[Rehabilitation need in sociomedical assessment--development and evaluation of an algorithm of decision-making]

For sociomedical assessment on application for medical rehabilitation by the German Statutory Pension Insurance no valid diagnostic assessment tools exist. In the present study two expert surveys and current guidelines were used to develop an algorithm of decision-making which is based on 13 criteria for rehabilitation need. The algorithm was evaluated in a prospective, longitudinal survey in the sociomedical services of five Bavarian Statutory Pension Insurance agencies on a sample of applicants with primary musculoskeletal disease (n=243). The medical experts rated the criteria after they had made their decision on the applications. To examine validity of the algorithm, concordance with the medical expert decisions and individual need for rehabilitation was analyzed. Respectively results of the algorithm were compared with data in course of the measure (rating by the clinician, patient questionnaire). Analyzing data, there is evidence for problems in rating the criteria by assessment on medical records; on average 9.2 (s=2.4) of 13 criteria could be rated. There is statistically significant concordance between decisions by the algorithm and sociomedical decisions based on assessment of medical records (kappa=0.37, p<0.001) or examination (kappa=0.71, p<0.001). Applications with positive recommendation by the algorithm have a significantly higher degree of subjective rehabilitation need (d=0.38). The results illustrate the need for measures to improve the basis of information by medical records. By reason of heterogeneity in practical application of criteria for rehabilitation need, operationalisation and explicit diagnostic algorithms are necessary to improve objectivity of making decisions.