Roughly one in nine Americans aged 65 and older is exposed to a prescription drug combination flagged as a major drug-drug interaction, a rate that has barely budged over the past several years. Between 2015-2016 and 2021-2023, the share of older adults potentially at risk dropped from 12.3 percent to 10.8 percent, a difference so narrow it did not reach conventional statistical significance. The persistence of that risk, even as clinical guidelines have been updated and electronic prescribing alerts have become standard, raises hard questions about whether the health system is doing enough to protect its most medicated population.
Why the 10.8 percent rate has not fallen faster
The most recent nationally representative estimate comes from an analysis of NHANES prescription data collected between August 2021 and August 2023. That analysis found that 10.8 percent of older adults were potentially at risk for a major drug-drug interaction, compared with 12.3 percent in the 2015-2016 cycle. The p-value for that decline was approximately 0.051, just above the threshold typically used to declare a statistically meaningful change.
A plausible explanation for the modest improvement is that prescribers have shifted away from certain interacting cardiovascular drug pairs. Older adults with cardiovascular disease who take five or more medications face sharply elevated rates of severe potential interactions, according to research published in BMC Geriatrics. That study tied polypharmacy directly to severe potential drug-drug interactions in this high-risk group. If fewer patients are being started on legacy combinations such as warfarin paired with certain antiplatelet agents or statins combined with particular antifungals, the national average would edge downward even without broader deprescribing efforts or widespread adoption of updated clinical criteria.
The trouble is that a population-level dip of roughly 1.5 percentage points still leaves millions of older adults exposed. And the statistical borderline result means the decline could reflect sampling variability rather than genuine clinical progress. For a patient taking a dangerous combination, the distinction between 12.3 percent and 10.8 percent is academic. The interaction risk is binary: either the combination is in the medicine cabinet or it is not.
Beers Criteria updates and what NHANES data actually show
The American Geriatrics Society published its 2023 update to the widely used Beers Criteria for potentially inappropriate medication use in older adults. That update includes a dedicated table of clinically important drug-drug interactions to avoid, covering combinations such as multiple central nervous system–active drugs, anticholinergic pairings, and warfarin interactions with specific agents. The Beers list gives pharmacists and physicians a consensus-backed reference for flagging risky prescriptions, but no available data in the primary sources track how consistently prescribers actually follow these recommendations at the point of care.
The NHANES prescription medication data file for the 2021-2023 cycle, documented by the CDC’s National Center for Health Statistics, records what participants reported taking during the survey period. It captures medications in use but does not link them to clinical outcomes such as hospitalizations, bleeding events, or falls caused by the flagged interactions. A systematic review and meta-analysis of drug-drug interaction prevalence in older community-dwelling adults found that estimates vary significantly depending on which interaction database researchers use, whether Micromedex, Lexi-Interact, or the Beers criteria themselves. That methodological variability means the 10.8 percent figure is one defensible estimate among several possible ones, not a single settled truth.
Separate research on U.S. nursing home residents between 2018 and 2020 measured both the prevalence and the duration of potential drug-drug interactions in institutional settings. That work, published in the Journal of the American Geriatrics Society, found that exposure to interacting drug pairs persisted for months in some cases. Nursing home residents, who typically take more medications than community-dwelling adults and rely on facility-level medication review, represent a population where the consequences of missed interactions can be especially severe.
Gaps in evidence and what older adults should watch for
Several important questions remain unanswered. First, no linked clinical outcome data in the available research connect the flagged interactions to actual adverse events at a national scale. Knowing that 10.8 percent of older adults take a risky combination is different from knowing how many of them end up in emergency departments because of it. Without that link, policymakers cannot easily calculate the cost-benefit ratio of more aggressive intervention programs.
Second, the hypothesis that reduced cardiovascular drug pairing is driving the national decline cannot be confirmed or rejected with the data at hand. The NHANES trend analysis reports an overall prevalence shift but does not break down which specific drug classes account for the change or how prescribing patterns evolved over time. Without that level of detail, it is difficult to design targeted interventions that focus on the combinations doing the most harm.
Third, the national estimates say little about disparities. Older adults with multiple chronic conditions, limited access to primary care, or fragmented specialist involvement are more likely to be exposed to complex regimens. Yet the available survey data do not fully illuminate how interaction risk differs by race, income, geography, or insurance status. That leaves open the possibility that some groups bear a disproportionate share of the danger while others see more benefit from updated guidelines and safety technologies.
For individual patients and families, the absence of perfect data does not mean they are powerless. Experts generally advise that older adults keep an up-to-date list of all prescription medications, over-the-counter drugs, and supplements, and share it with every clinician involved in their care. Asking a pharmacist or physician to review the entire list at least once a year, and after every hospitalization, can help catch combinations that automated systems may have missed or that were started years ago under different standards.
Particular attention should go to drugs that affect blood clotting, blood pressure, blood sugar, or brain function. Combinations involving anticoagulants, antiplatelet agents, insulin or other glucose-lowering drugs, opioids, benzodiazepines, and certain antidepressants are frequently implicated in serious drug-drug interactions. Patients and caregivers can watch for warning signs such as unusual bruising, confusion, dizziness, extreme fatigue, or sudden changes in balance, and report them promptly rather than assuming they are just part of aging.
Where policy and practice could go next
The stubbornly high prevalence of potential major interactions suggests that current safeguards are not enough. Electronic prescribing systems routinely generate interaction alerts, but clinicians often experience “alert fatigue” and override messages they perceive as low value. Refining these systems to prioritize the most clinically consequential combinations, while suppressing minor or theoretical ones, could make the remaining warnings harder to ignore.
Medication review programs represent another underused tool. Structured pharmacist-led reviews-whether in primary care clinics, community pharmacies, or nursing homes-can identify opportunities to discontinue unnecessary drugs, substitute safer alternatives, or adjust doses to reduce interaction risk. Aligning reimbursement policies so that clinicians are paid for this time-intensive work would signal that deprescribing and interaction prevention are core elements of quality care, not optional extras.
On the research side, linking prescription data to hospitalization and claims records could provide the missing outcome evidence. If investigators can show, for example, that a specific set of interactions accounts for a measurable share of preventable emergency visits or inpatient stays, health systems would have a clearer target for intervention. That kind of evidence could also inform updates to criteria like Beers, ensuring that the combinations flagged as “major” are those with the strongest real-world impact.
Until such data are available, the 10.8 percent figure should be read as both a warning and a benchmark. It signals that, despite incremental improvements and evolving guidelines, a substantial minority of older Americans remain exposed to potentially dangerous drug combinations. It also provides a starting point against which future policy changes, technology upgrades, and clinical initiatives can be measured. The goal is not merely to shift the percentage by a fraction of a point, but to make sure that fewer older adults ever find out what a major drug-drug interaction feels like firsthand.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
