Background: In 2007, for the first time, psychiatric disorders were the most common cause (29%) of new cases of disability pension in Austria. Against the backdrop of an expected increase in mental health disorders, inpatient medical- psychiatric rehabilitation is supposed to contribute to a comprehensive mental health care. The aim of this review is a systematic analysis of the sustainability of inpatient psychiatric/psychosomatic rehabilitation covering an observation period of at least 12 months after discharge from rehabilitation.

Methods: In August 2013, we conducted a systematic literature search in six databases. In addition, we searched manually for relevant articles. After deduplication we screened a total of 1.872 references. We ordered 176 full-text articles to verify the fulfillment of the inclusion criteria.

Results: We included 1 meta-analysis, 4 RCTs, 2 pre-post studies with, and 14 without a control group. We extracted a total of 180 individual results on changes between admission (T0), discharge (T1) and catamnesis (K) 12 months (min.) after discharge from rehabilitation, which had been measured by 38 (in controlled studies) or 49 (in studies without control group) instruments and/or relevant subscales respectively. Overall, results showed similar trends across studies and outcome measures: the improvement between admission and discharge deteriorated to varying degrees during the catamnestic period, although they were still better than at the time of admission. Consequently, effect sizes were larger from admission to discharge than from admission to catamnesis. The reported effect sizes showed substantial differences, even if the same instruments had been applied. Therefore, no general conclusions can be drawn regarding specific domains in which psychiatric/psychosomatic rehabilitation leads to better results than in others. In the most recent studies on specific aftercare, a trend towards an improved stabilisation of treatment effects can be observed. Except for one RCT, the risk of bias of included studies is high due to high drop-out rates. Limited patients' willingness to participate in specific programmes (e.g. integrated vocational rehabilitation or aftercare measures) should be considered as a critical factor for the applicability of the results.

Conclusion: The disorder/recovery profiles of patients with psychiatric/psychosomatic rehabilitation are well described on the basis of numerous observational studies. Due to the lack of comparative studies, no conclusions can be drawn on the effectiveness and sustainability of psychiatric/psychosomatic rehabilitation compared to no rehabilitation. In recent years, however, a trend towards (randomised) controlled trials, which at least allow statements about the comparative effectiveness of different interventions, can be observed. Especially specific aftercare programmes (e.g. web-based interventions) and integrated vocational rehabilitation seem to be an important research focus. Instead of querying the overall effectiveness and sustainability of psychiatric/psychosomatic rehabilitation, future research should target critical success factors and differential effects of rehabilitation.