Strategic Investment Board

Correction: BMC Public Health 25, 3211 (2025)https://doi.org/10.1186/s12889-025-24075-3<br/><br/>Following publication of the original article [1], the authors identified an error in Pamela Whitaker’s name and an error in the affiliation listed for both Pamela Whitaker and Saul Golden.<br/><br/>The incorrect author name is: Pamela Whittaker.<br/><br/>The correct author name is: Pamela Whitaker. Pamela Whitaker is affiliated to Belfast School of Art, Ulster University, Belfast, UK.<br/><br/>Saul Golden is affiliated to Belfast School of Architecture and the Built Environment, Ulster University, Belfast, UK.<br/><br/>The author group has been updated above and the original article [1] has been corrected

IntroductionHealth care systems have failed to address the poor physical health outcomes of people living with severe mental illness. Interventions that focus on specific health behaviours and/or lack a co-design basis show little promise. There is a need for whole systems approaches that tackle the complex issues, including social isolation, discrimination, stigma, and low motivation, that influence poor health in this population. A social prescribing model that accommodates the needs and preferences may be a way forward.MethodsA mixed methods approach that assesses the CHOICE model (Challenging Health Outcomes Integrating Care Environments) in relation to (a) the social exclusion, loneliness and social support of a cohort of people living in the community; (b) participants’ experience of social prescribing and potential improvements to the intervention; (c) understanding the implementation factors, mechanisms and outcomes; (d) the engagement and sustainment of community partnerships; (e) institutional changes in policy and practice.DiscussionCodesigned and community-based participatory interventions may be crucial in tackling the health and social inequalities experienced by people with severe mental illness. However, given the complexity of such interventions, the social prescribing model that we describe in this paper, requires considerable implementation data prior to a full trial.