Background Transport accidents result in Mental Health, pain and return-to-work (RTW) difficulties. To screen for risk of future poor outcomes assists early identification and effective recovery management for compensation systems that provide compensation for health care services post-accident. The purpose of this study was to evaluate the measurement properties of a predictive risk screening tool known as the Client Conversational Tool-Revised (CCT-R) within the recovery branch of the Transport Accident Commission (TAC), Victoria, Australia.

Methods De-identified data from 630 TAC clients who claimed compensation between April 2012 and October 2014 was analysed. Claims and payments data of later health service use was linked with associated CCT-R ’risk scores’ for mental health, pain and RTW. Chronbach’s alpha and Guttman Split-Half reliability analysis was used to evaluate internal consistency. Validity was assessed with analysis of receiver operating characteristice curves to estimate the area under the curve (AUC) for each CCT-R domain.

Results The mental health domain performed adequately with good reliability (Chronbach α = 0.841) and reasonable validity (AUC = 0.716; specificity, 76.84%; sensitivity, 66.28%). Poor validity evidence was observed for pain (AUC = 0.608; sensitivity, 37.93%; specificity, 83.74%) and RTW domains (AUC = 0.534; sensitivity, 18.99%; specificity, 88.56%). The RTW domain presented with good reliability (Chronbach α = 0.847). The Pain domain could not be described as reliable (Chronbach α = 0.039).

Conclusions As a risk screening tool the CCT-R currently presents with discernible deficiencies relating to its measurement properties in two of the three domains. To facilitate early intervention decision management, CCT-R mental health scores could be useful, while pain and RTW domains require further revision.