Assessing medication use patterns in patients hospitalised with COVID-19: a retrospective study

Abstract

Objective To describe patterns of medication use—that is, dexamethasone; remdesivir; and tocilizumab—in the management of patients hospitalised with COVID-19.Design and setting Retrospective observational study, using routinely collected, linked electronic data from clinical practice in Scotland. Data on drug exposure in secondary care has been obtained from the Hospital Electronic Prescribing and Medicines Administration System.Participants Patients being treated with the drugs of interest and hospitalised for COVID-19 between 1 March 2020 and 10 November 2021.Outcomes Identification of patients subject to the treatments of interest; summary of patients’ baseline characteristics; description of medication use patterns and treatment episodes. Analyses were descriptive in nature.Results Overall, 4063 patients matching the inclusion criteria were identified in Scotland, with a median (IQR) age of 64 years (52–76). Among all patients, 81.4% (n=3307) and 17.8% (n=725) were treated with one or two medicines, respectively; dexamethasone monotherapy accounted for the majority (n=3094, 76.2%) followed by dexamethasone in combination with tocilizumab (n=530, 13.0%). Treatment patterns were variable over time but roughly followed the waves of COVID-19 infections; however, the different drugs were used to varying degrees during the study period.The median (IQR) treatment duration differed by medicine: dexamethasone 5 days (2–9); remdesivir 5 days (2–5); and tocilizumab 1 day (1–1). The overall median (IQR) length of hospital stay among all patients included in the study cohort was 9 days (5–17); 24.7% of patients died in hospital.Conclusion The use of adjuvant medicines in patients hospitalised with COVID-19 appears in line with evolving evidence and changing treatment guidelines. In-hospital electronic prescribing systems are a valuable source of information, providing detailed patient-level data on in-hospital drug use.