Methods: SMART2D was an implementation research project and was structured in three phases or learning cycles: Theory of Change (including the formative studies), intervention framework and adaptive trial (including the implementation process). The work was guided by a framework for knowledge translation, allowing for a balance between evidence, stakeholder interaction and contextual adaptation. Each phase included multiple sub-studies which answered key research questions on their own; and also fed into the core activity of developing and implementing a self-management support intervention.
Learning Cycle 1: Theory of Change: A series of literature review and qualitative exploratory studies were conducted using a generic guide which was subsequently contextualized to site-specific interview and focus group guides. This phase extended over year 1 and 2 and looked at understanding current practices among participants and their perceptions of different aspects of self-management. In addition, the data from the formative phase in the three sites was further used to develop a theory of change for SMART2D (hereafter referred to as SMART2D Framework).
Learning Cycle 2: Intervention framework: The intervention program design was conducted in three steps facilitated by a coordinating team: (a) situational analysis based on the SMART2D Framework and definition of intervention objectives and core strategies; (b) designing generic tools for the strategies; and (c) contextual translation of the generic tools and their delivery. The situational analysis was done through an exploratory multiple-case study design where data from the interviews, focus group discussions, and observations were integrated and analyzed.
Learning Cycle 3: Adaptive trial: An adaptive implementation cluster randomized trial was implemented in two rural districts in Uganda with three clusters per study arm and in an urban township in South Africa with one cluster per study arm. A feasibility trial was implemented in socioeconomically disadvantaged suburbs in Stockholm, Sweden with two clusters per study arm. There were two study arms comprising a facility plus community arm and a facility-only arm. Uganda had a third arm comprising usual care. The primary outcome was controlled plasma glucose among participants with T2D and reduction in plasma glucose among those at high-risk or with pre-diabetes.
Results: The formative phase resulted in several studies that improved our understanding of current practices and perceptions, such as dietary practices in relation to diabetes or notions of well-being among participants in rural Uganda; perceptions of self-management among participants and community health workers in South Africa; and the meaning of community, the relevance of high-risk status and mismatch in perceptions about self-management between providers and patients in Sweden to name a few. Situational analyses showed that objectives and key functions addressing the mediators of self-management from the SMART2D framework could be shared between the three contexts.
The contextualized intervention framework included strategies focused on organization of care, linkage between health facility and the community, and strengthening patient role in self-management, community mobilization and a supportive environment as a complement to existing facility care. Facility only intervention was limited to ensuring adequate supplies of medicines and diagnostic material and the use of treatment guidelines in Uganda; and a check for standardized delivery of facility care in participating health centres in Sweden and South Africa. Contextualization was achieved by cross-site exchanges and local stakeholder interaction to balance intervention fidelity with local adaptation.
Uganda and South Africa implemented adaptive trials while Sweden implemented a feasibility trial. Overall, the trial enrolled a total of 1532 participants, 801 in Uganda, 584 in South Africa, and 147 in Sweden. Of 1532 participants, 775 were participants with diabetes, and 757 were at high risk for diabetes. Duration of intervention was 12 months in Uganda and 6 months each in South Africa and Sweden. UG had a usual care arm (enrolled: 266, completion: 42.9%) where diagnostics, equipment and medication were provided for the trial duration. Follow-up completion varied across the country sites, and diagnostic groups and was the lowest for usual care arm in Uganda [diabetes: UG: 75.0%, SA: 38.7%, SW: 81.4%; and high risk: UG: 57.6%, SA: 41.1%, SW: 76.1%]. Summary results of the SMART2D project are described in the attached figure.
