Electronic assessment of cardiovascular potentially inappropriate medications in an administrative population...

To explore the feasibility of the electronic assessment of potentially inappropriate medication (PIM) criteria in a large administrative database and to explore the validity of the cardiovascular subset of PIM criteria, by studying the association with relevant outcome.

A cohort study using administrative data from Stockholm County, Sweden (VAL database). Eligible for inclusion were community-dwelling older people (≥65 years), alive in Stockholm County on 31 December 2015. We applied PIM criteria pertaining to the cardiovascular medication group (first-level ATC C group), and we assessed the association between PIM use and mortality and hospitalisation.

Patients' (n = 315 120) mean age was 74.0 years (range 65-114), and 54.7% were women. There were 111 cardiovascular PIM criteria in the repository, from which 44 were not registered or prescribed in our population. We excluded another 43 requiring information not available in the database, or duplicates, resulting in 24 applicable criteria. The prevalence of polypharmacy (≥ five medications) was 25.5% and the prevalence of at least one PIM use was 8.3%, including 2.8% underuse and 5.3% misuse. Patients with intake of ≥10 medications had 38% increased mortality risk compared to those with 0-4 medications. Unplanned hospitalisation and emergency department visits were positively associated with underuse (12% and 25%, respectively) and misuse (13% and 12%, respectively).

It was feasible to select a subset of cardiovascular PIM criteria originating from different PIM lists and to apply this subset in an administrative database. Additionally, by applying this subset, we showed significant associations with clinical outcome.