The Risk Approach (RA) is a framework for selective provision of health care services advocated by the World Health Organization (WHO). Although the potential utility of RA is stressed in the WHO literature, the framework remains largely untested as a basis for resource deployment. Despite this, advocates claim the logic of RA is 'theoretically unassailable'. The purpose of this paper is to challenge this claim by critically examining some of the assumptions and proposed methods contained in the WHO literature concerning RA. Two areas of concern are discussed, relating to (i) prediction and the nature of relations between markers and outcomes, and (ii) marker measurement and scoring.