Harnessing ChatGPT and GPT-4 for evaluating the rheumatology questions of the Spanish access exam to specialized medical training.

The emergence of large language models (LLM) with remarkable performance such as ChatGPT and GPT-4, has led to an unprecedented uptake in the population. One of their most promising and studied applications concerns education due to their ability to understand and generate human-like text, creating a multitude of opportunities for enhancing educational practices and outcomes. The objective of this study is twofold: to assess the accuracy of ChatGPT/GPT-4 in answering rheumatology questions from the access exam to specialized medical training in Spain (MIR), and to evaluate the medical reasoning followed by these LLM to answer those questions. A dataset, RheumaMIR, of 145 rheumatology-related questions, extracted from the exams held between 2010 and 2023, was created for that purpose, used as a prompt for the LLM, and was publicly distributed. Six rheumatologists with clinical and teaching experience evaluated the clinical reasoning of the chatbots using a 5-point Likert scale and their degree of agreement was analyzed. The association between variables that could influence the models' accuracy (i.e., year of the exam question, disease addressed, type of question and genre) was studied. ChatGPT demonstrated a high level of performance in both accuracy, 66.43%, and clinical reasoning, median (Q1-Q3), 4.5 (2.33-4.67). However, GPT-4 showed better performance with an accuracy score of 93.71% and a median clinical reasoning value of 4.67 (4.5-4.83). These findings suggest that LLM may serve as valuable tools in rheumatology education, aiding in exam preparation and supplementing traditional teaching methods.

Understanding the role and adoption of artificial intelligence techniques in rheumatology research: An in-depth review of the literature.

The major and upward trend in the number of published research related to rheumatic and musculoskeletal diseases, in which artificial intelligence plays a key role, has exhibited the interest of rheumatology researchers in using these techniques to answer their research questions. In this review, we analyse the original research articles that combine both worlds in a five- year period (2017-2021). In contrast to other published papers on the same topic, we first studied the review and recommendation articles that were published during that period, including up to October 2022, as well as the publication trends. Secondly, we review the published research articles and classify them into one of the following categories: disease identification and prediction, disease classification, patient stratification and disease subtype identification, disease progression and activity, treatment response, and predictors of outcomes. Thirdly, we provide a table with illustrative studies in which artificial intelligence techniques have played a central role in more than twenty rheumatic and musculoskeletal diseases. Finally, the findings of the research articles, in terms of disease and/or data science techniques employed, are highlighted in a discussion. Therefore, the present review aims to characterise how researchers are applying data science techniques in the rheumatology medical field. The most immediate conclusions that can be drawn from this work are: multiple and novel data science techniques have been used in a wide range of rheumatic and musculoskeletal diseases including rare diseases; the sample size and the data type used are heterogeneous, and new technical approaches are expected to arrive in the short-middle term.

Difficult-to-treat rheumatoid arthritis (D2T RA): clinical issues at early stages of disease.

OBJECTIVES: Most studies on difficult-to-treat rheumatoid arthritis (D2T RA) have focused on established RA. Here, we analyse whether disease activity in the early stages of RA could influence progression to a D2T RA under real-life conditions. Other clinical and treatment-related factors were also analysed. METHODS: A longitudinal multicentre study of patients with RA was conducted from 2009 to 2018. Patients were followed up until January 2021. D2T RA was defined based on EULAR criteria (treatment failure, signs suggestive of currently active/progressive disease and management being perceived as problematic by the rheumatologist and/or patient). The main variable was disease activity in the early stages. The covariates were sociodemographic, clinical and treatment-related factors. We ran a multivariable logistic regression analysis to investigate risk factors associated with progression to D2T RA. RESULTS: The study population comprised 631 patients and 35 (5.87%) developed D2T RA. At the time of diagnosis, the D2T RA group were younger, with a higher disability, 28-joint Disease Activity Score (DAS28) score, tender joint count and pain scores. In our final model, DAS28 was not statistically significantly associated with D2T RA. No differences were found between groups for therapy. Disability was independently associated with D2T RA (OR: 1.89; p=0.01). CONCLUSIONS: In this cohort of patients newly diagnosed with RA, our results do not allow us to prove the influence of active disease according to DAS28. However, we did find that younger patients and those with elevated initial disability scores are more likely to develop D2T RA regardless of other factors.

Therapeutic Response After Immunosuppressive Drug Prescription in Non-infectious Uveitis: A Survival Analysis.

INTRODUCTION: To identify factors affecting the response rate to immunosuppressive drugs (ISDs) in patients with non-infectious uveitis (NIU). METHODS: This longitudinal retrospective cohort study included patients from the Hospital Clinico San Carlos Uveitis Clinic diagnosed with NIU from 1992 to 2016. Subjects were followed up from ISD prescription until the achievement of good therapeutic response (GTR), ISD treatment change, or up to 12 months. GTR was defined as the complete resolution of the eye inflammatory manifestations with a corticosteroid dose ≤ 10 or ≤ 5 mg per day of prednisone or equivalent (GTR10 and GTR5, respectively) maintained for at least 28 days. Kaplan-Meier curves were estimated for GTR. Demographic, clinical, and treatment-related factors were analyzed using Cox robust regression. RESULTS: A total of 73 patients (100 episodes of ISD prescription) were analyzed. In 44 and 41 episodes, GTR10 and GTR5 were achieved, respectively. A lower hazard for both GTRs was associated with uveitic macular edema at prescription and with a higher "highest oral corticosteroid dose prescribed in the year before ISD prescription". GTR10 was higher if cyclosporine was prescribed (compared to other ISDs), and if a higher number of ISDs had been previously prescribed. GTR5 hazard was lower for patients with posterior uveitis or if the ISDs were prescribed before 2008, and higher if periocular corticosteroids had been administered before ISD prescription, or if the duration of the posterior segment activity was shorter. CONCLUSIONS: Factors associated with GTR to ISDs may help to identify patients with NIUs who could benefit from a thorough follow-up.

There and back again: analyzing the effect of outpatient readmission on the quality of life of patients attending a rheumatology clinic.

AIMS: The aim of this study was to assess the effect of "outpatient readmissions" on the health-related quality of life (HR-QoL) of outpatients from a rheumatology clinic, meaning the effect of the patient's return to the outpatient clinic after having received care and been discharged. METHODS: We conducted an observational longitudinal retrospective study, with patients selected from the Hospital Clínico San Carlos Musculoskeletal cohort, based on having received at least one discharge from the outpatient clinic and having returned (readmission) at least once after the discharge. The main outcomes were the patients' baseline HR-QoL (measured on the first visit of each episode) and the ΔHR-QoL (difference between the HR-QoL in the last and the first visit of each episode). Successive episodes of admission and readmission were chronologically ordered, paired and analyzed using nested linear mixed models, nested by patients and by admission-readmission tandem. We carried out bivariable and multivariable analyses to assess the effect of demographic, clinical, treatment and comorbidity-related variables in both main outcomes. RESULTS: For the first main outcome, 5887 patients (13,772 episodes) were analyzed. Based on the multivariable level, readmission showed no significant marginal effect on the baseline HR-QoL (p-value = 0.17). Conversely, when analyzing the ΔHR-QoL, we did observe a negative and significant marginal effect (p-value = 0.028), meaning that readmission was associated with a lower gain in the HR-QoL during the follow-up, compared with the previous episode. CONCLUSION: In the outpatient setting, readmission exerts a deleterious effect in patients undergoing this process. Identification of outpatients more likely to be readmitted could increase the value of the care provided.

Predictors of health-related quality of life in musculoskeletal disease patients: a longitudinal analysis.

INTRODUCTION: Rheumatic and musculoskeletal diseases (RMDs) have a significant impact on patients' health-related quality of life (HRQoL) exacerbating disability, reducing independence and work capacity, among others. Predictors' identification affecting HRQoL could help to place efforts that minimize the deleterious impact of these conditions on patients' wellbeing. This study evaluates the influence of demographic and clinical predictors on the HRQoL of a cohort of RMD patients, measured using the Rosser classification index (RCI). METHODS: We included patients attending the Hospital Clínico San Carlos (HCSC) rheumatology outpatient clinic from 1 April 2007 to 30 November 2017. The primary outcome was the HRQoL assessed in each of the patient's visits using the RCI. Demographic and clinical variables extracted from a departmental electronic health record (EHR) were used as predictors: RMD diagnoses, treatments, comorbidities, and averaged HRQoL values from previous periods (for this last variable, values were imputed if no information was available). Association between predictors and HRQoL was analyzed using penalized generalized estimating equations (PGEEs). To account for imputation bias, the PGEE model was repeated excluding averaged HRQoL predictors, and common predictors were considered. DISCUSSION: A total of 18,187 outpatients with 95,960 visits were included. From 410 initial predictors, 19 were independently associated with patients' HRQoL in both PGEE models. Chronic kidney disease (CKD), an episode of prescription of third level analgesics, monoarthritis, and fibromyalgia diagnoses were associated with worse HRQoL. Conversely, the prescription in the previous visit of acid-lowering medication, colchicine, and third level analgesics was associated with better HRQoL. CONCLUSION: We have identified several diagnoses, treatments, and comorbidities independently associated with HRQoL in a cohort of outpatients attending a rheumatology clinic.

Factors Associated with Health-Related Quality of Life in Psoriatic Arthritis Patients: A Longitudinal Analysis.

INTRODUCTION: Psoriatic arthritis (PsA) is considered a multifaceted disease, with patients reporting low health-related quality of life (HRQoL). Data on disease burden are substantial and there exists a need for properly designed studies to learn more about the evolution of HRQoL in this condition. This study aims to identify factors associated to HRQoL evolution in PsA patients followed-up in a real-world setting in Spain. METHODS: We conducted a retrospective longitudinal observational study including incident patients from the rheumatology outpatient clinic of Hospital Clínico San Carlos (Madrid, Spain), diagnosed for the first time of PsA, defined as having received any ICD9/ICD10 diagnosis code of PsA, from 2007 to 2016, and followed-up until loss of follow-up, death, or November 2017. The influence of demographic and clinical variables in baseline HRQoL [assessed with the Rosser Classification Index (RCI)] was analyzed using bivariable and multivariable generalized linear models. The influence of those variables and of treatment-related factors in repeated measures of HRQoL was analyzed using bivariable and multivariable generalized estimating equations (GEE) models nested by patient. RESULTS: Two hundred and thirty patients were included in the analysis, with 3384 registered visits. At baseline, older age, a previous diagnosis of obesity, and the presence of enthesitis were significantly associated with worse HRQoL. During follow-up, using an exchangeable working correlation structure, the presence of enthesitis was also associated with worse HRQoL, coefficient (95% CI) - 0.006 (- 0.01 to - 0.002), p = 1.00E-03; conversely, treatment with methotrexate or antimalarials was associated with better HRQoL with 0.007 (0.001-0.014), p = 0.020 and 0.003 (0.001-0.005), p = 3.00E-03, respectively. CONCLUSIONS: Musculoskeletal manifestations and comorbidities exert a deleterious effect in HRQoL of PsA patients. Therefore, the optimal management of this condition needs to also address these manifestations in order to try to restore the QoL of these patients.

Influence of colchicine prescription in COVID-19-related hospital admissions: a survival analysis.

AIMS: To analyze the association between colchicine prescription and COVID-19-related hospital admissions in patients with rheumatic and musculoskeletal diseases (RMDs). METHODS: Patients attending a rheumatology outpatient clinic from a tertiary care center in Madrid, Spain, from 1 September 2019 to 29 February 2020 were included. Patients were assigned as exposed or unexposed based on whether they were prescribed with colchicine in their last visit to the clinic during the 6 months before the start of the observation period. Treatment changes during the observation period were also considered. The primary outcome was COVID-19-related hospital admissions between 1 March and 20 May 2020. Secondary outcome included COVID-19-related mortality. Several weighting techniques for data balancing, based and non-based on the propensity score, followed by Cox regressions were performed to estimate the association of colchicine prescription on both outcomes. DISCUSSION: The number of patients entered in the study was 9379, with 406 and 9002 exposed and unexposed follow-up periods, respectively. Generalized Boosted Models (GBMs) and Empirical Balancing Calibration Weighting (EBCW) methods showed the best balance for COVID-19-related hospital admissions. Colchicine prescription did not show a statistically significant association after covariable balancing (p-value = 0.195 and 0.059 for GBM and EBCW, respectively). Regarding mortality, the low number of events prevented a success variable balancing and analysis. CONCLUSION: Colchicine prescription does not play a significant protective or risk role in RMD patients regarding COVID-19-related hospital admissions. Our observations could support the maintenance of colchicine prescription in those patients already being treated, as it is not associated with a worse prognosis. PLAIN LANGUAGE TITLE: Colchicine influence in COVID-19-related hospital admissions.

Risk factors for hospital admissions related to COVID-19 in patients with autoimmune inflammatory rheumatic diseases.

OBJECTIVES: To describe patients with autoimmune inflammatory rheumatic diseases (AIRD) who had COVID-19 disease; to compare patients who required hospital admission with those who did not and assess risk factors for hospital admission related to COVID-19. METHODS: An observational longitudinal study was conducted during the pandemic peak of severe acute respiratory syndrome coronavirus 2 (1 March 2020 to 24 April). All patients attended at the rheumatology outpatient clinic of a tertiary hospital in Madrid, Spain with a medical diagnosis of AIRD and with symptomatic COVID-19 were included. The main outcome was hospital admission related to COVID-19. The covariates were sociodemographic, clinical and treatments. We ran a multivariable logistic regression model to assess risk factors for the hospital admission. RESULTS: The study population included 123 patients with AIRD and COVID-19. Of these, 54 patients required hospital admission related to COVID-19. The mean age on admission was 69.7 (15.7) years, and the median time from onset of symptoms to hospital admission was 5 (3-10) days. The median length of stay was 9 (6-14) days. A total of 12 patients died (22%) during admission. Compared with outpatients, the factors independently associated with hospital admission were older age (OR: 1.08; p=0.00) and autoimmune systemic condition (vs chronic inflammatory arthritis) (OR: 3.55; p=0.01). No statistically significant findings for exposure to disease-modifying antirheumatic drugs were found in the final model. CONCLUSION: Our results suggest that age and having a systemic autoimmune condition increased the risk of hospital admission, whereas disease-modifying antirheumatic drugs were not associated with hospital admission.

Outpatient Readmission in Rheumatology: A Machine Learning Predictive Model of Patient's Return to the Clinic.

Our objective is to develop and validate a predictive model based on the random forest algorithm to estimate the readmission risk to an outpatient rheumatology clinic after discharge. We included patients from the Hospital Clínico San Carlos rheumatology outpatient clinic, from 1 April 2007 to 30 November 2016, and followed-up until 30 November 2017. Only readmissions between 2 and 12 months after the discharge were analyzed. Discharge episodes were chronologically split into training, validation, and test datasets. Clinical and demographic variables (diagnoses, treatments, quality of life (QoL), and comorbidities) were used as predictors. Models were developed in the training dataset, using a grid search approach, and performance was compared using the area under the receiver operating characteristic curve (AUC-ROC). A total of 18,662 discharge episodes were analyzed, out of which 2528 (13.5%) were followed by outpatient readmissions. Overall, 38,059 models were developed. AUC-ROC, sensitivity, and specificity of the reduced final model were 0.653, 0.385, and 0.794, respectively. The most important variables were related to follow-up duration, being prescribed with disease-modifying anti-rheumatic drugs and corticosteroids, being diagnosed with chronic polyarthritis, occupation, and QoL. We have developed a predictive model for outpatient readmission in a rheumatology setting. Identification of patients with higher risk can optimize the allocation of healthcare resources.