Drug Evaluation Committee Recommendations for Sensitivity Analysis in Observational Studies Practical Edition
April 2019
The Data Science Subcommittee of the Drug Evaluation Committee of the Japan Pharmaceutical Manufacturers Association (JPMA) Task Force 2 for Fiscal Year 2016 developed a practical guide—a sequel to the “Recommendations for Sensitivity Analysis in Observational Studies: Introduction” published in 2017—as well as the Excel tool ESATJ for performing quantitative sensitivity analysis.
In recent years, the infrastructure for observational studies using medical information databases has been developing in Japan, and pharmaceutical companies have begun utilizing them for purposes such as safety monitoring. Furthermore, regarding post-marketing surveillance, the GPSP Ordinance has been revised, and “comparative post-marketing surveillance studies” using a comparison group, as well as “post-marketing database studies” utilizing medical information databases, are now being conducted.
The results of observational studies conducted by companies on their own pharmaceutical products—particularly studies involving causal inference through comparison—may lead to decision-making regarding those products. However, biases and uncertainties associated with observational studies, especially those arising from unmeasured causes, make decision-making difficult.
The quantitative sensitivity analysis introduced in this book provides numerical results that have been adjusted for biases arising from such information gaps by incorporating evidence or assumptions. Simple methods require no special software and can be performed using spreadsheet software.
We hope this book will serve as a helpful resource for evaluating the robustness and uncertainty of observational study results.