Optimising insulin use in people living with type 2 diabetes at primary healthcare facilities : the Tshwane insulin project

Abstract

BACKGROUND : In South Africa (SA), glucose control for individuals with type 2 diabetes follows a stepwise approach. According to the guidelines, insulin therapy is started after using two oral agents. However, various challenges may delay the initiation of insulin. OBJECTIVES : To implement a nurse-led, telehealth-assisted programme to address these challenges, aiming to transition patients to insulin safely to achieve better glycaemic control. METHODS : From 2021 to 2023, we conducted a single-arm, unblinded before-and-after study in primary care facilities in Tshwane District, Gauteng Province, SA. Participants were on insulin or two oral agents at maximum doses. Study nurses monitored glycated haemoglobin (HbA1c) results, and participants with HbA1c levels of ≥8% (≥10% during the COVID‑19 pandemic) were counselled about insulin use. During an initiation visit, participants received demonstrations and education on using insulin and glucose meters. The participants then tested their glucose levels at home according to a fixed schedule. Over 14 weeks, we implemented monthly clinic visits supplemented by home visits facilitated by community healthcare worker teams. During these visits, glucose results were communicated to the clinic physician via the Vula mobile app, allowing timely adjustments to insulin therapy. RESULTS : Of the 293 participants, 65% (n=192) were women and 35% (n=101) were men. The mean (standard deviation (SD)) age was 53 (10) years, with a baseline mean (SD) HbA1c level of 12.1% (1.7%). Of those initiated, 169 (58%) were on oral agents and 124 (42%) were on insulin. Biphasic mixed human insulin was prescribed to 185 participants (63%) and intermediate human neutral protamine Hagedorn (NPH) insulin to 108 (37%). Immediately after baseline assessment and during the 14-week study period, 72 participants (23%) were lost to follow-up, and seven were hospitalised during the study period. Glucose values decreased over 14 weeks, with approximately one-third of participants having no insulin adjustments, one-third having one adjustment, and one-third having more than one adjustment. The mean (SD) HbA1c level decreased from 12.1% (1.6%) to 8.8% (1.6%) over the 14 weeks in 240 paired samples (p<0.001). Ten percent of these participants achieved HbA1c levels <7%, and 34% had levels <8%. CONCLUSION : The nurse-led, telehealth-supported intervention successfully transitioned participants onto twice-daily mixed insulin or night-time intermediate NPH insulin, resulting in a significant decrease in HbA1c from 12.1% to 8.8%. However, clinics will require additional resources to initiate or intensify insulin therapy in primary care settings.