Why we need large, simple studies of the clinical examination: the problem and a proposed solution