Relating process evaluation measures to complex intervention outcomes: findings from the PACE-UP primary care pedometer-based walking trial

Abstract

Background: The PACE-UP trial demonstrated positive effects of a pedometer-based walking intervention on objective physical activity (PA) outcomes at three and 12 months in 45–75-year-old primary care patients, in postal and nurse-supported trial arms compared with controls. We explored associations between process evaluation measures and change in PA outcomes. Methods: The MRC framework guided process evaluation. Three quantitative measures (nurse session attendance [dose delivered], PA diary completion [fidelity] and pedometer use [fidelity]) were selected as independent variables in multi-level models estimating intervention effectiveness on PA outcomes (changes in step-counts and time in moderate-to-vigorous PA [MVPA] levels in ≥ 10-min bouts). Results: Dose: attending all three nurse sessions compared with 0–2 sessions was associated with an increase in steps/day at three and 12 months of 1197 (95% confidence interval [CI] = 627–1766) and 605 (95% CI = 74–1137), respectively; and MVPA in bouts (min/week) at three and 12 months by 74 (95% CI = 45–103) and 30 (95% CI = 3–57), respectively. Fidelity: postal and nurse groups showed strong positive associations of diary return with steps/day at three months: postal 1458 (95% CI = 854–2061), nurse 873 (95% CI = 190–1555). MVPA in bouts (min/week): postal 64 (95% CI = 33–94), nurse 50 (95% CI = 15–85). At 12 months, only the postal group effects remained statistically significant: steps/day 1114 (95% CI = 538–1689), MVPA 47 (95% CI = 18–75). Regular pedometer use in the postal group only was associated with higher three-month and 12-month steps/day: 1029 (95% CI = 383–1675) and 606 (95% CI = 22–1190), respectively, and with MVPA in bouts at three months: 40 (95% CI = 6–73). Conclusion: Process evaluation measures demonstrated significant associations with PA outcomes at three and 12 months. We cannot infer causality, but the associations between the process measures and PA outcomes suggest that they were important in enabling the trial changes observed and should be considered core components of the PACE-UP nurse and postal interventions. We have shown the MRC framework to be a useful tool for process evaluation of intervention implementation.