Abstract
AIM: We aimed to determine whether glucose-lowering medication (GLM) prescribing differs by severe mental illness (SMI) status. MATERIALS AND METHODS: We conducted a population-based cohort study using routinely collected linked data, including people in Scotland diagnosed with type 2 diabetes between January 2004 and March 2022. We used Cox proportional hazards models to compare the time from diabetes diagnosis to the first prescription of each of metformin and insulin in people with and without a hospital admission record of SMI (schizophrenia, bipolar disorder or depression prior to diabetes diagnosis). We adjusted for age, sex, area-based socio-economic deprivation, smoking status and baseline HbA1c and estimated glomerular filtration rate. RESULTS: Among 317 761 people with type 2 diabetes, 14 600 (4.6%) had pre-existing SMI. During a median follow-up of 1.6 years, 10 859 (74%) of those with SMI and 219 823 (73%) without SMI were prescribed metformin. After accounting for confounders, people with SMI received their first metformin (HR 1.09, 95% CI: 1.07-1.11) and insulin (HR 1.24, 95% CI: 1.18-1.31) prescriptions at a faster rate than people without SMI. Body mass index (BMI) was missing in 21% of participants, but similar estimates were obtained when we additionally adjusted for BMI in analyses that only included people with BMI data. CONCLUSIONS: Among people with type 2 diabetes each of metformin and insulin is prescribed sooner in people with compared to those without SMI. Further research is needed to understand the reasons for this, and to investigate the implications of these differences in prescribing patterns.