Outcomes assessment is often difficult to accomplish in evaluation research studies in situations where the gathering of pretest data is impossible or prohibitively expensive. The purpose of this longitudinal study was to investigate the substitutability of retrospective pretest ratings for actual pretest ratings in indexing change in patient satisfaction with health status. The sample consisted of 251 women receiving medical treatment for breast cancer enrolled in the Self-Help Intervention Project (SHIP). ANOVA, ordinary least-squares regression, and pooled time-series regression analysis revealed that retrospective assessments were not significantly different from their prospective counterparts in means and variances and that they differed from current assessments taken at the same time (p<.01). In addition, prospective assessments emerged as a significant independent predictor of corresponding retrospective scores (p<.01), accounting for up to 30% of the recall scores. These findings have implications for inclusion of retrospective pre-post comparisons in outcomes evaluation research.
Copyright 2002 Wiley Periodicals, Inc. Res Nurs Health 25:233-241, 2002