Feasibility and impact of a patient support group care model on diabetes and hypertension care in informal settlements in Nairobi, Kenya: a quasi-experimental study

Abstract

Background A support group care model including self-financing is a promising strategy to improve care for patients with diabetes or hypertension in resource-constrained settings. Objectives We investigated the uptake, feasibility, and impact of a self-financing patient support group care model on cardiometabolic parameters among adult patients with uncontrolled diabetes or hypertension in informal settlements in Nairobi, Kenya. Methods A two-group prospective quasi-experimental study was conducted. The outcomes were changes in mean glycated haemoglobin (HbA1c), systolic blood pressure (SBP), diastolic blood pressure (DBP), body mass index, and waist–hip ratio in control versus intervention communities, assessed 6 months after intervention implementation. Results At baseline, 118 patients with diabetes (intervention, 60; control, 58) and 176 with hypertension (intervention, 87; control, 89) were enrolled. At endline, 81 patients with diabetes and 137 with hypertension were surveyed. In the intervention arm, HbA1c decreased from 10.8% to 9.0% (mean difference [95% CI]: −1.7 [−2.4, −0.9] p < 0.001) and in the control arm from 10.6% to 9.9% (−0.9 [−1.5, −0.3] p = 0.005). Difference-in-difference analysis showed a notably greater reduction in HbA1c in the intervention arm (−0.942 [0.463] p < 0.05). In the intervention arm, SBP decreased from 155.0 mmHg to 148.7 mmHg (−6.3 [−11.7, -0.9] p = 0.022) and in the control arm, from 160.1 mmHg to 152.5 mmHg (−7.6 [−12.9, −2.3] p = 0.005). DBP in the intervention arm changed from 99.1 mmHg to 97.9 mmHg (−1.1 [4.2, 1.9] p = 0.462) and in the control arm from 99.7 mmHg to 94.8 mmHg (−4.9 [7.8, −2.0] p = 0.001). Conclusions A self-financing patient support group care model is feasible, improves cardiometabolic parameters and can be a strategy to manage diabetes, hypertension, and other chronic diseases in low-resource settings.