Statistical Properties of Stepped Wedge Cluster-Randomized Trials in Infectious Disease Outbreaks

Abstract

<jats:title>Abstract</jats:title> <jats:p>Randomized controlled trials are crucial for the evaluation of interventions such as vaccinations, but the design and analysis of these studies during infectious disease outbreaks is complicated by statistical, ethical, and logistical factors. Attempts to resolve these complexities have led to the proposal of a variety of trial designs, including individual randomization and several types of cluster randomization designs: parallel-arm, ring vaccination, and stepped wedge designs. Because of the strong time trends present in infectious disease incidence, however, methods generally used to analyze stepped wedge trials may not perform well in these settings. Using simulated outbreaks, we evaluate various designs and analysis methods, including recently proposed methods for analyzing stepped wedge trials, to determine the statistical properties of these methods. While new methods for analyzing stepped wedge trials can provide some improvement over previous methods, we find that they still lag behind parallel-arm cluster-randomized trials and individually-randomized trials in achieving adequate power to detect intervention effects. We also find that these methods are highly sensitive to the weighting of effect estimates across time periods. Despite the value of new methods, stepped wedge trials still have statistical disadvantages compared to other trial designs in epidemic settings.</jats:p>

The required size of cluster randomized trials of non-pharmaceutical interventions in epidemic settings

“…We assume that each exposed individual's incubation period is drawn from an exponential distribution with a mean of 5.51 days. 30 We match the simulations done elsewhere, 21,22 assuming the mean infectious period across individuals is 5 days (from an exponential rather than a gamma distribution). We assume an approximate negative binomial degree distribution for the network structure of each cluster with a mean of 15 contacts per individual and overdispersion parameter k. The contact structure does not exactly follow the negative binomial degree distribution because we delete self-loops after using a configuration model (CM) algorithm to create the contact structure.…”

“…37 Furthermore, because the direct effects continue and indirect effects may increase on short time scales, the effect size in cRCTs in epidemics can increase over time. 21,22 Eventually, however, the exhaustion of susceptible individuals will lead to a reduction in the effect size as incidence rates become more similar between intervention and control clusters. We explore the effects of the time interval used in our simulations by increasing the lag between intervention and evaluation.…”

“…Simulation approaches have been previously used to evaluate the statistical power of cRCTs evaluating vaccination at both individual and cluster scales. [19][20][21][22] This allows investigators to size cRCTs to have a desired power to detect a specified true effect size of interest (i.e., reduction in transmission). Although estimators have different properties and interpretations depending on the phase of the epidemic-e.g., by capturing more indirect effects or having less impact when there is more pre-existing immunitythis allows hypothesis tests to be appropriately powered.…”

“…Although estimators have different properties and interpretations depending on the phase of the epidemic-e.g., by capturing more indirect effects or having less impact when there is more pre-existing immunitythis allows hypothesis tests to be appropriately powered. 21,22 Here, we build on these results to provide estimates for the number of clusters and number of individuals measured within each cluster needed to test the effectiveness of an NPI, with the aim of bounding the feasibility of such analyses.…”

Big data and simple models used to track the spread of COVID-19 in cities

“…Furthermore, in many places, clusters and infections identified by contact tracing account for only a modest proportion of all transmissions, so inferences drawn from such examples might be of limited use for generalization. Randomized experiments to evaluate interventions (in which the introduction of an intervention is staggered over time in different populations) 6 are an alternative, but they are hard to do 7 , and such studies have been rare for COVID-19.…”