This study identified rates and correlates of comorbid disorders among Medicaid/Medicare beneficiary patients from small urban and rural communities receiving methadone treatment for OUD. This analysis highlights a high comorbidity burden, especially for affective disorders. Our findings also reveal a consistent role for past-year non-medical opioid use to detect patients in greater need of integrated care for comorbid disorders. To date, few studies have examined comorbid affective and substance use disorders among this understudied population.

Comorbid disorders were common in this sample, especially for affective disorders, with rates of 48, 41, and 34%, for anxiety, depression, and PTSD, respectively. Consistent with predictions, rates were lower for comorbid substance use disorders, though 28% of patients still screened positive for stimulant use disorder. When analyzing differences by past-year non-medical opioid use status, considerably higher rates for all seven comorbid disorders were demonstrated among patients reporting use before and/or while in treatment. As illustration, within the sub-sample of patients reporting past-year use that occurred before treatment, or while in treatment, 62 and 57%, respectively, screened positive for anxiety, compared to 39% of patients reporting no past-year use. Similarly, 62% of patients reporting past-year use while in treatment screened positive for PTSD, compared to 39% reporting past-year before treatment, and 23% reporting no past-year use. The difference was starkest for stimulant use disorder, where 46% (past-year use while in treatment) and 45% (past-year use before treatment), respectively, screened positive, compared to only 12% of patients reporting no past-year use. For all seven disorders assessed, patients reporting no past-year use had lower rates than either group reporting past-year use, regardless of whether it occurred before or while in treatment. Consistent with hypotheses, these relationships remained even when accounting for other demographic and opioid use characteristics, as past-year non-medical opioid use was the variable most consistently associated with comorbid disorders in multivariable analyses. This relationship was strongest for PTSD, stimulant use disorder, and sedative use disorder.

One other opioid use characteristic, opioid use consequences, was associated with an increased likelihood of having comorbid disorders. In bivariate analyses, a greater number of consequences was associated with higher rates for anxiety, PTSD, stimulant use disorder, and sedative use disorder, though in multivariable analyses, significant associations only remained for PTSD and stimulant use disorder. This finding extends prior work demonstrating that a greater accumulation of opioid use consequences increases the likelihood of comorbid affective disorder symptoms among people who regularly use heroin in a large urban area in the same state as this study (34). Injection use preference was not related to comorbid disorder status, despite prior research demonstrating an association between injection use and comorbid affective and substance use disorders among participants in large urban areas (35, 36). Similarly, fentanyl use was not related to comorbid disorders. Our analysis may highlight that even though fentanyl penetration to the drug supply and comorbidity are key contributors to overdose risk among Medicaid beneficiaries (37), in this geographic setting, the relationship of fentanyl use and comorbidity may be better understood as an interactive (vs. probabilistic) relationship. With regard to demographic characteristics, few relationships were observed. Younger age, lower educational attainment, and female gender were associated with a greater likelihood of PTSD, stimulant use disorder, and anxiety, respectively. Comorbid disorder status was unrelated to race/ethnicity in this sample. Similarly, rates did not differ for patients from rural areas compared to those residing in the small urban area where the OTP is located, suggesting that comorbid disorders among rural and small urban patients may be more similar than different.

The findings from this study highlight a few clinical implications. First, we suggest treatment authorities require OTPs to use validated screening tools for comorbid disorders, going a step further than current federal guidelines (9, 10). Second, we recommend that capacity and planning for coordinated care, particularly for affective disorders, be built into existing intake procedures. These strategies might include providing co-located services for affective disorders (15), developing partnerships for mental health service provision that build in accommodations for people living in small urban and rural areas (38), or evaluating the efficacy of evidence-based approaches for affective disorders, such as the Unified Protocol or Acceptance and Commitment Therapy (39, 40), when implemented through telehealth or adapted as computer-delivered interventions. Regardless of the strategy to assess and coordinate care, approaches should emphasize feasible innovations that mitigate implementation barriers and present financially sustainable changes at OTPs to facilitate adoption in a complex treatment system. The use of mixed methods implementation science protocols, such as NIATx (41), adapted to OTP settings in small urban and rural settings, may be a promising strategy to identify a more comprehensive understanding of patient and provider experiences as a means of improving and sustaining innovations to existing care models. This may be especially important at treatment intake, given that nearly all patients (e.g., transfers excluded), initiate methadone treatment following recent and ongoing non-medical opioid use, a robust correlate of affective disorder comorbidity in this sample. Third, we recommend future research that examines whether routinely screening patients (e.g., during clinical sessions and/or through short message services) engaged in long-term treatment for recent non-medical opioid use is feasible and improves linkage to care for comorbid disorders. Findings may provide support for the one-item measure presented here as a more efficient and less invasive method than current protocols (e.g., urine drug screens). Last, polysubstance use of stimulants and opioids has been rising nationally, and is a key determinant in the fourth wave of the opioid overdose crisis (42). While this study didn't directly assess polysubstance use, the high rates of comorbid stimulant use disorder in this sample of OUD patients suggests polysubstance use is occurring for a sizable portion of patients, who urgently need integrated approaches that can reduce overdose-related harm.

This study has limitations. First, the sample was a convenience sample with a heterogeneous length of care for their current treatment episode. While this does introduce important differences, many studies examine recent non-medical opioid use without inquiring about treatment engagement. Furthermore, we sought to counteract this heterogeneity and aid clinical interpretation by providing comorbid disorder rates within sub-samples categorized by past-year non-medical opioid use status. Second, we did not assess all possible comorbid affective and substance use disorders, in part due to time constraints to gather info on disorders with low base rates (e.g., schizophrenia, hallucinogen use disorder). Similarly, we did not assess addictions not commonly addressed at OTPs (e.g., tobacco use disorder, gambling disorder), nor did we assess comorbid health conditions, such as infectious diseases, that overlap with OUD (43), and represent other important avenues where integrated approaches improve treatment outcomes (44). Future studies should investigate a full spectrum of OUD-related comorbidities, which ostensibly would highlight an even higher comorbidity burden and need for integrated approaches than this analysis. Third, our sample, while innovative in many ways (small urban and rural setting, public insurance homogeneity), was not powered to conduct an in-depth comparison of racial differences or population-specific comorbidity patterns (e.g., comorbid disorder rates among Black/African American patients), which may have provided valuable information about health disparities. Future research should gather data in small urban and/or rural settings where there is a greater representation of Medicaid/Medicare beneficiaries from diverse racial groups (e.g., Black/African Americans in the Deep South, Hispanic/Latinos in the American Southwest, and Native Americans in the Great Plains) who are receiving methadone treatment. Last, our comorbid affective disorder screening measures, while using established administration instructions and timelines, did not assess whether the patient would've screened positive at other time points in the past-year. As a result, our analysis may underestimate the rates of affective disorders compared to other studies (3, 45).

In conclusion, this study highlights a high rate of comorbid disorders, especially affective disorders, among publicly-insured methadone patients from small urban and rural areas. This burden is especially high for patients reporting recent non-medical opioid use, regardless of whether that use occurred before or during their current methadone treatment episode. Innovative and feasible approaches that assess patients for comorbid disorders and recent non-medical opioid use are needed to improve care coordination. We encourage local and federal treatment authorities, OTP directors, and methadone treatment researchers to consider our findings when developing screening, implementation, and coordinated care strategies.