When it comes time to visit a doctor, it’s common to have many priorities. Maybe it’s getting relief for that aggravated pickleball injury, taming a lingering cough or finally having that weird mole checked out [you really should]. Oftentimes the risks associated with our medications are not high on the priority list, but they should be.
While any medication has risks, understanding and managing those risks is more important for some prescription medications than others. High-risk (or high-alert) medications (HRMs) are those that present extreme danger either due to patient characteristics (e.g., age, chronic disease, etc.) or misuse. As such, HRMs require prescribers and health systems to employ a number of practices and tools to evaluate and mitigate risk towards improving patient safety.
Given their prevalence, there is mounting interest about prescribing practices of HRMs and the implications for health care.
Recent Study
In a study published in JAMA Network Open, evaluators from Harvard University and Boston University compared HRM prescribing trends between traditional fee-for-service Medicare (TM) and Medicare Advantage (MA), which are privately managed plans for Medicare eligible individuals that are publicly funded through a capitated payment arrangement.
To complete their analysis, the authors compared over 13.7 million matched pairs of beneficiaries taken from samples spanning 2013-2018. The study relied on several sources to obtain data on the sample population, including the Medicare Master Beneficiary Summary file, Social Vulnerability Index, U.S. Office of Management and Budget, and the Medicare Part D Master Beneficiary Summary File.
For its primary measure, the study relies on the Healthcare Effectiveness Data and Information Set (HEDIS) and its Use of High-Risk Medications in Older Adults (DAE) metric. As a primary outcome, the authors considered the total number of HRMs that were prescribed to the qualified enrollees. As a secondary outcome, the authors looked at the proportion of older enrollees who had been prescribed at least 1 HRM per year. Other outcomes included the proportion of enrollees who had received 2 or more HRMs per year or the same HRM twice in the same year.
In addition to the primary variable of Medicare insurance type (i.e., enrollment in TM vs. MA), the study also examined a number of covariates. The researchers considered age, sex, race and ethnicity, dual-eligibility status, rurality, social vulnerability, eligibility for Medicare’s low-income subsidy, and a patient health indicator that factors the number of non-HRM medications.
The authors first used linear regressions to construct their primary model, and after accounting for covariates and other effects (fixed and random), they plotted the adjusted rate of unique HRM prescriptions. After the secondary outcomes were plotted similarly, sensitivity analyses were completed according to a range of criteria.
Ultimately, the study found that the rate of HRM use decreased in each year of the study period (2013-2018) – this was true for both TM and MA alike. Consistent with previously observed prescribing trends, HRM use in MA was significantly lower than in TM, but the gap between the two had narrowed. In the final year of the study period, the rate of HRM use in TM was still 56.9 HRMs (per 1000 beneficiaries) compared to 41.5 in MA. Similar patterns were observed in the analyses of the secondary outcome of the proportion of enrollees who had been prescribed at least 1 HRM per year. When compared with TM, MA performed better with a lower adjusted proportion of beneficiaries who had been prescribed at least 1 HRM (3.9%) versus 5.3% in TM. Relative to patient characteristics, the study observed higher rates of HRM use for certain population subgroups, including those who were female, American Indian or Alaska Native, or White.
Conclusion
The authors note several key limitations, including limiting analyses to only those medications identified by the DAE measure during the study period. The study was also unable to assess the extent that the HRM prescribing were clinically appropriate. The authors also explain that this work is limited by the use of MA as a single exposure and that only filled prescriptions were included in the analyses.
Despite these limitations, this study has implications for both medical practice and health care policy. As the study found that certain populations (female, American Indian or Alaska Native, and White individuals) received HRMs with greater frequency, there is a need to better understand how prescribers assess clinical presentation of these populations. The study’s findings also highlight how the mechanisms responsible for the overall decrease in HRM use in TM are not entirely known. The authors recommend that the Centers for Medicare & Medicaid Services explore additional avenues (e.g., tying HRM rates to reimbursement models) to narrow the gap between TM and MA relative to HRM rates.
Given their potential for harm, further research into HRM medication management strategies is an essential component of improving patient care and safety for older adults.
