The European Alliance of Associations for Rheumatology (EULAR) released 3 overarching principles and 10 points-to-consider when analyzing and reporting comparative effectiveness research using observational rheumatology data. The guidance was published in Annals of the Rheumatic Diseases.
Observational studies with longer follow-up periods complement randomized controlled trials (RCTs). However, observational studies are subject to inherent limitations, such as confounding and attrition bias. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines offer a framework for researchers reporting observational data. The EULAR sought to build on these guidelines and offer more detailed points-to-consider specific to rheumatology studies.
An 18-member task force including rheumatologists, epidemiologists, statisticians, patient representatives, and health professionals were convened. A total of 2 in-person meetings were conducted, the first of which established the focus of the task force and scope of a systematic literature review of methods. The systematic review resulted in 211 articles that included observational rheumatology studies comparing treatments on various outcomes. Draft points-to-consider were developed based on the systematic review and a simulation study. The task force refined and finalized the points-to-consider at a second meeting and during a subsequent 2-round online Delphi process.
Overall, the task force established 3 overarching principles and 10 recommendations.
- Treatment effectiveness should be assessed in the context of all treatments available in clinical practice.
- Observational studies are invaluable companions to RCTs; however, they may be subject to limitations, including confounding and missing data.
- Results from observational studies were noted to be more trustworthy when the methods are robust, transparent, and detailed.
- The STROBE guidelines should be followed when reporting comparative effectiveness in observational studies.
- Several outcomes across multiple health domains should be compared when assessing effectiveness.
- Data from patients lost to follow-up should be reported.
- Data from patients who stop or change therapies, the reasons for discontinuation, and the characteristics of patients who discontinued by reason should also be reported.
- The task force noted that covariates based on known potential confounders must be selected and model selection be justified to avoid biases and errors.
- A study baseline must be established at treatment initiation and reported when covariates are measured.
- Data from all patients who are being initiated on treatment must be analyzed.
- The EULAR task force said that attrition must be accounted for in analyses by considering the use of multiple imputation techniques and/or causal inference models, such as inverse probability weighting.
- Sensitivity analyses should be used to check assumptions and reasonable alternative covariates can be considered, especially where there is attrition.
- A statistical analysis plan must be prepared in advance.
The task force concluded, “To improve the reliability of an increasing number of real-world comparative effectiveness studies in rheumatology, special attention is required to reduce potential biases. Adherence to clear recommendations for the analysis and reporting of observational comparative effectiveness studies will improve the trustworthiness of their results.”
Courvoisier DS, Lauper K, Kedra J, et al. EULAR points to consider when analysing and reporting comparative effectiveness research using observational data in rheumatology. Ann Rheum Dis. Published online January 20, 2022. doi:10.1136/annrheumdis-2021-221307