A Randomized Controlled Trial of Social Prescribing: Comparing a Comprehensive Navigation Model With Signposting
Comprehensive navigation helped primary care patients access needed health or social resources more often than signposting.
*Randomized controlled trial; Level 2 (OCEBM).
Citation
Dahrouge S, Gauthier AP, Durand F, et al. A Randomized Controlled Trial of Social Prescribing: Comparing a Comprehensive Navigation Model With Signposting. Ann Fam Med. 2026;24:301-310. doi:10.1370/afm.250265
Background
Social prescribing connects patients with community services to address health and social needs. Evidence is mixed, and prior trials had not directly compared different levels of navigation support.
Patients
326 adults with health and/or social needs referred from primary care practices in two Ontario regions with large Francophone populations. Community Health Centres were excluded; patients needing immediate attention for uncontrolled medical illness were not eligible.
Intervention
Access to Resources in the Community: bilingual, comprehensive, patient-centered navigation with information, practical help, emotional support, and follow-up for up to 3 months.
Control
Signposting to 211-Ontario, an existing free telephone and online community-resource information service.
Outcome
Primary outcome: self-reported access to at least 1 needed health or social resource.
Follow-up Period
3 months.
Results
| Outcome | Navigation | Signposting | Effect |
|---|---|---|---|
| Accessed at least 1 needed resource (primary) | 50.3% | 35.8% | Absolute increase 14.5%; number needed to treat 7; adjusted odds ratio 1.82 (95% CI, 1.13-2.94) |
| Mean number of resources accessed | 0.93 | 0.67 | Higher with navigation |
| Overall satisfaction score | 90.4% | 61.5% | Higher with navigation |
| French-language services among Francophones wanting French | 92% | 36% | Higher with navigation |
| Ability to engage in care | Improved 2.4% | Declined 4.9% | Difference 7.3% (95% CI, 1.5-13.1) |
OCEBM = Oxford Centre for Evidence-Based Medicine; CI = confidence interval.
The primary analysis included all randomized patients, with those lost to follow-up counted as not accessing services. Findings were similar among completers. The control was an available current community-resource service.
Limitations
Access was self-reported and not verified. Follow-up was short, so effects on health outcomes or health care use are unknown. Attrition was 27%. Patient-level randomization could cause contamination, likely reducing apparent benefit. The study occurred in selected Ontario regions and may not generalize to all settings.
Funding
Ontario public research grant; no conflicts reported.
Clinical Application
Use comprehensive navigation for patients with complex social needs, especially when language-concordant services matter; signposting alone may be insufficient.
Discussion
Sign in to join the discussion.
In this randomized controlled trial, comprehensive ARC navigation improved self-reported access to at least one needed health/social resource versus 211-Ontario signposting (50.3% vs 35.8%; absolute difference 14.5%; aOR 1.82). Given self-reported outcomes and 27% attrition, is the result valid/applicable enough to change primary care practice? In this RCT, access to at least one needed resource was 50.3% with ARC navigation and 35.8% with 211-Ontario signposting. Which statement best distinguishes the absolute and relative effects?