RESEARCH DOSSIER
Opioid Use Disorder, Overdose Risk, and Continuity of Care
Abstract. Opioid use disorder, overdose risk, and continuity of care form a connected longitudinal research problem across reinforcement, drug exposure, treatment access, and health-system transitions. This critical integrative narrative monograph reviews historical developments in opioid dependence science, diagnostic boundaries, mechanisms of craving and withdrawal, overdose toxicology, mortality cohort evidence, randomized emergency department trials, naloxone implementation, and care-continuity research. The synthesis analyzes tolerance loss after detoxification or institutional transitions, post-overdose mortality, illicit-market change, treatment disparities, and the distinction between initiation and sustained treatment exposure. Comparative evidence consistently associates ongoing opioid agonist treatment with lower mortality, while controlled studies show context-specific effects of care initiation and mixed findings for peer-delivered navigation. The analysis distinguishes randomized intervention contrasts from observational mortality associations, identifies confounding and measurement limitations, and develops a longitudinal research agenda for transitions, equity, and patient-centered outcomes. No new empirical study, registered systematic review, or independently conducted meta-analysis is claimed.
Keywords: opioid use disorder; opioid overdose; overdose mortality; continuity of care; opioid dependence; craving; withdrawal; loss of tolerance; fentanyl; methadone; buprenorphine; naltrexone; naloxone; medication for opioid use disorder; post-overdose care; treatment retention; emergency department; prison reentry; harm reduction; social determinants of health
Scientific Object, Research Questions, and Evidentiary Method
Opioid use disorder (OUD), overdose risk, and continuity of care form a coupled psychological and public-health research object. OUD denotes a clinically consequential pattern of opioid use characterized by impaired control, craving, continued use in the face of harmful consequences, and functional impairment or distress. Overdose risk denotes the probability of an acute potentially fatal toxic exposure in a specified person, circumstance, drug environment, and observation interval. Continuity of care denotes the preservation of clinically meaningful access, therapeutic relationships, treatment exposure, harm-reduction resources, and follow-up across time and institutional transitions. A scientific account must explain how reinforcement and dependence interact with immediate toxicity, and how opportunities to receive and retain care alter these risks. These constructs overlap but cannot be measured by the same indicator. CDC, 2024 Koob & Volkow, 2016
The governing questions concern the historical development of these constructs; the neurobehavioral and social mechanisms maintaining recurrent opioid use; the distribution and changing causes of fatal overdose; the consequences of tolerance loss, polysubstance exposure, interrupted treatment, and transitions between settings; the causal evidence for sustained care; and the methods best suited to estimating population-level benefit. The core unit of analysis is a person moving through changing states of exposure and care. A hospital visit, a discharge, a prison release, or a missed prescription may be a discrete episode in administrative data while being experienced by the person as a decisive disruption of a continuous life and treatment trajectory. Consequently, an episodic outcome measure can mischaracterize a longitudinal process. Larochelle et al., 2018 Binswanger et al., 2007
This monograph is a critical integrative narrative synthesis of seminal clinical work, original randomized trials, longitudinal and record-linked cohorts, comparative-effectiveness studies, systematic reviews, surveillance reports, and official guidance. Evidence is grouped by design, outcome, target population, and inferential scope. Primary studies are identified alongside their original bibliographic records; current guidance is treated as guidance rather than experimental discovery. The synthesis makes no claim of registered systematic search, exhaustive study capture, new participant-level data, or independent meta-analysis. Relative associations in observational cohorts are interpreted with attention to confounding, exposure ascertainment, selection, and timing. Quantitative details are given only where the primary records or their openly accessible manuscripts support them. Sordo et al., 2017 Wakeman et al., 2020
Diagnostic Boundaries: Use, Dependence, Disorder, and Toxicity
Opioid exposure and opioid use disorder should be represented on separate axes. Exposure specifies what drug was consumed, by which route, in what context, and with what co-exposures. Disorder classification specifies a clinically evaluated pattern of impaired control, craving, hazardous use, interpersonal or role consequences, and persistence despite harm. Tolerance and withdrawal describe physiological adaptation and can occur in patients receiving supervised opioid therapy without establishing the broader behavioral syndrome. Conversely, an individual who receives an unexpected opioid-adulterated drug can experience overdose without a preceding OUD diagnosis. Clinical classification is thus informative about sustained risk and care needs but is never an exhaustive model of every acute exposure scenario. CDC, 2024
Diagnostic instruments obtain accounts of behavior, consequences, duration, and impairment. The evidence base combines interviews, health-record coding, toxicology and longitudinal observation, each with different error structures. A claim-based definition depends on access to services and coding decisions. A urine drug screen measures detectable exposure within the assay's window and may miss relevant compounds. Patient self-report can capture context and changes in motivation but also responds to stigma, legal concern, recall error, and therapeutic alliance. Construct validity improves when independent indicators are triangulated rather than when one indicator is treated as a decisive truth. Clinical assessment must also address co-occurring pain, mental illness, medications, and socioeconomic conditions without assuming that these explain the disorder causally.
Severity has multiple dimensions. Frequency of opioid use, craving, withdrawal intensity, social disruption, overdose vulnerability, and capacity to obtain treatment can diverge. A person stable in long-term medication care may have substantial historical disorder severity while having low current illicit-opioid exposure. Another person with intermittent use may be at severe overdose risk because of reduced tolerance and an unpredictable supply. For scientific comparisons, the estimand must state whether the result concerns prevalence of OUD, exposure to drugs, engagement with care, treatment discontinuation, nonfatal overdose, or mortality. Interpretation becomes especially important when datasets mix prescription-opioid and illicit-opioid populations under one coded diagnosis. Strang et al., 2003 Sordo et al., 2017
Historical Formation of Opioid Dependence Research
Historical explanations of opioid dependence moved from moralized accounts of recurring drug use toward medical, behavioral, neurobiological, and social research programs. Early clinical observations recorded withdrawal and compulsive patterns, but their diagnostic vocabularies and populations differed markedly from contemporary clinical classifications. An important milestone was Dole and Nyswander's 1965 study of methadone treatment in a group of 22 people previously dependent on heroin. Its central proposition was that sustained pharmacological stabilization could relieve craving and support social rehabilitation rather than making each treatment encounter a short isolated detoxification episode. The small uncontrolled historical trial established a new research and service model; it does not by itself estimate modern comparative effectiveness. Dole & Nyswander, 1965
Subsequent work broadened the question from physiological dependence to retention, infectious-disease exposure, criminal-legal outcomes, employment, and mortality. Marsch's 1998 meta-analysis integrated evidence about methadone maintenance and illicit opioid use, HIV risk behavior, and criminality. Later randomized trials and Cochrane syntheses compared buprenorphine maintenance with placebo and methadone, emphasizing that retention and opioid-use outcomes depend on comparator and treatment regimen. This historical trajectory changed how outcomes are interpreted: remaining in care, maintaining functioning, and reducing lethal exposure are scientifically relevant even when abstinence is incomplete. Marsch, 1998 Mattick et al., 2014
Research on fatal overdose further transformed the field. Strang and colleagues' 2003 follow-up of patients leaving inpatient detoxification documented deaths after treatment and raised the problem of lost tolerance. Binswanger and colleagues' 2007 population-based prison-release cohort identified markedly elevated mortality shortly after release. The incorporation of linked mortality records, drug-detection surveillance, and longitudinal treatment exposure allowed investigators to examine transitions rather than merely compare those who enter treatment with those who do not. More recently, changing illegal drug markets and widespread synthetic-opioid involvement required renewed attention to exposure dynamics and the limits of earlier heroin-era risk estimates. Strang et al., 2003 Binswanger et al., 2007 CDC NCHS, 2026
Reinforcement, Craving, Withdrawal, and Neurobehavioral Adaptation
Opioid reinforcement involves pharmacological actions on opioid receptor systems and their interaction with reward, learning, stress, and action-selection networks. Koob and Volkow's neurocircuitry analysis provides a general addiction framework: intoxication and incentive processes, withdrawal and negative affect, and preoccupation and anticipatory control interact over repeated exposure. The model is a cross-substance theoretical synthesis rather than a direct proof that all people with OUD follow identical trajectories. For opioids, acute relief of distress and withdrawal can become a strong immediate contingency. The same action may be pursued to obtain pleasure, stop discomfort, regulate fear of withdrawal, or preserve capacity to perform ordinary tasks. Separating these motives is necessary for psychologically credible research. Koob & Volkow, 2016
Learning accounts emphasize state-dependent valuation and reinforcement history. Environmental cues, anticipated relief, social networks, stressors, and access to drugs can acquire motivational significance. Craving should be operationalized as a time-varying self-reported and behavioral phenomenon, not inferred automatically from any positive drug test. Withdrawal can increase the immediate cost of abstaining and narrow the range of feasible choices. Behavioral experiments, experience sampling, cue-reactivity paradigms, and ecological longitudinal observation can test specific components of this account. Yet laboratory cue reactivity is not interchangeable with real-world overdose prediction; translation requires prospective validation and attention to baseline risk and access to treatment.
An integrative model links reinforcement to service organization through constraints. A missed appointment, transportation disruption, stigma-inducing encounter, or sudden medication shortage can convert a relatively stable treatment period into renewed withdrawal or exposure. Such events interact with biological adaptation but arise from institutional arrangements. This is a testable multilevel proposition: measured service discontinuities should predict subsequent changes in treatment exposure, reported withdrawal, and overdoses, conditional on baseline severity, local supply, and competing causes. The distinction between proximate withdrawal relief and long-term health priorities helps explain why care design matters in behavior change and why motivational narratives alone are insufficient. Wakeman et al., 2020 CDC, 2024
Overdose as an Acute Toxicological and Social Event
An opioid overdose is a toxicological event in which opioid exposure produces dangerous central nervous system and respiratory depression; it can result in hypoxic injury or death. Fatality depends on dose relative to tolerance, compound potency, route, co-exposures, availability of witnesses, timing of rescue, and medical response. Naloxone antagonizes opioid effects and is an emergency reversal intervention; access to it and ability to respond are therefore determinants of whether an overdose is fatal. The World Health Organization's community-management guidance emphasizes providing naloxone and response training to people likely to witness overdose. These recommendations locate emergency survival within a social network and public response system. WHO, 2014 FDA, 2023
Overdose risk is conditional and dynamic. Exposure to the same nominal amount of an opioid does not confer the same risk after prolonged treatment, following a period of nonuse, or amid simultaneous sedative exposure. Illicit-market uncertainty creates additional risk because purchasers may lack reliable information about identity or potency. Risk is shaped by whether someone is alone, housed, reachable by emergency services, familiar with overdose signs, and able to obtain rescue medication. Public education therefore requires an ecological framework: knowledge is useful only when resources, timing, and circumstances permit protective action. An overdose-response intervention is evaluated by actual access, distribution, witnessed response, survival, and longitudinal connection to care, rather than awareness surveys alone.
Nonfatal overdose constitutes both an adverse outcome and an important observation opportunity. Survivors may experience respiratory or neurological complications and remain at elevated risk of future events. A nonfatal overdose code can signal contact with emergency services while missing episodes that never receive formal treatment. Claims-based recurrence rates are consequently lower-bound indicators of some types of subsequent events rather than complete counts of exposure. Records should differentiate all-cause death from opioid-specific death, confirmed from suspected poisoning, fatal from nonfatal events, and incident from recurrent episodes. Researchers must explicitly address ascertainment and competing-risk problems before inferring whether a service model prevents death. Larochelle et al., 2018
Epidemiology, Surveillance, and the Changing Opioid Market
Population surveillance depends on case definitions, toxicology, death-certification practice, and reporting completeness. In the United States, provisional CDC National Center for Health Statistics figures released May 13, 2026 estimated 69,973 drug overdose deaths in 2025, compared with 81,313 in 2024; deaths involving opioids were estimated at 44,564 in 2025 compared with 55,296 in 2024. These are provisional national all-drug and opioid-involved counts, not OUD prevalence estimates. They should be presented with the reporting year and revision status, and not construed as evidence that individual risk has disappeared or that a specific policy caused the decline. Different countries and jurisdictions have distinct reporting systems and exposure mixtures. CDC NCHS, 2026
Survey estimates also require precise denominators. CDC's analysis of 2022 U.S. national survey data reported that only about one quarter of adults who needed OUD treatment received recommended medications. The denominator was a defined treatment-need population derived from a survey; it is not all people who ever used opioids, and it is not a global treatment coverage estimate. The survey identified disparities by race, ethnicity, gender, and age. These findings support examining barriers in service supply and acceptability, but cross-sectional differences alone cannot specify the causal contributions of each institution or patient preference. Dowell et al., 2024
Fentanyl-related poisonings and stimulant co-exposure complicate both research and messaging. Fitzgerald and colleagues' U.S. poison-center analysis documented increasing reported co-exposures involving fentanyl and cocaine or methamphetamine from 2015 through 2023; these data describe reported nonfatal exposures selected into poison-center records. They cannot yield population prevalence among all drug users or independently identify a mechanism for national mortality changes. Supply composition, treatment access, socioeconomic conditions, and emergency-response capacity vary over geography and time. The most productive epidemiological framework links evolving market exposure to a longitudinal care cascade while maintaining separate measures of disorder, treatment, and poisoning. Fitzgerald et al., 2025
Tolerance Loss and the Temporal Concentration of Overdose Risk
Loss of opioid tolerance is a time-dependent process with immediate implications for overdose vulnerability. Physiological adaptation during repeated exposure changes the effects of subsequent dosing. Following abstinence or markedly reduced exposure, tolerance may diminish, and a return to previously familiar quantities can produce unexpectedly severe toxicity. The empirical importance of this process is illustrated by Strang and colleagues' follow-up of people after inpatient opioid detoxification: the observed overdose deaths clustered among people classified as having completed detoxification and achieved loss of tolerance. This was a small observational cohort, useful for identifying a possible hazard mechanism rather than estimating a stable population-wide causal effect. Strang et al., 2003
Treatment transitions also create exposure and measurement discontinuities. A patient who stops maintenance medication, a person released from an institution where opioid access was limited, and a patient discharged after overdose each experiences a different combination of altered tolerance, available supply, psychosocial stress, and access to rescue. The causal question is whether avoidable interruptions in treatment increase overdose risk, and through which pathways. Estimating this requires repeated measures of actual medication exposure, competing drug use, confounding by clinical instability, and temporal proximity to disruption. Administrative indicators such as enrollment or a pharmacy fill may imperfectly represent ingestion and cannot alone confirm pharmacological protection. Sordo et al., 2017 Binswanger et al., 2007
The practical implication is longitudinal risk assessment. The same individual may occupy periods of relative protection and periods of acute vulnerability across a single year, even with an unchanged diagnostic label. Survival models should treat exposure status as time-varying and examine the first days and weeks after discontinuation separately from later periods. Researchers must avoid assuming that every post-treatment death was caused by tolerance loss: socioeconomic instability, mental illness, co-use, treatment selection, and changing drug composition may also contribute. A careful mechanism account describes conditional vulnerability and requires empirical tests that discriminate physiological change from institutional and environmental processes. Sordo et al., 2017
Mortality During and After Opioid Agonist Treatment
The most influential synthesis of mortality across opioid substitution treatment was Sordo and colleagues' 2017 systematic review and meta-analysis of cohort studies. Across methadone cohorts the authors reported pooled all-cause mortality rates of 11.3 deaths per 1,000 person-years during treatment and 36.1 outside treatment, with an unadjusted rate ratio of 3.20 for periods outside versus inside methadone care. The direction was consistent across studies, while heterogeneity in baseline risk, recruitment, and health systems was considerable. These results are compatible with a substantial protective association of continued treatment; they remain observations of treatment-exposure periods, not individually randomized assignments to continued and discontinued care. Sordo et al., 2017
Time-varying health status threatens simple comparisons. People may disengage from treatment during crisis, lose housing, change insurance, enter custody, or become severely ill, and the same conditions can increase mortality. Conversely, entry into treatment may follow a dangerous episode and temporarily select higher-risk patients. The observed effect therefore combines pharmacological exposure, service contact, baseline differences, and changing selection processes. Appropriate approaches include explicitly defined treatment regimes, risk-set alignment, inverse probability adjustment where assumptions are credible, sensitivity analyses, and outcomes measured outside the original treating institution. A stronger argument arises when observational associations align with biological plausibility, clinically relevant trials, and distinct datasets. Sordo et al., 2017 Wakeman et al., 2020
The policy-relevant outcome is continuity rather than mere program entry. Mortality studies suggest that treatment discontinuation is an analytically consequential event; the same work does not license a universal claim that any recorded treatment encounter prevents overdose or that all medications have equivalent mortality effects. Methadone and buprenorphine have stronger population-level survival evidence than treatments for which initiation and retention barriers are substantial or mortality events are sparse. The useful distinction is between the existence of an evidence-based intervention and a person remaining exposed to its protective mechanism under real-world conditions. That distinction directs attention to retention, transferability, follow-up, and rapid re-entry after interruptions. Sordo et al., 2017 Wakeman et al., 2020
The Post-Overdose Period as a Distinct Risk State
Survival following an overdose identifies a group at serious subsequent risk, but it also constitutes a point of contact at which intervention is possible. Larochelle and colleagues used linked Massachusetts data to follow adults who survived nonfatal opioid overdose and assessed subsequent mortality according to receipt of methadone, buprenorphine, or naltrexone treatment. Their study showed associations between medication exposure and lower mortality for methadone and buprenorphine, while noting the limited size of some medication-exposure groups and the possibility of residual confounding. This design improves on a simple cross-sectional comparison because it defines a clinically meaningful starting event and observes later outcomes. Larochelle et al., 2018
The post-overdose trajectory includes immediate emergency stabilization, assessment of co-occurring injuries and substance exposure, communication of overdose prevention information, access to naloxone, readiness to pursue or continue treatment, and subsequent service contact. The timing of these steps matters. A referral delivered before discharge is a process indicator; attending an appointment, receiving a medication, and remaining in contact over months are different outcomes. Overdose survivors often interact with multiple service systems, so apparent treatment absence may represent failures of linkage as well as preferences or structural barriers. A credible care cascade should distinguish a documented offer, its acceptance, actual access, sustained engagement, and mortality. Larochelle et al., 2018 D’Onofrio et al., 2015
The ethically important distinction between an acute rescue and recovery is also analytically important. Naloxone may restore respiration during an emergency without treating the persistent syndrome of opioid use disorder; sustained therapy can reduce recurrent exposure while still leaving a need for emergency rescue in unpredictable circumstances. Therefore, overdose reversal, continuing medication treatment, and psychosocial support should be assessed as complementary components with separate outcome measures. Research should include repeated nonfatal events, injuries, quality of life, and mortality, while accounting for events outside health care administrative systems. WHO, 2014 Wakeman et al., 2020
Incarceration, Reentry, and Institutional Discontinuity
Release from incarceration is one of the clearest examples of an institutionally structured risk transition. Binswanger and colleagues' population-based study of people released from Washington State prisons examined mortality after release and found a sharply elevated risk of death, particularly during the early period following release, relative to the general population. Drug overdose was an important cause. Differences between formerly incarcerated people and the general population preclude treating the comparison as a pure causal effect of release, yet the temporal pattern identifies a period of concentrated service needs. Binswanger et al., 2007
Post-release exposure depends on changes in tolerance, housing, social ties, legal obligations, employment, health insurance, and access to medications. Where treatment begun during custody is interrupted at the gate, a person may confront simultaneous withdrawal risk and uncertain drug supply. Continuity should therefore be modeled across custody and community services rather than as separate institutional programs. Relevant research indicators include verified medication provision before discharge, time to a community treatment encounter, transfer of medical records with consent, availability of naloxone, transportation, and deaths occurring shortly after release. The purpose of this framework is to identify which cross-system processes explain outcomes rather than attributing every adverse event to individual motivation. Binswanger et al., 2007 Sordo et al., 2017
Risk extends beyond prisons to residential detoxification, psychiatric admission, medical hospitalization, and discharge from treatment programs. Each setting has its own rules about medication, communication, consent, and follow-up. The scientific question is whether continuity-promoting policies increase sustained exposure to evidence-based care and reduce repeated emergencies. Randomized trials may be feasible for specific transition services, while observational studies require careful comparison groups and consistent definitions of treatment initiation and interruption. Structural differences across countries make jurisdiction-specific estimates essential, even when the core principle of continuity has broad applicability.
Fentanyl, Polysubstance Exposure, and Changing Risk Environments
Changes in the illicit opioid supply alter the relationship between behavioral history and acute risk. Illicitly manufactured fentanyl and related synthetic opioids may be encountered with uncertain potency, alone or with other substances. The pharmacological and toxicological environment must therefore be measured alongside frequency of use and clinical OUD severity. A person using stimulant products may experience opioid toxicity when fentanyl is present, which makes an exclusive focus on self-identified opioid users insufficient for public-health prevention. Drug checking, local toxicology, and overdose surveillance address different slices of this changing exposure problem. Fitzgerald et al., 2025
Fitzgerald and colleagues' analysis of U.S. poison-center reports studied nonfatal fentanyl-associated exposures involving cocaine or methamphetamine between 2015 and 2023. It found rising proportions of stimulant co-exposures within the selected fentanyl-related poison-center case set. The sample was drawn from poison-center reports among individuals aged 13 or older with specified intentional exposures and nonfatal effects; it did not enumerate all fentanyl users or provide the denominator for national overdose incidence. Such evidence can identify an emerging clinical pattern without directly estimating the causal effect of stimulant co-use on an individual's fatal overdose risk. Fitzgerald et al., 2025
Concurrent sedatives may exacerbate respiratory depression, while stimulant involvement complicates presentation and the interpretation of toxicology. The relevant prevention framework includes polysubstance assessment, accessible rescue, sustained OUD treatment where diagnosed, and population communication calibrated to current local supply. Longitudinal research should differentiate self-reported exposure, confirmed analytical detection, poisoning severity, and subsequent care. Treatments developed in earlier heroin-dominant settings require continuing evaluation as withdrawal experience, supply composition, and patient preferences change. WHO, 2014 D’Onofrio et al., 2026
Continuity of Care as a Longitudinal Scientific Construct
Continuity of care refers to the preservation of a clinically meaningful therapeutic process as people move across providers, settings, and periods of vulnerability. It contains at least three distinguishable dimensions: informational continuity, in which clinical knowledge follows the patient; relational continuity, in which trustworthy therapeutic contact is maintained; and management continuity, in which interventions and follow-up remain coherent despite institutional transitions. In OUD care these dimensions can diverge. A patient may have electronic records without a timely appointment, a trusted counselor without access to medication, or medication access without help managing housing and comorbid illness. Researchers must specify which dimensions are measured and why they are expected to affect mortality or functioning. Larochelle et al., 2018 Wakeman et al., 2020
The full cascade begins before a person enters care. Recognition of need, perceived safety of seeking care, an offer of evidence-based treatment, actual treatment initiation, continued exposure, return after interruption, and sustained life outcomes form a sequence with competing exits and reentries. Treatment attrition can be voluntary, imposed, caused by transfer, or caused by death; treating all attrition as equivalent masks mechanisms. Measures should distinguish an administrative lapse from a clinically verified stop and an unsuccessful follow-up from an unavailable service. Longitudinal data with patient-reported outcomes, pharmacy and service use records, and mortality linkage can better resolve these processes than single-visit encounter counts.
Continuity is also a service-system property. A hospital cannot ensure uninterrupted treatment merely by recording that a referral was given; an accepting clinic must have capacity, offer appropriate treatment, and communicate with the patient. From a measurement perspective, process completion should include verified appointment availability, contact, and meaningful access. The scientific framework should separate service-level fidelity, intermediate outcomes such as treatment days, and ultimate outcomes such as survival. This layered model makes it possible to identify which component of an intervention is effective and whether it is maintained in ordinary practice. D’Onofrio et al., 2015 Wakeman et al., 2020
Clinical Pathways, Medication Access, and Treatment Retention
Medication for opioid use disorder can alter withdrawal, craving, and exposure while supporting longer-term clinical stabilization. Methadone and buprenorphine provide opioid agonist-based treatment through different receptor pharmacology and delivery arrangements, while naltrexone is an antagonist that requires an opioid-free initiation period. These differences become relevant here because they shape access and continuity, rather than because this dossier duplicates the adjacent pharmacotherapy monograph. Clinical programs must evaluate whether eligible patients can begin a treatment they choose, sustain it through foreseeable disruptions, and re-enter when they disengage. A medication that is effective under maintained exposure cannot protect someone who is structurally unable to obtain it. Mattick et al., 2014 Lee et al., 2018
The X:BOT randomized trial compared extended-release naltrexone with buprenorphine–naloxone for relapse prevention among participants recruited after inpatient opioid treatment. Its intention-to-treat results were strongly affected by the greater difficulty initiating extended-release naltrexone, demonstrating why a treatment comparison must specify whether it begins at assignment or after successful induction. Analyses restricted to people who successfully begin treatment answer a different question and may break the protection of randomization when induction success depends on patient characteristics. This distinction is especially important for continuity: the barriers to starting care are part of real-world effectiveness, not an ignorable administrative detail. Lee et al., 2018
Wakeman and colleagues examined six treatment pathways in a large U.S. administrative claims cohort of 40,885 adults with recorded OUD. Only methadone or buprenorphine pathways were associated with reduced risk of overdose and serious opioid-related acute care use relative to no treatment during follow-up. The study's observational design, insurance-linked sampling and treatment categorization constrain causal interpretation, but the results demonstrate why a care system's success should be assessed by the treatments people actually receive and maintain. They also underscore the weakness of evaluating a service solely by whether it offers counseling or detoxification. Retention is an important intermediate outcome whose ultimate significance depends on patient-important benefits. Wakeman et al., 2020
Emergency Departments: From Stabilization to Initiation
Emergency departments occupy an important position in the OUD care system because they encounter patients during overdoses, withdrawal, injuries, and unrelated medical crises. A 2015 randomized clinical trial by D'Onofrio and colleagues compared referral alone, a brief intervention with referral, and emergency department–initiated buprenorphine with arranged follow-up. At 30 days, engagement in formal addiction treatment was higher in the buprenorphine-initiation group than in either comparison group. This design directly supports the causal effect of the combined initiation-and-follow-up strategy on the measured engagement outcome in its study setting, while leaving open the independent contribution of each intervention component and longer-term effects on mortality. D’Onofrio et al., 2015
Beginning treatment in the emergency department changes the temporal organization of care. Instead of relying only on a patient to locate medication after leaving the hospital, the intervention can bridge the interval between urgent presentation and outpatient services. Yet early treatment engagement should be measured separately from persistent retention, nonfatal overdose, and death. Studies should distinguish the effect of the medication initiation protocol, availability of follow-up appointments, patient preference, and hospital-to-community communication. The intervention may be effective in a supported trial infrastructure while requiring considerable system redesign in resource-constrained emergency departments. D’Onofrio et al., 2015
The ED INNOVATION randomized trial published by D'Onofrio and colleagues in 2026 compared a seven-day extended-release injectable buprenorphine strategy with sublingual buprenorphine among adults presenting with untreated OUD across 29 U.S. emergency departments. The trial's primary seven-day engagement rates were 40.5% and 38.5%, respectively, without a statistically significant between-group difference. The comparison was between two active treatment strategies with enhanced linkage support at study sites, so it does not contradict the earlier trial's comparison of initiation with referral-based pathways. The newer result instead shows that a more sustained formulation did not independently improve this specific short-term engagement measure under the studied conditions. D’Onofrio et al., 2026
Overdose Education, Naloxone, and Community Rescue
Community naloxone distribution recognizes that opioid overdose unfolds in settings where medical professionals may not be present when respiratory depression begins. The World Health Organization's 2014 guidance recommends that people likely to witness an opioid overdose should have access to naloxone and instruction in its use. This intervention depends on distributing rescue medication to people positioned to act and on enabling recognition, activation of emergency services, and appropriate follow-up. Naloxone reverses opioid-mediated effects acutely; its presence does not in itself resolve the chronic behavioral and clinical processes associated with OUD. The system therefore needs both immediate rescue capacity and continuing pathways into care. WHO, 2014
Walley and colleagues evaluated the implementation of overdose education and nasal naloxone distribution in Massachusetts using a geographically stratified interrupted time-series design. Communities with greater program implementation had lower observed rates of opioid overdose deaths than communities without implementation at the comparison level. The study provides important real-world implementation evidence, but communities were not individually randomized and multiple contemporaneous factors may influence mortality. Its value lies in connecting a plausible community mechanism—available rescue among witnesses—to population-level events while showing the importance of strong time-series methods. Walley et al., 2013
Emergency reversal infrastructure should be judged by functional reach. Distribution totals, awareness, and training attendance are intermediate indicators; the scientific goals involve timely recognition, actual ability to administer naloxone, survival, and later access to longitudinal services. Pharmaceutical access also changes over time: the U.S. Food and Drug Administration approved the first naloxone nasal spray for over-the-counter use in March 2023, expanding one legal pathway to public access in that jurisdiction. Availability in retail markets does not demonstrate affordability or geographic reach, and medication policy varies internationally. The relationship between overdose rescue and care continuity remains an empirical rather than purely regulatory question. FDA, 2023 WHO, 2014
Recovery, Patient Autonomy, and Long-Term Outcomes
Recovery is a multidimensional trajectory encompassing health, agency, social connection, stability, and personally meaningful functioning. Abstinence from illicit opioids is an important outcome for many people, but a complete longitudinal assessment also records survival, reduced overdose, continuity of medication, infectious disease prevention, social participation, housing stability, quality of life, and the person's own goals. A one-dimensional outcome can obscure progress or risk: repeated medication retention may be protective even when occasional exposure occurs, whereas apparent abstinence after discharge may coincide with increased risk if tolerance declines and emergency resources are absent. Sordo et al., 2017 Strang et al., 2003
Patient autonomy matters for both ethics and effectiveness because treatment preferences affect initiation, acceptance, and continued participation. A service may improve uptake when it offers accurate information, interprets withdrawal and craving experiences without stigma, and provides feasible choices among evidence-supported interventions. However, preference itself is shaped by prior care experiences, supply constraints, and historical restrictions. Treatment choice cannot be assumed equivalent to an unconstrained free decision; investigators must observe what options were actually accessible. Treatment outcome research therefore benefits from patient-reported decision quality and care experience alongside pharmacological measures and mortality linkage.
Long-term clinical evaluation should model re-entry as a normal feature of a recurrent condition. A patient who pauses treatment and later re-establishes care contributes information about resilience, barriers, and protective systems. Counting this pathway as a single permanent failure at the first exit would misdescribe the treatment process. Survival analyses with recurrent states, repeated measures and multiple follow-up windows can express this complexity. When outcomes differ—such as high satisfaction but unchanged measured overdose—a mature synthesis preserves both findings and examines the mechanisms by which they might eventually converge. Wakeman et al., 2020 Larochelle et al., 2018
Integrated Care Systems and the Transition From Episode to Trajectory
Integrated OUD care requires coordination across emergency medicine, primary care, specialized addiction services, mental health care, harm reduction, social services, and, where relevant, correctional institutions. The core scientific proposition is that the health effects of these services depend on how reliably people pass between them while retaining access to critical interventions. An emergency department may deliver successful overdose reversal but leave medication access unresolved; a treatment clinic may achieve initial retention but be unable to respond to hospitalization or sudden housing loss. An integrated model therefore defines outcomes and handoffs at the interfaces, including who is responsible for the next contact, whether consented information is transferred, and whether the chosen treatment remains immediately available. D’Onofrio et al., 2015 Larochelle et al., 2018
Care coordination needs measurable organizational mechanisms. A warm handoff is a facilitated connection in which the receiving service and patient establish contact rather than relying solely on a printed referral. Its success requires a destination with actual capacity, patient consent, an accessible time and place, and a plan for contingencies when the appointment fails. In an experiment, one might randomize the support arrangement while holding medication availability and comparator contact constant. In a natural experiment, change in clinic hours, prescribing regulations, or telehealth access may alter service reach; inference then requires examination of concurrent changes and plausible alternative explanations.
Retention should be interpreted at multiple levels. A service may maintain high appointment attendance because it restricts enrollment to people with stable housing, while another serving medically complex patients may have lower recorded adherence yet reach more people otherwise excluded. Comparative system performance thus needs risk adjustment, equity measures, patient-reported access, and longitudinal mortality alongside simple utilization counts. A scientifically informative outcome system evaluates access, continuity, and patient-important outcomes as linked but separable achievements. Wakeman et al., 2020 Dowell et al., 2024
Measurement Architecture for Overdose and Care Continuity
An adequate measurement architecture begins with clear denominators, timestamps, and sources. OUD prevalence may be derived from structured clinical interviews or survey criteria, while overdose mortality may be defined from certified cause-of-death records. Nonfatal overdose can be captured in emergency records, ambulance data, poison-center reports, or self-report, each sampling different events. Treatment initiation can be indicated by a medication administration record, a dispensing event, or a patient report, while retention may be measured as days covered, uninterrupted episodes, or repeated clinical contact. Comparing these measures without a shared definition can generate apparent contradictions where researchers are estimating different quantities. Dowell et al., 2024 Fitzgerald et al., 2025
Temporal ordering is essential for causal claims. Treatment exposure after an overdose cannot prevent that index overdose, but may influence subsequent events. A record of medication at any time during a follow-up period must not be treated as if it were present at baseline, because the time before a person receives treatment may be incorrectly credited as exposed survival time. This immortal-time bias can make care appear protective even without a true effect. Designs should align the start of follow-up with eligibility and exposure assignment, update treatment states where justified, and address deaths, incarceration, and loss to observation as informative events. Larochelle et al., 2018 Sordo et al., 2017
Measurement should also include dimensions missing from administrative records. Real-world access, trust, coercion, withdrawal burden, experienced stigma, treatment preference, patient-defined recovery, and perceived safety require validated self-report or qualitative inquiry. Cross-cultural research must test the meaning and performance of instruments rather than assuming that literal translation guarantees comparability. Qualitative methods can reveal how institutional procedures generate breaks in care; quantitative designs can test whether those processes predict subsequent treatment and health outcomes. Triangulation is most useful when each method's limitations are acknowledged and the same outcome is defined consistently.
Critical Appraisal of Randomized, Observational, and Population Evidence
Randomized assignment is particularly valuable when evaluating whether a specific initiation or support strategy changes care engagement. D'Onofrio's 2015 trial demonstrated an advantage for emergency department–initiated buprenorphine with follow-up compared with referral-focused strategies on its primary engagement outcome. The 2026 ED INNOVATION trial found no improvement in seven-day engagement from a seven-day extended-release formulation relative to sublingual buprenorphine when both were administered within active treatment and linkage protocols. The trials answer different questions; their results support a practice of comparing the precise intervention and comparator rather than transferring conclusions across trials on the basis of a shared drug name. D’Onofrio et al., 2015 D’Onofrio et al., 2026
Observational mortality datasets allow the study of rare serious outcomes over long periods and in wider populations. Their principal vulnerabilities include confounding by indication, treatment-seeking selection, unmeasured exposure, differential observation, and changes in care status. The consistent association of opioid agonist treatment with lower mortality across cohorts is scientifically meaningful, particularly when viewed with established pharmacology and related randomized evidence, but a pooled estimate of observed exposure periods should not be narrated as an individually randomized causal effect. Sordo's meta-analysis also displayed marked heterogeneity in baseline death rates, supporting context-specific interpretation of absolute risks. Sordo et al., 2017 Larochelle et al., 2018
Population-level interventions such as naloxone distribution create different identification problems. Interrupted time series can account for pre-intervention trends, but coincident shifts in drug markets, treatment services, surveillance, or emergency response can still threaten attribution. Randomized peer-navigation trials, by contrast, provide direct treatment-versus-comparator contrasts yet may be underpowered for mortality and may assess only short-term or service-use outcomes. A high-quality evidence synthesis must preserve these distinctions and avoid choosing a preferred design by prestige alone; the best available design depends on the estimand, event frequency, ethical feasibility, and context. Walley et al., 2013 Beaudoin et al., 2022 Doran et al., 2026
Comparative Interpretation and Unresolved Scientific Disputes
Several apparently conflicting findings become intelligible when the underlying scientific questions are separated. Agonist-treatment cohorts associate ongoing medication exposure with lower mortality, while peer-navigation trials may show little incremental effect against an active behavioral comparator. The first evidence concerns treatment exposure and long-term risk; the second concerns whether a specific method of connecting people to services improves measured outcomes. Both can be true because the effectiveness of a protective intervention does not guarantee that every strategy intended to increase its uptake will succeed. This distinction turns a broad assertion about access into testable claims about the causal chain from contact to initiation, persistence, and survival. Sordo et al., 2017 Beaudoin et al., 2022
Another dispute concerns how recovery should be measured. Abstinence-only definitions may omit benefits in overdose survival and sustained care, while measures focused solely on medication retention risk ignoring quality of life and self-directed goals. Scientifically defensible evaluation defines each outcome separately, treats some as mediators of others, and reports patient-important benefits and harms together. This framework also avoids interpreting positive toxicology as total treatment failure or interpreting a claimed period of abstinence as conclusive evidence of safety. Research on person-centered outcomes should retain comparable clinical endpoints alongside carefully validated broader measures. Wakeman et al., 2020 Strang et al., 2003
Claims about fentanyl-era treatment and overdose trends require ongoing review. Changing supply can alter respiratory toxicity, withdrawal experience, and intervention feasibility. National provisional mortality declines in 2025 are important surveillance observations, but they do not isolate the contribution of medication expansion, harm reduction, policing, drug-market shifts, or broader social factors. Well-specified longitudinal and quasi-experimental studies, including comparison regions and measured service exposure, are needed to attribute change. The strongest synthesis distinguishes an observed outcome, a mechanistic account, a causal hypothesis, and an intervention effect demonstrated by a credible study design. CDC NCHS, 2026 D’Onofrio et al., 2026
Future Research: Dynamic Risk, Care Equity, and Artificial-Era Tools
The most important next step is to study OUD care as a changing multistate process. Prospective data should represent transitions among no treatment, initiation, sustained medication exposure, interruption, re-entry, overdose, hospitalization, incarceration, and death. Person-time must be attributed correctly, and models should allow recurrent episodes and competing events. Multi-system linkage may improve outcome ascertainment but requires privacy protection, patient involvement, explicit consent where relevant, and careful validation across institutions. The central hypothesis is that reducing preventable discontinuity at high-risk transitions will improve sustained exposure to effective interventions and ultimately lower serious harms. Evidence from existing cohorts supports prioritizing this hypothesis while leaving the precise mediating pathways for direct testing. Sordo et al., 2017 Larochelle et al., 2018
Evaluation of care-equity interventions should distinguish intervention availability from actual uptake and from benefit conditional on uptake. Randomized or credible quasi-experimental studies could test low-threshold clinic schedules, transport support, remote follow-up, pharmacy-based access, and discharge navigation in populations systematically excluded from conventional services. The study protocol should preregister primary outcomes, measure resource costs and treatment satisfaction, preserve a patient choice framework, and follow individuals long enough to assess continuity beyond early contact. Designs must test effect heterogeneity by geography, age, and social constraints without essentializing groups. Observed access disparities justify targeted investigation rather than assuming one universal solution. Dowell et al., 2024 Beaudoin et al., 2022
In the Artificial Era, digital and artificial systems may support appointment continuity, evidence retrieval, risk communication, translation, coordination, and recognition of vulnerable transitions. Their evaluation must be empirical and operational. A predictive system should be judged by calibration across groups, temporal generalization, prospective clinical utility, privacy impact, workflow reliability, and whether actions following an alert actually improve outcomes. The model's prediction of an overdose is not itself prevention. Human oversight, access to evidence-based services, protection against punitive secondary use, and longitudinal randomized implementation evaluation are central requirements. These are research directions grounded in service-system mechanisms rather than claims that artificial tools currently produce proven mortality benefits.
Conclusion
Opioid use disorder, overdose risk, and continuity of care constitute a connected but multidimensional research problem. Reinforcement and withdrawal explain important features of recurrent opioid use, while overdose mortality depends on a dynamic exposure environment, tolerance, co-use, rescue opportunity, and access to sustained interventions. Historical studies established the value of treating opioid dependence longitudinally. Later cohort mortality analyses consistently associated ongoing methadone or buprenorphine treatment with lower risk, and randomized emergency-department trials demonstrated that treatment initiation and linkage arrangements can improve early engagement under defined comparison conditions. The strongest evidence concerns particular interventions and measured outcomes, not an undifferentiated promise of treatment or recovery. Dole & Nyswander, 1965 Sordo et al., 2017 D’Onofrio et al., 2015
The evidence also shows where research is incomplete. Observational survival comparisons contain selection and time-varying confounding; short behavioral trials may be too small or too narrowly focused to evaluate recurrent overdose and death; changing synthetic-opioid markets challenge assumptions carried forward from earlier drug environments. High-risk transitions after detoxification, incarceration, hospitalization, and overdose require explicitly designed longitudinal responses. Prevention systems need immediate naloxone access and reliable paths into continuing care. Effective treatment must be attainable, uninterrupted where desired, and recoverable after disruptions. Binswanger et al., 2007 Larochelle et al., 2018 Beaudoin et al., 2022
A coherent scientific model follows individuals through changing risk states and holds each explanation accountable to its evidence. Research should distinguish diagnosis from exposure, initiation from sustained care, survival from recovery quality, and observed association from causal effect. The resulting agenda integrates psychological mechanisms, pharmacological protection, public-health surveillance, service systems, and the agency of people receiving care. Its governing practical question is how institutions can reliably preserve access to interventions whose benefits are supported by evidence, especially at precisely the moments when risk rises. Wakeman et al., 2020 WHO, 2014
References
- Beaudoin, F. L., Jacka, B. P., Li, Y., Samuels, E. A., Hallowell, B. D., Peachey, A. M., Newman, R. A., Daly, M. M., Langdon, K. J., & Marshall, B. D. L. (2022). Effect of a peer-led behavioral intervention for emergency department patients at high risk of fatal opioid overdose: A randomized clinical trial. JAMA Network Open, 5(8), e2225582. https://doi.org/10.1001/jamanetworkopen.2022.25582
- Binswanger, I. A., Stern, M. F., Deyo, R. A., Heagerty, P. J., Cheadle, A., Elmore, J. G., & Koepsell, T. D. (2007). Release from prison—A high risk of death for former inmates. New England Journal of Medicine, 356(2), 157–165. https://doi.org/10.1056/NEJMsa064115
- Centers for Disease Control and Prevention, National Center for Health Statistics. (2026, May 13). U.S. overdose deaths decrease for third consecutive year in 2025 [Provisional surveillance release]. https://www.cdc.gov/nchs/pressroom/releases/20260513.html
- Centers for Disease Control and Prevention. (2024). Opioid use disorder: Treating. https://www.cdc.gov/overdose-prevention/hcp/clinical-care/opioid-use-disorder-treating.html
- D’Onofrio, G., Herring, A. A., Hawk, K. F., Perrone, J., Cowan, E., McCormack, R. P., Dziura, J., Matthews, A. G., Pantalon, M. V., Owens, P., Martel, S., Coupet, E., Jr., Lofwall, M. R., Walsh, S. L., Edelman, E. J., Carpenter, J. E., Strout, T. D., Baumann, M. R., Anderson, E.,. Fiellin, D. A. (2026). Emergency department–initiated buprenorphine for opioid use disorder: A randomized clinical trial. JAMA, 335(11), 948–960. https://doi.org/10.1001/jama.2025.27019
- D’Onofrio, G., O’Connor, P. G., Pantalon, M. V., Chawarski, M. C., Busch, S. H., Owens, P. H., Bernstein, S. L., & Fiellin, D. A. (2015). Emergency department–initiated buprenorphine/naloxone treatment for opioid dependence: A randomized clinical trial. JAMA, 313(16), 1636–1644. https://doi.org/10.1001/jama.2015.3474
- Dole, V. P., & Nyswander, M. (1965). A medical treatment for diacetylmorphine (heroin) addiction: A clinical trial with methadone hydrochloride. JAMA, 193(8), 646–650. https://doi.org/10.1001/jama.1965.03090080008002
- Doran, K. M., Welch, A. E., Kepler, K. L., Jeffers, A., Chambless, D., Cowan, E., Wittman, I., Regina, A., Siu, K., Bailey, V. S., Rostam-Abadi, Y., Kennedy, J., Kunins, H. V., Gwadz, M., Shelley, D., Cleland, C. M., & McNeely, J. (2026). Peer navigator intervention and opioid-related adverse events for emergency department patients: A randomized clinical trial. JAMA Network Open, 9(2), e2555903. https://doi.org/10.1001/jamanetworkopen.2025.55903
- Dowell, D., Brown, S., Gyawali, S., Hoenig, J., Ko, J., Mikosz, C., Ussery, E., Baldwin, G., Jones, C. M., Olsen, Y., Tomoyasu, N., Han, B., Compton, W. M., & Volkow, N. D. (2024). Treatment for opioid use disorder: Population estimates—United States, 2022. MMWR Morbidity and Mortality Weekly Report, 73(25), 567–574. https://doi.org/10.15585/mmwr.mm7325a1
- Fitzgerald, N. D., Black, J. C., Cottler, L. B., Martins, S. S., & Palamar, J. J. (2025). Trends in nonfatal fentanyl exposures involving stimulants in the United States, 2015–2023. American Journal of Preventive Medicine, 69(3), 107742. https://doi.org/10.1016/j.amepre.2025.107742
- Koob, G. F., & Volkow, N. D. (2016). Neurobiology of addiction: A neurocircuitry analysis. The Lancet Psychiatry, 3(8), 760–773. https://doi.org/10.1016/S2215-0366(16)00104-8
- Larochelle, M. R., Bernson, D., Land, T., Stopka, T. J., Wang, N., Xuan, Z., Bagley, S. M., Liebschutz, J. M., & Walley, A. Y. (2018). Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: A cohort study. Annals of Internal Medicine, 169(3), 137–145. https://doi.org/10.7326/M17-3107
- Lee, J. D., Nunes, E. V., Novo, P., Bachrach, K., Bailey, G. L., Bhatt, S., Farkas, S., Fishman, M., Gauthier, P., Hodgkins, C. C., King, J., Lindblad, R., Liu, D., Matthews, A. G., May, J., Peavy, K. M., Ross, S., Salazar, D., Schkolnik, P.,. Rotrosen, J. (2018). Comparative effectiveness of extended-release naltrexone versus buprenorphine–naloxone for opioid relapse prevention (X:BOT): A multicentre, open-label, randomised controlled trial. The Lancet, 391(10118), 309–318. https://doi.org/10.1016/S0140-6736(17)32812-X
- Marsch, L. A. (1998). The efficacy of methadone maintenance interventions in reducing illicit opiate use, HIV risk behavior and criminality: A meta-analysis. Addiction, 93(4), 515–532. https://doi.org/10.1046/j.1360-0443.1998.9345157.x
- Mattick, R. P., Breen, C., Kimber, J., & Davoli, M. (2014). Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database of Systematic Reviews, (2), CD002207. https://doi.org/10.1002/14651858.CD002207.pub4
- Sordo, L., Barrio, G., Bravo, M. J., Indave, B. I., Degenhardt, L., Wiessing, L., Ferri, M., & Pastor-Barriuso, R. (2017). Mortality risk during and after opioid substitution treatment: Systematic review and meta-analysis of cohort studies. BMJ, 357, j1550. https://doi.org/10.1136/bmj.j1550
- Strang, J., McCambridge, J., Best, D., Beswick, T., Bearn, J., Rees, S., & Gossop, M. (2003). Loss of tolerance and overdose mortality after inpatient opiate detoxification: Follow up study. BMJ, 326(7396), 959–960. https://doi.org/10.1136/bmj.326.7396.959
- U.S. Food and Drug Administration. (2023). Information about naloxone and nalmefene [Regulatory and patient information; later updates possible]. https://www.fda.gov/drugs/postmarket-drug-safety-information-patients-and-providers/information-about-naloxone-and-nalmefene
- Wakeman, S. E., Larochelle, M. R., Ameli, O., Chaisson, C. E., McPheeters, J. T., Crown, W. H., Azocar, F., & Sanghavi, D. M. (2020). Comparative effectiveness of different treatment pathways for opioid use disorder. JAMA Network Open, 3(2), e1920622. https://doi.org/10.1001/jamanetworkopen.2019.20622
- Walley, A. Y., Xuan, Z., Hackman, H. H., Quinn, E., Doe-Simkins, M., Sorensen-Alawad, A., Ruiz, S., & Ozonoff, A. (2013). Opioid overdose rates and implementation of overdose education and nasal naloxone distribution in Massachusetts: Interrupted time series analysis. BMJ, 346, f174. https://doi.org/10.1136/bmj.f174
- Watson, D. P., Tillson, M., Taylor, L., Xu, H., Ouyang, F., Beaudoin, F. L., O’Donnell, D., & McGuire, A. B. (2024). Results from the POINT pragmatic randomized trial: An emergency department-based peer support specialist intervention to increase opioid use disorder treatment linkage and reduce recurrent overdose. Substance Use & Addiction Journal, 45(3), 378–389. https://doi.org/10.1177/29767342231221054
- World Health Organization. (2014). Community management of opioid overdose. https://www.who.int/publications/i/item/9789241548816
Social Inequality, Stigma, and Treatment Opportunity
Care continuity is conditioned by social and institutional resources. Transportation, paid leave, childcare, insurance, housing, documentation, language accessibility, and geographic service availability can determine whether a person receives treatment after agreeing to it. These factors also correlate with background mortality risk, complicating causal inference. Dowell and colleagues' analysis of the 2022 U.S. National Survey on Drug Use and Health estimated that 25.1% of adults classified as needing OUD treatment received medications for the disorder. Treatment receipt varied by demographic groups, and the survey's definition of need included respondents already successfully treated who did not meet active symptom criteria during the year. This makes the numerator and denominator explicit and prevents misleading generalization. Dowell et al., 2024
Stigma changes both behavior and measurement. Fear of blame, legal consequences, family separation, or denial of services may deter disclosure and reduce opportunities to establish trust. At the same time, stigma is embedded in policies and clinical communication, so its effects are not reducible to an individual's beliefs. Research can operationalize anticipated stigma, experienced discrimination, internalized attitudes, clinician practices, and institutional eligibility separately. Studies of a care-navigation intervention need to show whether such determinants mediate actual treatment uptake and retention, preferably through validated measurements and designs that address differential dropout. Without this discipline, the word stigma becomes a broad explanatory label with few falsifiable implications.
Age, pregnancy, chronic pain, rural location, homelessness, and co-occurring psychiatric conditions generate distinct treatment constraints. Clinical trials frequently exclude or undersample important subgroups, and observational estimates may differ across insurance environments and health systems. The presence of an observed demographic disparity should motivate investigation of service delivery, economic conditions, historical exclusion, and preference-sensitive care rather than an assumption about intrinsic group characteristics. A scientifically adequate framework evaluates equitable opportunity to initiate, sustain, and re-enter treatment, alongside outcomes that matter to participants. Dowell et al., 2024 Wakeman et al., 2020