Physical activity referral schemes (PARS) are composed of various components, such as a written prescription or a person-centered approach. The role of these components in their effectiveness is yet to be understood. Therefore, we aimed to explore the relationships between PARS components and physical activity, scheme uptake, and adherence rate; and to estimate the effect of PARS.
We searched Scopus, PubMed, Web of Science, CINAHL, ScienceDirect, SpringerLink, HTA, Wiley Online Library, SAGE Journals, Taylor & Francis, Google Scholar, OpenGrey, and CORE. Eligible studies were published between 1990 and November 2023 in English or German, investigated PARS with participants aged ≥ 16 years, and reported physical activity, scheme uptake, or scheme adherence. Separate random-effects meta-analysis by comparison group were conducted for physical activity. Scheme uptake and adherence rates were pooled using proportional meta-analysis. The components were analyzed via univariate meta-regression. We rated the risk of bias using RoB2 and ROBINS-I, and the certainty of evidence using GRADE.
Fifty-two studies were included. PARS were more effective in increasing physical activity than usual care (k = 11, n = 5046, Hedges' g = 0.18, 95%CI 0.12 to 0.25; high certainty of evidence). When PARS were compared with physical activity advice or enhanced scheme versions, the pooled Hedges' g values for physical activity were -0.06 (k = 5, n = 1082, 95%CI -0.21 to 0.10; low certainty of evidence), and 0.07 (k = 9, n = 2647, 95%CI -0.03 to 0.18; low certainty of evidence) respectively. Scheme uptake was 87% (95%CI 77% to 94%, k = 14, n = 5000) across experimental studies and 68% (95%CI 51% to 83%, k = 14, n = 25,048) across non-experimental studies. Pooled scheme adherence was 68% (95%CI 55% to 80%, k = 16, n = 3939) and 53% (95%CI 42% to 63%, k = 18, n = 14,605). The meta-regression did not detect any significant relationships between components and physical activity or scheme uptake. A person-centered approach, screening, and brief advice were positively associated with scheme adherence, while physical activity sessions were negatively associated.
PARS are more effective in increasing physical activity than usual care only. We did not identify any components as significant predictors of physical activity and scheme uptake. Four components predicted scheme adherence, indicating that the component-effectiveness relationship warrants further research.