Prevention of Missing Data in Pragmatic Clinical Trials of Nonpharmacologic Interventions for Pain Management
A working document from the Pain Management Collaboratory Biostatistics/Design work group
PMC’s Biostatistics/Design Work Group addresses the problem of missing data in pragmatic clinical trials, and offers a survey of methods that some of PMC’s trials are incorporating to prevent missing data. To mitigate and adapt for missing data from participant drop out, it is imperative that trials have prevention strategies at every stage of the trial—from planning to execution. The following working document delivers examples and recommendations.
Missing data in trials of pain management are common as dropout rates tend to be high1,2. In clinical trials, missing data can reduce power and bias study conclusions. Pragmatic trials differ from explanatory trials in that they aim to follow-up participants no more than would be the case in usual care3. In fact, trials at the most extreme end of the pragmatic spectrum have no follow-up contact with participants and obtain outcome data from existing sources such as the electronic health record (EHR). Participants in trials at this end of the spectrum may not even know they are in a trial. While approaches are available for analysis in the presence of missing data, plans for prevention of missing data are particularly vital for pragmatic pain trials. This white paper discusses some approaches used in pragmatic trials from the Pain Management Collaboratory (PMC3) to limit the extent of missing data.
- Gewandter JS, McDermott MP, McKeown A, et al. Reporting of missing data and methods used to accommodate them in recent analgesic clinical trials: ACTTION systematic review and recommendations. Pain. 2014;155(9):1871-1877.
- Katz N. Methodological issues in clinical trials of opioids for chronic pain. Neurology. 2005;65(12 Suppl 4):S32-49.
- Loudon K, Treweek S, Sullivan F, Donnan P, Thorpe KE, Zwarenstein M. The PRECIS-2 tool: designing trials that are fit for purpose. BMJ. 2015;350:h2147.