OBJECTIVES: Missing data in patient-reported outcomes (PRO) research may have a negative impact on the scoring of PRO instruments, analysis of data, interpretation of findings, and regulatory and reimbursement submissions and reviews. The objectives of this work were to 1) provide an overview of the issue of missing data, 2) discuss the impact of missing data from the regulatory and reimbursement perspectives, and 3) review the design and statistical methods in the monitoring, prevention, and handling of missing data.

METHODS: We reviewed published peer-reviewed articles, regulatory documents by the European Medicines Agency (EMA) and United States Food and Drug Administration Agency (FDA), and documents by the health technology assessment (HTA) bodies in Germany including the Federal Joint Committee (GBA) and Institute for Quality and Efficiency in Healthcare (IQWIG). Industry experience in handling missing data from a clinical operations perspective was also included.

RESULTS: There is a growing interest in preventing and managing missing PRO data (e.g., EMA, 2011; FDA, 2009). The US FDA encourages prespecified plans for handling item- and domain-level missing data. In German HTA reimbursement submissions, missing data is reported as a major non-acceptance reason. Commonly used statistical methods such as mean imputation, weighted sum score, and Cronbach’s alpha can be used to handle item-level missing data. Longitudinal modeling techniques can be used to handle score-/visit-level missing data. At the clinical study operational level, inclusion of data collection completion (site visit completion) form can provide valuable information on completeness of data and the reasons for missing data. Such data may help develop strategies to prevent missing data and may help the selection of appropriate statistical methods. Electronic data collection (i.e., ePRO) can also help ensure a complete data set.

CONCLUSIONS: Future research is needed to investigate the impact of these design and statistical procedures on regulatory approvals and reimbursement decisions.