Font Size: 

The West Philadelphia Asthma Care Implementation Program (WEPACC) was designed to improve asthma control in low-income children from communities that are disproportionately impacted by asthma and other serious pediatric health conditions. WEPACC is a randomized control trial that used a factorial design to compare usual care to community health worker (CHW) delivered interventions in primary care, school alone, and combined primary care-school settings. However, the delivery of the different components of the WEPACC intervention was severely impacted by the SARS-CoV-2 pandemic. The shutdown of Philadelphia schools in Spring 2020 (roughly halfway through the study period) prevented the study team from delivering the school components as designed, while social distancing measures led to a striking decrease in asthma morbidity. Some participants completed all study visits prior to the shutdown for the pandemic, while others completed some (or all) of their study visits after the shutdown and therefore did not receive the full “dose” of the intervention as planned. In this presentation I will describe the challenges that we faced in the statistical analysis of data from this trial. I will also discuss and demonstrate the application of some recently published recommendations regarding how to properly modify the statistical analysis plan to account for unplanned interruptions in a clinical trial. For this presentation, the focus will be on multivariate outcomes, in keeping with the  spirit of the thematic track,  Modeling Multivariate Data (organized by Prof. Anuradha Roy).

COVID, generalized estimating equations, intervention study, multivariate, quasi-least squares regression