"Wer profitiert?" - Patientenmerkmale als Erfolgsprädiktoren in der Psychosomatischen Rehabilitation
LicenseStandard (Fassung vom 03.05.2003)
In the German health system, many psychotherapeutic in-patient treatments are provided in form of a so called psychosomatic rehabilitation. This study investigated the characteristics of in-patients as predictors of outcome in psychosomatic rehabilitation. Information from a hospital quality management system and from the discharge documentation was used. 463 consecutive in-patients were assessed at intake and discharge, 329 of them at an additional 3 month follow-up. Potential predictor variables included demographic data, diagnosis, symptoms, duration of sick-leave, vocational issues and motivation. Various outcome measures were defined and combined. The main outcome criterion was related to the ability to work and the state of health. All criterions were dichotomized and overall criterion was generated. Depending on the chosen criterion about 40 - 60 % of the participants were successful. 52 % improved over-all. Predictor models were developed by correlation analysis and logistic regression. In the bivariate approach, long-term sick-leave, the desire for a pension and a high impairment at the beginning of the hospital treatment were particularly negative predictors. In the logistic regression, a substantial proportion of variance was explained by the above mentioned variables. The prediction of the ability to work showed a total correct classification rate of 84 %, the correct classification rate for the state of health amounted 72 % and for the over-all outcome measure 78 %. The results for the follow-up are slightly poorer. However, the over-all model performances are reasonable. As conclusions, attention should be given to vocational issues and to impairment. This study could initiate further research how to improve the treatment for in-patients with rather unfavourable characteristics.
Subject HeadingsPsychosomatik [GND]
Logistic regression [LCSH]
Treatment outcome [MeSH]