Effectiveness of direct-acting antiviral therapy for hepatitis C in difficult-to-treat patients in a safety-net health system: a retrospective cohort study

Chronic hepatitis C virus (HCV) infects an estimated 3.4–6.0 million people in the United States and approximately 71 million people worldwide [1, 2]. Deaths related to HCV infection have been on the rise for the past decade [3], with cirrhosis-related complications, including hepatocellular carcinoma (HCC), accounting for the increased morbidity and mortality [4, 5]. However, a sustained virologic response (SVR) is associated with improved quality of life as well as reduced rates of fibrosis progression, cirrhosis-related complications, HCC, and all-cause mortality [5, 6].

Prior to the advent of direct-acting antiviral (DAA) therapy, the mainstay of HCV therapy involved interferon (IFN)-based regimens that had frequent contraindications, were poorly tolerated, and achieved at best a 50% SVR rate [7]. Introduction of DAAs has since revolutionized the HCV treatment landscape. SVR rates for DAA therapy exceed 90% in registration trials and are better tolerated than IFN-based regimens [8]. Similarly, real-world data suggests high rates of effectiveness in clinical practice; however, studies thus far have described outcomes primarily in academic centers treating well-insured, primarily Caucasian populations [9–12].

A large burden of HCV infection falls on marginalized populations, including the uninsured, homeless, incarcerated, substance users, those with mental disorders, and those infected with HIV [1, 13]. However, there is a paucity of data regarding effectiveness of DAAs in these populations, who are largely treated in safety-net health systems. Further, DAAs may be unavailable for these patients due a myriad of barriers, including high drug costs, limited healthcare access, low health literacy, language barriers, unstable housing, and significant psychosocial and medical comorbidities [14]. Additionally, at-risk populations have been reported to have lower rates of follow-up, which may translate into undetected drug-related adverse events [15], poor adherence that could lead to treatment failure, and risky behavior such as ongoing substance abuse that may be associated with an increased likelihood of reinfection [16]. Consequently, these patients are often excluded from registration trials [17]. Thus, real-world studies in these vulnerable populations are necessary to inform interventions that address the specific challenges faced in their care.

The aim of our study is to describe the effectiveness of DAA-based HCV therapy in a racially diverse, socioeconomically disadvantaged, underinsured population seen in a large urban safety-net health system.

In this retrospective cohort study, we evaluated the effectiveness of DAA-based treatment for HCV in a predominantly underinsured, racial/ethnic minority population, with high rates of alcohol and substance use, mental health disorders, and high rates of cirrhosis – a group of patients who must be treated in order to achieve HCV eradication. We found that 90% of this traditionally “difficult-to-treat” population followed in a resource-constrained safety-net health system could achieve SVR with DAA-based therapy. There were no differences in SVR by insurance status, racial/ethnic group, or history of drug/alcohol use. The only factor negatively associated with SVR was the presence of decompensated cirrhosis, highlighting another benefit of treating patients with earlier stages of fibrosis prior to disease progression [21].

We found a high rate of treatment success in our cohort comparable to those reported in other real-world outcome studies at large academic centers and private clinics [9–12]. Our population is unique in that it consists of a large group of low socioeconomic status patients, of whom 56% were uninsured and 13% were Medicaid insured. This demographic has a considerably higher prevalence of chronic HCV infection than the general population [22], yet is consistently under-represented in existing real-world studies of DAA treatment outcomes. Only two smaller studies have described effectiveness of DAAs in indigent populations [14, 23]. Barocas et al. [23] reported SVR was observed in 62 (97%) of 64 homeless and marginally housed adults; however, included patients had been pre-selected based on likelihood of adherence to treatment and all but one were insured. Beck et al. [14] reported SVR in 183 (93%) of 189 patients receiving SOF-based HCV therapy within a safety-net health system in San Francisco; however, 95% of patients in this study were also insured. Of note, both prior studies were conducted in states that offer expanded Medicaid, through which access to HCV treatment is possible for a larger proportion of low-income patients. Data shows that the burden of HCV is 7.5% higher in those with Medicaid compared to commercial insurance [24]. In states without expanded Medicaid, such as Texas, access to HCV treatment is limited due to a smaller proportion of low-income patients being granted Medicaid coverage, as well as the existence of restrictions on treatment eligibility (i.e., advanced liver disease, negative toxicology screens, etc.) for those who do have Medicaid coverage. Our study data represent outcomes that can be achieved even in a state without expanded Medicaid. Further, we found that rates of SVR were unaffected by insurance status, suggesting that concerns regarding treatment of underinsured populations may be unfounded.

The success of our program is based on (1) utilizing resources available through pharmaceutical companies, namely through PAPs, which are designed to provide access to medications for those without insurance; and (2) coordinating comprehensive care through a multi-disciplinary team consisting of pharmacists, nurse navigators, and providers. In our set up, pharmacists or pharmacy technicians assisted patients in completing medication application forms, collating necessary documentation, obtaining insurance prior authorization, and arranging medication delivery to patient homes. Our nurse navigator coordinated follow-up blood draws and clinic visits. Several improvements were made to the process as the clinic expanded, wherein visits arranged after treatment to assess for SVR (“SVR visits”) were found to reduce loss to follow-up; linking of laboratory and appointment visits such that they occurred on the same day improved patient experience; education of primary care providers within the healthcare system on HCV therapies and the clinic referral process improved referral rates; and introduction of an order set that combined medication and laboratory orders allowed monthly medication dispensing and follow-up blood draws to occur in a timely fashion, thereby reducing gaps in treatment. Through strategic planning and resource utilization, our clinic effectively adapted minimal resources to screen and initiate DAAs at a rate of approximately 50 new patients per month. This model may be particularly helpful in centers caring for large numbers of uninsured patients, amongst whom the prevalence of chronic HCV infection is considerably higher than in the general population [22].

Another unique feature of our patient population is the high proportion of Black and Hispanic patients (44% and 16%, respectively) as compared to other real-world studies, reflecting the ethnic makeup of the patient population within the Parkland catchment area. African Americans and Hispanics in the USA are disproportionately affected by HCV [13]. Previous studies found that disease progression tends to be accelerated in Hispanic patients [25], and both African Americans and Hispanics are more likely to experience treatment failure with IFN-based therapy as compared to their White counterparts [26, 27]. A recent study of 21,095 Veterans treated with DAAs found Blacks and Hispanics were less likely to respond to DAAs compared to Whites after adjusting for baseline characteristics, especially if using shorter 8-week courses of treatment [28]. However, in our population, we found that treatment with DAA-based therapy yielded similar rates of success in Black, Hispanic, and White patient groups.

We found decompensated cirrhosis to be the only negative predictor of SVR in multivariate analysis, which has similarly been described in prior analyses [9–11]. Our uninsured patients were treated regardless of stage of liver disease because predictive models have demonstrated the cost efficiency of treating individuals with earlier stage disease [21] and that curative treatment improves quality of life and mortality [18]. However, Medicaid reimbursement criteria in Texas (and many other states) restrict DAA accessibility to patients with advanced liver disease. Additional Medicaid reimbursement criteria in other states include evidence of undetectable HIV RNA in co-infected patients and documented drug and alcohol abstinence [29]. Our study found similar rates of treatment success among patients with alcohol and drug use histories as well as those co-infected with HIV, although data quantifying active substance use and HIV RNA levels in co-infected patients was not collected. These data suggest that imposing restrictions on access to DAAs based on stage of liver disease, HIV status, and substance use is not evidence-based and contributes to socioeconomic disparity, disrupts prevention efforts, and ultimately leads to increased healthcare expenditure [29, 30]. Patients denied HCV treatment by Medicaid are unlikely to qualify for PAPs because they are considered to be otherwise “funded”, thereby creating a paradoxical situation in which some patients on Medicaid are disadvantaged in terms of HCV treatment access compared to those without insurance.

Our study population suffered a loss to follow-up rate of 5%, only marginally higher than the 2.5–3.8% reported in non-indigent real-world studies [9–11] and comparable with the 5–15% reported in previous studies of IFN-based therapy in indigent populations [24, 28, 31]. Interestingly, among those lost to liver clinic follow-up, approximately 1 in 5 were subsequently seen by their primary care providers within the same healthcare system. It is possible that knowledge gaps in non-specialist providers were a contributing factor to the lack of laboratory follow-up after treatment completion; thus, future development of targeted interventions, such as provider education on HCV screening and treatment, may help to overcome these barriers.

This study had several key limitations. Firstly, it was an observational, non-randomized retrospective study without a control group, and was underpowered to fully examine the relative benefits and drawbacks of different DAA regimens. Second, our clinic structure included specialist providers, and thus our findings may not translate to other primary care practice settings. Finally, long-term follow-up data with which to identify reinfection rates was unavailable; such data may play an important role in this high-risk population.

The advent of DAAs has led to improved outcomes amongst those with access to treatment. However, underinsured patients who are cared for by safety-net hospitals carry a significant burden of disease but may have limited access to these curative therapies. Our program demonstrates that access to HCV treatment is possible with a dedicated staff, and that outcomes in this socioeconomically disadvantaged population are comparable to those achieved in commercially insured, high-income settings.