[Evaluating the rehabilitation process by means of peer review: examination of the methods used and findings of the 2000/2001 data collection in the somatic indications].

This paper reports the results of a peer review system that was implemented in the context of the quality assurance programme of the statutory German Pension Insurance scheme. The data reported refer to the 2000/2001 data collection period for medical rehabilitation in the somatic indications. Examination of inter-rater reliability for judgements of individual raters shows satisfactory results only in orthopaedics. In the quality assurance programme, rehabilitation centres are usually evaluated by the mean of 20 rater judgements. The reliability of this aggregated measure is satisfactory in all indications. The results of 561 rehabilitation centres show that those quality criteria are in particular need of improvement that refer to subjective concepts of patients (e. g., subjective theories of illness). Between peer review procedures in 1998 and 1999, the quality scores of rehabilitation centres had improved whereas between 1999 and 2000/2001, no further improvement can be shown. However, those rehabilitation centres with a low quality score in 1999 (lowest quartile of the distribution) underwent a positive development between 1999 and 2000/2001. Reasons for this trend and possibilities for improving interrater reliability of the peer review process as an element of the quality assurance programme of the German Pension Insurance scheme are discussed.