(210) Prevalence and Predictors for potentially Adverse Drug Event-related Hospitalisations

Background: Polypharmacy has become the rule rather than an exception in older people and is a key risk factor for potentially adverse drug event (ADE)-related hospitalisations (pADEH). As a prevention strategy, clinical practice guidelines typically recommend an annual comprehensive medication review for all people with polypharmacy, which has substantial work load implications but unproven value. An unresolved question is, how best to prioritise patients for medication safety interventions in primary care.

Objectives: To examine the prevalence and relevant predictors for emergency hospital admissions with 15 types of adverse events (AE)-related hospitalisations (AEH) and the subgroup of pADEH in two Scottish health boards in 2019.

Methods: We conducted a population based cross-sectional analysis of AEH and pADEH among all 450,582 people aged 40 years or older resident in the Tayside and Fife regions of Scotland. AEH was defined as an emergency admission of a patient with a specific AE based on relevant ICD-10 codes documented as “main condition” or based on laboratory data. pADEH was defined as the subgroup of AEH, where a drug known to cause such an AE had been dispensed in the 90 days prior to admission. We calculated the proportion of the source population with one or more AEH and pADEH in 2019. We performed binomial logistic regression in order to examine associations of different predictors present on 1/1/2019 (e.g., age, sex, number of drugs, previous hospital stays) on the risk of incident AEH and pADEH in the following 12 months.

Results: Our preliminary analyses found that among people aged 40 years or older, 2.35% had one or more AEH and 0.86% had one or more pADEH. The prevalence of the 15 different AEH in 2019 ranged from 21 per 100,000 residents to 962 per 100,000 residents, while the prevalence of corresponding pADEH ranged from 1 to 297. Fall (injuries), acute kidney injury and syncope were the most common AEH, while fall (injuries), bleeding outside the GIT and syncope were the most common pADEH. The predictors included in multivariate regression models explained 19.2% of variation in the risk of AEH and 22.5% in the risk of pADEH. AEH were most strongly associated with length of stay in hospital in the previous year (≥ 31 vs. 0 days; adj. OR 5.27; 95% CI 4.06 - 6.85), whereas pADEH were most strongly associated with number of drugs present on 1/1/2019 (≥ 15 vs. 0 drugs; adj. OR 13.33; 95% CI 10.00 - 17.76).

Conclusions: Our preliminary analyses show that pADEH are common and that predictors for AEH and pADEH differ. A limited number of predictors explained 22.5% of variation in the risk of all-cause pADEH. Strategies to improve estimation of pADEH risk and prioritization of patients for medication safety interventions may be achieved by considering specific pADEH rather than all-cause pADEH.