Association Between Dual Use of Department of Veterans Affairs and Medicare Part D Drug Benefits and Potentially Unsafe Prescribing

Veterans 65 years and older with prescription drug benefits from the Department of Veterans Affairs (VA) are almost universally eligible for Medicare Part D (hereinafter referred to as Part D). Although dual eligibility may increase access to necessary medications, receiving prescriptions from 2 systems (dual use) may also fragment care and undermine prescribing safety. Previous work showed that dual VA–Part D prescription drug use is a risk factor for potentially unsafe medication (PUM) exposure in veterans with dementia¹ and opioid users.^1,2 However, whether these risks generalize to the entire Medicare-covered population of older VA users remains unknown. We evaluated the association of dual prescription use through the VA and Part D (vs VA-only use) with the prevalence of PUM exposure in a national cohort of dually eligible older veterans.

We linked national VA and Part D records of use of health care services and prescriptions in a cohort of 279 940 veterans who were continuously enrolled in VA and Part D and received at least 1 medication through the VA in 2015. We further limited the study to veterans 68 years or older on January 1, 2015.¹ We categorized these veterans as dual users (ie, ≥1 medication from the VA and Part D) or VA-only users. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines. This study was approved by the institutional review board of the VA, which waived the need for informed consent.

Dates of administrative health care data use and analysis ranged from January 1, 2012, through December 31, 2015. We examined 4 previously validated PUM measures in 2015: any prescription for a Healthcare Effectiveness Data and Information Set high-risk medication in the elderly (PUM-HEDIS); any exposure to prescriptions with an anticholinergic cognitive burden score of at least 3 (PUM-ACB); any overlapping days of exposure to drug combinations with high risk for severe interactions (PUM-DDI); and a composite measure of any type of PUM exposure.³ We also examined an alternative definition of dual use: the proportion of total prescriptions (VA plus Part D) received from the VA in 2015. Several covariates differed at baseline. We used entropy balancing to achieve covariate balance.⁴ Unlike propensity weighting, entropy balance weighting directly adjusts the weights to sample means and variances, reducing the effect of model misspecification (Table).

Among 279 940 Medicare-eligible older veterans receiving a prescription from the VA, 18.9% (95% CI, 18.7%-19.0%) were dual users and 44.3% (95% CI, 37.3%-51.4%) were exposed to at least 1 PUM in 2015. Among dual users, 49.7% (95% CI, 49.2%-50.0%) were exposed to any PUM type, including 38.6% (95% CI, 38.2%-39.0%) to PUM-ACB, 19.3% (95% CI, 18.9%-19.6%) to PUM-HEDIS, and 4.4% (95% CI, 4.2%-4.6%) to PUM-DDI.

In adjusted results, dual use was associated with increased odds of any PUM exposure (adjusted odds ratio [aOR], 1.84; 95% CI, 1.80-1.88) and an additional 19.4 days of exposure (95% CI, 18.1-20.8 days) (Table). Dual use was also associated with increased odds of PUM-HEDIS (aOR, 1.82; 95% CI, 1.77-1.88), PUM-ACB (aOR, 1.53; 95% CI, 1.50-1.57), and PUM-DDI (aOR, 3.25; 95% CI, 3.02-3.48). Dual use measured as the proportion of total 2015 prescriptions from the VA revealed that PUM exposure was lowest among VA-only users, and PUM exposure peaked in veterans receiving prescriptions in near-equal proportions (50:50) from the VA and Part D (Figure).

Dual use of VA and Part D prescription drug benefits was associated with an almost 2-fold increase in the odds of exposure to any PUM compared with VA-only use and more than 3 times the odds of exposure to severe drug-drug interactions. Despite the limitations of observational studies, our results show that the prescribing safety risks associated with dual VA–Part D use are not limited to high-risk subgroups.^1,2 Policies that increase veterans’ access to non-VA health care professionals therefore may unintentionally jeopardize patient safety. Furthermore, it is possible that the safety risks found in this study of veterans may extend to all patients who receive prescriptions across disconnected health care providers or systems. To mitigate these potential risks, policies intended to expand access to non-VA providers must ensure patient information is shared and integrated into routine practice for all patients seeking care across multiple health care systems.^5,6

Accepted for Publication: May 23, 2019.

Corresponding Author: Joshua M. Thorpe, PhD, MPH, Division of Pharmaceutical Outcomes and Policy, UNC-Chapel Hill Eshelman School of Pharmacy, 301 Pharmacy Ln, Chapel Hill, NC 27599 (joshua.thorpe@unc.edu).

Published Online: July 22, 2019. doi:10.1001/jamainternmed.2019.2788

Author Contributions: Dr Thorpe and Mr Cashy had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: JM Thorpe, CT Thorpe, Carico, Van Houtven.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: JM Thorpe, Cashy.

Critical revision of the manuscript for important intellectual content: JM Thorpe, CT Thorpe, Schleiden, Carico, Gellad, Van Houtven.

Statistical analysis: JM Thorpe, Cashy, Carico.

Obtained funding: JM Thorpe, CT Thorpe.

Administrative, technical, or material support: JM Thorpe, CT Thorpe, Schleiden, Van Houtven.

Supervision: JM Thorpe.

Conflict of Interest Disclosures: Dr JM Thorpe reported receiving a research grant from the Department of Veterans Affairs (VA) during the conduct of the study. Dr CT Thorpe reported receiving grants from the VA during the conduct of the study and grants from the VA, the Donaghue Foundation, and the Vasculitis Foundation outside the submitted work. Dr Gellad reported receiving grants from the VA during the conduct of the study. Dr Van Houtven reported receiving grants from the VA Health Services Research and Development Service (HSR&D) and the National Institutes of Health during the conduct of the study. No other disclosures were reported.

Funding/Support: This study was supported by the Merit Award IIR 12-379 from the VA HSR&D. Dr Carico was supported by a postdoctoral fellowship from the VA Office of Academicc Affiliations.

Role of the Funder/Sponsor: The sponsor had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Joseph T. Hanlon, PharmD, MS, and Maria K. Mor, PhD, were compensated for intellectual contributions made to the original grant. Chester B. Good, MD, contributed to the intellectual content of the original grant but, as a VA physician, was not compensated.