Deutsch | Imprint


Zoom = Maus über Bild bewegen | Bild hier öffnen


Final report Outcome



Prof. Dr. Oskar Mittag
Abteilung Qualitätsmanagement und Sozialmedizin (Direktor: Prof. Dr. W. H. Jäckel)
Universitätsklinikum Freiburg
Engelbergerstraße 21
79106 Freiburg
Tel.: 0761 / 270 73540

previous article  |  next article

Outcome measurement in medical rehabilitation: Comparison of different methods of outcome measurement using datasets from the rehabilitation sciences and cognitive interviews about change items

Oskar Mittag (a), Thomas Kohlmann (b), Thorsten Meyer (c), Heiner Raspe (d)
(a) Abteilung Qualitätsmanagement und Sozialmedizin (AQMS), Universitätsklinikum Freiburg
(b) Institut für Community Medicine, Abteilung Methoden der Community Medicine, Greifswald
(c) MHH, Forschungsbereich Integrative Rehabilitationsforschung, Institut für Epidemiologie, Sozialmedizin und Gesundheitssystemforschung, Hannover
(d) Institut für Sozialmedizin, Universitätsklinikum Schleswig-Holstein (Campus Lübeck)


Funding: Bundesministerium für Bildung und Forschung (FKZ 01GX0747)


Project start: 1.1.2008


Funding duration: 18 months (cost-neutral extension until March 2010)


Project leaders:


Primary cooperation partners:



The evaluation of medical rehabilitation measures mostly takes place in Germany through the analysis of changes in relevant outcomes in a simple pre-post comparison. To this aim, status measurements are carried out at different measurement time points and the extent of changes between pre- and post-measurement are depicted in the form of standardised mean differences (effect sizes) (=indirect change measurement). However, treatment effects can also be measured by means of explicit change judgements by the patients themselves, by, for example, measuring the changes at the post-time point in the sense of "improved – unchanged – worsened" (= direct change measurement) or in the form of global ratings of the treatment success. Results from a series of research projects from the rehabilitation sciences show that the direct change measurement or global success ratings predominantly depicts better successes than the consideration of pre-post effects on the mean value level.


Project aim

Based on this confusing contradiction between the predominantly good to very good evaluation of long-term rehabilitation success by the patients (global or direct success measurement) and the often only low effect sizes using indirect methods of change measurement, the aim of the project was to determine empirically founded solution approaches for central content-based and methodological problems of outcome measurement in medical rehabilitation. The project should contribute to an improved understanding of the mutual interdependence of the different methods of outcome measurement and permit substantiated decisions regarding which methods of change measurement reliably and validly depict the effects of rehab measures in a single-group pre-post comparison.



Four available datasets from recently concluded rehab research projects (total N=5592 rehabilitees) were subjected to secondary analyses, with the following assessment and evaluation methods:

In parallel to this, in a second investigation approach, the process of the patients in evaluating the rehab measure was reconstructed by means of cognitive interviews in order to explain existing divergences between indirect and direct change measurement.


Current status

For the evaluation of (medical) interventions, different methodological approaches are available, between which decisions have to be made in concrete cases. In the framework of the research project, we examined methodological and content-based problems of change measurement. Our goal was to derive empirically supported recommendations regarding which type of change measurement is suitable for which questions and under which framework conditions. The overall project is subdivided into five subprojects with different questions and differing scopes.

The Freiburg subproject (Prof. Dr. Oskar Mittag) was concerned above all with the contradiction that rehabilitees retrospectively evaluate the success of rehab measures as good or very good, but that in a comparison of the pre- and post-values on the mean level (indirect change measurement), no or only small effects are shown. Furthermore, the Freiburg work group sought paths with which to depict individual change progressions as simply and clearly as possible. A further subproject (Prof. Dr. Erik Farin-Glattacker) used, in contrast to this, complex statistical methods to model progression types with multiple status measurement.

The Greifswald subproject (Prof. Dr. Thomas Kohlmann) examined differences in the content-based structure of data that were gathered by means of direct and indirect change measurement. Dr. Nikolous Gerdes was concerned with problems of the clinical significance of the measured changes.

In the Lübeck subproject (Prof. Dr. Thorsten Meyer), in a second investigation approach, explanatory models for the discrepancies between the results of direct and indirect change measurement were developed. With the help of qualitative interviews, it was examined here how rehabilitees understand questions on direct and indirect change measurement, which cognitive processes determine their response behaviour, and what explanations there are for discrepancies of the two measurement methods.

These subprojects were supported by an expert group which met regularly during the running time of the projects at workshops and symposia. As one fundamental result, we have formulated joint recommendations for the patient-based change measurement in medical rehabilitation, which are currently in press. Moreover, further publications from the project are available. Overall, we hope that our results have contributed to an improved understanding of the occurrence of change in rehabilitation and to the measurement thereof.

< Development of a standardised ... - Kilian  |  QSFK - Trojan >