Title : Measuring the responsiveness of the SF-36 in the cardiac rehabilitation literature: Combined results of a comparative effectiveness review and meta-analysis
Objective:To investigate the responsiveness of Patient Outcome Measures (POMs) which are essential to map treatment effectiveness of Cardiac (CR) rehabilitation programs, this is an area often overlooked and may be an influence in the variability of evidence currently available.
Data Source: A comparative effectiveness literature review was conducted of studies with a pre to post POM assessment of responsiveness (change in health status over time). This was followed by a meta-analysis whose purpose was to gather pooled information on the
responsiveness of the main POMs.
Methods: A quality review of this literature not only included RCTs but also parallel studies as well as all observational and retrospective trials. This review included a list of articles and their characteristics, a quality assessment of the literature and a list of POMs utilised in this setting were assessed for responsiveness. For the meta-analysis a correlation analysis was undertaken between SF-36 pre and post “within” mean scores in the CR literature to define the degree to which the two sets of variables were related, this was for each domain of the SF-36 and for the PCS and MCS scores.
Results: There was inconsistency in the literature with the measurement of responsiveness or effect-size. The most responsive POM in this setting was the Global Mood Scale (GMS). The most commonly used POM was the SF-36 with the SF-36 PCS domain being the most responsive of the composite SF-36 domains, however the PCS shows less ability to discriminate in the higher SF-36 scores. In the individual domains Role Physical, Role Emotional and Physical function are closely grouped as being the most responsive.
Conclusion: The surveyed literature found no “gold standard” POM for CR. The domains of the SF-36 are not responsive to CR, and this is true particularly in the area of more severe disease. Overall it was found found that a Likert scale is more responsive than a dichotomous scale and that a simple questionnaire is more responsive in a pre to post setting than a complex questionnaire.