Trends in Viral Hepatitis Cost-Effectiveness Studies
Published Online: December 18, 2012
Sammy Saab, MD; Young Choi, BS; Harman Rahal; Kun Li, MS; and Myron Tong, MD
Over the past 10 to 15 years, there have been an increasing number of treatment strategies for hepatitis B and C. Overall, every generational succession in viral hepatitis treatment has been associated with incremental increase in efficacy and occasionally costs.1-3 Along with the introduction of new treatment strategies, cost-effectiveness studies have been used to explore whether these strategies compare favorably with other therapies on an economic basis. Cost-effectiveness studies are often used to justify a more expensive treatment or diagnostic strategy on the basis that its incremental increase in costs is within an arbitrary threshold ratio. In the United States, a benchmark of $50,000 per quality- adjusted life-year (QALY) is widely used as the upper limit of cost-effectiveness.4
A concern with cost-effectiveness studies is that there is a tendency to publish favorable results.5-7 This concern has been discussed with respect to a variety of disease categories, including cardiovascular, endocrine, infectious, musculoskeletal, orthopedic, neoplastic, and neurologic/psychiatric diseases.8-11 Factors shown to potentially bias publication include the country where the study was performed12,13 and whether the study was sponsored by a pharmaceutical company.8,9,11,14,15
Viral hepatitis is a medical and public health concern, with continually evolving treatment and diagnostic strategies. Given the presence of bias toward publishing studies with favorable results, we sought to determine whether there was a trend toward publishing studies that show the cost-effectiveness of new treatments and/or diagnostic strategies. Our hypothesis was that there would be a trend to publish studies that demonstrated a favorable cost-effectiveness result. In addition, we examined the predictors of publishing favorable results.
Two researchers independently conducted a systematic literature search of PubMed and the Tufts Medical Center Cost-Effectiveness Analysis Registry16 to find all original cost-effectiveness studies regarding the antiviral treatment of hepatitis B or hepatitis C (YC and HR). The search terms were cost-effectiveness, antiviral drug, hepatitis B treatment, and hepatitis C treatment. All articles were published in English between 1987 and June 2011 (Figure 1).
Cost-effectiveness ratios were recorded from the main conclusion of each study. If there was more than 1 result in the study’s conclusion, the mean value was calculated. Studies that did not report their cost-effectiveness analyses in dollars per QALY were identified as “not reported in dollars per QALY.” The conclusion of the author(s) about whether or not the study showed cost-effectiveness was documented as yes or no. Studies with conclusions stating both cost-effectiveness and non–cost-effectiveness were identified as yes/no but were included in the “no” category to bias against our hypothesis. All non-US currencies were converted to US dollars at the exchange rate of the year of publication.17
The source of funding for each study was documented as not reported (no mention of funding), private industry (partial or complete funding by a private industry), academic (government or public agency funding to an academic institution), or other (article explicitly mentioned that funding was neither nonindustry nor nonacademic).
Each article was judged on its quality using criteria for grading cost-effectiveness studies.18 Whether or not the article disclosed any conflict of interest was also documented as yes or not reported. The number of sensitivity analyses presented in each study was recorded, as well as the country of origin, the year of publication, and the journal impact factor at the time of publication.19
The corresponding authors of all the studies were contacted by e-mail and were e-mailed up to 3 times if no response was received. Alternative e-mails for the corresponding authors were also used in the event of no response. The authors were asked whether they (1) had conducted cost-effectiveness studies that they submitted but were never published (whether favorable or unfavorable) and (2) had conducted cost-effectiveness studies that they never submitted (whether favorable or unfavorable).
We systematically analyzed reported cost-effectiveness ratios and the conclusions of the author(s) from each study to examine whether there was any bias between positive predictors and negative predictors. We distinguished positive predictors from negative predictors using 2 methods. With the first method we directly cited the conclusion of the author(s) about whether the study showed favorable costeffectiveness. An answer of yes indicated positive predictors, while an answer of no implied negative predictors. Of the 77 studies, 71 implied favorable cost-effectiveness ratios. With the second method, we defined $50,000/QALY as a threshold by using reported cost-effectiveness ratios from each study. The reported cost-effectiveness ratios below the threshold were treated as positive predictors, while the rest were considered negative predictors. We also explored whether any study characteristic predicted the reported cost-effectiveness ratio. The reported cost-effectiveness ratios were directly used as our outcome instead of authors’ conclusions. For those studies with more than 1 number reported, the mean of multiple numbers was taken and used for our data analysis. The reports without any number were excluded.
We used a binomial test to examine the existence of bias among all the published papers, with the null hypothesis being that there is no bias between positive and negative predictors. The nonparametric Kruskal-Wallis test was used to determine whether each study characteristic had an effect on the reported cost-effectiveness ratio. We used SAS statistical software version 9.2 (SAS Institute Inc, Cary, North Carolina) for all analyses. P values were all 2-sided with significance set at .05.
PDF is available on the last page.