Starting Two or More Drugs Concurrently in Primary Care: How Often Is It Done, How Often Is It Needed?

Abstract

Background: There is growing awareness of the need for more cautious, conservative prescribing. One conservative prescribing principle urges prescribers, whenever possible, to start only one new medication at a time. Little is known about how often primary care physicians (PCPs) start multiple medications at the same time, and when that is needed.

Objective: To describe how frequently PCPs start multiple prescriptions at the same time, evaluate evidence supporting the necessity of initiating multiple prescriptions concurrently, and describe PCP and clinical sites' prescribing variability.

Design: Retrospective cohort study.

Participants: PCPs at four sites who wrote prescriptions during January 2017-December 2018.

Main measures: Frequency of initiating two or more new prescriptions during the same session.

Key results: Across the four sites, 4646 PCPs wrote 7,849,914 new prescriptions. The Veterans Administration (VA) site had the highest percentage of encounters with multiple concurrent new drug starts (27.2%), followed by Northwestern (NW) (19.7%), Brigham and Women's Hospital (BWH) (16.1%), and University of Illinois Chicago (UIC) (14.0%). Within each site, there was wide variation among PCPs in percentage of encounters where they prescribed multiple new medications. Interquartile range varied: 11.0-18.5% (BWH), 15.1-22% (NW), 11.0-15.8% (UIC), and 22.9-31.0% (VA). Reviewing the most frequent combinations, only 0.6% had strong evidence for starting them concurrently. Most were drugs either recommended to be taken together (16.8%) or reasonable to be taken together, but with no evidence supporting starting them simultaneously (71.5%). A smaller percentage of concurrent starts were potentially problematic (10.4%) or contraindicated (0.7%) due to overlapping side effects or drug-drug interactions.

Conclusions: PCPs frequently started multiple medications concurrently, often without compelling evidence, with notable variations across prescribers and institutions. Although we could not conduct detailed chart review for each encounter, classification of the most frequent drug pairs concurrently prescribed in our study suggests opportunities to potentially improve prescribing safety.