OBJECTIVES

The Centers for Medicare and Medicaid Services Polypharmacy-CNS (Poly-CNS) quality metric identifies individuals taking multiple medications with central nervous system side effects. The Poly-CNS measure may reduce risky prescribing for persons living with dementia (PLWD).

DESIGN

Secondary analysis of the Optimize cluster randomized pragmatic trial of deprescribing education in primary care conducted between April 1, 2019, and March 31, 2020.

SETTING AND PARTICIPANTS

Two overlapping cohorts of PLWD taking 5+ chronic medications. The "full intervention" cohort was randomized at the clinic level to receive patient and clinician deprescribing education materials. The "clinician education cohort" included additional individuals whose clinicians received educational materials but who did not receive patient materials.

METHODS

Description of individuals meeting Poly-CNS criteria based on pharmacy dispensing records and their highest frequency combinations of medications. Generalized linear mixed models compared the odds of meeting Poly-CNS criteria by intervention or control group and modeled associations between patient characteristics and CNS polypharmacy at any time. Longitudinal analysis using generalized linear mixed models assessed trends of meeting Poly-CNS criteria over time.

RESULTS

The proportion meeting Poly-CNS criteria at any time during the observation period in the full intervention cohort (N = 2908) was 12.3% and 10.8% in the clinician education cohort (N = 5226). Adjusted P value for trend over time in the combined full intervention cohort was .0587, and <.0001 in the clinician education cohort. There was no effect of the intervention on odds of meeting Poly-CNS criteria in either cohort. Characteristics in both cohorts independently associated with meeting Poly-CNS criteria were younger age, greater morbidity burden, and female sex plus hospice use at baseline in the full intervention cohort. The most commonly combined medications were opioids plus antidepressants plus gabapentin.

CONCLUSIONS AND IMPLICATIONS

Individuals with cognitive limitations and multimorbidity pharmacologically treated for mood and/or pain symptoms may be at risk for CNS polypharmacy.