Hepatitis C Treatment in Kentucky Medicaid Recipients with Concurrent Opioid Use Disorder: A Cross-Sectional Study

Abstract

Background

Hepatitis C virus (HCV) infections and injection drug use have concurrently increased in the last decade. Evidence supports simultaneously treating chronic HCV and opioid use disorder (OUD) with medication. Kentucky is a hard-hit state for both conditions that has undertaken policy and practice efforts to increase access to both types of medications.

Objective

To examine receipt of direct-acting antivirals (DAAs) for patients living with HCV-OUD and received any vs. no medications for opioid use disorder (MOUD).

Design

We conducted a cross-sectional study using a proprietary dataset from HealthVerity of health claims between 1/1/2020 and 12/31/2021.

Patients

Kentucky Medicaid beneficiaries aged ≥ 18 with concurrent chronic HCV-OUD diagnoses.

Main Measures

Multivariable logistic regression models were used to calculate adjusted proportions of HCV DAA receipt based on receipt of MOUD, adjusting for patient characteristics and region.

Key Results

Of 2149 patients, 36% (n = 780) received HCV DAAs; 84% (n = 1804) received any MOUD during the study period. Buprenorphine was the most common MOUD type used (n = 1414, 66%). Adjusting for covariates, HCV DAA receipt was lower among people who received any vs. no MOUD (33% vs. 46%, p < 0.0001). Methadone (vs. no MOUD, 29% vs. 46%, p = 0.0002) had the greatest difference in odds of HCV DAA receipt.

Conclusions

Gaps in HCV treatment among Kentucky Medicaid recipients with OUD were pervasive. Despite evidence supporting HCV-OUD co-treatment, patients receiving MOUD were significantly less likely to receive curative HCV treatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11606-025-09356-2.

INTRODUCTION

The hepatitis C (HCV)-opioid use disorder (OUD) syndemic has worsened in the last decade because of a shift to injection drug use among people who use opioids.^1–5 HCV is a blood-transmissible disease, meaning people who inject drugs are at higher risk of acquiring HCV. In the majority of cases, acute HCV develops into a chronic infection, leading to liver cirrhosis and liver cancer. From 2014 to 2022, the number of new HCV infections in the USA more than doubled,² over half of which were attributed to injection drug use.³

Medications to treat HCV and OUD are each highly effective,^6–8 and there is growing evidence that co-treatment has several clinical and public health benefits.^9–14 New innovations in HCV direct-acting antiretroviral (DAAs) can now cure chronic HCV infections in 95% of cases.^6,7 Providing medications for opioid use disorder (MOUD; buprenorphine, methadone, naltrexone) are highly effective at reducing drug use, preventing future overdose, and reducing HIV and HCV incidence.⁸ Co-treatment with MOUD and DAAs results in greater rates of medication retention for both DAAs and MOUD.^9,11,15,16 For instance, a study by Rosenthal and colleagues showed that over 80% of patients who completed integrated HCV and OUD treatment achieved HCV cure and longer MOUD retention.^9,11 Similarly, models of synergistic co-treatment and harm reduction practices (i.e., syringe service programs offering sterile syringes) show the greatest public health and clinical gains in preventing acute HCV infections and overdoses.^15,16

Given these benefits, there are clinical evidence, renewed attention, and new policy opportunities to co-treat HCV and MOUD. Multiple studies show that there are no contraindications to treating HCV in people receiving MOUD or people who inject drugs, but lack of knowledge, misinformation, and stigma may persist among providers.^17,18 Co-treatment is also a focus of state and federal Hepatitis Elimination Plans^19–21 and is called out explicitly in research and practice priorities^14,22 to reduce barriers to HCV treatment, with particular focus on people who inject drugs and people insured by Medicaid.^23,24 There have also been several policy efforts to lower barriers to HCV DAAs and MOUD for Medicaid beneficiaries. Some states have removed Medicaid restrictions on HCV DAAs including prior authorizations, limiting HCV DAA prescribers to specialists, and barriers to retreatment for people who use drugs. ^24–27 At the same time, policy barriers to MOUD were substantially lowered during the COVID-19 pandemic, including allowances for initiating buprenorphine by telemedicine,²⁸ flexibilities in take-home methadone doses,²⁹ and Medicaid coverage for methadone received in outpatient settings.²⁸

Kentucky has been hit hard by the HCV-OUD syndemic, making it an important case study for concurrent OUD-HCV medication access through Medicaid.^30–33 In 2018, Kentucky removed Medicaid requirements that patients remain abstinent from alcohol and drugs and/or undergo drug testing before starting HCV treatment, also referred to as “sobriety requirements.”²⁷ However, some barriers remain in place, such as sobriety requirements for retreatment and prior authorizations for DAAs.²⁷ Additionally, effective August 2020, the Kentucky Department of Medicaid Services lowered the threshold for DAA access by allowing primary care providers to prescribe DAAs without consulting a specialist or requiring additional liver imaging or testing.³¹

While there is growing evidence in support of HCV-OUD co-treatment, little is known about the current landscape of concurrent HCV DAA and MOUD uptake among Medicaid members. To fill this gap, the purpose of this study was to explore HCV-OUD medication co-treatment among Kentucky Medicaid beneficiaries. Given lowered state and federal policy barriers and public health emphasis on HCV elimination in PWUD, we hypothesize that people who receive MOUD are already engaged in health care and are thus more likely to receive HCV DAAs.

METHODS

We conducted a retrospective cross-sectional analysis of a proprietary claims dataset from Health Verity, a data warehousing company that specializes in linking multi-source datasets across payer types. The dataset spanned health claims made between January 2020 and December 2021 among people who were living in Kentucky, had a preexisting OUD diagnosis, and were at least 18 years old.

Study Population

Our sample included Medicaid beneficiaries with a chronic HCV diagnosis who had not previously received DAA treatment. Chronic HCV diagnoses were determined using a validated algorithm.³⁴ This algorithm classified people as living with chronic HCV if they had a medical claim for an HCV RNA test followed by three or more chronic HCV codes on different services dates.³⁴ Validation tests showed that this algorithm had 97% specificity and 35% sensitivity.³⁴ For each individual in the sample, we identified an index HCV RNA test and followed participants for the remainder of the study period. We excluded individuals who received HCV DAA treatment during the 6-month period prior to the index HCV RNA test (6-month lookback period); received an index HCV RNA test after June 30, 2021 (6-month washout period); had missing or unknown sex, age, or pharmacy data; or lived outside Kentucky. A flow diagram is available in Appendix Fig. 1.

Measures

The outcome variable for this study was HCV DAA receipt (1/0). To determine whether patients received curative HCV DAAs, we used National Drug Center codes (see Appendix Table 1) to identify HCV DAA prescription fills from pharmacy claims. Patients were considered to have received curative HCV DAAs following a chronic HCV diagnosis if the prescription fill date for the DAA occurred after the patient’s index HCV RNA test.

The primary predictor variable was MOUD receipt during the study period (1/0). Patients were determined to have received any MOUD if they received buprenorphine, methadone, or naltrexone (see Appendix Table 1 for NDC codes) at any time during the study period. We also created an MOUD type variable that included mutually exclusive categories of MOUD receipt (buprenorphine, methadone, naltrexone, no MOUD). If a person received more than one kind of MOUD, the MOUD with the greatest number fills during the study period was determined as the primary MOUD type used.

Patient demographic variables included in the dataset were as follows: patient age category (18–30, 31–40, 41–50, 51 +), sex (female, male), and patient residential region (Western/Southern, Central, Eastern) defined by the first two digits of the patient’s three-digit zip code (Appendix Fig. 2). Other variables were Elixhauser comorbidity index (continuous, range 1–31)^35–37 and receipt and type of HCV DAA medication (glecaprevir/pibrentasvir [most frequently used DAA in our sample], or another type of DAA [elbasvir/grazoprevir, ledipasvir/sofosbuvir, sofosbuvir/velpatasvir, sofosbuvir/velpatasvir/voxilaprevir]).

Analysis

First, we calculated descriptive statistics of all variables by subgroup and compared differences between DAA subgroups using chi-squared tests for categorical variables and t-tests comparing means for continuous variables. Second, to determine the individual association of each covariate, we created separate bivariate logistic regression models with HCV DAA receipt as the outcome. Third, we constructed a multivariable logistic regression model of HCV DAA receipt using patient demographic variables (age, sex, residential region), comorbidity score, and any MOUD receipt as covariates. As a subanalysis, we constructed a similar second model that examined MOUD type used as the primary predictor. For both logistic regression models, we calculated regression-adjusted predicted proportions of HCV DAA receipt for each covariate holding all covariates at their mean values.

Sensitivity Analysis

We conducted two additional sensitivity analyses. First, we adjusted for Medicaid coverage gaps of 30 days or more in the multivariable model. Second, we tested for potential demographic differences in MOUD receipt by calculating descriptive statistics, chi-square, and t-tests. Characteristics that were significantly associated with MOUD receipt were included in an interaction term in the multivariable model. Per our hypothesis, significant differences in MOUD receipt would result in disproportionate HCV DAA receipt.

The threshold for statistical significance was p < 0.05. Descriptive statistics for HCV DAA type and MOUD type were included for additional detail but were not included in the final model. Analyses were completed in SAS 6.4. This study met the Johns Hopkins Bloomberg School of Public Health Institutional Review Board criteria for non-human subjects research. We followed Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.³⁸

Role of Funding Source

This study was funded by Bloomberg Philanthropies as part of the Bloomberg Overdose Prevention Initiative. The funder had no role in the study’s design, conduct, or reporting.

RESULTS

Population Characteristics

The analytic study sample included 2149 people with concurrent chronic HCV and OUD (Table 1). The minority, 36% (n = 780), received curative HCV DAAs. The greatest proportions of people were ages 31–40 (n = 934; 43%), were female (n = 1094; 51%), lived in Eastern Kentucky (n = 954; 44%), and received any MOUD (n = 1804; 84%). Of people who received any MOUD, buprenorphine was the most common MOUD type used (n = 1470/1804, 78%), followed by methadone (n = 205/1804; 11%), then naltrexone (n = 185/1804; 10%). Elixhauser’s comorbidity score was 6.2, on average (SD = 3.2) across the study sample. Of people who received curative HCV DAAs, glecaprevir/pibrentasvir was the most common HCV DAA (n = 762/780; 98%).

Table 1.

Characteristics of Kentucky Medicaid Recipients Living with Opioid Use Disorder and Chronic HCV by HCV DAA Receipt, 2020–2021

Characteristic	Total population (n = 2149)		Did not receive DAA (n = 1369; 64%)		Received HCV DAA (n = 780; 36%)		p
	n	%	n	%	n	%
Age category							0.0777
18–30	437	20%	293	21%	144	18%
31–40	934	43%	586	43%	348	45%
41–50	469	22%	282	21%	187	24%
51 +	309	14%	208	15%	101	13%
Sex							0.0384
Female	1094	51%	720	53%	374	48%
Male	1055	49%	649	47%	406	52%
Patient regiona							0.7024
Central	940	44%	607	44%	333	43%
Eastern	954	44%	604	44%	350	45%
Western and Southern	255	12%	158	12%	97	12%
Comorbidities (mean, SD)	6.1	3.2	6.5	3.4	5.2	5.4	< 0.0001b
HCV medications
Glecaprevir/pibrentasvir (Mavyret)	762	35%	-	-	-	-
All other DAAs	18	1%	-	-	-	-
Received MOUD							< 0.0001
No MOUD	345	16%	187	14%	158	20%
Any MOUD	1804	84%	1182	86%	622	80%
MOUD typec							0.0003
Buprenorphine	1414	66%	923	67%	491	63%
Methadone	205	10%	144	11%	61	8%
Naltrexone	185	9%	115	8%	70	9%
No MOUD	345	16%	187	14%	158	20%

^aRegions defined by 2-digit zip code. West and South = 42, Central = 40, East = 41

^bp-value reflects t-test significance. All others are chi-square significance

^cMOUD type determined by primary MOUD type used during the study period. See Appendix Table 2 for details about variable construction

There was a significant difference in HCV DAA receipt by MOUD status (p < 0.0001), sex (p = 0.0384), and mean comorbidity score (p < 0.0001). A smaller proportion of people who received curative HCV treatment also received any MOUD (n = 622/780; 80%, vs. no curative HCV treatment, n = 1182/1369; 86%). On average, people who received HCV DAAs had lower comorbidity scores (mean = 5.2, SD = 2.6; no HCV DAA receipt, mean = 6.5, SD = 3.4). There were no statistically significant differences in DAA subgroups by age category (p = 0.0777), or residential region (p = 0.7024).

Regression Analysis

There were significant differences in adjusted percentages of HCV DAA receipt by MOUD receipt, age category, and comorbidity score (Table 2). The adjusted percentage of HCV DAA receipt was 12.32 percentage points lower (p < 0.0001) among people who received any MOUD compared to people who did not receive MOUD (46% vs. 33%; aOR = 0.596, 95% CI = 0.468, 0.758). Odds of curative HCV treatment receipt significantly decreased for each point increase in comorbidity score (aOR = 0.85, 95% CI 0.821, 0.88; p < 0.0001). Compared to people aged 18–30, curative HCV treatment receipt was significantly higher for people aged 41–50 (41% vs. 29%; aOR = 1.708, 95% CI 1.283, 2.273; p = 0.0002) and 51 or older (39% vs. 29%; aOR = 1.549, 95% CI 1.105, 2.171; p = 0.0111). There was no significant difference in HCV DAA receipt by sex or patient residential region. Unadjusted analysis results are in Appendix Table 3.

Table 2.

Logistic Regression-Adjusted Odds Ratios and Predicted Proportions of HCV DAA Treatment Among Kentucky Medicaid Recipients Living with OUD and Chronic HCV

Variables	AOR	95% confidence interval		Pred. proportion-HCV DAA receipt	Percentage point difference	p-value
Received MOUD
No MOUD (ref)	-	-	-	46%	-	-
Any MOUD	0.596	0.468	0.758	33%		< 0.0001
Age category
18–30 (ref)	-	-	-	29%	-	-
31–40	1.253	0.98	1.603	34%	4.88	0.0716
41–50	1.708	1.283	2.273	41%	12.13	0.0002
51 +	1.549	1.105	2.171	39%	9.78	0.0111
Sex
Female	0.88	0.733	1.057	34%	− 2.92	0.1712
Male (ref)	-	-	-	37%	-	-
Patient regiona
Central	0.961	0.715	1.291	35%	− 0.91	0.7907
Eastern	0.963	0.718	1.292	35%	− 1.85	0.802
Western/Southern (ref)	-	-	-	36%	-	-
Comorbiditiesb	0.85	0.821	0.88	-	-	< 0.0001
Primary MOUD typec,d
No MOUD (ref)	-	-	-	46%	-	-
Buprenorphine	0.599	0.468	0.766	34%	− 12.20	< 0.0001
Methadone	0.483	0.331	0.705	29%	− 16.82	0.0002
Naltrexone	0.715	0.491	1.041	38%	− 8.13	0.0799

AOR, adjusted odds ratio

^aComorbidities were calculated as a continuous variable using Elixhauser’s comorbidity scores

^bPatient region determined by first two digits of residential zip code

^cMOUD type determined by primary MOUD type used during the study period. See Appendix Table 2 for details about variable construction

^dResults presented for primary MOUD type were obtained from a separate model. Results for the full model are presented in Appendix Table 4

The subanalysis of MOUD type showed that HCV DAA receipt was significantly lower for people who primarily used buprenorphine and methadone, but not naltrexone, when compared to people who did not receive MOUD (Table 2). There was a 16.82 percentage point adjusted difference in HCV DAA receipt among people who received methadone versus no MOUD (29% vs. 46%; aOR = 0.483, 95% CI = 0.331, 0.705; p = 0.0002), and a 12.20 percentage point difference in people who received buprenorphine versus no MOUD (34% vs. 46%, aOR = 0.599, 95% CI 0.468, 0.766; p < 0.0001). There was no significant difference in HCV DAA receipt among people who received naltrexone compared to people who received no MOUD (p = 0.0799), or between people who received buprenorphine versus methadone (p = 0.1992). Detailed regression analysis results are in Appendix Table 4.

Sensitivity analyses yielded no major differences from our main findings and conclusions (Appendix Tables 5a–6b).

DISCUSSION

We examined HCV DAA receipt among Kentucky Medicaid recipients with OUD and HCV and found that only 36% of patients with dual HCV and OUD diagnoses received curative HCV treatment. The low rates are especially notable given that the sample included many people who had health care interactions to receive MOUD, a setting where guidelines and federal health authorities recommend that patients receive testing and counseling related to HCV.^39–41 Paradoxically, odds of receiving HCV treatment were significantly lower for patients who also received MOUD during the study period, especially among people who received methadone or buprenorphine.

While the pervasive gap in HCV DAA receipt among Kentucky Medicaid recipients with OUD in Kentucky is consistent with findings in other areas of the country,^25,42,43 the finding that MOUD receipt was negatively associated with HCV treatment receipt was unexpected. Buprenorphine, the most commonly used MOUD in our sample, is often prescribed in primary care settings. In forming our hypothesis, we assumed that there would be new opportunities for co-treatment, especially in primary care settings, because of the August 2020 change that allowed primary care providers to more easily prescribe DAAs along with telemedicine flexibilities for buprenorphine initiation. Our findings point to improvements in MOUD access for Medicaid beneficiaries, but significant lags in simultaneously addressing the HCV continuum of care through co-treatment. Understanding the HCV-OUD co-treatment landscape in hard-hit states like Kentucky is critical to swiftly and efficiently end the HCV-OUD syndemic, eliminate HCV, and combat the overdose crisis.

Innovative health services approaches to HCV-OUD co-treatment are needed to enhance HCV treatment points of entry, especially among people already engaged in OUD care. In this study, only about a third of people in this sample received HCV treatment, but over 80% received any form of MOUD during the study period, pointing to the need to capitalize on existing patient care engagement. Focused efforts on primary care providers are crucial as restrictions on prescribing medications for both HCV and buprenorphine to treat OUD continue to be reduced. Since data for this study were collected, training requirements to prescribe buprenorphine through the X-waiver have been removed and restrictions have been removed on HCV DAA prescribers in 46 states⁴⁸ which make primary care environments a vital opportunity for co-treatment.⁴⁹ Telemedicine is also a potential venue for effective and efficient co-treatment. A clinical trial by Talal and colleagues⁴⁷ showed that 90% of people who received HCV and OUD care by telemedicine achieved cure compared to 39% of patients with referrals to HCV care only. Additionally, primary care providers might partner with harm reduction and opioid treatment programs. These are effective sites for HCV testing and treatment,⁴⁴ with similar HCV cure rates to members of the general population.^45,46

Clinician education and patient education are also important to closing co-treatment gaps. Clinicians and patients alike may assume that drug use or treatment with MOUD is contraindicated for HCV DAAs,⁵⁰ may be unaware of the policy change that allows non-specialists to prescribe HCV DAAs,³¹or do not interpret treatment with MOUD as achieving sobriety.⁵¹ Clinician stigma against people who use drugs may also play a role in co-treatment gaps. Additionally, patients may elect to defer HCV treatment to prioritize OUD treatment because they do not perceive HCV as a severe health risk, believe HCV treatment is too costly, or that they are not able to receive HCV medication because of their drug use or MOUD treatment.⁵² Refining clinician HCV treatment curricula through projects such as the Kentucky Hepatitis Academic Mentorship Program (KHAMP),⁵³ or Project ECHO,^54,55 may combat clinician stigma against people who use drugs and MOUD, inform clinicians of policy changes to enhance HCV DAA access, and negate misinformation about contraindications of MOUD with HCV DAAs.

Limitations

This study has several limitations. First, our sample is limited only to Kentucky Medicaid members, so results are not generalizable to other states or people not insured by Medicaid, but could be used to inform future efforts in other states. Second, our study period occurs during the height of the COVID-19 pandemic, when OUD and HCV treatment access may have been limited. Third, we lack measures about whether someone was actively using or injecting drugs which may be relevant to understanding barriers to HCV or MOUD care. Fourth, our data do not include Medicaid eligibility criteria (e.g., social security receipt, Medicaid expansion eligibility), which may point to important socio-economic barriers to HCV and OUD care. Finally, our data do not include blood tests for HCV seropositivity. Our definition of chronic HCV diagnoses relies solely on claims data and is not verifiable using seropositivity tests or state HCV surveillance data as HCV is only reportable in Kentucky in cases among pregnant women and children. For the same reason, we cannot verify DAA retention or if someone achieved cure after being prescribed a DAA.

CONCLUSION

We detected significant gaps in curative HCV DAA receipt among Kentucky Medicaid members people with concurrent chronic HCV and OUD diagnoses. People who received any MOUD were significantly less likely to have received a curative HCV DAA. Concurrent chronic HCV and OUD prevalence was highest in Eastern Kentucky, and treatment gaps were detected throughout the state. Clinical, policy efforts, and state plans to eliminate HCV must continue to innovatively engage and co-treat people living with OUD and HCV as a National Hepatitis C Elimination plan approaches implementation.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This project was funded by Bloomberg Philanthropies as part of the Bloomberg Overdose Prevention Initiative, a collaborative partnership aimed at combatting the nation’s opioid epidemic. Bloomberg Philanthropies had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Declarations:

Conflict of Interest:

Dr. Sugarman, Dr. Saloner, Dr. Harris, Dr. Bandara, and Ms. Flanagan reported receiving grants from Bloomberg Philanthropies during the conduct of the study. Dr. Bandara reported receiving grants from the National Institutes of Health and Centers for Disease Prevention and Control outside the submitted work. Dr. Irvin reported part-time work on federal detail at the National Institutes of Health on the national hepatitis C elimination plan outside the submitted work. Dr. Saloner reported personal fees from Susman Godfrey outside the submitted work.