Objective : To evaluate the risk for hepatocellular carcinoma (HCC) and the effect of interferon therapy on the incidence of HCC in patients with chronic hepatitis C by combining results from different studies in Japan.
Design : Literature-based meta-analysis
Methods : Thirteen follow-up studies for patients with chronic hepatitis C conducted in Japan were selected by systematic review of MEDLINE, EMBASE and manual searching. An unadjusted incidence rate ratio of HCC between treatment groups was calculated and an adjusted incidence rate ratio was estimated after adjustments using estimates for the degree of confounding from some of the studies.
Results : A total of 7 observational studies (1, 498 patients in the non-interferon group and 5, 451 patients in the interferon group) were included in this meta-analysis. The summary estimate of unadjusted incidence rate ratio was 0.37 (95% confidence interval [CI], 0.30 to 0.46). For 4 studies that had performed a multivariate analysis to adjust confounding factors such as sex, age, histological stage of hepatitis, the correction factor, which was estimated from the ratio of an unadjusted incident rate ratio to an adjusted incident rate ratio, was 0.64. The adjusted incidence rate ratio for incidence of HCC estimated using the correction factor was 0.58 (95% CI, 0.46 to 0.73) for interferon treatment to non-interferon treatment. After including all studies, the pooled 5-year estimated risk for HCC was 0.107 in the non-interferon group (n=2, 016) and the risk was 0.051 in the interferon group (n=6, 691).
Conclusion : Treatment with interferon reduces the risk for HCC in patients with chronic hepatitis C.

Objective : To examine the cost-effectiveness of interferon αcon-1 [consensus interferon (CIFN)] for chronic hepatitis C patients with genotype lb and high-titer.
Design : Cost-effectiveness analysis.
Methods : Data from a randomized clinical trial comparing the efficacy of CIFN to interferon-αn 1 (IFN-αn 1) for chronic hepatitis C patients were applied to a cohort simulation by Markov model to project lifelong clinical and economic outcomes from the payer's perspective. Natural history model and decision analytical model were built based on published literature and actual healthcare reimbursement data.
Results : From the randomized trial, sustained response proportion and biochemical response proportion were 16.7%and 18.2%for patients receiving CIFN, compared with 3.3%and 18.0%for IFN-αn 1, respectively. The simulation model showed that CIFN should prolong life expectancy by 0.4 year at negative incremental costs, compared to IFN-αn 1 strategy. Compared to no IFN strategy, CIFN should prolong life expectancy by 1.2year at an incremental cost-effectiveness ratio of ¥1, 320, 000 per life year gained. The results were robust, with CIFN remaining cost-effective in sensitivity analysis compared to IFN-αn 1 and no IFN treatment.
Conclusion : For chronic hepatitis C patients with genotype 1b and high-titer, CIFN should prolong life and be cost effective in comparison with IFN-αn 1 and no IFN treatment.

The fundamental issues which underlie research misconduct of Diovan case are 1）lack of data-quality control system managed by a competent data manager, 2）absence of trial statistician with sufficient knowledge of methodology (from all the aspects such as scientific, ethical and operating), 3）lack of standard operating procedures (SOPs) to secure blinding of assessors and independence of interim data analysis, which eventually allowed the investigators and statisticians to improperly check and edit crucial data in the midst of the trial. The Biometric Society of Japan, the community of Japanese biostatisticians, issued “The Biometric Society of Japan statement for clinical trials” , established “Statisticians standard of conducts” , and started the Accreditation for Trial Statistician from 2017. Many clinical trials sponsored by universities or research institutes (excluding investigator-initiated TIKEN) are not equipped with SOPs, adequate logistics, and clear responsibility. Moreover, a fair number of so-called “trial statisticians” only provide sample-size calculation and technical aspects of data analysis as statistical consultation. Academia should learn from Diovan case and pursue its role of educating creditable trial statisticians who bear social responsibility and establishing their social status.