Introduction Glycaemic control associates with better outcomes for hospitalised patients. Whether GLP‐1 receptor agonists (GLP‐1 RA) are suitable and effective drugs for inpatients is unclear. Methods A retrospective, single centre, observational study using data from the electronic health record. Patients admitted using GLP‐1 RA as outpatients, from 2016 to 2019, were identified. Outcomes were compared to those admitted using twice‐daily (BD) mixed insulin. Capillary glucose, medication use, creatinine, and demographic data were collected. As drugs may be discontinued/not administered in hospital, days when GLP‐1 RA was administered were ‘GLP‐1 RA active’ and, for insulin, ‘insulin active’. The primary comparison was rate of hypoglycaemia (<4 mmol/L) and severe hypoglycaemia (<3 mmol/L). A logistic regression model examined variables for hypoglycaemia. Results GLP‐1 RA comprised n = 262 admissions and BD insulin n = 166. The ‘insulin active’ cohort (n = 957 patient days) had higher risk of hypoglycaemia than ‘GLP‐1 RA active’ (n = 806 days); occurring on 14.7% of days; 95% confidence interval [CI] 12.6–17.1 versus 9.9% days; 95% CI 8.0–12.2; p = 0.002, and severe hypoglycaemia 4.0% of days (95% CI 2.8–5.4) versus 2.0% (95% CI 1.1%–3.2%; p = 0.005). Daily glucose (mean ± standard deviation) was 10.8 ± 5.2 mmol/L in insulin active versus 9.6 ± 4.7 mmol/L in GLP‐1 RA active; p < 0.001. Insulin use, age, and acute admissions predicted hypoglycaemia. The odds ratio for hypoglycaemia was 2.15 times greater (95% CI, 1.14–4.08; p = 0.019) with insulin than with GLP‐1 RA. Conclusions GLP‐1 RA provided better glycaemic control than BD mixed insulin and should be continued during hospitalisation unless there is a clear indication for cessation.