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

Highlights

• Agreement between physiotherapist- and whiplash CPR-determined prognostic risk classification was very low.

• Physiotherapists were overly optimistic regarding patient outcomes.

• The whiplash CPR may aid physiotherapist prognostic judgements.

• Strategies to enhance use of the whiplash CPR and prognostic indicators are needed.

Abstract

Background

A 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.

Objectives

To 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.

Design

Pragmatic, observational.

Method

38 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.

Results

Physiotherapist 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).

Conclusions

Physiotherapists should consider incorporating the whiplash CPR into current assessment processes to enhance accuracy in prognostic decision-making.

Introduction

Substantial heterogeneity in the population, complexity in the condition, and modest effect sizes from treatment trials (Lamb et al., 2013; Jull et al., 2013), means that clinicians may be faced with uncertainty when making decisions regarding the best course of management for individuals with acute whiplash associated disorders (WAD). The ability to gauge the likely prognosis of patients with acute WAD is important given that up to 50% of those injured will not fully recover, but will develop persistent pain and disability (Carroll et al., 2009; Sterling et al., 2006). Appropriate treatment in the early post injury period will likely be critical to facilitate recovery given most recovery (if it occurs) takes place in the first 2–3 months following injury (Sterling et al., 2010).

To our knowledge, physiotherapists' ability to identify risk of poor recovery has not been well investigated in people with acute WAD. Predictions of return to work status have been demonstrated to be improved by the addition of physiotherapist judgements to a predictive model that included demographic, psychological and pain variables (Scott and Sullivan, 2010). However, this study included patients with sub-acute to chronic WAD undergoing a rehabilitation intervention and did not provide indication of the accuracy of predictions about later recovery made in the acute injury stage. Evidence of physiotherapists' accuracy in identifying the prognosis of people with other spinal pain conditions has been inconsistent (Hancock et al., 2009; Dagfinrud et al., 2013; Abbott and Kingan, 2014; Cook et al., 2015). Cook and colleagues investigated physiotherapist estimation of three patient outcomes two weeks after initial contact, in people with any duration neck and back pain and found significant associations between clinician-predicted and actual disability and patient-reported recovery outcomes, but not resolution of pain (Cook et al., 2015). Alternatively, Dagfinrud and colleagues investigated physiotherapist prediction of eight week disability outcome in a neck pain population and concluded that clinician predictions did not add value to the prognostic model (Dagfinrud et al., 2013).

Research indicates that clinicians frequently rely on past experience, pattern recognition and intuition to guide early decision-making (McGinn et al., 2000), which can be misleading and may result in inconsistency (Liao and Mark, 2003) and bias (Kleinmuntz, 1990). Physiotherapists have a positive attitude towards using evidence based practice (Iles and Davidson, 2006) and consider clinical prediction rules (CPR) as part of best practice (Knox et al., 2015). However there seems to be very low clinical adoption rates of CPRs relating to musculoskeletal pain (Knox et al., 2015; Haskins et al., 2014). A clinical prediction rule (whiplash CPR), to improve the precision of clinicians' prognostic judgements for individuals with acute WAD, was derived and has successfully undergone a retrospective, external validation (Ritchie et al., 2013, 2015). Low, medium, and high risk prognostic groups are determined using cut-off scores for an individual's, Neck Disability Index score (Vernon and Mior, 1991), age and hyperarousal symptoms (Foa et al., 1997) (Fig. 1.). Although the final step in the development of a CPR, impact analysis, is currently being investigated (Rebbeck et al., 2016), the validated CPR may be considered for use with an acute WAD population (Kelly et al., 2017a). Determining the agreement between physiotherapists' estimation and the whiplash CPR's risk groupings is important in informing the utility of the CPR (Brehaut et al., 2006). For example, the whiplash CPR's perceived usefulness and subsequent adoption may be enhanced if agreement is found to be poor, as this would indicate the potential benefit of using the CPR in order to sub-group patients for treatment based on predicted recovery. Conversely, if agreement is good the effort required to implement such a tool may outweigh any potential benefit of use in the clinical setting (McGinn et al., 2000). The primary aim of this study was to determine the level of agreement between physiotherapist estimated and whiplash CPR determined prognostic risk classification of people with acute WAD. Secondary aims were to investigate the clinical findings most commonly used by physiotherapists to classify prognostic risk, and gain insight into whether or not physiotherapists plan to provide a different number of treatment sessions based on patient prognostic risk grouping.