Potentially Inappropriate Prescribing in Long-Term Care and its Relationship With Probable Delirium

This study examined potentially inappropriate prescribing (PIP) of medication and its association with probable delirium among long-term care (LTC) residents in Ontario, Canada.

Population-based cross-sectional study using provincial health administrative data, including LTC assessment data via the Resident Assessment Instrument-Minimum Dataset version 2.0 (RAI-MDS 2.0).

LTC residents in Ontario between January 1, 2016, and December 31, 2019.

We used residents' first RAI-MDS 2.0 assessment in the study period as the index assessment. Probable delirium was identified via the delirium Clinical Assessment Protocol. Medication use in the 2 weeks preceding assessment was captured using medication claims data. PIP was measured using the STOPP/START criteria and 2015 Beers criteria, with residents classified as having 0, 1, 2, or 3+ instances of PIP. Relationships between PIP and probable delirium was assessed via bivariate and multivariable logistic regression models.

The study population included 171,190 LTC residents (mean age 84.5 years, 66.8% female, 62.9% with dementia). More than half (51.8%) of residents had 1+ instances of PIP and 21% had 3+ instances of PIP according to the STOPP/START criteria; PIP prevalence was slightly lower when assessed using Beers criteria (36.5% with 1+, 11.1% with 3+). Overall, 3.7% of residents had probable delirium. The prevalence of probable delirium increased as the number of instances of PIP increased, with residents with 3+ instances of STOPP/START PIP being 1.66 times more likely (95% CI 1.56-1.77) to have probable delirium compared to those with no instances of PIP. Similar findings were observed when PIP was measured using the Beers criteria. Central nervous system (CNS)-related PIP criteria showed a stronger association with probable delirium than non-CNS-related PIP criteria.

This population-based study highlighted that PIP was highly prevalent in long-term care residents and was associated with an increased prevalence of probable delirium.