Type 2 diabetes self management education and support in Thailand: a cluster randomized trial (2019-2021)

Background:<br/>Diabetes self-management education and support (DSMES) improves glycemia via improved care coordination and self-management behaviors driven by educational programs tailored to individual needs, but scalable models for its delivery in low-and-middle-income countries are lacking. We evaluated the effectiveness of two scalable models of DSMES delivery: two sessions with a nurse (nurse-led) or one each with a nurse and a lay volunteer (peer-assisted).<br/>Methods:<br/>A three-arm cluster randomized trial was conducted comparing two DSMES interventions (nurse-led or peer-assisted) to routine care. The interventions consist of two DSMES sessions (once at baseline and a refresher at 6 months). The nurse-led group was delivered only by nurses, while the peer-assisted group was delivered by nurses with peer assistance . Seven primary care units were allocated to each arm. Newly diagnosed (at enrollment) or high risk glycemia (baseline HbA1c &gt;10%) people with type 2 diabetes were eligible. Primary outcomes were differences in HbA1c and Thai-CV risk score (estimates ten-year risk of cardiovascular events in Thai population based on various clinical parameters) at 12 months relative to routine care. Secondary outcomes were individual cardiovascular risk factors.<br/>Findings:<br/>693 individuals were enrolled (mean age 59 years, 269 (39%) male, and 424 (61%) female) and 664 completed the study (4% total drop-outs). For the control, nurse-led, and peer-assisted arms respectively, the means (SD) for primary outcomes are as follows. For HbA1c: at baseline; 8·5% (1·4), 8·5% (1·9), and 8·8% (2·0), at 12-month; 7·8% (1·6), 7·7% (1·3), and 8·1% (1·6). For Thai-CV risk scores: at baseline; 22 (15), 23 (15), and 23 (14), at 12-month; 21 (14), 23 (15), and 22 (13). Mean differences at 12-month from control (95% confidence intervals) for nurse-led and peer-assisted arms respectively were: HbA1c; –0·1% (–0·4 to 0·1) and 0.1% (–0·2 to 0·5), Thai-CV risk scores; 0% (–2 to 2) and –0% (–2 to 2).<br/>Interpretation:<br/>While there were no improvements in HbA1c or CV risk scores, low drop-out rates observed demonstrate the feasibility of the models. Future explorations of the appropriate balance between the intensity and feasibility of program delivery are needed.<br/>Funding:<br/>UK Medical Research Council (MR/R020876/1). Thailand Research Fund (DBG6180007).<br/>