Adoption of New Agents and Changes in Treatment Patterns for Hepatitis C: 2010-2014
Xiaoxi Yao, PhD; Lindsey R. Sangaralingham, MPH; Joseph S. Ross, MD; Nilay D. Shah, PhD; and Jayant A. Talwalkar, MD
Chronic hepatitis C virus (HCV) is an infection that affects nearly 3.2 million individuals in the United States.1
HCV is associated with increased risk of developing liver cirrhosis, hepatocellular carcinoma, liver failure, and death.2
Over the past 2 decades, the mainstay of treatment of HCV has been interferon alfa and ribavirin, but the treatment causes numerous adverse events, including influenza-like symptoms, depression, cytopenia, hemolytic anemia, fatigue, pruritus, and rash.3,4
It also requires long treatment duration (typically 48 weeks), and results in an average low cure rate—defined by sustained viral response (SVR)—of less than 50%.5,6
Additionally, many patients are not eligible for, or cannot tolerate, interferon-based treatment.7,8
As a result of these challenges, until recently, only 10% to 30% of adults with chronic HCV received treatment.9
Since 2011, significant advances have been made among the options for treatment of chronic HCV.
The new generation of medications, such as ledipasvir/sofosbuvir (Harvoni), sofosbuvir (Sovaldi), and simeprevir (Olysio), increase SVR to over 90% and shorten the treatment period to 12 to 24 weeks.10-14
The incidence and severity of adverse events are also lower in these new agents.11,15
Considering the efficacy, safety, tolerability, and convenience, these new treatment regimens are recommended by the clinical practice guidelines developed by the American Association for the Study of Liver Diseases and the Infectious Diseases Society of America in collaboration with the International Antiviral Society–USA.16
There is hope that these new medications may increase the historically low treatment rate in patients with chronic HCV; however, to date, little is known about how these new treatments have been adopted into practice and what their impact has been on patients’ receipt of treatment. In this study, we sought to examine whether the introduction of new HCV medications was associated with an increased treatment rate, especially among the elderly and patients with other complex chronic conditions who are likely to be ineligible for or intolerant of previous regimens.
A secondary aim of this study was to estimate the financial burden of HCV treatment on patients. The cost of the newest agents has been very controversial, with estimates ranging from $66,000 per person for simeprevir to $84,000 for sofosbuvir per person for a 12-week course of treatment.17
The high costs of the new drug regimens have raised significant concerns about insurance plan affordability and cost sharing for patients. Little is known about the patients’ cost sharing; therefore, this study aimed at describing out-of-pocket (OOP) costs of patients undergoing different treatments.
We conducted a retrospective analysis of administrative claims data from Optum Labs Data Warehouse (OLDW), which includes privately insured and Medicare Advantage enrollees throughout the United States. The database contains health information on over 100 million enrollees over the last 20 years from geographically diverse regions across the United States, with greatest representation from the South and Midwest.18
The included health plans provide claims for professional (eg, physician), facility (eg, hospital), and outpatient prescription medication services.19
The Mayo Clinic Institutional Review Board exempted this study from approval as it represents research on preexisting, deidentified data.
We identified all adults (18 years or older) who were diagnosed with chronic HCV between January 1, 2010, and December 31, 2014. Patients were considered to have chronic HCV if they had at least 1 chronic HCV diagnosis (International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM)
codes: 070.44 and 070.54), or at least 2 unspecified HCV diagnoses (ICD-9-CM
: 070.70, 070.71 and V02.62) on different dates. Patients entered the cohort once they had at least 12-month continuous enrollment in medical and drug health plans and at least 1 HCV diagnosis during these 12 months. Due to the changes in enrollment and disenrollment in health plans, patients entered and left the cohort at different times. The date of entry was defined as the index date, and the 12 months prior to the index date was defined as baseline. We excluded patients who received any HCV treatment at baseline. A detailed flowchart illustrating the cohort creation process can be found in eAppendix Figure 1 (eAppendices
available at www.ajmc.com
The primary outcome of interest was whether patients received any HCV treatment and which treatment they received. We followed the patients from the index date until they were no longer enrolled in the health plans or until end of the study period (December 31, 2014). We assessed 6 treatment regimens available during the study period: 1) interferon/ribavirin, 2) boceprevir (Victrelis), 3) telaprevir (Incivek), 4) simeprevir or sofosbuvir (not combined), 5) simeprevir and sofosbuvir (combined), and 6) ledipasvir/sofosbuvir (see eAppendix Table 1 for the generic names, brand names, Current Procedural Terminology, Version 4 (CPT-4)/Healthcare Common Procedure Coding System codes, and the FDA approval dates). We separated patients who received the combined sofosbuvir and simeprevir from those who received only 1 of the 2 agents due to highly differing costs. We did not separate patients who received only eprevir from those who received only sofosbuvir, because a single patient in the cohort received simeprevir alone.
Treatment regimens 2 through 4 (from above) include the use of ribavirin and/or interferon. If patients received interferon/ribavirin with another medication at the same time, they were considered receiving the other treatment regimen. For example, if patients received boceprevir with interferon and ribavirin, they were considered to be on the boceprevir regimen. We calculated the percentage of patients receiving treatment, and the percentages of patients receiving each of the 6 regimens per study quarter. Patients were considered being treated or being on a specific treatment regimen during the time from the fill date of a medication through the end of this medication fill (fill date + days of supply – 1). Many patients may be cured after treatment, and patients need to be monitored for at least 12 to 24 weeks to determine whether the treatment was successful. If the treatment failed, some patients may not be eligible or willing to receive another treatment11,20
; therefore, most treated patients no longer needed treatment—at least in the short term—so we excluded them from both the numerator and denominator of the percentages after the conclusion of their last treatment. Patients were not counted prior to the index date or after they disenrolled from the health plan.
The secondary outcome was patients’ OOP cost, defined as the sum of the deductible and copayment of the medications. All the costs were adjusted to 2014 dollars using the personal consumption expenditures inflation factor.21
We calculated the descriptive statistics of costs stratified by patients’ most recent treatment regimens.
We used multivariable logistic regression analysis to estimate the predicted probability of receiving treatment prior to and after the new medications became available. We divided patients into 2 groups: 1) “pre-group,” which includes patients whose enrollment ended before December 1, 2013, approximately when simeprevir and sofosbuvir became available; and 2) “post group,” which includes patients who had enrollment on or after December 1, 2013. In this analysis, we excluded 1565 patients who were in the post group but finished treatment prior to December 2013.
Other covariates included demographics and socioeconomic characteristics at index date (ie, age, gender, race, household income, and residence region), baseline comorbidities, Charlson-Deyo Comorbidity Index (CDCI),22
and the length of follow-up. We adjusted for comorbidities that can influence patients’ treatment decisions, including depression (ICD-9-CM
codes: 296.2x, 296.3x, 311), liver cancer (ICD-9-CM
: 155.0, 155.2), liver disease or cirrhosis (ICD-9-CM
: 571.2, 571.5, 571.6, 571.8, 571.9), and liver transplant (ICD-9-CM
: V42.7, 996.82; ICD-9-CM
procedure: 50.5, 50.51, 50.59; CPT-4 codes: 47135, 47136). The CDCI was used to assess a patient’s overall comorbidity burden. Statistically significant interactions between the main independent variable and other covariates (ie, age, gender, region, and comorbidities) were also included in the model. All analyses were conducted using SAS 9.3 (SAS Institute Inc, Cary, North Carolina) and Stata 13.1 (Stata Corp, College Station, Texas).
From January 1, 2010, to December 31, 2014, we identified 56,116 patients with both a diagnosis of chronic HCV and no receipt of any treatment at baseline. Patients’ demographic, socioeconomic, and health characteristics, stratified by their most recent treatment regimens, are presented in Table 1
. The mean age was 55.4 years; the majority of the cohort were male (59.8%) and white (63.3%); 13.4% of patients had depression at baseline, 2.0% had liver cancer, 17.2% had liver disease or cirrhosis, and 1.6% had liver transplant. Compared with patients who were not treated, patients who received new treatment regimens were generally older and had more comorbidities, whereas patients who received the old therapies were younger and had fewer comorbidities.
A total of 5436 (9.7%) patients in our cohort received HCV treatment during an average of 1.8 years of follow-up. The Figure
shows the quarterly trend of the percentages of patients who used each of the treatments. In each quarter of 2010, approximately 2.5% to 3% of the patients with chronic HCV were treated using interferon/ribavirin. The use of interferon/ribavirin as the primary treatment gradually decreased from 2.4% in the first quarter of 2010, to 0.10% in the last quarter of 2014. Following the market debut of boceprevir and telaprevir in May 2011, the use of these 2 medications quickly increased. In the last quarter of 2011, 2.5% of patients with HCV received telaprevir and 0.7% received boceprevir, but neither was used in the last quarter of 2014. After simeprevir and sofosbuvir became available at the end of 2013, their use increased and peaked in the second quarter of 2014: 3.3% received 1 of the 2 agents and 1.8% received them together. In the last quarter of 2014, ledipasvir/sofosbuvir became available, and nearly every treated patient was using the new regimens: ledipasvir/sofosbuvir (2.0%), sofosbuvir or simeprevir (1.4%), and sofosbuvir combined with simeprevir (1.1%) (eAppendix Table 2).
We also examined the trends of treatment among different subpopulations and found the trend among the elderly (65 years or older) was very different than the rest of the cohort (eAppendix Figure 2 and eAppendix Table 3). Prior to 2014, only 0.5% to 1.5% of elderly patients were on treatment each quarter, but after the new medications became available, the treatment rate tremendously increased. In the last quarter of 2014, the percentage of elderly patients who were being treated was comparable to the overall population (4.4% vs 4.6%).
The predicted probabilities of receiving treatment prior and after new medications became available are shown in Table 2
. Overall, the treatment rate was estimated to increase from 5.39% to 6.77%, with a significant absolute increase of 1.38%. The absolute increase was greatest among patients who had a past liver transplant (9.11%): from 5.54% to 14.65%. The absolute increase in treatment rate was also high among patients 65 years or older (4.34%); those with liver cancer (5.47%), liver disease, or cirrhosis (5.15%); and a CDCI score of 2 or more (2.95%). The predicted probability of receiving treatment significantly decreased among patients aged 18 to 40 years and 41 to 50 years, and patients with a CDCI score of 0%, –3.54%, –2.24%, and –1.12%, respectively. The increase in treatment rate was greater among male than female (1.75% vs 0.90%), black than white (1.62% vs 1.31%), poor (annual household income <$40,000) than high-income (>$100,000; 2.04% vs 0.77%), and patients living in the Midwest and Northeast than those living in the South and West (2.47%, 2.76%, 0.36%, 0.97%, respectively).
shows the OOP costs among the 5436 treated patients. The median OOP costs of patients receiving new regimens were relatively low ($112-$340), but there were great variations and the mean costs were high ($1982-$2127). Some patients did not pay any cost OOP, whereas others paid as high as $75,831.
To the best of our knowledge, no contemporary data exists on the trends and patterns of HCV treatment in real-world clinical practice, especially in the new era of HCV treatment since December 2013. Our study is the first to examine the early adoption of new HCV treatment in a large, national, privately insured population. We observed substantial changes in treatment patterns over the past 5 years. The older treatments, ribavirin, interferon, boceprevir, and telaprevir, are essentially obsolete now, with few patients receiving these therapies at the end of 2014. Instead, patients are now almost exclusively using therapies only made available since 2013, because of the efficacy, safety, and convenience of the new medications.
Our study also found the introduction of new medications was associated with increased treatment rates, especially among those who were older and had other chronic conditions that necessitated urgent treatment. We also observed different patterns of the uptake of new medications among different subpopulations of patients. Although white and high-income patients had the highest treatment rate both before and after the new medications became available, the gaps between white and black and high- and low-income individuals were closing. The Midwest and Northeast seemed to adopt the new treatment more quickly than the South and West. Patients who were young or had few chronic conditions might be less likely to receive care. Due to the high costs of new medications, some insurers developed prior authorization programs and asked physicians to treat only patients who require immediate therapy23
; this makes it difficult for young and healthy patients to obtain treatment. These patients may also be willing to hold off on treatment, with the expectation that more new HCV medications would become available and the price would drop.
In the cohort, 1948 patients received prescriptions for sofosbuvir, 781 for simeprevir, and 780 for both agents. This translates to 40% of patients receiving prescriptions for sofosbuvir and nearly everyone who received prescriptions for simeprevir was treated by the combination of the 2 agents. Although the combination of sofosbuvir and simeprevir was approved by the FDA in November 2014,24
the off-label use started right after the 2 agents were individually approved. The efficacy and safety of this combination was supported by the combination of simeprevir and sofosbuvir in a trial of patients with HCV infection,12
and may be attractive for patients who are unwilling or unable to undergo interferon-based regimens. The most recent American guideline includes the combined therapy as one of the recommended regimens.16
Nearly half the patients receiving the combination had other complex chronic conditions, such as depression (14.5%), liver cancer (2.9%), liver disease or cirrhosis (30.2%), and liver transplant (5.3%). These comorbidities were captured during the 12-month baseline, so it is possible that patients’ diseases information was not fully captured. It is also possible that patients developed new conditions during follow-up. Despite the potentially underestimated prevalence of comorbidities, it is clear many patients receiving the combination did not have any severe chronic conditions other than HCV. Interestingly, the payers seemed to cover costs of the combined regimen very well; the average OOP costs were only $100 higher for the combined regimen than the regimen that was not combined. In fact, the median cost was lower in the combined regimen ($300 vs $340).
Overall, the OOP costs were affordable for many of the patients who received the new therapies. The median cost for the ledipasvir/sofosbuvir regimen was only $112—much lower than any previous regimen. Despite the high wholesale prices, due to the effectiveness of treatment as well as the expected reductions in downstream costs associated with averted progression of liver disease, the new standard HCV treatments might be cost-effective in most patients.25-28
Therefore, some insurers might be willing to cover most of the costs. However, the OOP costs varied a lot among patients, and some patients had to pay approximately $35,000 to $76,000 for the new regimens. In this study, we did not use a multivariable regression model to assess the independent role of new medications on OOP costs. One reason is that new medications are available to broader patient populations than old regimens. Moreover, patients who received newer agents right after the medications became available are often those who had good insurance coverage.
One limitation of this study is that we could not fully capture patients’ past treatment history. A patient whose previous treatment failed might make decisions differently regarding whether to receive another treatment and which treatment to receive, compared with a treatment-naïve patient. Although we excluded patients who received treatment at baseline, it is possible some patients may have been treated prior to the baseline or prior to their enrollment in the health plan. However, the percentage of patients who underwent HCV treatment has been very low, especially before 2011, so there should be very few patients in our cohort who were previously treated.
Another limitation is that when estimating the OOP costs, we only considered the medication costs. We did not consider the OOP costs associated with interferon injection performed by providers, so we may have underestimated the costs of the interferon. However, only approximately 1% of claims were injections performed by health professionals. Additionally, the estimated OOP costs may not accurately reflect the costs per treatment regimen. Some patients may receive multiple treatments during follow-up, some may discontinue treatment, not comply with treatment, or their follow-up may end before finishing treatment. We conducted a sensitivity test that included only compliant patients who received a single treatment episode and finished at least 12 weeks of treatment. We excluded patients who received multiple regimens and who had a very short (<77days) or long treatment period (>365 days), and who had a proportion of days covered less than 80% (eAppendix Table 4). The estimated costs per regimen using this method were similar to the results in Table 3, except for ledipasvir/sofosbuvir. The median costs of ledipasvir/sofosbuvir increased from $112 to $180, and the mean increased from $1982 to $3030. Because ledipasvir/sofosbuvir became available in October 2014, patients who initiated treatment in November and December 2014 would not be able to finish treatment by the end of the year. Therefore, we did not capture the whole costs for some patients on ledipasvir/sofosbuvir.
The new HCV antiviral medications have been quickly adopted into practice; they are associated with increased treatment rates, especially among the elderly and patients with multiple chronic conditions that necessitate urgent treatment, but also make them intolerant or ineligible for previous interferon-based regimens. In this study, we did not include Viekira Pak, another new HCV medication approved by the FDA in December 2014. This new drug, developed by AbbVie Inc, has become a strong competitor of Harvoni and Sovaldi, which were developed by Gilead Sciences Inc. With more new medications available and increasing competition, the changes in treatment patterns and affordability will continue to evolve.