Introduction: Surgical treatment, discharge planning and rehabilitation procedures are rarely based upon defined assessment procedures. It might therefore be useful to develop simple and reliable screening tools to identify patients for early discharge, intensified rehabilitation and limited treatment.
Patients and methods: 234 patients were initially contacted. From these 217 gave informed consent. The reported study included the 134 home dwelling elderly. All patients could be contacted or died after six month. Thus,data from more than 90% could be used for the analysis. The six month mortality was 10%. A population based cohort of elderly patients referred to five local hospitals was tested. Mortality, institutionalisation and mobility were defined as major outcome criteria. Only information that was available during the first week of treatment was used in the model. Predictors expressed as odd ratios (OR) were calculated using logistic regression with variable selection.
Results: The most important predictors for institutionalisation were age, inability to eat without assistance, postoperative night time confusion, stroke history, cognitive deficit and fear of falling. Outdoor mobility was strongly associated to the preoperative ADL performance measured as the Barthel-Index, history of malignancy and fear of falling. Six month mortality was associated with male sex, the Barthel-Index and fear of falling.
Discussion: It seems feasible to improve postoperative resource allocation by predictor led stratification. This need to be tested in intervention trials under the specific condition of the German health care system.