ASSOCIATION FOR Worksite Health R 0 M 0 0 Formerly AFB Practitioners’ Forum APPLIED RESEARCH BRIEFS Evaluating HealthPromotion Technology: Consistency of Results Among Health Risk Appraisals PURPOSE In recent years, there has been an explosion in both the availability and use of Health Risk Appraisals (HRAs), as well as controversy regarding their validity, reliability, and appropriate use. ~-’~ There appears, how- ever, to be a consensus that HRAs are most appropriate and valid when used with white, middle-aged, middle- class males. 4 Additionally, the database~ typically used for HRAs are strongest and most detailed [or identifying cardiovascular disease risk. ~ | When used as a health education/he~alth promotion tool, Schoenbach et al. state that it is crucial that HRAs "deal with appropriate risk characteristics and produce appropriate recommendations for change. ’’6 In this analysis the appropriateness and consistency of cardiovas- cular disease risk characteristic measures and behavioral recommendations generated by eight microcomputer HRAs for a white, middle-aged man are evaluated. METHODS Eight popular microcomputer HRAs (including the Centers for Disease Control and Prevention/Carter Center HRA)were selected for use in this study using personal knowledge of the programs, references from practitioners, and information from government agen- cies. Microcomputer HRAs were chosen for several reasons. First is the greater availability of these pro- grams. Second, we had observed a sharp increase in the utilization of these programs by local organizations (health departments, hospitals, etc.) as microcomputers have become more commonplace. Third, we expect microcomputer HRAuse to expand since it allows for immediate feedback of results, as opposed to waiting for forms and results to be exchanged through the mail. Also, the CDC HRAfor microcomputers being available to state and local health departments for low cost will encourage its adoption. The databases on which the HRAs rested were not divulged by all proprietors, even upon telephone follow-up. For those that were, the most common response was the CDC database. The HRAs that used a time frame for making risk predictions generally used the probability of dying within the next 10 years. The prices for the HRAs used in this study ranged from $100 to $800. Kathleen D. Mullen, Ph.D., and Daniel L. Bibeau, Ph.D., are both Associate Professors in the Department of Public Health Education, University of North Carolina at Greens- boro. In order to evaluate the consistency and appropriate- ness of behavioral recommendations across the eight microcomputer HRAs, a hypothetical case was con- structed that would represent the population for which current HRAs are most valid. Thus, data were entered into each HRAmicrocomputer program for a white, 45- year-old, middle-class male who was 177 cm tall and weighed 68.2 kg. There was no history of heart disease (personal or family), however the subject smoked one pack of cigarettes a day, had a blood pressure of 138/ 110 mmHg, a blood cholesterol level of 280 mg/dl, and was sedentary. The lowest level for physical activity was always selected for consistency. Thirty-two additional variables, such as responses to questions on drug use, driving habits, etc., were held constant across the eight t-IRAs. Printouts from the eight HRAs were analyzed for consistency of results for "risk age" or "risk score," and "achievable age." Consistency and appropriateness of behavioral recommendations regarding smoking, blood pressure, blood cholesterol, and exercise were also analyzed. RESULTS The results of this study indicate that a wide variety of results and recommendations can be anticipated for an identical case when assessed by various HRAs. For the hypothetical case used in this study--a white, middle- aged male with four major lifestyle-related risk character- istics for cardiovascular disease--there was a range of 15.2 years among four appraisals on the estimate of "appraised" or "risk" age (see Table 1). Special attention should be given to the fact that the lowest estimated risk age was below the case’s actual age of 45, while the highest risk age estimate was well over the actual age. A similar range was found for the reported "achievable" Table 1 Differencesin "Appraised" Age, "Achievable" Age, and Risk Score Among HRAs N Actual Scores Appraised or Risk Age 4 Achievable Age 4 Risk Score** 4 43.5, 48, 55, 58.7 years* 38.5, 40, 46, 52.9 years 50, 73, 77, and 86 out of 100 points *Note: The subject’s actual age was 45 **Note: Four of the eight HRAs did not compute a risk "age" score. Rather, they used a scale of 100 possible points. 260 American Journal of Health Promotion