CoChat
Sign Up
Cited by 511Opioid Use Disorder TreatmentSubstance Abuse Treatment and OutcomesHealthcare Policy and Management

The economic burden of opioid use disorder and fatal opioid overdose in the United States, 2017

Open access
Florence, Curtis, Luo, Feijun, Rice, Ketra·2021·Drug and Alcohol Dependence
DOI 10.1016/j.drugalcdep.2020.108350PMC PMC8091480
Open this paper in CoChat — free account

Abstract

Background The United States (U.S.) is experiencing an ongoing opioid crisis. Economic burden estimates that describe the impact of the crisis are needed when considering federal and state resources devoted to addressing overdoses. In this study, we estimate the societal costs for opioid use disorder and fatal overdose from all opioids in 2017. Methods We estimated costs of fatal overdose from all opioids and opioid use disorder based on the incidence of overdose deaths and the prevalence of past-year opioid use disorder for 2017. Incidence of fatal opioid overdose was obtained from the National Vital Statistics System; prevalence of past-year opioid use disorder was estimated from the National Survey of Drug Use and Health. Costs were estimated for health care, criminal justice and lost productivity. Costs for the reduced quality of life for opioid use disorder and life lost due to fatal opioid overdose were valued using U.S. Department of Health and Human Services guidelines for valuing reductions in morbidity and mortality. Results Costs for opioid use disorder and fatal opioid overdose in 2017 were estimated to be $1.02 trillion. The majority of the economic burden is due to reduced quality of life from opioid use disorder and the value of life lost due to fatal opioid overdose. Conclusions These estimates can assist decision makers in understanding the magnitude of opioid use disorder and fatal overdose. Knowing the magnitude and distribution of the economic burden can inform public policy, clinical practice, research, and prevention and response activities.

Integrity signals

Registry

Crossref · 10.1016/j.drugalcdep.2020.108350

Cites

33 of 33 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

11 of 22 cited sources checked · 7 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    The population requiring treatment for OUD is much larger than the current treatment capacity.

    “However, the population requiring treatment has been estimated to be much larger than the current treatment capacity (Jones et al., 2015).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Jones et al., 2015 source
      Experimental — accuracy not yet measured
  • Premisecitation

    Naloxone distribution programs reduce opioid mortality rates.

    “Naloxone distribution programs have been shown to reduce opioid mortality rates (Walley et al., 2013).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Walley et al., 2013 source
  • Interpretationreasoning

    This is a change from previously published estimates, where a large share of costs was attributable to OUD.

    “This is a change from previously published estimates, where a large share of costs was attributable to opioid use disorder (Birnbaum et al., 2011; Florence et al., 2016; Rice et al., 2014).”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Birnbaum et al., 2011 source
    • cites Florence et al., 2016 source
    • cites Rice et al., 2014 source
  • Premisecitation

    The HRQOL weights used for OUD were based on a survey of 1,178 participants from the US population.

    “These HRQOL weights are based on a survey of 1,178 participants from the US population to assess the reduction in quality of life based on whether or not a person with opioid use disorder was currently using opioids either by injection or prescription opioid misuse”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Wittenberg et al., 2016 source

      The Results report 1,178 respondents in the individual-focused survey, which assessed opioid-misuse health states. But the survey had 2,054 respondents overall, and the Methods (Study design) say responses from the 876-person spouse-spillover version also supplemented the individual utility data. Thus 1,178 is not the full survey sample underlying the weights. Page numbers are unavailable in our copy. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • Findingown data

    Using an alternative source of health care cost data had the smallest impact on the total estimate, changing it from $1,020.7 billion to $1,020.9 billion.

    “Using and alternative source of health care cost data has the smallest impact on the estimates (from $1,020.7 billion to $1,020.9 billion).”
    Quote in the paperNot a substantial claim
  • May combine two claims. 16
  • Classified as probably not a substantial claim. 15
  • The quote only partly states this claim. 1

  • May combine two claims.

    Premisecitation

    Mandatory provider review of prescription drug monitoring programs before prescribing opioids and laws governing pain clinics reduce the amount of opioids prescribed and prescription overdose death rates.

    “Mandatory provider review of prescription drug monitoring programs before prescribing opioids and laws that govern the ownership and operation of pain clinics have been shown to reduce the amount of opioids prescribed and prescription overdose death rates (Dowell et al., 2016).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Dowell et al., 2016 source

  • May combine two claims.

    Interpretationcitation

    The study valued lost health-related quality of life only for individuals with OUD; there is likely a substantial reduction in quality of life for friends and family members who care for loved ones with use disorder or lose loved ones to overdose.

    “We also only value lost HRQL for individuals with opioid use disorder. There is likely a substantial reduction in quality of life for friends and family members who care for loved with use disorder, and lose loved ones to overdose (Wittenberg et al., 2016).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Wittenberg et al., 2016 source

      Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The original source study used a matched case-control design to estimate the impact of OUD diagnoses on health care spending.

    “The original source study used a matched case-control design to estimate the impact of opioid use disorder diagnoses on health care spending.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Florence et al., 2016 source

  • May combine two claims.

    Premisecitation

    Previous research shows that both fatal opioid overdose and OUD are likely undercounted in the data sources used.

    “Finally, previous research has shown that both fatal opioid overdose (Ruhm, 2017) and opioid use disorder (Barocas et al., 2018) is likely undercounted in the data sources we used.”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Ruhm, 2017 source
    • cites Barocas et al., 2018 source

  • May combine two claims.

    Premisecitation

    OUD can be undercounted because the NSDUH surveys the noninstitutionalized population, which excludes incarcerated persons, and homeless persons are also likely missed by household-based surveys.

    “Opioid use disorder can be undercounted because the NSDUH surveys the noninstitutionalized population, which will exclude incarcerated persons. The homeless population will also likely be missed by household-based surveys.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Barocas et al., 2018 source

      The Discussion (page unavailable) says survey-based approaches can underreport OUD and reports prevalence estimates higher than national estimates. It does not say that NSDUH excludes incarcerated people or that household surveys likely miss homeless people; searches for those survey-population claims found no supporting passage. (LLM, conf=0.83)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Both incarcerated and homeless populations have been shown to have high rates of substance use disorder.

    “Both of these populations have been shown to have high rates of substance use disorder (Barocas et al., 2018).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Barocas et al., 2018 source

      The Methods describe identifying opioid use disorder in linked Massachusetts databases, and the Discussion reports prevalence by county and age group. Neither section reports substance use disorder rates specifically among incarcerated or homeless people; searches for “incarcerated” and “homeless” found no relevant passage. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductcitation

    In the source study, matching controlled for basic demographic variables, health status was matched based on the Charlson comorbidity index and prior health care costs, and regression models controlled for skewness in expenditure data.

    “Matching controlled for basic demographic variables, health status was matched based on the Charlson comorbidity index and health care costs prior to use disorder diagnosis, and regression models controlled for skewness in the expenditure data.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Florence et al., 2016 source

  • May combine two claims.

    Findingown data

    OUD accounted for $471.0 billion (95% PI $417.8-$525.7 billion) of the 2017 burden, slightly less than half of the total.

    “Slightly less than half of these costs were attributable to opioid use disorder ($471.0 billion (95% PI $417.8 to $525.7billion)).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The sum of health care, criminal justice, lost productivity, and HRQOL/VSL valuations captures the total societal economic burden of OUD and fatal opioid overdose.

  • May combine two claims.

    Findingown data

    The results were most sensitive to the choice of VSL and VSLY values; at the low HHS value the total burden was $556.0 billion, and at the high value it was over $1,480.4 billion.

    “Our results are most sensitive to the choice of values for the VSL and VSLY. At the low estimate of a VSL/VSLY from the HHS guidelines, the overall economic burden is $556.0 billion, and at the high value of the VSL/VSLY, the overall burden is over $1,480.4 billion.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The general-purpose VSL and VSLY values from HHS guidelines reflect the U.S. population's willingness to pay to prevent opioid overdose deaths and morbidity.

  • May combine two claims.

    Study design and conductown data

    Health care costs were estimated only for diagnosed cases of OUD in persons age 12 and older, so costs to other persons, such as those with neonatal abstinence syndrome, were not estimated.

    “These costs are also only estimated for diagnosed cases of opioid use disorder in persons age 12+, so we do not estimate costs to other persons, such as in the case of neonatal abstinence syndrome.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Criminal justice costs were apportioned from government budget data and may not reflect the costs of all criminal justice activity related to opioids.

    “Criminal justice costs were apportioned from government budget data and may not reflect the costs of all criminal justice activity related to opioids.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Medication-assisted treatment for OUD reduces opioid misuse among patients and the rate of fatal opioid overdose.

    “Medication-assisted treatment for opioid use disorder has been shown to reduce the rate of opioid misuse among patients with opioid use disorder and the rate of fatal opioid overdose (Mattick et al., 2009, 2014; Schwartz et al., 2013).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Mattick et al., 2009, 2014

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Schwartz et al., 2013 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Health care costs were estimated with a convenience sample of health insurers and public insurance beneficiaries and may not be representative of the U.S. population.

    “Health care costs were estimated with a convenience sample of health insurers and public insurance beneficiaries and may not be representative of the U.S. population.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    The study used probabilistic simulations with 10,000 replications to assess sensitivity and reported 95% prediction intervals.

    “We then conducted a simulation with 10,000 replications and report a 95% prediction interval [PI] that is at the 2.5% and 97.5% cutoffs of the distribution of estimates in the simulation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    A traditional confidence interval could not be directly estimated because the economic burden calculation used data from a variety of sources.

    “Calculation of the economic burden in this study uses data from a variety of sources, so it is not possible to directly estimate a traditional confidence interval for the estimates.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    This is a broader range of outcomes, and a more comprehensive estimate, than the most recent studies.

    “This is a broader range of outcomes, and a more comprehensive estimate, than the most recent studies (Davenport et al., 2019).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Davenport et al., 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    Many of the cost estimation methods used methods from a previous study, so the limitations of that study are also present in these results.

    “Many of the cost estimation methods used methods from a previous study, so the limitations of that study are also present in these results.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The resulting cost estimate from the source study is the increase in annual health care costs associated with OUD.

    “The resulting cost estimate is the increase in annual health care costs associated with opioid use disorder.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The case definition using underlying and multiple cause ICD-10 codes includes all opioid-related overdose deaths regardless of intent.

    “This includes all opioid-related overdose deaths regardless of intent (intentional, unintentional, homicide, or undetermined).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The HHS guidelines provide consensus estimates of a VSL discounted to present value at a rate of 3%.

    “However, the guidelines provide consensus estimates of a VSL, discounted to present value at a rate of 3%.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Office of the Assistant Secretary for Planning and Evaluation, 2016

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Survey weights were used to estimate nationally representative population totals for cases of OUD.

    “Survey weights were used to estimate nationally representative population totals for cases of opioid use disorder.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The NSDUH is a nationally representative sample of the U.S. civilian noninstitutionalized population ages 12 and older.

    “The NSDUH is a nationally representative sample of the U.S. civilian noninstitutionalized population ages 12 and older.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Center for Behavioral Health Statistics and Quality, 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    Despite limitations, the study's estimates provide an up-to-date understanding of the overall economic impact of the crisis.

    “Despite these limitations, these estimates provide an up-to-date understanding of the overall economic impact of the crisis.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductreasoning

    The alternative HRQOL instruments used in sensitivity analysis lack specificity to the condition and the populations studied were limited to subsets of the OUD population.

    “We use these alternative measures as a sensitivity analysis, and not as our primary results, because both instruments lacks specificity to the condition being studied, and the populations studied were limited to subsets of the population of persons with opioid use disorder (e.g. only those who use heroin).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    HHS guidelines recommend using measures of lost quality of life and the VSL specific to the outcome being analyzed when feasible, because preferences for reducing morbidity and mortality risk may differ by cause and population.

    “The guidelines recommend using measures of lost quality of life and the VSL that are specific to the outcome being analyzed, when feasible. The rationale for this is that the preferences of the U.S. population for reducing the risk of morbidity and mortality may differ by the cause and population affected.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Office of the Assistant Secretary for Planning and Evaluation, 2016

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims. The quote only partly states this claim.

    Study design and conductreasoning

    A primary limitation is that the study measures the cost of OUD in a single year; estimating overall economic costs would require observing cases from inception through life, but sufficient information to develop lifetime costs is not available.

    “A primary limitation of our methodology is that we only measure the cost of opioid use disorder in a single year.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The NSDUH questionnaire identifies OUD (opioid abuse or dependence) based on questions similar to the DSM-IV.

    “The survey collects detailed information on substance use, including a questionnaire that identifies opioid use disorder (opioid abuse or dependence) based on questions similar to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) (American Psychiatric Association, 1994).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites American Psychiatric Association, 1994

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

27 of 27

  • The paper relies on this without support.

    The general-purpose VSL and VSLY values from HHS guidelines reflect the U.S. population's willingness to pay to prevent opioid overdose deaths and morbidity.

    “We did not have a VSL value specific to the willingness to pay to prevent an opioid overdose death and instead used a general-purpose set of values based on results from studies that examined the value of risk reduction over a diverse set of risks.”

  • The paper relies on this without support.

    The sum of health care, criminal justice, lost productivity, and HRQOL/VSL valuations captures the total societal economic burden of OUD and fatal opioid overdose.

    “Our total valuation of economic burden of opioid use disorder and fatal overdose is therefore the sum of the HRQOL/VSL valuation, health care and substance use disorder treatment cost, criminal justice and lost productivity.”
2 of 2

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Cost per case of opioid use disorder was $221,219, derived by dividing the total U.S. cost of opioid use disorder ($470,975 million) during 2017 by the number of opioid use disorder cases the same year (2.129 million).

    “Cost per case of opioid use disorder ($221,219) was derived by dividing the total U.S. cost of opioid use disorder ($470,975 million) during 2017 by the number of opioid use disorder cases the same year (2.129 million) (3).”

    From State-Level Economic Costs of Opioid Use Disorder and Fatal Opioid Overdose — United States, 2017 (2021) · cites this paper as [3]

    The Results report approximately $471.0 billion in 2017 opioid use disorder costs and approximately 2.1 million people with the disorder, consistent with the claimed calculation in broad terms. Our copy does not give the more precise $470,975 million or 2.129 million figures, or state the resulting $221,219 cost per case. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Cost per case of fatal opioid overdose was $11.548 million, derived by dividing the total cost of fatal opioid overdose ($549,691 million) by the number of fatal opioid overdose cases (47,600).

    “Cost per case of fatal opioid overdose ($11.548 million) was derived by dividing the total cost of fatal opioid overdose ($549,691 million) by the number of fatal opioid overdose cases (47,600) (3).”

    From State-Level Economic Costs of Opioid Use Disorder and Fatal Opioid Overdose — United States, 2017 (2021) · cites this paper as [3]

  • The economic cost of the U.S. opioid epidemic in 2017 was estimated at $1,021 billion, including $471 billion for opioid use disorder and $550 billion for fatal opioid overdose.

    “The economic cost of the U.S. opioid epidemic in 2017 was estimated at $1,021 billion, including cost of opioid use disorder estimated at $471 billion and cost of fatal opioid overdose estimated at $550 billion (3).”

    From State-Level Economic Costs of Opioid Use Disorder and Fatal Opioid Overdose — United States, 2017 (2021) · cites this paper as [3]

Cites

Flagged first, then by citations

5 of the 38 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Related

Authors
Florence, C., Luo, F., & Rice, K.
Affiliations
Centers for Disease Control and Prevention (US)
Corresponding author
Florence, Curtis
ORCID
Venue
Drug and Alcohol Dependence
Publisher
Elsevier BV
Published
2021-01
Volume
218
Pages
108350
ISSN
0376-8716 (ISSN-L)
DOI
10.1016/j.drugalcdep.2020.108350
PMC
PMC8091480
OpenAlex
W3095492073
Cited by
511
References
38
Citations by year
2026: 69 · 2025: 122 · 2024: 103 · 2023: 110 · 2022: 68 · 2021: 36 · 2020: 2 · 2019: 1
FWCI
40.44
Type
journal article
Language
en
Open access
Licence
Locations
  • Drug and Alcohol Dependence (publishedVersion)
  • PubMed (publishedVersion)
  • PubMed Central (acceptedVersion)
Funders
National Institutes of Health (CC999999)
Topics
  • Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Opioid Use Disorder Treatment
  • Health Sciences › Medicine › Epidemiology › Substance Abuse Treatment and Outcomes
  • Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Healthcare Policy and Management
Keywords
opioid use disorder, economic burden, social costs, health care costs, criminal justice costs, productivity loss, quality of life, opioid epidemic, public policy
MeSH terms
Opioid Epidemic, Opiate Overdose, Analgesics, Opioid, Criminal Law, Delivery of Health Care, Humans, Opioid-Related Disorders, Quality of Life, United States, Prevalence, Cost of Illness, Drug Overdose

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-09-28.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-09-28.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review