A new method for bounding selection bias in case-control studies was proposed by researchers at National Taiwan University (NTU). The study was published in the American Journal of Epidemiology.
Conducted by Ms. Tzu-Hsuan Huang, a doctoral student at NTU, and her advisor Dr. Wen-Chung Lee, professor of Institute of Epidemiology and Preventive Medicine in the College of Public Health, this study proposed bounding formulas for selection bias.
Researchers conducting observational studies need to consider three types of biases: selection bias, information bias, and confounding bias. A whole arsenal of statistical tools can be used to deal with information and confounding biases. However, methods for addressing selection bias and unmeasured confounding are less developed. In this paper, the authors extended Prof. Lee’s prior work on unmeasured confounding to propose formulas for bounding selection bias. By using the formulas, researchers are informed about whether an observed positive finding can be explained away by the elusive selection bias. Easy-to-use R code was also provided in the paper.
“The method proposed in the paper should help researchers make more prudent interpretations of their (potentially biased) results.” said Prof. Lee.