Objectives 1) To evaluate how frequently subjects with knee pain seek care from their general practitioner (GP) and from physiotherapy (PT); and 2) to evaluate the factors associated with GP care seeking and PT use.

Methods Population-based longitudinal cohort study of subjects, age 40-79, with knee pain. Inclusion criteria:1) pain or discomfort in/around the knee on most days at any time in the past and 2) any pain or discomfort in/around the knee in the past year. Exclusions were inflammatory arthritis, recent knee surgery or injury, inability to undergo MRI or x-ray. At the 3-year follow-up, subjects were evaluated with comprehensive questionnaires to assess clinical variables including self-reported knee swelling and pain severity using the Western Ontario and McMaster's Universities (WOMAC) Osteoarthritis Index (VAS 3.1). WOMAC Pain at night was used in this analysis as an indicator of severe pain. Weight and height were measured and body mass index (BMI) calculated (kg/m²). Self-reported GP care seeking and PT use for knee discomfort was ascertained. Fixed-flexion AP knee radiographs were read by 2 independent readers using Kellgren-Lawrence (KL) 0-4 grading, where KL≥2 indicates radiographic OA. Multivariable logistic regression analysis was used to evaluate the association of clinical factors with care seeking at 3-year follow-up. Separate models were used for GP and PT care seeking. A priori selected factors in both models included age, sex, BMI, education (≤grade 12, post-secondary), WOMAC Pain at night (0-100), self-reported knee swelling and KL grade (<2/≥2). In addition, in the PT use model, GP care seeking for knee discomfort (no/yes) was also included as an independent variable.

Results Of 163 subjects, 47.2% were female, 28.2% had no post-secondary education. Mean (SD) age was 60.0 (10.8), BMI was 27.0 (4.1), pain at night was 12.4 (19.0). KL grades 0-4 were seen in 20.9%, 19.0%, 27.0%, 12.3%, and 20.9%, respectively, i.e. 60.2% had radiographic OA. 58.4% of subjects reported having seen their GP for knee discomfort, while only 36.3% reported using PT. In multivariable logistic regression, variables significantly associated with GP care seeking included male sex (OR 2.5, 95% CI 1.2-5.3), self-reported knee swelling (OR 5.0, 95% CI 2.1-11.9), and KL grade ≥2 (OR 2.8, 95% CI 1.2-6.6). Age, education, BMI and pain at night were not significant in this model. Variables significantly associated with PT use included female sex (OR 2.8, 95% CI 1.3-6.3), pain at night (OR 0.97, 95% CI 0.95-1.0), and GP care seeking (OR 3.6, 95% CI 1.5-8.3), while post-secondary education was of borderline significance (OR 2.4, 95% CI 0.98-6.1). Age, BMI, self-reported knee swelling and KL grade were not significantly associated with PT use.

Conclusions In this population-based cohort of subjects with knee pain, GP care seeking for knee discomfort and management with PT was underutilized. Males, those with self-reported knee swelling and with radiographic OA were more likely to seek GP care. Subjects who had sought GP care were three times more likely to use PT, than those who had not seen a GP, suggesting that GP referral may play an important role for PT use. This study highlights the need for patient education to enhance appropriate health care utilization in those with knee pain, including patients with earlier stages of OA, where therapeutic interventions may have a greater impact.

Disclosure of Interest None declared