To provide statistical leadership for strengthening the use of appropriate statistical methodology in health economics and outcomes research and improve the analytic techniques used in real world data analysis.
- To review and recommend statistical methods used in outcomes research studies
- To provide insight on the impact of using appropriate statistical methodology in a rigorous manner to enhance credible research
- To better understand the limitations of analyses and methods used by the HEOR community
Statistics are used in many outcomes research studies, in which understanding the proper use of the methods and approaches that have been developed around certain types of analyses is crucial to producing reliable and valid results that may be reproducible to the rest of the research community. Appropriate application of statistics is the foundation of credible research.
To answer a number of research questions, there are various statistical methods used in real world data analysis such as multiplicity, Bayesian analyses, estimation in disease state changes, model averaging, machine learning/artificial intelligence, and causal inferences, to name a few. However, there are large gaps in the statistical knowledge base as it relates to real world data analysis including the effect of missing values in claims databases, their impact on analyses, and appropriate methods for imputation.
Manuscript and Reports
- Mukherjee, K., Gunsoy, N.B., Kristy, R.M. et al. Handling Missing Data in Health Economics and Outcomes Research (HEOR): A Systematic Review and Practical Recommendations. PharmacoEconomics (2023).
- Baio G, Gunsoy N.B, Onwudiwe N, Vanness D J. Are Missing Data Properly Accounted for in Health Economics and Outcomes Research? Value & Outcome Spotlight 2020;6(2):27-30.
- Statistical Terms and Definitions Monographs
Emma Hawe, MSc, BSc
Helene Karcher, PhD
Kumar Mukherjee, MS, PhD
Missing Data in Health Economics and Outcomes Research