Morphine milligram equivalents · overdose risk

Morphine Milligram Equivalents Are Not a Risk Assessment

A morphine milligram equivalent total is a screening flag, not a risk assessment. In a Medicare cohort, federal dose-based measures captured 29.29 percent of overdose episodes against 90.57 percent for models that weighed dose with other predictors, so the record needs more than the milligram.

A daily dose total is the easiest number on a pain chart to extract, which is why quality measures count it. It is also among the weakest predictors that chart holds.

A morphine milligram equivalent total is the easiest risk variable to pull from a pharmacy claim, which is why it anchors health-plan quality measures and point-of-sale edits. It is also a conversion with wide error bars, applied to patients whose metabolism changes what a milligram does. The chapter video A Dose Is Not a Diagnosis presents Chapter 17 of The Pained Brain by Dr. Gurpreet Singh Padda, MD, MBA, MHP, and Dr. KrisJay Fucanan, MD; the study-level evidence, with each paper’s limits, is in the Chapter 17 book companion. For the ordering physician, the practical question is what belongs in the record next to the dose.

Where did the 90 MME threshold come from?

The 2016 federal guideline told prescribers that any increase to 90 MME or beyond should be avoided or carefully justified, and it described its recommendations as voluntary. A quality-measurement body then specified high-dosage use as an average of 90 MME or more for at least 15 days, instructed that "a lower rate indicates better performance," and had moved its line down from 120 to 90 in 2019 for alignment. Medicare directed drug plans to install a care-coordination edit at 90. The agency’s 2024 guidance says those edits "should not be implemented as a prescribing limit or as a substitute for clinical judgment," while a 2026 memorandum reports high-dosage use monthly to plan sponsors, with several such measures displayed as Part D Star Rating measures. That is the structural driver: a plan’s public rating moves with the milligram total, and the incentive flows down to the practice. The practice position concedes that prescribing ran too high and that the milligram is a real hazard. The objection is to treating a conversion as a diagnosis.

How accurate is a morphine milligram equivalent conversion?

Conversion variance is large enough to move a patient across the line. Asked to convert identical doses, 319 clinicians produced a mean of 176 milligrams for a 75-microgram fentanyl patch, with a standard deviation of 117, and 193 milligrams for 40 milligrams of methadone, with a standard deviation of 201. The most recent systematic evaluation of conversion factors found 24 studies supporting factors across 29 mapped opioids and called for continued validation. The 2022 guideline took the specific number out of its recommendation statement, stated that no dosage threshold eliminates risk, and directed that buprenorphine not be counted in the daily total because of its ceiling effect on respiratory depression. A metric that cannot represent a partial agonist, and cannot be reproduced by two pharmacists reading one chart, is a screening flag. It is not a risk assessment.

What predicts opioid overdose better than dose?

The cleanest head-to-head comparison used 560,057 fee-for-service Medicare beneficiaries without cancer. The federal dose-based measures captured 29.29 percent of overdose episodes over twelve months; machine-learning models captured 90.57 percent. Total dose ranked among the top ten predictors, next to substance use disorder diagnoses, age, disability status and benzodiazepine fills, out of 268 candidates. The limit is real: positive predictive value ran 0.18 to 0.22 percent because overdose is rare. A simpler five-variable model validated at a c-statistic of 0.75.

Single variables carry larger effects than dose bands. In opioid-naive adults starting a first prescription, comorbid substance use disorder carried an adjusted hazard ratio of 2.74, and age 75 or older 3.22. Concurrent benzodiazepines carried a hazard of 5.05 in the first 90 days of overlap, falling to 1.87 on days 91 to 180, so a new co-prescription is a different exposure from a long-stable one. At the population level, high-dose prescribing alone accounted for an attributable fraction of 0.07 of fatal overdoses, while critical encounters such as a nonfatal overdose or release from incarceration accounted for 0.37.

Terrain variables a dose total cannot hold

The milligram enters a person whose physiology changes its effect. CYP2D6 poor metabolizers form 96 percent less morphine from codeine; ultrarapid metabolizers form 45 percent more. After hip and knee arthroplasty, the direction of an opioid receptor variant’s effect on morphine dose reversed with diabetes status, even though diabetes, genotype and their interaction explained only 2.7 percent of dosing variance. A perioperative review describes excess adiposity altering distribution, hepatic clearance and renal elimination, and notes that sleep apnea is common in the same population. Those are two biological drivers, hepatic enzyme capacity and an insulin-resistant, adipose terrain, sitting under the structural one above.

Measura [Cardiometabolic and Autonomic Health Analysis] does not offer pharmacogenetic genotyping or sleep studies; both are different tests done elsewhere. What it adds is the terrain that the chart otherwise infers:

• Metabolic status. Laboratory panels document the insulin and glucose picture. The patient described in the book had a fasting insulin of 31 that six years of dose-centered care never recorded.
• Body composition. Bioimpedance body composition measures fat and lean compartments rather than inferring them from weight.
• Cognition and balance in older patients. Age 75 or older is itself among the stronger overdose predictors above. A cognitive assessment and vestibular and balance testing give that patient a documented baseline, and pairing cognitive assessment with fall prevention keeps both in one visit.

When a dose change is already under discussion

None of this is a dosing recommendation, and the evidence on changes cuts both ways. What it shows consistently is that the manner matters. Monthly reductions above 30 percent carried an adjusted overdose hazard of 5.33 the following month, while reductions of 10 percent or less carried none. In a trial emulation of stable patients, eleven-month overdose or suicide risk was 0.96 percent with no change, 1.10 percent with tapering and 1.28 percent with abrupt discontinuation. After tapering, adherence to antihypertensive and diabetes medications fell to 0.60 and 0.69 of baseline, a measurable reason to hold a metabolic baseline before any plan changes. In high-risk veterans, a mandated interdisciplinary case review in place of a cut reduced discontinuation by 11.16 percentage points and all-cause mortality by 3.31. Measurement belongs in that review.

Documentation, standing orders and quality reporting

A high-dosage measure counts milligrams; nothing in it records whether anyone looked for the generator or the terrain. The chart can. A standing order can attach metabolic labs and body composition to patients on long-term therapy for chronic non-cancer pain, and cognitive and balance testing to those 75 and older, so the decision is reproducible across clinicians. Findings entered as structured fields in the chart sit next to the dose when a plan review or a pharmacy edit arrives. The same findings support the cardiometabolic documentation that MIPS and quality reporting already expects. Locating the generator is its own workup, covered in documentation before the diagnostic nerve block, and what the easiest first treatment takes from a patient over years is taken up in quality-adjusted life years in spine care.

Frequently asked questions

References

• National Committee for Quality Assurance (2026). Use of Opioids at High Dosage (HDO) – HEDIS measure page. NCQA HEDIS Measure Library. https://www.ncqa.org/report-cards/health-plans/state-of-health-care-quality-report/use-of-opioids-at-high-dosage-hdo/
• Centers for Medicare & Medicaid Services (2024). Frequently Asked Questions about Formulary-Level Opioid Point-of-Sale Safety Edits (July 5, 2024). CMS.gov (guidance document). https://www.cms.gov/files/document/frequently-asked-questions-about-formulary-level-opioid-point-sale-safety-edits-july-5-2024.pdf
• Dowell, D., Ragan, K. R., Jones, C. M., Baldwin, G. T., & Chou, R. (2022). CDC Clinical Practice Guideline for Prescribing Opioids for Pain – United States, 2022. MMWR. Recommendations and Reports, 71(3), 1–95. https://doi.org/10.15585/mmwr.rr7103a1
• Rennick, A., Atkinson, T., Cimino, N. M., Strassels, S. A., McPherson, M. L., & Fudin, J. (2016). Variability in Opioid Equivalence Calculations. Pain Medicine, 17(5), 892–898. https://doi.org/10.1111/pme.12920
• Lo-Ciganic, W.-H., Huang, J. L., Zhang, H. H., Weiss, J. C., Wu, Y., Kwoh, C. K., Donohue, J. M., Cochran, G., Gordon, A. J., Malone, D. C., Kuza, C. C., & Gellad, W. F. (2019). Evaluation of Machine-Learning Algorithms for Predicting Opioid Overdose Risk Among Medicare Beneficiaries With Opioid Prescriptions. JAMA Network Open, 2(3), e190968. https://doi.org/10.1001/jamanetworkopen.2019.0968
• Larochelle, M. R., Bernstein, R., Bernson, D., Land, T., Stopka, T. J., Rose, A. J., Bharel, M., Liebschutz, J. M., & Walley, A. Y. (2019). Touchpoints – Opportunities to predict and prevent opioid overdose: A cohort study. Drug and Alcohol Dependence, 204, 107537. https://doi.org/10.1016/j.drugalcdep.2019.06.039
• Hernandez, I., He, M., Brooks, M. M., & Zhang, Y. (2018). Exposure-Response Association Between Concurrent Opioid and Benzodiazepine Use and Risk of Opioid-Related Overdose in Medicare Part D Beneficiaries. JAMA Network Open, 1(2), e180919. https://doi.org/10.1001/jamanetworkopen.2018.0919
• Jurewicz, A., Gasiorowska, A., Leźnicka, K., Maciejewska-Skrendo, A., Pawlak, M., Machoy-Mokrzyńska, A., Bohatyrewicz, A., & Tarnowski, M. (2025). Do Diabetes and Genetic Polymorphisms in the OPRM1 and COMT Genes Modulate the Postoperative Opioid Demand and Pain Perception in Osteoarthritis Patients After Total Knee and Hip Arthroplasty? Journal of Clinical Medicine, 14(13), 4634. https://doi.org/10.3390/jcm14134634
• Glanz, J. M., Xu, S., Narwaney, K. J., McClure, D. L., Rinehart, D. J., Ford, M. A., Nguyen, A. P., & Binswanger, I. A. (2023). Association Between Opioid Dose Reduction Rates and Overdose Among Patients Prescribed Long-Term Opioid Therapy. Substance Abuse, 44(3), 209–219. https://doi.org/10.1177/08897077231186216
• Li, Y., Barr, K. D., Trafton, J. A., Oliva, E. M., Garrido, M. M., Frakt, A. B., & Strombotne, K. L. (2023). Impact of Mandated Case Review Policy on Opioid Discontinuation and Mortality Among High-Risk Long-Term Opioid Therapy Patients: The STORM Stepped-Wedge Cluster Randomized Controlled Trial. Subst Abus, 44(4), 292–300. https://doi.org/10.1177/08897077231198299

Related reading

• Delirium in Elderly Patients: Rule Out Pain Before Sedation
• Quality-Adjusted Life Years: Measuring What Spine Care Returns
• Laboratory Panels

Medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP, medical director of Measura. Last reviewed September 12, 2026.