatric rheumatology remains limited despite the expanding availability of RWD. This review details the most recent methodological developments and guidelines, outlining the core concepts needed to apply the framework appropriately. The aim of this review is also to summarize recent applications in adult and pediatric immune-mediated diseases, and to explore the opportunities and limitations of this approach with a specific focus on pediatric rheumatology.
To evaluate contemporary global treatment of diffuse cutaneous systemic sclerosis (dcSSc) skin by SSc specialists. An anonymous survey was distributed via the Scleroderma Clinical Trials Consortium, European Scleroderma Trials and Research Group (EUSTAR), and Collaborative National Quality and Efficacy Registry (CONQUER) distribution lists between June and September 2025. The survey comprised four sections: Demographics, Treatment of dcSSc Skin, Impact of Recent Guidelines, and Clinical Trials. Responses included 103 physicians (93 completed): 43% North America, 31% Europe, 15% South America and 12% elsewhere. The majority practice within an SSc centre (80%) and participate in clinical trials (84%). Mycophenolate mofetil (MMF) was preferred first-line treatment (71%) for dcSSc without ILD, rising to 92% for dcSSc with mild/non-progressive ILD. For active skin involvement despite MMF, trial referral was preferred as next step irrespective of ILD (59% without ILD and 58% with ILD). Treatment preferences have evolved, with 40% increasing MMF use and 47% decreasing methotrexate use, over the previous 2-3 years. Biologic use has increased, with 56% and 37% reporting increasing rituximab and tocilizumab use, respectively. For dcSSc without ILD, 85% have prescribed rituximab, and 65% have prescribed tocilizumab. Biologic availability has impacted trial enrolment (73% agreement). Treatment decision-making for dcSSc skin involvement has been materially influenced by recent BSR and/or EULAR guidelines (43% agreement). MMF is preferred first-line for dcSSc. In the absence of compelling scientific justification for combination or step-up immunomodulatory approaches, trial referral is currently the preferred next treatment option. Biologic use is increasing, which impacts trial enrolment.
Inflammatory rheumatic and musculoskeletal diseases (RMDs) are hallmarked by an increased risk of cardiovascular (CV) disease compared with the general population. Evidence suggests that CV risk remains under-recognised and underdiscussed in routine care, although data on patient experience is lacking. We aimed to assess awareness, perceptions, and experiences regarding CV risk and prevention in patients with RMDs. An anonymous, online survey was codesigned by rheumatologists, researchers, and patient research partners in English language and translated into 9 languages. This was distributed through patient organisations/forums and social media. A total of 951 patients with RMDs from 43 countries (mainly females, aged 50-70 years, and with a diagnosis of rheumatoid arthritis, spondyloarthritis, or systemic lupus erythematosus) completed the survey. Although 68% of respondents acknowledged an increased CV risk in RMDs, and traditional CV risk factors were recognised by up to 60% of the participants, knowledge of disease-specific contributors (eg, disease activity) to CV risk was limited. Only 35% of respondents had ever discussed CV risk with a healthcare provider, mainly rheumatologists or cardiologists, and the content primarily focused on lifestyle advice. Lack of awareness on the topic relevance (38%), perceived complexity of the topic (25%), limited time of clinical visits (35%), and lack of informative materials (33%) emerged as the main key barriers among those poorly/not informed. This large, multinational survey demonstrates substantial gaps in CV risk awareness, knowledge, and patient-provider communication among patients with RMDs. These figures support the development of structured, patient-centred educational strategies and enhanced multidisciplinary communication schemes to improve CV prevention in RMDs.
Over the last two years, inflammatory dermatology has undergone a transition from morphology-based classification toward pathway-driven immune endotyping. Atopic dermatitis, psoriatic disease, and hidradenitis suppurativa have emerged as paradigmatic models of this transition, illustrating how epithelial dysfunction, involving autoinflammation and adaptive immunity, may lead to systemic inflammatory crosstalks, connecting cutaneous and musculoskeletal manifestations. Recent studies have refined the understanding of progressive immune maturation in atopic dermatitis, while psoriasis and juvenile psoriatic arthritis support skin-to-joint inflammatory models involving IL-23/IL-17 signaling. Hidradenitis suppurativa is increasingly recognized as a systemic autoinflammatory disease within a broader neutrophilic inflammatory spectrum. Parallel therapeutic advances involving IL-17, IL-23, TNF, and JAK inhibition reinforced the translational importance of these discoveries and accelerated the development of precision dermatology strategies.
OBJECTIVE: To compare patients with systemic lupus erythematosus (SLE) from different centers with respect to demographics and Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index (SLICC/ACR DI) scores, and to assess whether the SLICC/ACR DI changed over time, and whether initial DI scores were related to outcome. METHODS: Members of SLICC completed DI scores and patient demographics on patients followed in their centers. Information was provided at 2, 5-10, and > 10 years of followup. Data were entered on computer and analyzed on SPSS/PC+ and SAS using descriptive statistics and analysis of variance. RESULTS: Information for 1297 patients within 2 years of first clinic visit was submitted from 8 centers. There were 1187 women and 110 men with a mean age at diagnosis of SLE of 32 years. Seven hundred sixty-two were Caucasian, 423 were black, and the remainder were of other races. There were more blacks in the American centers than in Canadian or European centers. Five centers provided information for the 3 time periods. The DI increased over time. Ninety-nine patients had died. Higher SLICC/ACR DI scores were documented in patients who went on to die. CONCLUSION: The SLICC/ACR DI is a valid measure for damage in SLE.
OBJECTIVE: We previously reported high Systemic Lupus International Collaborating Clinics (SLICC) scores in fatal cases of systemic lupus erythematosus (SLE) from our inception cohort. This study was done to clarify if the SLICC damage scores 5 years after diagnosis predicted the outcome. METHODS: We studied 80 patients with SLE (70 women, 10 men), all enrolled and diagnosed during the years 1981 through 1991 in our inception cohort, and all alive 5 years after inclusion into the cohort. In all patients the SLICC/American College of Rheumatology (ACR) damage index (DI) was scored at 5 years after SLE diagnosis, and these scores were tested for predictive value. The outcomes were survival or late mortality within the following median observation period of 7 years. All surviving patients were followed through 1999, and no patient was lost to followup. RESULTS: At study entry, 5 years after the diagnosis of SLE, 37 patients had no damage to score with SLICC. Of the remaining 43 patients, 25 had a score of 1 and 18 had a score of 2 or more. In total, 14 fatalities occurred within 7 years after study entry, 7 among the 18 with initial SLICC/ACR DI of 2 or more compared with 7 fatalities among the 62 with less or no damage (p < 0.01). Cardiovascular or cerebrovascular SLICC/ACR DI items were more common in fatal cases than in survivors (p < 0.001). A SLICC score at 5 years of 2 or more increased the relative risk for fatality by 3.4 (95% CI 1.5-14.4), and had a predictive value of 38%. A SLICC score of 0 at 5 years gave an odds ratio in favor of survival of 0.06 (95% CI 0.0-0.5) and had a predictive value for survival of 97%. During an extended followup for one more year the predictive value of damage for fatalities was even more pronounced (p = 0.003, log-rank). CONCLUSION: SLICC damage scores registered 5 years after SLE diagnosis have a high predictive value for survival during the following median observation time of 7 years. These data provide strong evidence that the items included in the SLICC score are clinically relevant.
Idiopathic intracranial hypertension (IIH) is characterised by raised intracranial pressure (ICP) and typically affects young women with obesity. Patients are at risk of permanent visual loss due to papilloedema. Some require emergency intervention to rapidly reduce papilloedema and preserve vision. The international standard of care for patients with sight-threatening IIH is cerebrospinal fluid (CSF) shunting. However, dural venous sinus stenting (DVSS) is an emerging procedure that is offered at many neuroscience centres internationally as the primary intervention. Currently, there are no randomised controlled trial data supporting the efficacy of any interventional approach for preserving vision in sight-threatening IIH. IIH Intervention is a UK-based two-arm, open-label, multicentre, randomised controlled phase IIb clinical trial with integrated health economic evaluation to compare CSF shunting with DVSS in patients who have confirmed IIH and are at risk of permanent visual loss due to severe papilloedema. The primary outcome is the global thickness of the peripapillary retinal nerve fibre layer (RNFL), an indicator of papilloedema, measured by optical coherence tomography (OCT) over a 6-month period. Secondary outcomes are global thickness of the RNFL over 12 and 24 months, as well as macular ganglion cell layer volume, perimetric mean deviation, headache outcomes, intervention reporting measures (including complications and revisions) and patient-reported outcomes over 6, 12 and 24 months. The protocol was approved initially on 12 December 2022 by West Midlands-South Birmingham Research Ethics Committee (ref: 22/WM/0230). Participants will be required to provide written informed consent. The results of this trial will be disseminated through national and international presentations and peer-reviewed publications. ISRCTN57142415.
The International Association for the Study of Pain (IASP) has defined three mechanisms for chronic pain: nociceptive, neuropathic, and nociplastic. Accurate differentiation between these mechanisms is essential for treatment selection but remains challenging due to symptom overlap and reliance on clinical judgment. Patient-reported questionnaires may support diagnosis, though their diagnostic accuracy relative to physician assessment has not been fully established. This study aimed to investigate the prevalence and overlap of chronic pain mechanisms in a tertiary pain clinic population and to compare physician-assigned diagnoses with results from patient-reported questionnaires. This cross-sectional study recruited 89 adult patients with chronic pain from the Rambam Institute for Pain Medicine at Rambam Health Care Campus in Haifa, Israel. Each participant underwent a comprehensive physician evaluation based on the criteria of the International Classification of Diseases and Related Health Problems, 11th revision, and completed the Central Sensitization Inventory (CSI), PainDETECT, and 2016 American College of Rheumatology (ACR) Fibromyalgia Diagnostic Criteria surveys. Diagnostic agreement, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated using physician diagnosis as the reference standard. Physician diagnoses identified neuropathic pain in 51 patients (57.3%), nociceptive pain in 38 (42.7%), and nociplastic pain in 25 (28.1%), with substantial diagnostic overlap (32.6% with ≥2 mechanisms). The CSI demonstrated high sensitivity (92.0%) but low specificity (45.3%) for nociplastic pain. PainDETECT showed moderate accuracy for neuropathic pain (sensitivity 68.6%, specificity 52.6%) and higher specificity for nociceptive pain (86.3%). The 2016 ACR Fibromyalgia criteria provided a more balanced diagnostic profile (sensitivity 88.0%, specificity 68.8%). Questionnaire-based measures are useful for screening chronic pain mechanisms but demonstrate variable accuracy when compared to physician diagnoses, particularly in the context of overlapping pain phenotypes. Refinement of diagnostic thresholds, combined assessment strategies, and updated questionnaires may improve differentiation of pain subtypes and enhance clinical decision-making.
The WHO/ILAR core set of endpoints for rheumatoid arthritis clinical trials signifies progress in a continuing worldwide effort. This core set includes the following measures: pain, patient global assessment, physical disability, swollen joints, tender joints, acute phase reactants, and physician global assessment; in studies of one or more years' duration, radiographs of joints should be performed.
ObjectivesThe importance of low detectable C-reactive protein (CRP) levels in patients with systemic lupus erythematosus (SLE) remains unclear. This study aimed to examine the association between CRP levels and damage accrual in SLE patients with low disease activity.MethodsPatients with SLE who met the criteria for lupus low disease activity state (LLDAS) at registration were divided into two groups: serum CRP levels consistently >1.0 mg/L (high CRP group) or ≤1.0 mg/L (low CRP group) at both registration and the first year. Damage accrual was assessed by evaluating the increase in the Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index (SDI).ResultsAmong the 276 patients, 81 were in the high CRP group and 195 were in the low CRP group. Patients in the high CRP group were older, included a lower proportion of women and individuals using hydroxychloroquine, and had higher body mass index and SDI values at baseline. CRP categories were not associated with SDI progression (HR: 1.40, 95% CI: 0.85-2.31), whereas older patients were more likely to experience damage accrual (HR: 1.03, 95% CI: 1.01-1.05, p = 0.002). In the subgroup analysis limited to younger patients, the high CRP group showed more SDI progression (HR: 2.91, 95% CI: 1.21-7.03, p = 0.017).ConclusionCRP levels were associated with SDI progression in younger patients with LLDAS. Even in patients with inactive SLE, positive CRP levels may indicate the need to modify the treatment for SLE and reassess atherosclerotic risk factors.
Introduction Systemic lupus erythematosus frequently manifests as lupus nephritis (LN), a major cause of renal morbidity. Standard markers lack specificity, making renal biopsy essential. Complement component 4D (C4d) staining, reflecting classical complement activation, correlates with histopathological lesions and outcomes. This study evaluates C4d's clinical relevance in biopsy-proven LN, emphasizing its potential in detecting disease severity. Materials & methods This observational study at Bangabandhu Sheikh Mujib Medical University (BSMMU), Dhaka, conducted from September 2021 to August 2022, enrolled 45 adults with biopsy-proven LN; 43 completed follow-up. Participants were selected based on revised American College of Rheumatology (ACR) criteria and underwent thorough clinical evaluation, serological testing, and ultrasound-guided renal biopsy. Histopathological assessment followed the International Society of Nephrology and the Renal Pathology Society (ISN/RPS) 2003 classification, including activity and chronicity indices and immunohistochemical C4d staining. Follow-up investigations were performed at two and six months. Results Among the 43 patients with LN (mean age 28.4 ± 7.7 years; 86% female subjects), 42 (97.7%) patients had hematuria and 25 (58.1%) had renal impairment. Class IV was the predominant histological subtype in 18 (41.9%) cases, with mild chronicity in 20 (71.4%) cases. C4d deposition, primarily arteriolar in 24 (55.8%) cases, correlated significantly with Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) scores and histological activity indices (p<0.05). There is a significant relationship between C4d intensity and SLEDAI score, highlighting the prognostic utility of C4d staining. Conclusion The present study identifies C4d staining to be associated with active renal damage and higher SLEDAI scores, reflecting immune-mediated injury. These findings highlight its value as a potential marker that may enhance current histological assessment. Therefore, staining with C4d on renal biopsy may be emphasized for early detection of severity in patients with lupus.
The transition from pediatric to adult care is a complex and challenging issue for increasing numbers of adolescents and young adults (AYA). Despite increasing attention to the transition of AYA with chronic medical conditions, there remain difficulties which vary by local resources, disease processes, and individual factors. There are also poorly understood factors regarding transition involving rare diseases, including glomerular disease. Specialties and organizations are working to evaluate and enhance the transition process; however, the variability in resources limits formation of a true consensus. AYA with glomerular disease, broadly as chronic kidney disease and specifically with respect to diagnoses such as lupus nephritis, are vulnerable during the health care transition (HCT). Though young, many of these patients have severe disease with the loss of kidney function over their disease course and risk significant complications around the time of transition. Many facets make HCT challenging related to personal, caregiver, provider, and system factors. Despite difficulties, some groups have found success, largely in dedicated clinics designed to promote successful transition from pediatric to adult models of care. Society guidelines focus on a structured, intentional process providing continuity, with or without a dedicated transition clinic. Preparation, communication across the transition, and support for patients and providers with additional staff are all recommended aspects of transition from the International Society of Nephrology and International Pediatric Nephrology Association. Additionally, transition toolkits are available from several sources which can aid in developing a structured transition program for patients with glomerular disease. (1) There are many challenges in transitioning AYA patients with glomerular disease from pediatric to adult care. (2) HCT is a vulnerable period with increased risk of poor outcomes. (3) A successful HCT may be promoted by developing a structured process and/or program focused on preparation, communication, and support.
This is the initial report of the generic health OMERACT study concerned with the sensitivity to change of generic quality of life (QOL) measures. Our objective was to determine which QOL instrument is best able to show a statistically significant improvement in patients with rheumatoid arthritis (RA) demonstrating relevant improvement in a core set of disease activity and disease-specific disability measures. A multicenter controlled trial of a single group with repeated measurements at 0 (baseline), 3, and 6 months was conducted. All participating centers recruited 10 patients with RA who were about to start methotrexate therapy for the first time because of active disease. Assessments included disease activity measures, disease-specific disability measures, and generic QOL measures. To date, 40 patients have been recruited from 4 centers for the study. After 6 months of treatment many of the generic QOL measures showed a 20% improvement from baseline and medium standardized response means around 0.5. In particular, the Nottingham Health Profile (NHP) and the Rheumatoid Arthritis Quality of Life (RAQOL) measures had the largest percentage improvement (22 and 29%, respectively) and standardized response means (both with 0.54). Early results on the sensitivity of generic health QOL measures are promising, in particular for the NHP and RAQOL measures.
Hereditary angioedema is a rare but potentially life-threatening disorder in which outcomes depend not only on correct diagnosis and effective medicines, but also on how health systems organize referral, laboratory confirmation, emergency pathways, reimbursement, home treatment, and long-term follow-up. Selected health systems in the Balkan Peninsula area provide a particularly informative setting for health-system comparison because neighboring countries with active hereditary angioedema expertise differ substantially in rare-disease governance, registry maturity, diagnostic infrastructure, treatment coverage, and patient-organization capacity. This Policy and Practice Review synthesizes international guidance, published regional literature, comparative country information from the Balkan Experts in Angioedema: Consensus and Ongoing Navigation (BEACON) initiative, and advocacy-informed implementation insights to assess current care delivery in Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Greece, Romania, Serbia, Slovenia, and Türkiye. Across the region, the most consistent problems are prolonged diagnostic delay, unequal access to complement and genetic testing, approved-but-not-reimbursed modern therapies, hospital-only access to rescue medication, uneven use of home treatment and self-administration, incomplete emergency preparedness, and variable registry and advocacy infrastructure. Countries with stronger alignment between policy frameworks, specialist centers, registries, reimbursement pathways, and patient organizations appear better positioned to deliver guideline-concordant care, whereas fragmentation at any point in the care pathway reduces the practical value of therapeutic advances. The review argues that the main barriers to equitable hereditary angioedema care across the included health systems are now predominantly regulatory, financing, organizational, and educational rather than scientific. We therefore propose actionable recommendations for ministries and payers, specialist centers and professional societies, emergency-care systems, registry stakeholders, and patient organizations, with the goal of converting regional heterogeneity into a structured quality-improvement agenda.
Behçet's syndrome (BS) is a chronic, multisystem variable vessel vasculitis defined by recurrent oral and genital ulcers, diverse mucocutaneous lesions, and potential involvement of the eyes, joints, vasculature, central nervous system, and gastrointestinal tract. Diagnosis remains a clinical challenge given the absence of pathognomonic laboratory or histological findings. We present a case of a 36-year-old Caucasian male patient with hypothyroidism who developed a severe, rapidly progressive first episode of BS characterized by hemorrhagic vesicular and bullous skin lesions, oral ulceration, and necrotic genital ulceration requiring surgical debridement. Extensive infectious evaluation, including plasma cell-free metagenomic next-generation sequencing (cf-mNGS), was entirely negative. Serologic workup was unremarkable; HLA-B51 was negative, and pathergy was equivocal. Skin punch biopsy demonstrated pan-dermal neutrophilic inflammation with acute vasculitis and focal epidermal necrosis - a critical histopathological feature distinguishing BS from Sweet syndrome, in which true vasculitis is characteristically absent. Under the International Criteria for Behçet's Disease (ICBD), the patient scored ≥4 points (oral ulcers: 2 points; genital ulcers: 2 points; skin lesions: 1 point). He responded to high-dose corticosteroids (prednisone 50 mg daily) and colchicine, achieving full remission within nine weeks with no recurrence. This case illustrates the diagnostic complexity of BS in the absence of classic genetic markers, emphasizes histopathology as the critical discriminator from neutrophilic dermatosis mimics, and underscores the importance of systematic multidisciplinary evaluation before initiating immunosuppressive therapy.
This study aimed to evaluate the efficacy of ivarmacitinib in patients with active ankylosing spondylitis (AS) according to baseline characteristics. Data were derived from a phase II/III trial (NCT04481139). Patients with active AS who received either ivarmacitinib 4 mg (n = 187) or placebo (n = 186) were included. Subgroup analyses were performed based on age, sex, body mass index (BMI), AS duration, history of biological or Janus kinase (JAK) inhibitor use, C-reactive protein (CRP) level, total back pain visual analogue scale (VAS) score, and night pain VAS score. At week 12 (W12), Assessment of SpondyloArthritis international Society 20% improvement (ASAS20) response rates were significantly higher in the ivarmacitinib arm compared to the placebo arm across subgroups stratified by sex, BMI, AS duration, history of biological/JAK inhibitor use, total back pain VAS score, and night pain VAS score. ASAS5/6 response rates at W12 were higher in the ivarmacitinib arm across all subgroups. Greater improvements in Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) scores at W12 were observed in the ivarmacitinib arm across subgroups defined by age, sex, BMI, AS duration, total back pain VAS score, and night pain VAS score. The ivarmacitinib arm achieved greater improvements in Ankylosing Spondylitis Disease Activity Score (ASDAS) at W12 across all subgroups. Efficacy outcomes improved through W24 across all subgroups with continuous ivarmacitinib treatment and after switching from placebo at W12. Ivarmacitinib 4 mg demonstrated consistent efficacy in patients with active AS across diverse demographic and clinical subgroups. Key Points • Ivarmacitinib improved treatment response rates vs. placebo at W12 across multiple subgroups. • Efficacy benefits were sustained through 24 weeks after treatment with ivarmacitinib. • Efficacy improvements were observed after switching from placebo to ivarmacitinib.
We extracted a validated disease activity measure in rheumatoid arthritis (RA), the Clinical Disease Activity Index (CDAI), from a large tertiary academic medical center electronic health record (EHR) using an automated large language model (LLM)-based approach without requiring model pretraining. The New York Presbyterian/Columbia University Medical Center Clinical Data Warehouse contains EHR data for over 4.5 million patients. RA patients were identified using International Classification of Disease-9 (ICD-9) and ICD-10 codes. Expert-curated CDAI keywords were extracted from unstructured notes using an automated natural language processing (NLP) pipeline leveraging GPT-4o API, a HIPAA-compliant, institutionally approved LLM platform. Performance was evaluated against expert chart review. Among 2756 RA patients with notes, 1038 (37.7%) were seropositive, 796 (28.9%) were seronegative, and 922 (33.4%) had unknown serostatus. Clinical Disease Activity Index and its components were extracted in 15.4% (160/1038) of seropositive patients indicating remission or low disease activity. Clinical Disease Activity Index documentation was more frequent among patients with multiple notes and among faculty, with high extraction accuracy (precision/recall/F1 = 0.97). This represents the first attempt to employ a zero-shot, ChatGPT-powered LLM platform to extract RA disease activity measures from real-world EHR data. Although a low prevalence of documentation was noted, important distinctions were observed when patients were subgrouped by serostatus, level of training, and number of visits. An LLM-based pipeline accurately extracted CDAI from a single large academic EHR, revealing infrequent real-world documentation.