Using data from Multidimensional Pain Inventory subscales to assess functioning in pain rehabilitation

The aim of this study was to examine whether Multidimensional Pain Inventory (MPI) subscale score changes can be used for monitoring interdisciplinary cognitive behavioural pain rehabilitation programmes, using the Psychological General Well-Being (PGWB) index as an independent variable of rehabilitation outcome. Data from 434 consecutively referred patients disabled by chronic pain were analysed. The intervention was a 4-week interdisciplinary pain rehabilitation group programme (5 h/day), based on biopsychosocial and cognitive behavioural principles. Mean PGWB total scores improved after rehabilitation (P<0.0001) with clinically relevant effect sizes for patients with 'dysfunctional' and 'interpersonally distressed' MPI profiles. Substantial correlations (r=0.7-0.3; P<0.001) were found between the changes in PGWB total scores and four of the MPI subscale scores. These were combined into a composite variable ['pain severity', 'interference', 'life control' (given reversed scores) and 'affective distress'], and were labelled as the Pain Rehabilitation Index. The subscales, 'support' and 'general activity level', were omitted, as changes were ambiguous with respect to functioning. 'Dysfunctional' and 'interpersonally distressed' profile patients showed a marked improvement in Pain Rehabilitation Index after rehabilitation (effect sizes of 0.77 and 0.43; P<0.0001, respectively). Conversely, the 'adaptive copers' may have deteriorated somewhat (effect size -0.28; P=0.036). We propose that scores from four MPI subscales are integrated and the difference pre-post rehabilitation is used to indicate composite rehabilitation outcomes, making it possible to interpret all included MPI subscales in the same direction. Psychometric evaluation of the index is warranted.