Economic Burden of Hepatitis C-associated Diseases in the United States

From Journal of Viral Hepatitis
Economic Burden of Hepatitis C-associated Diseases in the United States
Source Medscape
A. C. El Khoury; W. K. Klimack; C. Wallace; H. Razavi

Posted: 03/14/2012; J Viral Hepat. 2012;19(3):153-160. © 2012 Blackwell Publishing

Discussion Only
Complete Text Available At Medscape
According to a database of drugs in development by Thomson Pharma,^[45] there are approximately 100 drugs in development to treat hepatitis C. Over the next decade, a number of new therapies will become available. A good understanding of the cost of hepatitis C sequelae is important to assess the value of the new treatments.

Top-down studies use national survey results to estimate the total cost attributed to chronic liver diseases, hepatitis C, and HCC. The NHDS is an annual survey conducted by the Center for Disease Control and Prevention that collects data on the characteristics of inpatients discharged from non-Federal short-stay (<30 days) hospitals in the United States.^[46] Data are collected from 239 hospitals and include demographic information and length of stay, as well as patients' diagnoses and procedures coded to International Classifications of Diseases (ICD). The NAMCS collects demographic data, reason for the visit, services provided, diagnostic procedures, diagnosis by ICD code, drugs prescribed, patient management, and planned future treatment from office based physicians.^[47] The survey includes results from 1600 responding physicians and about 42 000 patient visits. The NHAMCS gathers data from hospital emergency and outpatient departments as well as ambulatory surgery centres.^[47] All of the above surveys exclude Federal, military, and Veteran's administration hospitals. These surveys do have limitations. Extrapolation is carried out from a small sample. One case may be reported multiple times if it involves multiple visits, and they under-report cases that do not involve visits to hospitals, outpatient centres, or physician offices (e.g., veterans who have a higher prevalence of hepatitis C).^[48] However, despite these limitations, they have been used to estimate the number of cases and duration of treatment by disease.

The studies took different approaches for estimating the associated cost. Work by Leigh et al. ^[20] started with the total national expenditure of $1035 billion on medical care in 1996, portioned it to hepatitis C based on hospital days, and adjusted it by including assumptions for outpatient visits. The NIH report relied on nationally reported costs, Medicare costs, and audit sales data for prescription drugs.^[19] Hospital costs were estimated from the Healthcare Cost and Utilization Project (HCUP NIS), which collects data on hospital inpatient stays.^[49] It includes data from 37 states reporting approximately 8 million hospital stays each year. The data collected include demographic, admission and discharge status, length of stay, total charges and payment source as well as diagnosis using ICD codes. In the top-down approach, billed charges were used as a surrogate for the total cost of patient services. For hospital charges, cost-to-charge ratio from the Center for Medicare & Medicaid Services^[50] was used to calculate total cost. Medicare reimbursement rates were used to estimate physician charges. Prescription drug costs were estimated from audited data provided by Verispan.^[19] The American Gastroenterological Association (AGA) also published a study reporting the cost burden of digestive diseases, but their study was later updated by the NIH report.^[19,21]
 
All top-down studies relied on the same data set, and yet, there was considerable variability in the projected cost (e.g., $694–$1660 million for chronic hepatitis C as shown in Table 1). The large difference in the forecast was because of two factors: methodology and number of cases over time. Not all studies used the same methodology to estimate total number of cases or cost. For example, the study by Leigh et al. used the ICD code for liver diseases (and adjusted for percentage that can be attributed to hepatitis C) rather than the code for hepatitis C because they felt that hepatitis C cases were under-recorded by the hospitals. In their cost estimates, they also included health care expenditure cost such as construction of new hospitals, arguing that without hepatitis C, some portion of these new expenses would not be necessary. Overall, their estimated direct cost attributed to chronic hepatitis C ($1660) in 1997 was substantially higher than estimates by the NIH ($1065) in 2004.^[19,20] Top-down studies captured costs at a particular point in time. Comparison across years and studies is feasible for diseases whose population is steady over time. However, this was not the case with HCV infection and its sequelae. As an example, HCC cases doubled between 1985 and 1998, making comparison across years difficult.^[23]
 
Comparison across top-down studies was not possible. There were too many variables changing across studies to allow direct comparison. That said, the NIH report was the gold standard in this group of studies. It was the most recent analysis, the editor was involved in the earlier AGA study and would have adjusted for the shortcoming of the earlier study, and it provided the most robust methodology. It should be noted that top-down analyses captured total costs including prescription drugs. In the NIH report, 47% of the annual cost associated with chronic hepatitis C was attributed to drug costs.^[19] Top-down studies are based on a set number of cases at a point in time. Coupled with an epidemiology study, the bottom-up analysis lends itself more readily for estimating the cost of the disease burden in the absence of treatment or the economic benefits of new treatment.
There was considerably more consistency among the bottom-up analyses. However, further research into the origins of the assumptions found that nearly all studies were using data reported by Bennett et al. in 1997.^[8] In this original study, the incremental cost (cost for one additional patient) was assessed in a study of 126 hospital patients at the University of Florida. Outpatient costs were estimated by applying an assumed cost/charge ratio to fee schedules and wholesale medication costs. A panel provided frequency estimates for services and medications. The fixed costs (cost of purchasing new building, equipment, etc.) were not considered. On an individual patient level, this is a safe assumption. However, at an infected population level, this approach will underestimate the total cost as the incidence of morbidities increases. Finally, this study reported the variable cost and diagnosis-related group cost. Nearly all subsequent reports used the variable cost only, which again underestimated the total cost associated with the sequelae. A couple of investigators did develop independent cost analyses, which were substantially different. A study by Younossi et al. ^[26] used Medicare fee schedules and physician assessment of cost frequency to estimate the cost of decompensated cirrhosis ($585 per year in 2010 US dollars vs$1110 from Bennett et al.). However, the same study used cost estimates from the Bennett study for HCC and liver transplantation. Kim et al. ^[29] used HCC annual costs from a 1994 NIH report based on a top-down approach. The top-down cost was $23 755 per year in 2010 US dollars vs$44 200 from Bennett et al. The cost of liver transplantation was studied by many authors.^[16,30–43] A recent meta-analysis by van der Hilst et al. consolidated all previous work to report a total cost of $201 110 (2010 US dollar) per liver transplant with a 95% confidence interval of $178 60–$223 460 (2010 US dollar). They showed that the cost of liver transplantation was 57% higher in the United States as compared to the Organization for Economic Cooperation and Development (OECD) countries. This analysis is considered the gold standard owing to its wide scope and the robustness of the methodology. It excluded the work by Bennett et al., which estimated at $174 935 (2010 dollar), lower than the range provided by the meta-analysis. The cost in subsequent years was reported by two studies:^[8,44] mean of $37 535 per year in 2010 dollars with range of $30 550–$46 750. Again, the cost estimate by Bennett et al. ($30 550 per year) was at the lower end of the range. The higher end of the range was defined by change in cost liver transplantation over the years.

Our analysis highlights the need for more updated cost studies. The historical cost burden analyses relied on data that were over 10 years old, and may not reflect the current clinical practice. The old data most likely underestimate the current costs even after adjustment for inflation. As an example, the variable cost associated with mild chronic hepatitis was estimated at $146 per patient per year in 2010 dollars.^[8] A more recent study reported that, excluding the outpatient visit, the cost of HCV antibody testing and HCV RNA PCR testing was $107 and $300, respectively, in 2006.^[51] Future studies could leverage insurance claim data, which have been shown to be an important source to reflect current standard practices.^[52,53]
 
Finally, the cost burden of hepatitis C also has an indirect cost component. A number of studies estimated the indirect cost defined as the cost of forgone earnings or production owing to hospitalization, ambulatory care, work loss owing to acute or chronic infection, and premature death.^{[19–21,24,29]} In most cases, the indirect cost was higher than the projected direct expenses. The total annual indirect cost was estimated at $51–$3370 million in the reported year (see Table 1). At the lower end of the range, only productivity loss owing to time away from paid labour was considered.^[21,29] At the high end of the range, production loss owing to early death was added. The most recent study for indirect cost was published by the NIH, which estimated the indirect cost at $2847 in 2004.^[19] Hospital and ambulatory data were used to estimate foregone earnings. Loss owing to early death was estimated using forgone lifetime earnings to age 75. Overall, hospital stays, ambulatory care, and mortality accounted for $46.6, $51.2, and $1685.7 million respectively in 2004. However, this technique underestimated the cost associated with work and leisure loss that did not result in hospitalization or physician visits. Using a completely different approach, Su et al. ^[24] estimated the indirect cost associated with disease-related absence using data from multiple large employers in the United States by comparing record of employees with and without HCV. They attributed an increase of $490 (2007 dollar) per person per year in indirect costs to HCV infection.

Over the next decade, a number of new therapies will become available to treat hepatitis C. The debate over the cost effectiveness of these new therapies will depend on the comparison of no treatment, treatment with pegylated interferon and ribavirin, and treatment with new therapies/combinations. This review reported all relevant cost studies in the United States by different categories: top-down, bottom-up, direct, and indirect. The top-down analyses could not be compared directly with bottom-up estimates, as the former represented the cost and duration and number of hospital and physician visits in a particular year while the latter focused on cost per year for each sequelae. The indirect cost, cost associated with loss of productivity and early death, can be substantially higher than direct costs—cost of physician visits, hospitalization, diagnosis, and treatment. Although bottom-up estimates are more appropriate for scenario-based cost burden analysis, most sequelae costs date back to analyses in the 1990s. Updated cost analyses would help stakeholders make informed decisions when choosing among future therapeutic alternatives.