Clinical Prediction Models For Medication Adverse Events In Patients With Rheumatic And Musculoskeletal Conditions: A Systematic Literature Review

Objectives: This systematic review aims to identify, summarize, and evaluate the methodological quality of existing clinical prediction models (CPMs) that predict adverse events (AEs) associated with medications prescribed for rheumatic and musculoskeletal diseases (RMDs).

Methods: We searched PubMed, Embase, and Medline databases up to March 2024. Studies were included if they developed multivariable CPM predicting AEs in adult patients using RMD medications. Data extraction and quality assessment were conducted using the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) and Prediction model Risk Of Bias Assessment Tool (PROBAST) checklists to ensure consistent reporting and assess the risk of bias (ROB).

Results: Of 2406 studies identified, 1734 titles/abstracts were screened, and 38 were reviewed in full. Twelve studies reporting 17 CPMs met eligibility criteria. Most CPMs (76.4%) focused on rheumatoid arthritis and disease modifying anti-rheumatic drugs (DMARDs) such as methotrexate (69.2%) and biologic drugs (15.3%). Cox proportional hazards or logistic regression models were commonly used. Twelve models (70.5%) had high overall ROB due to inappropriate variable selection methods and sample size.

Conclusions: This is the first systematic review summarising CPMs for AEs associated with RMD medications. It highlights that existing CPMs are affected by methodological pitfalls, including inappropriate variable selection and lack of clear sample size justification. Future models could consider a broader range of RMDs and medications. Emerging methods such as machine learning with the ability to model complex interactions, and multi-outcome CPMs to predict several AEs to one class of drug may improve predictions.