AGREEMENT BETWEEN THE KEELE START MSK TOOL AND THE SF-ÖMSPQ IN PEOPLE WITH MUSCULOSKELETAL PAIN IN AUSTRALIAN PRIMARY CARE

K. Kang¹, K. Evans², S. Kobayashi³, M. Simic², D. Beales⁴, K. Bennell⁵, M. Nicholas⁶, M. Sterling⁷, N.E. Foster⁸, S. Coates⁹, T. Rebbeck¹

¹The University of Sydney, Faculty of Medicine and Health, Camperdown, Australia, ²University of Sydney, Faculty of Medicine and Health, Camperdown, Australia, ³Australian Catholic University, School of Allied Health, Faculty of Health Sciences, Camperdown, Australia, ⁴Curtin University, School of Physiotherapy and Exercise Science, Bentley, Australia, ⁵University of Melbourne, Melbourne Physiotherapy School, Melbourne, Australia, ⁶Kolling Institute, Pain Management, St Leonards, Australia, ⁷NHMRC Centre of Research Excellence, Better Health Outcomes for Compensable Injury, Brisbane, Australia, ⁸University of Queensland, Surgical Treatment and Rehabilitation Service, Brisbane, Australia, ⁹Australian Catholic University, Discipline of Physiotherapy, Camperdown, Australia

Background: Risk stratified models of care, whereby treatment is matched with risk subgroups, may reduce the burden on the healthcare system. Two risk screening tools, the Keele STarT MSK Tool and the short-form Örebro Musculoskeletal Pain Screening Questionnaire (SF-ÖMSPQ), are frequently used in different populations and settings to identify people at risk of poor outcome. The Keele STarT MSK Tool classifies people into three risk subgroups (low, medium, and high risk) whilst the SF-ÖMSPQ classifies people into two subgroups (low and high risk). No study has compared the use of these tools in an Australian primary care setting.

Purpose: To evaluate:
1) Agreement between the two risk screening tools for people with common musculoskeletal (MSK) conditions.
2) Discriminative validity of the two tools with respect to self-reported disability.

Methods: In this cross-sectional study, 767 participants with neck pain (NP), low back pain (LBP) and knee osteoarthritis (knee OA) who were within 4 weeks of presenting for Australian primary care completed the two risk screening tools, socio-demographic questions and health outcomes. Descriptive statistics were used to compare the number of people in each risk subgroup by each tool. Cohen’s kappa test and Spearman's rho were used to determine the agreement between the two tools. One-sample t-tests were used to calculate differences between self-reported disability and risk subgroups.

Results: The proportion of participants stratified as low, medium and high-risk subgroups by the Keele STarT MSK Tool were 16.9%, 56.8% and 25.9% respectively. The proportion of participants stratified as low and high-risk subgroups by the SF-ÖMSPQ were 57.4% and 42.6% respectively. Most participants were medium risk using the Keele STarT MSK Tool irrespective of their MSK conditions (NP 63.7%; LBP 51.5%, knee OA 59.5%) and low risk using the SF-ÖMSPQ (NP 57.9%, LBP: 51.8%, knee OA 63.7%). A higher proportion of participants with knee OA (63.7%) were stratified as low risk using the SF-ÖMPSQ compared with other conditions (e.g., NP 57.9%, LBP: 51.8%). There was “moderately strong” agreement between the two tools across the three MSK conditions (r=0.72-0.74, p<0.01). Discriminative validity was demonstrated in both tools, with significant differences in self-reported disability between risk subgroups (e.g., WHODAS, mean (95%CI) for SF-ÖMSPQ; low risk 6.98 (6.46 to 7.51), high risk 16.28 (15.35 to 17.21), p<0.001).

Conclusions: In our Australian cohort, most participants were medium risk using the Keele STarT MSK Tool and were low risk using the SF-ÖMSPQ. More participants with knee OA were stratified as low risk than those with spinal pain conditions. There was “moderately strong” agreement between the tools. Risk subgroups for both tools discriminated well for self-reported disability.

Implications: The tools performed similarly in the Australian primary care setting as the original derivation cohorts, hence providing confidence that either tool could be used in an Australian primary care setting. The choice of which tool to use may depend on the available matched care pathways for each risk subgroup in the particular healthcare setting. Future longitudinal data will evaluate accuracy of the tools in predicting long term outcomes.

Funding acknowledgements: This study analysed the baseline data from the PACE MSK trial supported by a NHRMC project grant (APP1141377).

Keywords:
Risk screening tools
Agreement
Australian primary care

Topics:
Musculoskeletal

Did this work require ethics approval? Yes

Institution: USYD, UQ, UMEL, Curtin Uni, RNSH

Committee: USYD, UQ, UMEL, Curtin Uni, Northern Sydney Local Health District

Ethics number: 2018/926, 2019000700/2018/926, 1954239, HRE2019-0263 and 2019/ETH03632