Incidence of Potentially Inappropriate Prescribing: Longitudinal Investigation of Outpatient EHR Prescriptions.

IMPORTANCE

The incidence of potentially inappropriate medication (PIM) prescribing among older adults is not as well studied as its prevalence. Estimates of factors associated with PIM incidence, such as patient age, sex, race-ethnicity, medication subsidy support, and comorbidity, are also limited.

OBJECTIVE

To estimate the incidence of PIM prescribing in older adult outpatients, as well as the incidence and predictors for each PIM class, in a large outpatient electronic health records (EHR) cohort.

DESIGN

Retrospective study of PIM prescribing among outpatients with encounters leading to prescription orders, 2015-2018, excluding prevalent cases.

SETTING

Outpatients receiving care from a multi-site health system in western Pennsylvania.

PARTICIPANTS

342,405 patients, contributing 893,754 person-years of follow-up.

MAIN OUTCOMES AND MEASURES

The incidence of PIM prescribing based on automated coding of 2019 Beers criteria. A multivariable Poisson regression model was estimated to assess the impact of age, sex, race-ethnicity, comorbidity, and medication subsidy (PACE/PACENET) on PIM risk. For each PIM class, the association between predictors and time to PIM prescribing was evaluated using proportional hazard models.

RESULTS

The incidence rate (IR) for 1 or more PIM was 193.5 per 1000 person-years, led by short- and intermediate-acting benzodiazepines (37.6), first-generation antihistamines (32.8), and skeletal muscle relaxants (22.0). The incidence of PIM prescribing was 15% higher among white patients and 35% lower among males. High comorbidity (Charlson score ≥ 3) was associated with a 59% higher risk. Participation in the PACE/PACENET program, a medication subsidy program, was associated with an 83% increase in incidence. Each additional year of age was associated with a 1.2% reduction in incidence.

CONCLUSIONS AND RELEVANCE

This study establishes benchmarks for the incidence of PIM prescribing in outpatients and identifies important disparities in PIM risk, which vary by PIM class.