Drug Authorization for Sofosbuvir/Ledipasvir (Harvoni) for Chronic HCV Infection in a Real-World Cohort: A New Barrier in the HCV Care Cascade

Background New treatments for hepatitis C (HCV) infection hold great promise for cure, but numerous challenges to diagnosing, establishing care, and receiving therapy exist. There are limited data on insurance authorization for these medications. Materials and Methods We performed a retrospective chart review of patients receiving sofosbuvir/ledipasvir (SOF/LED) from October 11-December 31, 2014 to determine rates and timing of drug authorization. We also determined predictors of approval, and those factors associated with faster decision and approval times. Results Of 174 patients prescribed HCV therapy during this period, 129 requests were made for SOF/LED, of whom 100 (77.5%) received initial approval, and an additional 17 patients (13.9%) ultimately received approval through the appeals process. Faster approval times were seen in patients with Child-Pugh Class B disease (14.4 vs. 24.7 days, p = 0.048). A higher proportion of patients were initially approved in those with Medicare/Medicaid coverage (92.2% vs. 71.4%, p = 0.002) and those with baseline viral load ≥6 million IU/mL (84.1% vs. 62.5%, p = 0.040). Linear regression modeling identified advanced fibrosis, high Model of End Stage Liver Disease (MELD) score, and female gender as significant predictors of shorter decision and approval times. On logistic regression, Medicare/Medicaid coverage (OR 5.96, 95% CI 1.66–21.48) and high viral load (OR 4.52, 95% CI 1.08–19.08) were significant predictors for initial approval. Conclusions Early analysis of real-world drug authorization outcomes between October-December 2014 reveals that nearly one in four patients are initially denied access to SOF/LED upon initial prescription, although most patients are eventually approved through appeal, which delays treatment initiation. Having Medicare/Medicaid and advanced liver disease resulted in a higher likelihood of approval as well as earlier decision and approval times. More studies are needed to determine factors resulting in higher likelihood of denial and to evaluate approval rates and times after implementation of restrictive prior authorization guidelines.


Introduction
Treatment of chronic hepatitis C (HCV) infection in the United States has been revolutionized with the development of novel direct-acting antiviral (DAA) therapies. DAA therapy has demonstrated better tolerability, adherence, as well as rates of sustained virologic response (SVR) and cure compared to antecedent interferon (IFN)-based therapies [1][2][3][4]. This advance has expanded the population of individuals with HCV infection who are potentially treatable. Owing to its efficacy, the American Association for the Study of Liver Disease (AASLD) and the Infectious Disease Society of America (IDSA) have modified their recommendations to include the combination of sofosbuvir and ledipasvir (SOF/LED) as first-line therapy for HCV genotype-1 infection, the most prevalent strain seen in the United States [3,[5][6][7][8].
However, care provision requires successful completion of numerous steps along a care continuum [9]. It has been recently estimated that only 16% of chronic HCV-infected individuals are prescribed antiviral treatment and only 9% achieve SVR [10] although this represents data from the interferon era and precedes the advent of all-oral anti-HCV regimens. The concept of a care cascade (diagnosis, linkage to care, retention in care, prescription of antiretroviral therapy, and viral suppression) has been utilized as a means for identifying care gaps and setting goals in patients with human immunodeficiency virus (HIV) infection, HCV-HIV co-infection, and recently for HCV mono-infected individuals as well [11][12][13][14][15]. Recently, barriers to completion of therapy have been reported which include but are not limited to diagnosis, knowledge of treatment options, completion of pre-treatment paperwork, lack of insurance coverage, medical eligibility, lack of program infrastructure for vulnerable populations, and medication costs [16][17][18][19][20]. To this effect, some interventions to improve access have also been proposed, such as provision for self-referral and shortened treatment duration [21,22].
Among the steps in HCV treatment provision, pre-authorization (also known as prior authorization, prior approval, or pre-certification) is the process by which a health insurance provider determines that specific treatment is medically necessary, and which allows for insurance coverage of treatment cost. It is currently known that DAA therapy is expensive with prices ranging from $63,000 to $300,000 per treatment course. The wholesale cost of a 12-week treatment course of SOF/LED is $94,500, amounting to $1,125 per pill [23]. As this results in prohibitive cost and limited availability, pre-authorization often requires that patients have advanced fibrosis (grade F3 or beyond) or cirrhosis to be given treatment priority [24].
Currently there are limited data on rate and timing of insurance pre-authorization after SOF/ LED prescriptions are written. In this study, we aim to perform a retrospective observational study reporting real-life data of drug approval rates in a cohort of patients with HCV infection who received prescription for SOF/LED treatment over a 3-month period. We also aim to determine factors associated with pre-authorization approval, time to pre-authorization decision, and time to pre-authorization approval. We hypothesize that the majority of patients for whom a pre-authorization request is filed will ultimately receive approval, and that insurance pre-authorization will be within the recommended guidelines for treatment for those with the highest need (i.e. advanced liver disease). However, we also hypothesize that there will be a proportion of patients who are ultimately not approved, as well as some who are approved only after appeal.

Study Subjects
As part of the SOF/LED acquisition process, all patients had pre-authorization requests sent to their insurance coverage providers. We reviewed the medical charts of all patients at Yale Liver Center who had an insurance pre-authorization request for SOF/LED filed between October 11, 2014 and December 31, 2014. Patients were then excluded if they received a prescription for HCV treatment other than combination SOF/LED.

Outcomes
For each patient, we recorded the insurance provider of pre-authorization request. Those without Medicare or Medicaid insurance carriers were categorized as having private insurance coverage. If a patient was listed as having both Medicaid/Medicare and another insurance provider, they were considered to have a private insurance provider. We recorded approval, denial, or pending status of pre-authorization initial request and appeal as of March 1 st, 2015.
If an individual was denied treatment and appeal was sought, date of appeal request and date of appeal decision were recorded.

Covariates
Patient characteristics included age, race, body mass index, co-morbid hypertension, psychiatric illness, diabetes, renal disease, hepatitis B or HIV co-infection, and baseline biochemical markers (total bilirubin, serum creatinine, and serum international normalized ratio, INR). We recorded HCV viral characteristics, including genotype, viral load, IL28B gene variant and prior treatment regimens. Severity of HCV infection was determined by progression of hepatic fibrosis. Those with METAVIR stage 4 fibrosis on liver biopsy, clinical hepatic decompensation, or imaging findings suggesting cirrhosis with portal hypertension were classified as having cirrhosis. Advanced fibrosis included those with cirrhosis and included individuals with grade 3 fibrosis on liver biopsy, advanced fibrosis by tissue elastography, and/or an elevated FIB-4 score (>3.25) [25]. In patients with cirrhosis, Child-Pugh class and MELD scores were recorded using laboratory testing closest to the pre-authorization filing date. A subset of patients received their pre-authorization request through their transplant clinic provider. It was noted for those whom this was the case.

Statistical Analysis
T-testing was used to compare continuous variables and chi-square testing was used for categorical variables. Univariate and multivariate analyses were performed using linear and logistic regression modeling with forward selection logistic regression to identify significant predictors of preauthorization approval and times-to-decision or approval. All data were analyzed using SAS 9.4 statistical software (Cary, NC). Full dataset with SAS code used for this analysis is available at S1 Appendix.

Ethics Statement
We obtained approval for conduct of this study by our institutional review board.

Results
A total of 174 patients with chronic HCV infection seen at the Yale Liver Center were prescribed antiviral therapy between October 11 th and December 31 st 2014, of whom 129 were prescribed SOF/LED. Tables 1-3 summarize demographic characteristics of this patient population. The mean age was 57.0 ± 9.9 years with 61.2% being males. 60.5% of the population had cirrhosis. Table 4 summarizes the outcomes of pre-authorization. Of the 128 for whom preauthorization status was determined, 100 (77.5%) received initial approval for pre-authorization. 117 (91.4%) of 129 received approval including those who required appeal. Initially, 19 patients (14.7%) required appeal and ultimately 6 (4.7%) were denied. As of March 1 st , 2015, the pre-authorization status of 5 (3.9%) are pending and 1 (0.0%) is unknown. The average time to final decision (approval or denial) was 26.1 ± 25.2 days, and in those approved the average time to decision was 22.9 ± 21.2 days. Table 5 summarizes the time-to-decision in all subjects with outcomes data and time-toapproval in those who were approved for pre-authorization. Females were found to have a significantly lower time-to-decision than males (19.8 vs. 30.0 days, p = 0.01) with a similar but non-significant finding in time-to-approval. Those with a Medicare/Medicaid had a shorter  Table 6 summarizes proportions of unapproved and initially approved for those whom preauthorization was sent categorized by patient characteristics. A significantly higher proportion of patients with Medicare/Medicaid were initially approved compared to those with private insurance (92.2% vs. 71.4%, p = 0.002). In addition, a significantly higher proportion of patients with a viral load 6 million were initially approved compared to individuals with viral load < 6 million (84.1% vs. 62.5%, p = 0.04).
Univariate linear regression modeling results are shown in Table 7. Significant associations to shorter times-to-decision and times-to-approval were noted with psychiatric disease, high FIB-4 score, and pre-authorization request from transplant clinic. Also, significantly shorter times were noted with increases in total bilirubin, INR, FIB-4 score, and MELD score. Table 8 summarizes univariate logistic regression model results. This analysis revealed that having Medicare/Medicaid (OR 4.72, 95% CI 1.63-13.67) and a high viral load (OR 3.17, 1.02-9.81) were associated with higher odds of initial approval compared to private insurance and low viremia, respectively.
Multivariate linear and logistic models are shown in Tables 9-16. In multivariate linear models, forward stepwise addition revealed that MELD score, female gender, and advanced fibrosis were significant predictors of a shorter time-to-decision and time-to-approval, while psychiatric disease was found to be a significant predictor of a shorter time-to-approval. These  (Tables 15 and  16).

Discussion
In our cohort of patients receiving pre-authorization request for SOF/LED over a threemonth period, we found that nearly one in four were denied initial approval, although most patients eventually obtained drug authorization through the appeals process. Female gender, advanced Child-Pugh class, and liver transplant clinic were associated with shorter decision or approval times. Finally, having Medicare/Medicaid and a high viral load were significant predictors for initial approval, with findings persisting after controlling for demographic covariates.
The cascade of care model for HCV treatment involves numerous steps from diagnosis to successful treatment and viral eradication with patient drop-out observed at every step [11]. This analysis focused on one specific process: pre-authorization request and approval in those with a known diagnosis of HCV infection prescribed SOF/LED. Fewer than 10% patients ultimately failed to obtain access to therapy, although the appeals process led to further delay to treatment initiation. Importantly, the proportion of patients with access to drug therapy may be overestimated as this analysis was largely restricted to insured patients, all of whom had already successfully linked to specialty care in a major tertiary care university liver clinic, completed a series of pre-treatment evaluations and a formal structured HCV class, and were deemed by a specialty provider to represent an appropriate candidate with adequate motivation to initiate treatment.   Table 11. Multivariate analyses for time-to-decision (n = 123). Prediction of time-to-decision based on linear multivariate model with selected variables (from stepwise selection as above) and demographic covariates (n = 123).  The higher approval rates in patients with Medicare/Medicaid was unexpected, and could not be explained by other patient or medical variables, as this association remained significant in the multivariate model. Following Food and Drug Administration (FDA) approval of SOF/ LED on October 10, 2014, updates to AASLD-IDSA HCV treatment recommendations affirmed that treatment be considered for all patients regardless of disease severity, although with the highest priority given to patients with advanced fibrosis, transplant recipients, or those with severe renal insufficiency [3]. Our hypothesis is that the higher than expected authorization rates by Medicare/Medicaid represented a time-limited anomaly driven by the absence of prior authorization guidelines until December 2014 and January 2015, through which Harvoni has been restricted by state Medicaid to patients with advanced liver fibrosis or cirrhosis (F3/F4), and selected patients at high risk for disease progression, and must be prescribed by specialty physicians [26,27]. Restrictive prior authorizationz guidelines were established by many public and private payors in this state by early 2015 (Table 17). As nearly half of patients prescribed SOF/LED in this analysis had Medicare/Medicaid coverage, drug authorization rates would be expected to be lower beyond January 2015.

Variable
In our cohort, patients in the liver transplant clinic were found to have shorter approval times, which may be attributable in part to overrepresentation of advanced liver disease in this population, and therefore likely be given initial approval through the pre authorization process with both public and private payors. We could not exclude the potential effect of variable access to certified specialty pharmacies with capacity to directly dispense SOF/LED medications to patients.
This is the first study to our knowledge assessing real-world access to interferon-free DAA regimens in established cohorts of patients with chronic HCV seeking antiviral therapy. These results contribute to the limited data available addressing proportion of patients successfully obtaining drug authorization through public and private insurance carriers, time to approval, and predictors for approval. Several limitations of our study warrant further investigation. We did not record data on proportion of treatment-eligible patients seeking treatment who declined to pursue SOF/LED prescription due to absence of insurance coverage, or perception of difficulty in accessing treatment due to mild liver fibrosis or other factors. Although the analysis was performed for consecutive unselected patients prescribed SOF/LED, this cohort represented a subset of patients who were deemed to be excellent candidates for treatment, and therefore selection bias by prescribing providers for individuals with anticipated approval could not be excluded. This study is also limited to authorization data in Connecticut, and state Medicaid and Medicare approval rates likely differ by states as well. Furthermore this study is focused exclusively on SOF/LED, and authorization results may be different for other FDA-approved interferon-free regimens such as sofosbuvir/simeprevir and paritaprevir/ritonavir, ombitasvir, dasabuvir, and ribavirin. Future studies are needed to clarify the variance in public and private insurance access to HCV regimens across states, stratified by liver fibrosis and other patient characteristics, the outcome of appeal requests, and approval of requests which are beyond FDA label or AASLD/IDSA recommendations.
In conclusion, we found that most patients filing a pre-authorization request for SOF/LED are eventually approved, but nearly 1 in 4 were denied access upon initial request, which may represent a barrier within the HCV care cascade. On multivariate analysis, advanced liver disease was associated with faster approval time, while Medicare/Medicaid and high viremia were associated with insurance approval. Further studies are warranted to investigate the impact of evolving drug authorization policies by Medicare/Medicaid and private payers on access to curative HCV therapies such as SOF/LED. Table 17. Select information requested for pre-authorization for specified insurance providers.
Exact criteria should be found in appropriate insurance pre-authorization form.