Exploring the association between medication, patient-specific factors and delirium in hospitalised patients: a 10-year retrospective, population-based cohort study in Iceland

Objectives This study aimed to identify potential significant statistical associations between the development of delirium, medication and several other patient-specific factors across a 10-year retrospective longitudinal real-world dataset. Design A retrospective, population-based cohort study. Setting The National University Hospital of Iceland. Participants All patients≥18 years hospitalised between 2010 and 2020. Primary and secondary outcome measures The primary outcome was a diagnosis of delirium within 2 years of a patient filling a prescription for a specific medication. The exposure was pre-admission medication use or a pre-existing condition. Secondary outcomes included mortality, length of hospital stay, readmission rate and frequency of additional delirium diagnosis. Results A total of 85 942 admissions and readmissions were included in the dataset, which comprised 1066 variables. The cohort comprised 55 495 patients (51.5% male) with a median IQR age of 71 years 58, 82. Throughout the study, 3533 patients were diagnosed with delirium at least once over the 10-year study period. A statistically significant association (χ 2 test, p <0.001) between the delirium (n=3533) and the non-delirium group (n=51 926) was identified for hyperpolypharmacy (≥10 medications), prescribed proton pump inhibitors, opioids, beta-lactam antibiotics, anticoagulants, statins and selective beta-blockers. Delirium was associated with an extended hospital stay, an increased readmission rate and increased mortality. No significant statistical association was found between delirium and the use of tricyclic antidepressants (p=0.121), nor with the presence of type I diabetes (p=0.216). On an annual average, the prevalence of patients who had obtained a delirium diagnosis within the previous year and were subsequently diagnosed again with delirium during their admission was 25.7%. Conclusion Numerous medications and other patient-specific risk factors have been identified, which underlines the complexity of delirium development. To strengthen patient safety and hinder the development of delirium, an increased emphasis is needed to review medication use and improve our understanding of medication risk associated with delirium cause, treatment and prevention. Specific prescribing/deprescribing advice should be developed to inform prescribing practice and patient safety.