Health risk appraisal: review of evidence for effectiveness.

Since its introduction some two decades ago, health risk appraisal (HRA) has become a standard offering in the health promotion repertoire. The technique's distinctive feature is its use of epidemiologic data to generate quantitative risk messages for the client. Yet despite the dedication and considerable investments that have gone into HRA's development, dissemination, and use, there is only limited empirical evidence that these quantitative risk messages have any effect on clients. There do not appear to be any formal studies of HRA's effect on participation in health promotion programs, although increasing recruitment is regarded as a major benefit of using HRA. There are few indications of HRA effects on health beliefs. Most positive reports of effects on behavior change come from uncontrolled studies; several randomized controlled trials have yielded ambiguous findings. Virtually no data exist concerning the impact of the quantitative risk messages that distinguish HRA from other assessment techniques and that have motivated the substantial efforts toward developing and refining HRA. HRA has evident appeal and is probably a useful health education device for middle-class, middle-aged, nonminority clients. It may well have desirable effects on health-related beliefs, attitudes, and behaviors when accompanied by counseling or education, but available evidence has not established its effectiveness. Given the difficulty of obtaining definitive evidence of the effectiveness of HRA and specifically of its use of quantitative risk projections, the need for such evidence is debatable. An adequately funded and reviewed research program to examine whether projections of absolute risk affect knowledge, beliefs, attitudes, and intention to change is recommended as the most fruitful next step. Epidemiologically based HRA procedures that provide feedback in terms of qualitative statements or relative risk may be a promising approach to prospective health assessment.