Rz. Goetzel et al., DIFFERENCES BETWEEN DESCRIPTIVE AND MULTIVARIATE ESTIMATES OF THE IMPACT OF CHEVRON-CORPORATIONS HEALTH QUEST PROGRAM ON MEDICAL EXPENDITURES, Journal of occupational and environmental medicine, 40(6), 1998, pp. 538-545
This investigation focused on alternative methods for evaluating the i
mpact of Chevron Corporation's Health Quest Fitness Center program on
Chevron Corporation's Health Quest Fitness center program on medical e
xpenditures, comparing descriptive and multivariate research designs.
Many uncontrolled studies of corporate health management programs base
estimates of program effectiveness on descriptive analyses such as St
udent's t tests. Unlike more sophisticated multivariate analyses, desc
riptive analyses often produce biased estimates of program cost saving
s. To test alternative research design methods, the investigators comp
ared inpatient and pharmacy expenditures for program participants and
non-participants over a 2.5-year period, using descriptive and multiva
riate regression analyses. Results showed that compared with non-parti
cipants, expenditures for participants were significantly lower for su
bjects who used a Health Quest fitness center at least twice weekly. P
revious descriptive studies suggested a much broader impact. The resul
ts underscore the need to use multivariate analyses when evaluating th
e financial impact of corporate health management programs, especially
when randomization cannot be used to assign participation status.