Pharmaceutical industry payments and prescriptions of direct acting antiviral drugs for hepatitis C virus infection

Hepatitis C virus (HCV) is a serious viral infection causing chronic HCV infection. Chronic HCV infection is the leading cause of hepatocellular carcinoma and liver transplantation in the United States (US).¹ In 2020, over 100 000 new cases of HCV and approximately 15 000 HCV-related deaths were reported in the US.² The treatment landscape for HCV has dramatically improved since the first direct-acting antiviral (DAA) was approved in 2011, achieving an eradication response rate exceeding 95%.³ However, the expensive cost of DAA treatment imposes a substantial financial burden on the US healthcare spending and patients. The recent introduction of several newer, less expensive DAAs has resulted in heightened competition among pharmaceutical companies and increased marketing activities of DAAs to physicians including HCV guideline authors.^{4, 5}

Considering the competitive nature of the DAA market and findings from previous studies showing that the marketing payments to physicians are associated with increased prescriptions, we hypothesized that physician prescription patterns for DAAs would be associated with the payments from the pharmaceutical companies. Nevertheless, the associations of industry payments with physician DAA prescription practices had not been investigated.

This analysis used data from three publicly accessible databases: the Centers for Medicare & Medicaid Services Medicare Part D, the Open Payments, and the Physician Compare databases. We focused on four DAAs currently recommended for initial HCV treatment in the US³: ledipasvir/sofosbuvir (Harvoni from Gilead Sciences, approved in October 2014), elbasvir/grazoprevir (Zepatier from Merck, approved in January 2016), sofosbuvir/velpatasvir (Epclusa from Gilead Sciences, approved in June 2016), and glecaprevir/pibrentasvir (Mavyret from AbbVie, approved in August 2017). Physician demographic characteristics were extracted from the Physician Compare database. The publicly accessible Medicare Part D database contains physician identifiers reporting more than 10 claims of a drug in a single year to protect patients' privacy. To encompass physicians who could prescribe DAAs, we extracted data on the 30-day standardized number of claims (including refills), Medicare spending, and total number of supply days associated with each DAA reported by physician prescribers from the Medicare Part D database for the years 2014 to 2021. General payments related to the four DAAs were collected from the Open Payments Database for the same period. This study included only physician prescribers, as the Open Payments Database covered payments to physicians only. The Mount Sinai institutional review board exempted this study from formal review, since this study only collected data from publicly available information and met the definition of non-human participant research.

Associations between the receipt of payments associated with each DAA and the prescription of respective DAAs following each DAA's approval were evaluated using a multivariable logistic generalized estimating equation (GEE) model, adjusting for covariates at the individual physician-year level. Covariates included gender, primary specialty, years in practice, hospital affiliation, practice region, regional healthcare spending, and payment/claim year, as conducted in previous studies.^6-8 Additionally, we evaluated associations between the annual number of payments and the total annual number of claims (30-day standardized, including refills) and Medicare spending using a multivariable linear GEE model, adjusting for the same covariates. Payments were considered relevant if a physician received them from a company in the year the drug was prescribed, as conducted in previous research.^{7, 9} As there were no claims and payments for elbasvir/grazoprevir in 2021, we examined the associations for elbasvir/grazoprevir between 2016 and 2020.

In the sensitivity analyses, we also explored the associations between payments made one year prior to claims and prescribing patterns. To standardize payment duration, we excluded the payment data for the first year following each drug's approval from our sensitivity analysis, as the payment duration in the first year varied depending on each drug's approval date. For instance, payments for ledipasvir/sofosbuvir, approved in October 2014, were recorded only for the last three months of that year, whereas payments for glecaprevir/pibrentasvir, approved in August 2017, covered five months in 2017. Therefore, to ensure comparability across different drugs and years, our sensitivity analysis examined the associations between payments made in the year following the approval year and the prescribing trends in the year thereafter for ledipasvir/sofosbuvir (payments after 2015 and claims after 2016), sofosbuvir/velpatasvir (payments after 2017 and claims after 2018), and glecaprevir/pibrentasvir (payments after 2018 and claims after 2019).

In all these analyses, the independent variables were whether physicians received payments for each DAA (for the logistic GEE model) and the number of payments (for linear GEE models). The dependent variables were whether physicians prescribed each DAA (for the logistic GEE models) and the number of claims and Medicare spending (for the linear GEE models). The associations for each drug were examined separately in all analyses.

Of 5044 eligible physicians who prescribed more than 10 claims per year for any of the DAAs, 73.0% were gastroenterologists/hepatologists (Supplemental Material 1). Regarding the prescription patterns, more than 91.3% of the physicians prescribed ledipasvir/sofosbuvir after its 2014 approval (Table 1). In contrast, only 7.3% and 14.7% prescribed elbasvir/grazoprevir and glecaprevir/pibrentasvir, respectively. Of a total of $10.1 billion Medicare spending for the four DAAs, 84.0% ($8.4 billion) were for ledipasvir/sofosbuvir and 11.5% ($1.2 billion) were for sofosbuvir/velpatasvir. Annual spending for ledipasvir/sofosbuvir peaked at $4.3 billion in 2015 and $2.3 billion in 2016, subsequently decreasing to $136.5 million in 2019 (Supplemental Material 2).

Of all eligible physicians, 94.3% received at least one payment from the DAA manufacturers over the eight years (Table 1). Gilead Sciences made payments to 32.3% of physicians for ledipasvir/sofosbuvir, totalling $12.2 million, and to 35.6% of physicians for sofosbuvir/velpatasvir, amounting to $19.0 million. In contrast, 52.4% of physicians received payments for elbasvir/grazoprevir, totalling $5.1 million, and 61.5% received payments for glecaprevir/pibrentasvir, amounting to $11.3 million. Following the 2016 approval of elbasvir/grazoprevir, Merck compensated physicians $2.7 million in 2016 and $2.1 million in 2017; however, these payments were substantially reduced to $338 706 in 2018, with no payments reported after 2020 (Supplemental Material 3). A notable difference was observed in the payment patterns by the DAA manufacturers. Merck and AbbVie made over 93.2% and 92.3% of their payments as gift payments (Supplemental Material 4A and 4C). In contrast, Gilead Sciences allocated about one-third of their payments for ledipasvir/sofosbuvir and sofosbuvir/velpatasvir as compensation payments to a limited number of physicians (Supplemental Material 4B and 4D). The median (interquartile range) annual payments per physician were higher for ledipasvir/sofosbuvir ($141 ($105–$4638)) and sofosbuvir/velpatasvir ($136 ($194–$6353)) than those for elbasvir/grazoprevir ($37 ($18–$91)) and glecaprevir/pibrentasvir ($43 ($25–$116)).

Multivariable logistic GEE models revealed that physicians who received payments related to each DAA were significantly more likely to prescribe ledipasvir/sofosbuvir (odds ratio (OR): 1.81; 95% confidence interval (95% CI): 1.62–2.02, p < .001), elbasvir/grazoprevir (OR: 1.96; 95% CI: 1.57–2.43, p < .001), sofosbuvir/velpatasvir (OR: 2.24; 95% CI: 1.96–2.57, p < .001), and glecaprevir/pibrentasvir (OR: 1.68; 95% CI: 1.46–1.94, p < .001) compared with those who did not receive (Table 2 and Supplemental Material 5). There were significant positive associations between the number of general payments and both the number of claims and Medicare spending across the four DAAs. Furthermore, our sensitivity analyses also indicated significant associations between the payments made to physicians one year prior to their claims and their prescription patterns for all four DAAs (Table 2 and Supplemental Material 6).

This study found that the payments related to DAAs from the manufacturers to the physicians who frequently prescribed DAAs to Medicare beneficiaries were significantly associated with their prescribing patterns for DAAs in the US. More than $6.8 billion Medicare spending was reported for ledipasvir/sofosbuvir for 2014–2016, and less expensive alternatives such as elbasvir/grazoprevir and glecaprevir/pibrentasvir were underrepresented among the physician prescribers. Gilead Sciences made larger proportion of their payments for compensations to a limited number of physicians since ledipasvir/sofosbuvir approval, whereas the other two DAA manufacturers distributed their payments to more physicians for meals and education payments. Additionally, the number of these payments to the physician prescribers had a significant positive association with increased number of claims and Medicare spending for each DAA. These findings were consistent with previous studies.^7-10

This study has several limitations including the possibility that payment data reported by manufacturers may be inaccurate and that payments related to DAAs may have been underreported. Furthermore, due to the limitations of publicly accessible Medicare database, we could not include physicians who prescribed 10 or less claims for DAAs under Medicare Part D. The findings of this study might not be generalized to non-Medicare beneficiaries.

Anju Murayama: conceptualization; methodology; resource; software; formal analysis; investigation; writing – original draft; writing – review & editing; visualization; study administration. Keith M. Sigel: conceptualization; methodology; writing – review & editing; supervision. Elizabeth S. Tarras: conceptualization; methodology; writing – review & editing. Deborah C. Marshall: conceptualization; methodology; writing – original draft; writing – review & editing; supervision.

The authors did not receive any financial support from industry for this study.

Dr. Marshall was supported by the Office of the Director, National Institutes of Health under Award Number DP50D031876. The remaining authors declare that there are no conflicts of interest for this study.

The Mount Sinai institutional review board exempted this study from formal review, since this study only collected data from publicly available information and met the definition of non-human participant research.

No informed consent was required.

Not applicable.

Not applicable.

Not applicable.