Comparative effectiveness research considered methodological insights from simulation studies in physician's...

Area of Science:

• Pharmacoepidemiology
• Comparative Effectiveness Research (CER)
• Instrumental Variable (IV) Methods

Background:

• Physician's prescribing preference instrumental variable (PPP IV) is increasingly utilized in pharmacoepidemiology for causal inference in comparative effectiveness research (CER).
• Methodological validation through simulation studies is crucial for understanding the performance and limitations of PPP IV in diverse CER settings.
• A comprehensive review is needed to synthesize findings from simulation and applied CER studies using PPP IV.

Approach:

• Systematic literature review of simulation and CER studies from 2005-2020.
• Databases searched include Ovid, PubMed, and Google Scholar.
• Analysis focused on simulation findings regarding optimal PPP IV application and heterogeneity in CER study methodologies.

Key Points:

• Six simulation studies provided guidance on using PPP IV across various settings (e.g., outcome types, sample size, outcome prevalence).
• Eighteen CER studies demonstrated heterogeneity in validation assumptions, estimation methods, and sample sizes.
• Despite heterogeneity, all identified CER studies concluded that PPP is a valid instrumental variable.

Conclusions:

• Future CER studies employing PPP IV should address methodological considerations like statistical methods, preference proxies, time-varying exposures, and outcome types.
• Integrating insights from simulation studies can enhance the robustness and validity of PPP IV in pharmacoepidemiological research.
• Continued methodological refinement is essential for maximizing the utility of PPP IV in generating reliable CER findings.