The aim of this study was to assess whether assigning a geriatrician to provide daily medical care to geriatric patients in the orthopedic ward can improve the prognosis and reduce the length of stay. Time series analysis was performed in two parts: 1) prospective analysis of two years' workload, and 2) retrospective analysis of data collected over the 4 years prior to the intervention. Intervention and control populations were pooled, and the effects of geriatric care and patient-related factors on outcome measures were assessed by logistic regression analysis. All subjects were patients aged > or = 70 years who attended the orthopedic ward in a university hospital in years 1989-90 (studied group: 287 cases) and in years 1985-88 (control group: 474 cases). In the study period, mortality was 8.4% compared to 18% in 1985-86 (p < 0.0006) and 14% in 1987-88 (p < 0.01). The operation rate in the study period was 89.9% vs 83.8% in 1985-86 (p < 0.02) and 81.8% in 1987-88 (p < 0.005). Length of stay was 26.2 +/- 14.4 days vs 32.9 +/- 30.9 days in 1985-86 (p < 0.05) and 26.9 +/- 16.5 days in 1987-88 (NS). Length of stay was more strikingly shortened in the subset of patients with femoral fracture undergoing surgical management (28.5 +/- 12.7 vs 37.6 +/- 32.6 days in 1985-86, p < 0.003, and 30.8 +/- 15 days in 1987-88, p < 0.02). Given the positive relationship between geriatric care and operation rate (o.r. = 1.5, CI = 1.1-1.9), the protective effect of surgical treatment on mortality (o.r. = 0.6, CI = 0.4-0.8) to some extent may mask the collinear effect of geriatric care. We conclude that assigning a geriatrician to assist with the medical care of elderly orthopedic patients in orthopedic wards is associated with increased operation rate, decreased mortality and shortened length of stay.