Racial and ethnic disparities in potentially inappropriate medication use in patients with dementia

Abstract

Introduction: Racial and ethnic disparities in potentially inappropriate medication (PIM) use among older adults with dementia are unclear.

Methods: Data were drawn from the baseline visits of participants who were ≥60 years old and diagnosed with dementia in the National Alzheimer's Coordinating Center Uniform Data Set (NACCUDS) recruited from National Institute on Aging (NIA)-funded Alzheimer's Disease Research Centers (ADCs) throughout the United States. PIM utilization was evaluated using the 2019 American Geriatrics Society Beers Criteria for PIM Use in Older Adults. We estimated the association between race and ethnicity and the following outcomes and estimation models: (1) any PIM use, any PIM in each drug class, and any PIM best avoided in dementia patients using logistic regression models, (2) total number of medications, total number of PIMs, and anticholinergic burden scale (ACBS) using Poisson or negative binomial regression models, and (3) proportion of total medications that were PIMs using generalized linear models (GLM).

Results: Compared to White participants, Black, Hispanic, and Asian participants reported taking fewer total medications (incidence rate ratio [IRR] ± standard error[SE] = 0.903 ± 0.017, 0.875 ± 0.021, and 0.912 ± 0.041, respectively, all p < 0.01). Asian participants were less likely to be exposed to any PIM (odds ratio [OR] ± SE = 0.619 ± 0.118, p < 0.05). Compared to White participants, Black participants were less likely to be exposed to benzodiazepine (OR ± SE = 0.609 ± 0.094, p < 0.01) and antidepressant (OR ± SE = 0.416 ± 0.103, p < 0.001) PIMs, but greater antipsychotic (OR ± SE = 1.496 ± 0.204, p < 0.01), cardiovascular (OR ± SE = 2.193 ± 0.255, p < 0.001), and skeletal muscle relaxant (OR ± SE = 2.977 ± 0.860, p < 0.001) PIMs. Hispanic participants were exposed to greater skeletal muscle relaxant PIMs and had lower anticholinergic burden. Asian participants were exposed to fewer benzodiazepine PIMs.

Discussion: Significant racial and ethnic disparities in exposure to PIMs and PIMs by medication category in dementia research participants who have access to dementia experts found in the study suggest that disparities may be wider in the larger community.

Keywords: American Geriatrics Society Beers Criteria; dementia; potentially inappropriate medications (PIM); racial and ethnic disparities.

Figure 1. Multivariable Analysis: Differences in Potentially Inappropriate Medications (PIMs) by Race and Ethnicity, Adjusted Incidence Rate Ratio (IRR) estimated from count models or Odds Ratio (OR) estimated from logistic regression models, Compared to Non-Hispanic White Participants.

All models controlled for demographic characteristics (age, sex, education, level of independence), clinical characteristics (history of diabetes, cardiovascular diseases, cerebrovascular disease, and psychiatric disorders), Mini-Mental State Examination (MMSE), Functional Activities Questionnaire (FAQ), Neuropsychiatric Inventory-Questionnaire (NPI-Q), Geriatric Depression Scale (GDS), and Body Mass Index (BMI). ACBS: anticholinergic burden scale.

Figure 2. Multivariable Logistic Regression Analysis: Differences in Potentially Inappropriate Medications (PIMs) by Medication categories by Race and Ethnicity, Adjusted Odds Ratio (OR) Estimated from Logistic Regression Models, Compared to Non-Hispanic White Participants.

All models controlled for demographic characteristics (age, sex, education, level of independence), clinical characteristics (history of diabetes, cardiovascular diseases, cerebrovascular disease, and psychiatric disorders), Mini-Mental State Examination (MMSE), Functional Activities Questionnaire (FAQ), Neuropsychiatric Inventory-Questionnaire (NPI-Q), Geriatric Depression Scale (GDS), and Body Mass Index (BMI). PPI, proton pump inhibitors. Barbiturates were not examined in multivariable estimations because of low utilization rates.