Related Experiment Videos
Propensity Score-Based Methods in Comparative Effectiveness Research on Coronary Artery Disease.
Alexandra G Ellis1,2, Thomas A Trikalinos1,2, Benjamin S Wessler3,4
1Center for Evidence Synthesis in Health, School of Public Health, Brown University, Providence, Rhode Island.
American Journal of Epidemiology
|October 10, 2017
Summary
Reporting of propensity score methods in observational studies of coronary artery disease treatments like coronary artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) is often incomplete and inconsistent with standards.
Area of Science:
- Cardiovascular Medicine
- Biostatistics
- Health Services Research
Background:
- Observational studies are crucial for comparing treatments like coronary artery bypass grafting (CABG), percutaneous coronary intervention (PCI), and medical therapy in coronary artery disease (CAD).
- Propensity score-based methods are frequently used in these studies to adjust for confounding variables.
- The quality of reporting and adherence to methodological standards in these analyses are critical for reliable results.
Purpose of the Study:
- To review and assess the conduct and reporting of observational studies employing propensity score methods for comparing CAD treatments.
- To identify common practices, inconsistencies, and deviations from methodological standards in the application of propensity scores.
Main Methods:
- Systematic selection of observational studies comparing CABG, PCI, or medical therapy for CAD.
- Analysis of 48 identified studies focusing on variable selection, covariate balance assessment, and propensity score model adjustments.
- Classification of over 400 variables used in propensity score models across 12 categories and 60 subcategories.
Main Results:
- Variable selection in propensity score models often relied solely on statistical criteria, with limited use for nonlinear or product terms.
- While 52% of studies assessed covariate balance, only one reported modifying the model based on this assessment.
- Reporting was frequently incomplete, and methods used were often inconsistent with current best practices in propensity score analysis.
Conclusions:
- The conduct and reporting of propensity score-based methods in observational studies of cardiovascular interventions are suboptimal.
- Inconsistent application and incomplete reporting hinder the reliability and comparability of findings from these crucial studies.
- There is a need for improved adherence to methodological standards and transparent reporting to enhance the validity of observational research in cardiology.