Medication Engagement, Determinants of Health, and A1C Levels Among Adults With Type 2 Diabetes within a Tribal Health System.

The findings come from 3,787 patients out of 10,506 patients with type 2 diabetes in the dataset, so many patients were excluded and the sample may not represent all patients. The study was correlational, so it cannot show that medication engagement caused lower A1C. Diet, exercise, and other factors that affect A1C were not available in the electronic health record. Medication engagement was based on pharmacy dispensing records and cannot confirm that patients actually took the medication. The data covered the height of the COVID-19 pandemic, which likely affected visits, medication engagement, and A1C testing. Medicaid status was used as a proxy for socioeconomic status and may miss patients with low income who are not enrolled. Patients who were dispensed only insulin and patients with end-stage renal disease were excluded, limiting generalizability. Telehealth use was minimal because of unreliable internet service. Hospitalization and survival data were not available, and some gaps in medication engagement may have been due to hospitalizations. Housing, food insecurity, and other social determinants were not measured.

The authors declared no conflicts of interest. The paper is listed as supported by NIH extramural research.

Do not read the lower A1C among engaged patients as proof that increasing medication pickup will lower A1C. This was a correlational electronic health record study, and diet, exercise, and whether medications were actually taken were not measured.