Generalizability of the Progressive Resistance Training versus Total Hip Arthroplasty (PROHIP) trial: a cross-sectional study of 402 patients in Denmark

Abstract

Background and purpose: There is ongoing debate over whether results from randomized trials assigning patients to surgery or first-line treatment can be generalized to clinical practice. Therefore, we aimed to compare patients with hip osteoarthritis accepting enrollment in the Progressive Resistance Training versus Total Hip Arthroplasty (PROHIP) trial in Denmark with those declining (enrolled in an observational cohort [non-PROHIP]).

Methods: We used a cross-sectional study design to compare demographics and patient-reported outcomes among patients eligible for enrollment in the PROHIP trial. We used the standardized difference (StdDiff), the absolute difference with 95% confidence interval (CI), and the propensity (odds ratio [OR]) of accepting participation in the PROHIP trial to assess imbalances between groups. We pre-specified that StdDiff values < 0.2 indicated a negligible difference, whereas values ≥ 0.8 indicated incomparability.

Results: 402 patients were included, with 109 in the PROHIP trial and 293 in the non-PROHIP cohort. Patients enrolled in the PROHIP trial had a mean (standard deviation [SD]) Oxford Hip Score at baseline of 25.1 (SD 5.9) compared with 22.6 (SD 6.9) in the non-PROHIP cohort (between-group difference, 2.5 points [CI 1.1-4.0], StdDiff 0.4, OR 1.06 [CI 1.02-1.10]). This pattern was consistent across almost all secondary patient-reported outcomes applied in the PROHIP trial. For most demographic variables, there were negligible between-group differences at baseline.

Conclusion: We found minimal imbalances in some baseline demographic variables and most patient-reported outcomes, with those who accepted enrollment in the PROHIP trial having more favorable outcomes at recruitment than those who declined. However, most differences were not clinically important.

Figure 1

Flowchart of the patients in the study. Progressive Resistance Training versus Total Hip Arthroplasty (PROHIP) trial. Parallel prospective non-interventional observational cohort (non-PROHIP).

Figure 2

(A) Number of patients assessed for eligibility; (B) number of patients enrolled in the Progressive Resistance Training versus Total Hip Arthroplasty (PROHIP) trial; (C) number of patients enrolled in the parallel prospective non-interventional observational cohort (non-PROHIP); (D) enrollment rate in PROHIP; and (E) enrollment rate in non-PROHIP.

Figure 3

Standardized differences of the means across the different patient-reported outcomes score variables between patients in the Progressive Resistance Training versus Total Hip Arthroplasty (PROHIP) trial and parallel prospective non-interventional observational cohort (non-PROHIP) at baseline. Standardized differences of less than 0.2 indicate a negligible difference, while 0.8 or greater indicates definitive incomparability (i.e., indicative of low generalizability) in the patient-reported outcome score variables between the 2 groups. OHS: Oxford Hip Score; HOOS: Hip disability and Osteoarthritis Outcome Score; UCLA: University of California Los Angeles (UCLA) activity score; VAS: Visual Analogue Scale; EQ-5D-5L: EuroQol Group 5-dimensions 5-levels; and EQ-VAS: the EuroQol Group Visual Analogue Scale.