Agreement is very low between a clinical prediction rule and physiotherapist assessment for classifying the risk of poor recovery of individuals with acute whiplash injury

Abstract

BackgroundA prognostic clinical prediction rule (whiplash CPR) has been validated for use in individuals with acute whiplash associated disorders (WAD). The clinical utility of this tool is unknown. ObjectivesTo investigate: 1) the level of agreement between physiotherapist- and whiplash CPR-determined prognostic risk classification of people with acute WAD; 2) which clinical findings are used by physiotherapists to classify prognostic risk; and 3) whether physiotherapists plan to differ the number of treatment sessions provided based on prognostic risk classification. DesignPragmatic, observational. Method38 adults with acute WAD were classified as low, medium, or high risk of poor recovery by their treating physiotherapist (n = 24) at the conclusion of the initial consultation. A weighted Cohen's kappa examined the agreement between physiotherapist estimated risk classification and the whiplash CPR. Physiotherapists' reasons for classification were provided and summarised descriptively. Kruskal-Wallis and post-hoc Dunn's tests compared projected number of treatment sessions between risk subgroups. ResultsPhysiotherapist agreement with the whiplash CPR occurred in 29% of cases (n = 11/38), which was less than what is expected by chance (K = −0.03; 95%CI -0.17 to 0.12). Physiotherapists most frequently considered range of movement (n = 23/38, 61%), a premorbid pain condition (n = 14/38, 37%), response to initial physiotherapy treatment (n = 12/38, 32%), and pain intensity (n = 12/38, 32%) when classifying prognostic risk. The projected number of treatment sessions was not different between risk groups using classifications provided by the physiotherapists (χ2(2) = 2.69, p = 0.26). ConclusionsPhysiotherapists should consider incorporating the whiplash CPR into current assessment processes to enhance accuracy in prognostic decision-making.