A0723
Title: Missing data: A key challenge for digital outcomes in clinical trials
Authors: Mia Sato Tackney - University of Cambridge (United Kingdom) [presenting]
Abstract: Clinical trials increasingly use digital devices to measure the impact of an intervention on health outcomes. For example, accelerometers in physical activity trials can measure step count in very fine intervals of time, typically in 5-second epochs. The data is usually aggregated to provide the primary analysis's daily or weekly step counts. Missing data is common as participants may not wear the device per protocol, or there may be device failure. Approaches to handling missing data in the literature have largely defined missingness on the day level using a threshold on wear time, which leads to loss of information on the time of day when data is missing. An approach is presented to identifying and classifying missingness at the finer epoch level. Missingness can then be handled using a non-parametric approach to Multiple Imputation (MI), where missing periods during the day are replaced by donor data from the same person where possible or data from a different person who is matched on demographic variables. An application of this approach is illustrated in the 2017 PACE-UP Trial. Finally, the key statistical opportunities and challenges to adopting digital endpoints in clinical trials are discussed.
