The scale of the underlying problem is enormous. Hospitalizations related to alcohol use disorder and opioid use disorder contribute to substantial illness and rising costs, collectively amounting to nearly 50 billion dollars in healthcare spending in 2022 alone. Randomized trials have long supported the use of guideline-recommended medications for these conditions, which meaningfully reduce alcohol or opioid use and improve treatment retention among patients with opioid use disorder. Observational studies have also suggested that these medications could reduce acute healthcare utilization, such as emergency department visits and hospital admissions. Against that evidence base, the treatment rates reported in national surveys are strikingly low: in 2025, only about 3 percent of Americans aged 12 years and older with alcohol use disorder reported receiving medication treatment, and only 16 percent of those with opioid use disorder reported receiving medication in the past year.

Pham and colleagues addressed the relationship between medication use and healthcare utilization using a cross-sectional analysis of US adults from the 2022–2023 National Survey on Drug Use and Health, a nationally representative survey that assesses substance use disorders based on formal diagnostic criteria. Their weighted sample represented 57.5 million adults with past-year alcohol use disorder and 2.9 million adults with past-year opioid use disorder. Within that population, only 2 percent of those with alcohol use disorder and 30 percent of those with opioid use disorder reported using medication treatment within the past year. Perhaps the most consequential single number in the analysis is this: among patients who did not receive medication treatment, the mean number of past-year outpatient visits was 3.3. In other words, many untreated patients were already seeing doctors regularly, which suggests that lack of contact with the healthcare system is not the primary barrier keeping them from evidence-based pharmacotherapy.

The utilization findings themselves were nuanced and, at first glance, counterintuitive. In stratified analyses, healthcare utilization was generally more frequent among patients who received medications for alcohol use disorder, but no consistent association emerged for those who received medications for opioid use disorder. In unadjusted analyses of patients with alcohol use disorder, past-year medication use was associated with more emergency department visits, with an incidence rate ratio of 3.10 (95 percent confidence interval 1.98–4.83), more nights hospitalized, with an incidence rate ratio of 4.10 (95 percent confidence interval 2.17–7.75), and more outpatient visits, with an incidence rate ratio of 1.76 (95 percent confidence interval 1.18–2.62). When the analysts controlled for sociodemographic factors and comorbidities, the associations with emergency department visits and nights hospitalized remained statistically significant but attenuated, at 1.91 (95 percent confidence interval 1.19–3.06) and 2.03 (95 percent confidence interval 1.09–3.79) respectively, while the association with outpatient visits was no longer significant at 1.22 (95 percent confidence interval 0.91–1.63).

For patients with opioid use disorder, the picture was different. The only significant adjusted association between medication treatment and healthcare utilization was for nights hospitalized, with an incidence rate ratio of 2.08 (95 percent confidence interval 1.22–3.55). The divergence between the two conditions is one of the study’s most intriguing puzzles, and the cross-sectional design of the dataset precludes any definitive explanation. One possibility is that patients with alcohol use disorder are offered treatment later in the course of their disease, after alcohol-associated chronic medical conditions have already developed that predispose them to higher rates of healthcare utilization. Another possibility is that the findings reflect a genuine differential effect of medication treatment across the two conditions, given that medications for opioid use disorder are generally considered more effective than those for alcohol use disorder. It also remains unknown how utilization differs depending on which specific medication is initiated and whether psychosocial treatments are delivered concurrently across different medical settings.

The demographic profile of who received treatment adds an important layer of interpretation. Adults who received medications were generally older and more socioeconomically disadvantaged than those who did not: they were less often married, less educated, more often unemployed, and had lower incomes. They also carried more medical comorbidities and had greater severity of their alcohol or opioid use disorder. To some extent, this pattern is reassuring, because it indicates that medications are reaching the population with the highest degree of social vulnerability and clinical need—the patients who stand to benefit from them the most. On the other hand, it suggests that patients are often not receiving treatment earlier in their disease course, at a point when intervention could prevent the downstream social and medical consequences of untreated addiction before they accumulate.

That late start may also help explain another persistent problem: extremely poor persistence with these medications. Prior research has documented high rates of discontinuation of medications for alcohol use disorder and limited retention in buprenorphine treatment for opioid use disorder. The greater socioeconomic disadvantage and disease severity observed among treated patients may themselves act as barriers to staying on treatment, compounding the challenge of retention. This points toward structured interventions that address social vulnerability directly—such as care navigation programs that help patients overcome logistical, financial, and coordination obstacles—as potentially necessary complements to medication prescribing if treatment is to be sustained over time.

Several methodological limitations shape how these findings should be read. Because the study is cross-sectional, the timing of medication use relative to healthcare encounters cannot be determined: it is impossible to know whether healthcare encounters preceded or followed medication initiation, so no causal relationship can be ascribed to the observation that patients receiving medications for alcohol use disorder had higher utilization. The utilization outcomes were self-reported and therefore prone to recall bias, a common trade-off in addiction health services research. National surveys like the National Survey on Drug Use and Health rely on self-reported data, but that approach yields reliable population-level prevalence estimates based on formal diagnostic criteria. Administrative data, by contrast, are compromised by selective and non-standardized screening for alcohol and opioid use disorder in many health systems, which makes case ascertainment less reliable. Future studies could explore novel data linkages or health systems with more standardized screening and documentation procedures to close this gap.

A final interpretive caution concerns the meaning of healthcare utilization itself. Even acute utilization is context-dependent and does not inherently represent a negative outcome. For a patient with an acute health condition, presenting to the emergency department may be a desirable outcome rather than a failure, particularly compared with delaying care and risking disease progression. The survey data used in this study cannot identify which healthcare encounters were preventable or directly attributable to alcohol or opioid use disorder, so higher utilization among treated patients cannot be automatically framed as harm or as treatment failure.

What the study cannot do is establish the effect of these medications on healthcare utilization, and the editorial emphasizes that any interpretation must be situated within decades of randomized trial evidence demonstrating meaningful clinical benefit. What the study does provide is a stark reminder of underutilization—even among patients who are already engaged with the healthcare system and seeing clinicians multiple times a year. The path forward, the editorial argues, lies in high-quality hybrid effectiveness-implementation trials that test strategies to increase uptake of medications for alcohol and opioid use disorder across emergency departments, inpatient wards, and outpatient settings, while simultaneously assessing clinical outcomes including subsequent healthcare utilization. With nearly 50 billion dollars in annual spending tied to alcohol- and opioid-related hospitalizations and treatment rates stuck in the single digits for alcohol use disorder, the economic and human stakes of closing that gap could hardly be higher.

Subject of Research: The association between medications for alcohol and opioid use disorder and healthcare utilization among US adults

Article Title: What is the Relationship Between Addiction Medications and Healthcare Utilization?

Article References: Bernstein, E. Y. (2026). What is the Relationship Between Addiction Medications and Healthcare Utilization?. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10885-7

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10885-7

Keywords: alcohol use disorder, opioid use disorder, medication treatment, healthcare utilization, NSDUH, emergency department visits, hospitalization, treatment retention, addiction medicine, health services research, Journal of General Internal Medicine, treatment gap