rwd-cohort-analysis

Propensity score matching and IPTW weighting adjust for confounding in comparative effectiveness research by estimating treatment probabilities and creating balanced cohorts. This skill provides balance diagnostics like SMD to ensure statistical validity in observational studies.

To calculate medication adherence for a new-user cohort, this skill computes PDC and MPR metrics over a specified period. It standardizes adherence measurement for real-world data, supporting analysis over a 365-day period or other customized timeframes.

Yes, you can use lifelines and statsmodels for survival analysis on EHR data. This skill requires these Python libraries to execute Kaplan-Meier curves and Cox proportional hazards models, ensuring statistical validity for real-world data analysis.

The best way to estimate comparative treatment effects in observational research is through a standardized pipeline that combines propensity score matching, IPTW weighting, and survival regression models. This approach reduces methodological errors when using real-world data.

This skill focuses on standardizing cohort identification, medication adherence, and treatment effect estimation. For limitations, it relies on administrative claims and EHR data, and requires washout period logic and propensity score adjustments to mitigate confounding risks in observational research.