(209) Potentially drug-related emergency hospital admissions and their influence on patient's long-term medication. Analysis of administrative data from a Scottish Database.

Background: Potentially adverse drug event related hospitalisations (pADEH) are key targets for medication safety inititatives in primary care. We have developed a set of 36 pADEH indicators (expert consensus on combinations of hospital admission diagnoses and preceding drug exposure), which can be measured in secondary data sources (e.g. claims data).

Objectives: First, to examine the incidence/prevalence of pADEH. Second, to assess to which extent drugs potentially causing pADEH are discontinued after hospitalisation for 5 selected adverse events. Third, to define characteristics associated with an increased likelihood of drug discontinuation.

Methods: Retrospective population-based cohort study of all 517,256 patients aged ≥40 years resident in the Tayside or Fife regions of Scotland between 2010–2020. The data source was provided by the NHS Tayside/University of Dundee Health Informatics Centre. A logistic regression analysis was conducted in order to examine associations of patient characteristics (including age, sex, length of stay, type of event, and type and number of drugs taken prior to admission) with discontinuation of one or more potentially ADE-causative drugs after pADEH.

Results: Over the study period of 10 years, there were 598,316 emergency hospitalisations and 212,742 incident hospitalisations, of which 90,594 (15.1%) were pADEH. A total of 39,420 incident hospitalisations had an admission diagnosis with 5 selected events: fall injuries, delirium, acute kidney injury (AKI), gastrointestinal bleeding (GIB), non- gastrointestinal bleeding (non-GIB)), of which 30,260 (76.8%) were pADEH and 25,208 (83.3%) were affected by polypharmacy. The pADEH rates for cases with the respective admission diagnoses were 72.2% for fall injuries, 84.7% for delirium, 91.1% for AKI, 73.9% for GIB, and 77.3% for non-GIB. In 17,742 (58.6%) of these 30,260 pADEH cases, at least one potentially causative drug was discontinued. Patients were more likely to experience drug discontinuation if they stayed longer in the hospital, had more associated drugs prescribed before admission and if the drug was a potentially inappropriate medication according to the FORTA list.

Conclusions: A set of 36 indicators implemented in a longitudinal electronic data base found rates of pADEH similar to those observed in prospective studies. For five types of pADEH identified by the indicators, potentially causative drugs were discontinued in more than half of the cases. Further research is required to validate the pADEH indicator set.