Can experts predict health risk from family genograms?

Abstract

BACKGROUND: The family genogram is sometimes used to aid diagnostic, therapeutic, and preventive care decisions. This study evaluated the efficacy of genograms for predicting health risk in comparison to predictions made using demographic and chart review data. METHODS: Six physicians with expertise in using genograms were asked to evaluate 20 actual patient cases and use three methods to predict the patients' chances, over the next three months, of: a) experiencing illness causing at least one disability day, b) making an unexpected physician visit for a new problem, or c) requiring hospitalization. The three methods were genogram evaluation, review of patient demographics, and review of patients' charts. Predictions from demographic data were always made first; the other two methods were used in varied order. Three months later, actual patient outcomes were reviewed and compared to predictions. RESULTS: Over the next three months, 44% of subjects experienced a disability day, 35% made an unexpected clinic visit, and none required hospitalization. Predictions of these events with genograms were no more accurate than predictions generated from chart review. The six genogram experts did not predict outcomes at better than chance levels. CONCLUSIONS: Genograms may be no more accurate than standard clinical chart review for predicting short-term (three month) health outcomes.