Optimizing measurement for neurobehavioural rehabilitation services: A multisite comparison study and response to UKROC.

To evaluate the efficacy of neurobehavioural rehabilitation (NbR) programmes, services should employ valid, reliable assessment tools; the ability to detect change on repeated assessment is a particular requirement. The United Kingdom Rehabilitation Outcomes Collaborative (UKROC) requires neurorehabilitation services to collect data using a standardized basket of measures, but the responsiveness and usefulness of using these in the context of NbR remains unknown. Anonymous data collected at two assessments for 123 people were examined using multiple methods to determine responsiveness of four outcome measures routinely used in NbR (HoNOS-ABI, FIM + FAM UK, MPAI-4, SASNOS). Predictive validity of two measures of rehabilitation complexity (RCS-E, SRS) regarding the extent of difference scores on these outcome measures at reassessment was also determined. All four outcome measures demonstrated responsiveness, with higher levels for SASNOS and MPAI-4 when only participants categorized as "most likely to change" at first assessment were analyzed. Predictive validity of the RCS-E and SRS in estimating the extent of change was variable. SRS was only predictive of improvement on the MPAI-4 whilst RCS-E was not predictive at all. Recommendations are made regarding ideal characteristics of NbR outcome measures, along with the need to develop measures of rehabilitation complexity specifically conceptualized for these programmes.