Keele Research Repository

BACKGROUND: There is no physiotherapy-specific quality indicator tool available to evaluate physiotherapy care for people with hip and/or knee osteoarthritis (OA). This study aimed to develop a patient-reported quality indicator tool (QUIPA) for physiotherapy management of hip and knee OA and to assess its reliability and validity.

METHODS: To develop the QUIPA tool, quality indicators were initially developed based on clinical guideline recommendations most relevant to physiotherapy practice and those of an existing generic OA quality indicator tool. Draft items were then further refined using patient focus groups. Test-retest reliability, construct validity (hypothesis testing) and criterion validity were then evaluated. Sixty-five people with hip and/or knee OA attended a single physiotherapy consultation and completed the QUIPA tool one, twelve- and thirteen-weeks after. Physiotherapists (n = 9) completed the tool post-consultation. Patient test-retest reliability was assessed between weeks twelve and thirteen. Construct validity was assessed with three predefined hypotheses and criterion validity was based on agreement between physiotherapists and participants at week one.

RESULTS: A draft list of 23 clinical guideline recommendations most relevant to physiotherapy was developed. Following feedback from three patient focus groups, the final QUIPA tool contained 18 items (three subscales) expressed in lay language. The test-retest reliability estimates (Cohen's Kappa) for single items ranged from 0.30-0.83 with observed agreement of 64-94%. The intraclass correlation coefficient (ICC) and 95% confidence interval (CI) for the Assessment and Management Planning subscale was 0.70 (0.54, 0.81), Core Recommended Treatments subscale was 0.84 (0.75, 0.90), Adjunctive Treatments subscale was 0.70 (0.39, 0.87) and for the total QUIPA score was 0.80 (0.69, 0.88). All predefined hypotheses regarding construct validity were confirmed. However, agreement between physiotherapists and participants for single items showed large measurement error (Cohen's Kappa estimates ranged from - 0.04-0.59) with the ICC (95% CI) for the total score being 0.11 (- 0.14, 0.34).

CONCLUSIONS: The QUIPA tool showed acceptable test-retest reliability for subscales and total score but inadequate reliability for individual items. Construct validity was confirmed but criterion validity for individual items, subscales and the total score was inadequate. Further research is needed to refine the QUIPA tool to improve its clinimetric properties before implementation.