Medicine Technology 🌱 Environment Space Energy Physics Engineering Social Science Earth Science Science
Science 2026-02-25 3 min read

Benzodiazepine Prescriptions Fell 28 Percent Over Four Years - But Dangerous Co-Prescribing Persists

Columbia University analysis of 104,000 adults finds benzodiazepine use dropped from 4.7 to 3.4 percent between 2018-2022, yet 41.6 percent of users still received other CNS depressants simultaneously

Benzodiazepines - a class of drugs that includes diazepam, alprazolam, and lorazepam - are effective short-term treatments for anxiety and insomnia but carry well-documented risks including dependence, withdrawal symptoms, falls, cognitive impairment, and overdose. For older adults, those risks are amplified. A study from Columbia University published in the Journal of Clinical Psychiatry tracks how prescribing patterns have shifted over four years and identifies a persistent safety concern that the overall decline in use has not resolved.

The Decline in Use

Researchers at the Columbia University Mailman School of Public Health and Columbia University Irving Medical Center analyzed data from 104,231 adults ages 18 and older who participated in the 2018-2022 Medical Expenditure Panel Survey, a nationally representative sample of the U.S. civilian, non-institutionalized population. Annual benzodiazepine use fell from 4.7 percent of adults in 2018 to 3.4 percent in 2022 - a reduction of roughly 28 percent over four years.

The decline was sharpest among adults 56 and older, the age group whose use had historically been highest and whose risk from these medications is greatest. Use in that group dropped from 7.0 percent to 4.7 percent. Adults ages 36-55 fell from 4.4 to 3.4 percent. The youngest group, ages 18-35, showed a smaller absolute change: 2.1 percent to 1.8 percent.

The American Geriatrics Society recommends avoiding benzodiazepines in older adults when possible, particularly in combination with other sedating medications. The decline in use among older adults is consistent with that guidance gaining traction in clinical practice, though the mechanism - whether due to physician behavior changes, patient preferences, formulary restrictions, or other factors - cannot be determined from this data.

The Co-Prescribing Problem

The more concerning finding is what is happening alongside benzodiazepine use. Among adults who received benzodiazepines in the study period, 41.6 percent also received at least one other central nervous system (CNS) depressant during the same year. CNS depressants include opioids, sleep medications, muscle relaxants, and other sedating drugs. Combining them with benzodiazepines can amplify respiratory depression and overdose risk.

The rate of co-prescribing varied by age and health status. Among adults 36-55, 44.6 percent of benzodiazepine users also received another CNS depressant; the figure was 43 percent for those 56 and older. Younger adults showed lower rates at 30 percent.

The co-prescribing rates among medically and psychiatrically vulnerable populations were substantially higher. Among benzodiazepine-treated adults in fair or poor health, 72 percent also received another CNS depressant. For those with serious psychological distress, the figure was 63 percent. These are precisely the patients most likely to be on multiple medications and most susceptible to adverse interactions.

"Although benzodiazepine use is declining in U.S. outpatient care, the high rate of co-prescribing with other CNS depressants underscores the need for careful monitoring," said lead author Mark Olfson, MD, MPH, professor of Epidemiology at Columbia Mailman School and professor of Psychiatry at Vagelos College of Physicians and Surgeons. "Before initiating a benzodiazepine, clinicians should review a patient's full medication list to assess potential drug-to-drug interactions - particularly for patients in poor health or experiencing serious psychological distress."

Context and Limitations

The study uses survey data, which has standard limitations: reliance on self-reported medication use, inability to capture actual daily dosing, and no information on the clinical reasons behind prescribing decisions. The data runs through 2022, so it does not capture more recent trends that may have emerged following subsequent prescribing guideline updates.

The historical context Olfson cites is worth noting: between 1990 and 2016, the percentage of U.S. office visits that included co-prescriptions of both benzodiazepines and opioids or other CNS depressants tripled. The current figures, while improved, reflect a prescribing culture that combined these drugs aggressively for decades.

Co-authors include Chandler McClellan and Samuel Zuvekas of the Agency for Healthcare Research and Quality, and Carlos Blanco of the National Institute on Drug Abuse.

Source: Columbia University Mailman School of Public Health and Columbia University Irving Medical Center. Published in the Journal of Clinical Psychiatry, February 25, 2026.