Rethinking approaches to analysis of global randomised controlled trials