Economic evaluations that use cluster trials
Policy-makers worldwide require cost-effectiveness analysis (CEA) to decide which health care programmes should be prioritised. CEA studies may use data from cluster randomised trials (CRTs), where randomisation is at the level of the cluster (for example the hospital) rather than the individual. Here, statistical methods are required which recognise that within-cluster costs and health outcomes may be correlated. However, most CEA alongside CRTs use methods that assume observations are independent, which may lead to incorrect inferences. We develop appropriate methods for CEA that use CRTs and provide guidance on their use, so that future studies can provide a firmer basis for allocating scarce resources.
This research considers multilevel models (MLM), seemingly unrelated regression (SUR), generalised estimating equations (GEEs), and a two-stage non-parametric bootstrap (TSB). The alternative approaches are contrasted in CEA that use cluster trials with different characteristics (e.g. few clusters, skewed costs, unequal numbers per cluster). As costs and outcomes are often missing, this research also investigates approaches for handling missing values that recognise the hierarchical structure of the data.
The research investigates the alternative methods in both simulations and case studies, and provides practical guidance, in particular software code, to help researchers apply the alternative methods.