[Skip to Navigation]
Sign In
Invited Commentary
Public Health
September 3, 2020

Gaps in Descriptive Epidemiology and Hepatitis C Virus Modeling Research

Author Affiliations
  • 1Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia
  • 2Department of Epidemiology and Biostatistics, School of Public Health, Georgia State University, Atlanta
JAMA Netw Open. 2020;3(9):e2016120. doi:10.1001/jamanetworkopen.2020.16120

There are an estimated 2.4 million people chronically infected with hepatitis C virus (HCV) in the United States, yet just more than half are aware of their infection.1 While most prevalent chronic HCV infections are still disproportionately concentrated among adults born between 1945 and 1965 (known in HCV epidemiology as the birth cohort), increasing HCV incidence among adults younger than 40 years as a result of injection drug use is concerning. Despite these troubling trends in HCV incidence, advancements in HCV treatment have resulted in several well-tolerated and highly successful oral therapy regimens that have the potential to significantly affect the current trajectory of the HCV epidemic. From a public health perspective, the underdiagnoses of prevalent infections and increasing incident infections, coupled with the availability of curative treatment, indicate that more HCV screening is needed. The recent paper by Tatar et al2 used a decision-analytic framework to assess the cost-effectiveness of universal screening of adults and targeted screening of people who inject drugs (PWID) for HCV infection. The authors present incremental cost-effectiveness ratios (ICERs), which represent the estimated cost per additional life-year gained, as their summary measure of interest. To evaluate ICERs, researchers typically define a willingness-to-pay threshold a priori, and resulting ICERs lower than the willingness-to-pay threshold are deemed cost-effective. Assuming HCV prevalence is 1% in the general population and 60% among PWID, the authors conclude that a targeted screening of PWID is a cost-effective approach (ICER, $45 465) to combat HCV infection in the United States, but universal screening was not cost-effective (ICER, $291 277).2

This study is part of a growing body of literature on economic evaluations of HCV screening programs, many of which informed a recent update to the HCV screening guidelines issued by the US Centers for Disease Control and Prevention (CDC). Prior to this update, the CDC recommended HCV screening for persons with known risk factors (including injection drug use) and all persons born between 1945 and 1965. In April 2020, the CDC updated previous guidance to include a 1-time HCV screening of all adults older than 18 years, except in settings in which HCV prevalence is less than 0.1%.3 Cost-effectiveness evidence cited in the CDC’s April 2020 report includes a cost-effectiveness analysis (CEA) comparing universal adult screening with screening the birth cohort, with an assumed HCV prevalence of 0.84% (ICER, $28 000); a similar CEA with an assumed HCV prevalence of 2.6% and 0.29% for birth cohort members and non–birth cohort members, respectively (ICER, $11 378); a CEA of a 1-time screening for persons aged 20 to 69 years (assumed HCV prevalence, 1.6%) compared with risk-based screening (ICER, $7900); and a CEA of programmatic variations of a 1-time HCV screening strategies among persons aged 15 to 30 years (resulting ICERs ranged from cost-saving to $71 000).

The CEA framework used by Tatar et al2 and others is a key component in the development of public health recommendations and policies. Economic analyses permit hypothetical comparisons of the potential costs and outcomes of public health interventions under specified assumptions defined by the model structure and quantitative inputs. As is the case with all model-based analyses, model outputs and their usefulness are dependent on the methods and assumptions used by the researchers. Public health decision-makers are encouraged to compare both economic and health outcome results across multiple scenarios and studies. To facilitate comparability across studies, the Panel on Cost-effectiveness in Health and Medicine endorses the use of a reference case, which establishes a standard set of methodologic practices across studies (eg, societal vs health care perspective, analytic horizon, summary outcomes, and so on).4 However, comparability of the epidemiologic representation of a health condition can be difficult to achieve or assess.

Often, the most influential inputs in cost-effectiveness models of public health interventions are basic epidemiologic measures of disease frequency and related assumptions. Systematic, accurate, and timely estimates of prevalence, incidence, and mortality are essential for describing population-level burden of disease and are also foundational parameters for modeling potential outcomes of policies and interventions. For many infectious diseases, these fundamental epidemiologic measures are estimated from surveillance data. Economic and epidemiologic modeling studies aiming to forecast the potential consequences of a public health policy or intervention on future infectious disease transmission, in particular, require inputs for transmission probabilities, which may also be estimated using surveillance data. For example, in the United States, a multisystem surveillance platform for HIV yields data that allow for robust estimation of HIV transmission probabilities.5 In contrast, as evidenced by the varying assumptions used in HCV transmission and CEA models, these measures are not reliably available or consistently used for HCV.

First, estimates of HCV incidence are lacking, primarily as a result of insufficient, underresourced disease surveillance. New HCV infections are reported as part of the CDC National Notifiable Disease Surveillance System, but most new infections go unreported for etiological reasons (eg, many cases are asymptomatic), differences in state laboratory testing requirements, inconsistent application of case definitions, and a lack of surveillance resources. To correct for this, published surveillance estimates and resulting model inputs are regularly adjusted using a single multiplication correction factor that is oversimplified and outdated.6 Furthermore, national HCV prevalence estimates are often based on data from large national serosurveys, but updates are reliant on long delays between survey cycles, and model-based extensions are required to try to account for high-risk populations not included in the sampling frame or to identify geographic variation.7 Estimates for specific high-risk populations (eg, PWID) are even less reliable. Many measures of disease frequency and transmission probabilities among PWID come from cohort studies that lack generalizability and have suboptimal retention rates.

As the HCV epidemic continues to change, more reliable, timely, and granular data inputs are needed to accurately capture epidemic dynamics and to facilitate more standardized modeling of potential outcomes of interventions. In the meantime, HCV cost-effectiveness researchers should continue to present, and emphasize, a wide range of scenarios and sensitivity analyses on the basic epidemiologic measures that inform their models. When evaluating HCV cost-effectiveness research, policy makers should critically assess epidemiologic model assumptions and evaluate the generalizability to their population of interest.

Back to top
Article Information

Published: September 3, 2020. doi:10.1001/jamanetworkopen.2020.16120

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Hall EW et al. JAMA Network Open.

Corresponding Author: Eric W. Hall, PhD, MPH, Department of Epidemiology, Rollins School of Public Health, Emory University, 1518 Clifton Rd, Atlanta, GA 30322 (eric.w.hall@emory.edu).

Conflict of Interest Disclosures: None reported.

Hofmeister  MG, Rosenthal  EM, Barker  LK,  et al.  Estimating prevalence of hepatitis C virus infection in the United States, 2013-2016.   Hepatology. 2019;69(3):1020-1031. doi:10.1002/hep.30297Google ScholarCrossref
Tatar  M, Keeshin  SW, Mailliard  M, Wilson  FA.  Cost-effectiveness of universal and targeted hepatitis C virus screening in the United States.   JAMA Netw Open. 2020;3(9):e2015756. doi:10.1001/jamanetworkopen.2020.15756Google Scholar
Schillie  S, Wester  C, Osborne  M, Wesolowski  L, Ryerson  AB.  CDC recommendations for hepatitis C screening among adults—United States, 2020.   MMWR Recommen Rep. 2020;69(2):1-17. doi:10.15585/mmwr.rr6902a1Google ScholarCrossref
Sanders  GD, Neumann  PJ, Basu  A,  et al.  Recommendations for conduct, methodological practices, and reporting of cost-effectiveness analyses: second panel on cost-effectiveness in health and medicine.   JAMA. 2016;316(10):1093-1103. doi:10.1001/jama.2016.12195PubMedGoogle ScholarCrossref
Skarbinski  J, Rosenberg  E, Paz-Bailey  G,  et al.  Human immunodeficiency virus transmission at each step of the care continuum in the United States.   JAMA Intern Med. 2015;175(4):588-596. doi:10.1001/jamainternmed.2014.8180PubMedGoogle ScholarCrossref
Klevens  RM, Liu  S, Roberts  H, Jiles  RB, Holmberg  SD.  Estimating acute viral hepatitis infections from nationally reported cases.   Am J Public Health. 2014;104(3):482-487. doi:10.2105/AJPH.2013.301601PubMedGoogle ScholarCrossref
Rosenberg  ES, Rosenthal  EM, Hall  EW,  et al.  Prevalence of hepatitis C virus infection in US States and the District of Columbia, 2013 to 2016.   JAMA Netw Open. 2018;1(8):e186371. doi:10.1001/jamanetworkopen.2018.6371PubMedGoogle Scholar