CDC Opioid Guideline Reshaped Prescribing for Older Adults, but Cancer Patients Were Largely Spared

When the Centers for Disease Control and Prevention released its landmark opioid prescribing guideline in 2016, the aim was clear: rein in dangerous prescribing practices that had fueled an epidemic of overdose deaths across the United States. The guideline urged clinicians to prescribe the lowest effective dose, to avoid abrupt discontinuation, and to exercise extreme caution when combining opioids with other sedating medications. A decade of debate has followed about who those recommendations helped and who they may have inadvertently harmed. Now a large new analysis of older Americans offers one of the clearest pictures yet of how the policy ripple effects played out differently for two populations that clinicians have long worried about conflating: patients living with cancer and patients without a cancer diagnosis.

The study, published in Supportive Care in Cancer, drew on the linked SEER-Medicare database, a powerful resource that merges population-based cancer registry data from the Surveillance, Epidemiology, and End Results program with billing claims from the federal health insurance program that covers most Americans aged sixty-five and older. The research team, led by Olajumoke A. Olateju of the University of Houston’s PREMIER Center with collaborators at Penn State College of Medicine and Sungkyunkwan University, identified 42,931 Medicare beneficiaries diagnosed with breast, colon, or lung cancer and a comparison cohort of 108,218 patients without cancer. All participants were continuously enrolled in Medicare Parts A, B, and D, the components covering hospital care, physician services, and outpatient prescription drugs, and none were enrolled in private Medicare Advantage plans, which would have obscured their prescription records.

To isolate the effect of the guideline, the investigators divided the study window into two periods. The pre-policy period ran from January 2014 through November 2015, capturing prescribing behavior before the recommendations appeared, while the post-policy period spanned April 2016 through September 2017, after clinicians had time to absorb and react to the new guidance. The three-month gap between the periods deliberately excluded the transition months when the guideline was being publicized. For each patient, the researchers measured opioid use within three months of a cancer diagnosis or an index insurance claim, tracking four distinct outcomes: the monthly rate of opioid prescribing, the monthly number of days of opioid supply, the average daily dose expressed in morphine milligram equivalents, or MME, and the frequency of high-risk co-prescribing with benzodiazepines, antidepressants, or non-benzodiazepine sedatives.

The statistical engine behind the analysis was a difference-in-differences design, a technique borrowed from health economics that compares how each group’s outcomes changed over time and then subtracts the change in the comparison group from the change in the group of interest. Because both groups experienced the same broad cultural and regulatory shifts during the study years, this approach filters out secular trends and isolates the differential impact of the guideline on cancer patients relative to everyone else. The models adjusted for baseline characteristics such as age, sex, race, comorbidities, and prior medication use, strengthening the case that the observed gaps reflect genuine differences in how the guideline was applied rather than pre-existing disparities between the cohorts.

The headline finding is striking in its asymmetry. After the guideline took effect, opioid prescribing declined in both groups, but the decline was far steeper among patients without cancer, while changes among cancer patients were minimal. The difference-in-difference estimates quantified that divergence precisely: cancer patients experienced reductions that were 0.14 percentage points smaller per month in opioid prescribing rates than their non-cancer counterparts, a difference that was highly statistically significant, and their monthly days of opioid supply fell 3.07 days less than the comparison group. In practical terms, the prescribing pullback that swept through general pain medicine largely bypassed oncology practices caring for newly diagnosed older adults.

Not every measure moved the same way, however. The average daily dose in morphine milligram equivalents did not change significantly differently between the groups, with a point estimate of 16.73 MME and a confidence interval stretching from minus 43.36 to 76.82, a result the authors report with a p-value of 0.5853. That null finding suggests that when clinicians did prescribe opioids for cancer-related pain, the intensity of those prescriptions, measured in potency per day, remained essentially stable across the policy transition. The guideline’s influence, in other words, showed up in whether and how long opioids were dispensed rather than in how strong the daily doses were for those who received them.

The co-prescribing results carry perhaps the most important safety implications. Combining opioids with benzodiazepines, the sedative class that includes drugs such as alprazolam and lorazepam, is well documented to elevate overdose risk, because both drug types depress the central nervous system and can synergistically suppress breathing. The Food and Drug Administration has issued its strongest boxed warning about the combination, and Medicare administrators have flagged concurrent opioid-benzodiazepine use as a quality concern. High-risk co-prescribing declined overall in the study population after the guideline, consistent with a broader national push to reduce these combinations. Yet here again the reductions were smaller among cancer patients: opioid-benzodiazepine, opioid-antidepressant, and opioid-non-benzodiazepine sedative combinations all decreased less in the cancer cohort, with p-values below 0.0001 for each comparison.

Why would cancer patients be relatively insulated from the guideline’s effects? The most likely explanation is intentional. The 2016 guideline was explicitly framed as applying to chronic pain outside of cancer care, palliative care, and end-of-life settings, and oncology societies such as the American Society of Clinical Oncology and the National Comprehensive Cancer Network maintain separate pain management frameworks that recognize cancer pain as a distinct clinical entity. Pain is extraordinarily common in cancer, affecting a majority of patients across the disease trajectory, and undertreated cancer pain remains a persistent problem associated with suffering, functional decline, and stigma that patients themselves report as a barrier to adequate relief. Clinicians caring for these patients appear to have heeded the carve-out, continuing to prescribe when the clinical situation demanded it.

But the persistence of high-risk co-prescribing among older cancer patients is not automatically reassuring. Older adults are particularly vulnerable to the sedative burden of combined medications, with elevated risks of falls, fractures, cognitive impairment, and respiratory depression. Antidepressants and benzodiazepines are frequently prescribed in this population for legitimate reasons, including depression, anxiety, neuropathic pain, and sleep disturbance, all of which are common complications of cancer and its treatment. The finding that these combinations declined less among cancer patients could reflect appropriate clinical prioritization of symptom control, or it could signal a blind spot in medication safety review for a group whose many prescriptions are managed by multiple specialists. Disentangling those two possibilities is a task the claims data used in this study cannot fully accomplish.

The study arrives at a moment when the pendulum of opioid policy is swinging again. The CDC issued a revised clinical practice guideline in 2022 that softened some of the 2016 recommendations and emphasized that opioids remain appropriate for many patients with serious illness, partly in response to reports of patients with legitimate pain being cut off abruptly. This new evidence from SEER-Medicare suggests that, at least among older adults with common cancers, the feared collateral damage of the 2016 guideline on cancer pain treatment did not materialize in prescribing patterns, while the intended safety improvements did reach the broader population. The challenge ahead, the authors’ work implies, is to preserve that distinction while extending medication safety scrutiny, particularly around sedative co-prescribing, to the cancer population without reigniting the undertreatment fears that have shadowed cancer pain management for decades. As opioid stewardship programs mature across hospitals and oncology practices, data like these will be essential for calibrating policies that protect patients from both overdose risk and unrelieved pain.

Subject of Research: Changes in opioid prescribing and high-risk co-prescribing among older Medicare beneficiaries with and without cancer following the 2016 CDC opioid guideline

Article Title: Changes in opioid prescribing and high-risk co-prescribing after implementation of the 2016 CDC opioid guideline among older adults with and without cancer

Article References: Olateju, O. A., Byun, J.-Y., Okeke, C. M., Shen, C., & Thornton, J. D. (2026). Changes in opioid prescribing and high-risk co-prescribing after implementation of the 2016 CDC opioid guideline among older adults with and without cancer. Supportive Care in Cancer, 34(11), Article 1076. https://doi.org/10.1007/s00520-026-11273-x

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11273-x

Keywords: opioid prescribing, CDC opioid guideline, cancer pain, SEER-Medicare, benzodiazepines, co-prescribing, older adults, health policy, morphine milligram equivalents, supportive care, medication safety, difference-in-differences