Persistence and spread of RADAR: A longitudinal evaluation of a diabetes care model

The RADAR (Reorganizing the Approach to Diabetes through the Application of Registries) project was launched to enhance type 2 diabetes (T2D) care and outcomes in on-reserve First Nations communities in Alberta, Canada. The objective of this study is to determine whether previously observed improvements in diabetes care outcomes were sustained over time in the communities that continued the program (continued communities) and whether these improvements were reproducible in new communities.

RADAR involved an innovative, culturally contextually grounded care model designed for First Nations people, which was deployed in a stepped-wedge design to communities. The primary outcome was a 10% improvement or persistence at target in A1C, systolic blood pressure (SBP), and/or LDL, using the intention-to-treat (ITT) framework. Outcomes were assessed in 3 continued communities at 3-, 4-, and 6- years post-implementation of the RADAR program. We provided the same intervention to 2 new communities and assessed outcomes at 1 year post-implementation (repeatability assessment).

In 2024, 256 T2D patients were registered in RADAR, ranging from 28 to 86 individuals per community. The average age was 61 years (SD=11.7) with 57% females (N = 145). ITT analysis showed all 3 continued communities maintained high rates of primary endpoint achievement, with 94%, 100% and 92% of participants in each community achieving the primary endpoint (n = 174), respectively, relative to baseline (p < 0.001). After 1 year of RADAR in the 2 new communities (n = 28 and n = 54), the ITT combined endpoint was achieved in 82% and 91% of participants, respectively (p < 0.001).

RADAR continues to be associated with high levels of achievement of the primary combined endpoint across communities and over time. This study shows promise that future RADAR implementation could be adapted, in collaboration with First Nation Health Managers, to other First Nations community settings.