The purpose of this paper is to review the types of missing data in patient registries, discuss design and operational strategies for avoiding or minimizing missing data, explore analytic strategies for handling missing data, and consider the impact of missingness on the interpretation and reporting of registry findings. These concepts are discussed in the context of internal validity of studies, or the ability to draw conclusions on the study population, rather than generalizability of results to broader populations who were not enrolled and therefore not represented in the study. The topics discussed are applicable to both retrospective and prospective designs and cover both primary and secondary data sources. Where appropriate, reference is made to other chapters in the document, Registries for Evaluating Patient Outcomes: A User’s Guide.
Click here for more information about this report.
