Mathematics
5 years ago

Ignoring balanced covariates in regression adjustment leads to biased results.

Epidemiology
Misuse of Regression Adjustment for Additional Confounders Following Insufficient Propensity Score Balancing
Tomohiro Shinozaki, Masanori Nojima

Paper Summary

Related papers

Theme background image
Mathematics
8 years ago

Improved survival estimates could save lives by overcoming censored data.

Regularization Methods
Missing Data
Causal Inference
Theme background image
Mathematics
5 years ago

New regression model improves accuracy of medical treatment decisions.

Causal Inference
Missing Data
Quality of Life
Theme background image
Mathematics
8 years ago

Propensity score methods not superior to conventional covariate adjustment in studies.

Quality of Life
Causal Inference
Clinical Trials

The article shows that adjusting for additional factors after balancing for certain variables can lead to bias in research results. This bias can affect the relationship between chemotherapy and recurrence-free survival in cancer patients. To avoid this bias, researchers should either adjust for the same variables used in the balancing process or reevaluate the balancing process with the additional factors included.