Improving the management of type 2 diabetes in China using a multifaceted digital health intervention in primary health care: the SMARTDiabetes cluster randomised controlled trial

Abstract

There is limited evidence, mainly from high-income countries, that digital health interventions improve type 2 diabetes (T2DM) care. Large-scale implementation studies are lacking. A multifaceted digital health intervention comprising: (1) a self-management application ('app') for patients and lay 'family health promotors' (FHPs); and (2) clinical decision support for primary care doctors was evaluated in an open-label, parallel, cluster randomized controlled trial in 80 communities (serviced by a primary care facility for >1000 residents) in Hebei Province, China. People >40 years with T2DM and a glycated haemoglobin (HbA1c) ≥7% were recruited (∼25/community). After baseline assessment, community clusters were randomly assigned to intervention or control groups (1:1) via a web-based system, stratified by locality (rural/urban). Control arm clusters received usual care without access to the digital health application or family health promoters. The primary outcome was at the participant level defined as the proportion with ≥2 "ABC" risk factor targets achieved (HbA1c<7.0%, blood pressure < 140/80mmHg and LDL-cholesterol < 2.6mmol/L) at 24 months. A total of 2072 people were recruited from the 80 community clusters (40 urban and 40 rural), with 1872 (90.3%) assessed at 24 months. In the intervention arm, patients used FHPs for support more in rural than urban communities (252 (48.6%) rural vs 92 (21.5%) urban, p<0.0001). The mean monthly proportion of active app users was 46.4% (SD 7.8%) with no significant difference between urban and rural usage rates. The intervention was associated with improved ABC control rates (339 [35.9%] intervention vs 276 [29.9%] usual care; RR 1.20, 95% CI 1.02-1.40; p=0.025), with significant heterogeneity by geography (rural 220 [42.6%] vs 158 [31.0%]; urban 119 [27.9%] vs 118 [28.6%]; p=0.022 for interaction). Risk factor reductions were mainly driven by improved glycaemic control (mean HbA1C difference-0.33%, 95% CI-0.48 to-0.17; p=0.00025 and mean fasting plasma glucose difference-0.58mmol, 95% CI-0.89 to-0.27; p=0.00013). There were no changes in blood pressure and LDL-cholesterol levels. A multifaceted digital health intervention improved T2DM risk factor control rates, particularly in rural communities where there may be stronger relationships between patients and doctors and greater family member support. National Health and Medical Research CouncilGlobal Alliance for Chronic Diseases (ID 1094712).