Reply
Journal
Hepatology
Journal Volume
61
Journal Issue
3
Pages
1092-1093
Date Issued
2015
Author(s)
Abstract
Potential conflict of interest: Nothing to report. We thank Cui et al. for their letter asking several questions about our article.1 We appreciate the points raised in the letter. First of all, we agree with the concern over whether it is cost‐effective to conduct mass screening for hepatocellular carcinoma (HCC) in areas with a high HCC incidence. We are tempted to demonstrate whether mass screening for HCC with abdominal ultrasonography (AUS) is effective in mortality reduction, but our study cannot answer whether it is cost‐effective in comparison with other screening strategies such as two‐stage screening (biochemical tests first and AUS for positive subjects) used previously in Taiwan2 because, in addition to effectiveness, other factors affecting the results of cost‐effectiveness also include the underlying incidence, performance of the screening tool, the efficacy of treatments and therapies, and cost involved with screening modalities and also treatments and therapies. Formal cost‐effectiveness analysis is thus required. We had completed this work and submitted it to Hepatology with a status of under review. The results show mass screening for HCC with AUS is more cost‐effective than that with two‐stage screening. The reason for it being cost‐effective for mass screening for HCC with AUS is due to the underlying high incidence of HCC in Taiwan. This does not mean our proposed mass screening for AUS is also cost‐effective in other countries. The interpretation and the generalization of our findings to other places or settings from the viewpoint of economical appraisal should be taken with great caution. Although the concern of Cui et al. about the aspect of economic evaluation is appreciated, it would not affect the main conclusion of our finding on the effectiveness of AUS applied to a general population aged 45‐69 years in our Changhua community in Taiwan. It should be noted that although our inclusion criteria was according to risk score of biochemical results and diabetes mellitus, it was conducted on the grounds of efficient study design with sampling a large proportion of a high‐risk group (such as hepatitis B and C virus infection) and a small proportion of intermediate and low‐risk groups in order to have a high yield of detecting HCC and also to test the effectiveness of covering eligible residents rather than high‐risk groups in the whole community. These sampling proportions were reweighted while assessing the effectiveness of mass screening for HCC with AUS, which was mentioned in the text of article: “The detailed methodology, which takes into account different risk groups, is provided in the Supporting Material” (see the detailed formula in the e‐Appendix of the article). The effectiveness of mass screening for HCC with AUS should be representative of the underlying population but should not be focused on a certain risk group. Therefore, including subjects by risk groups according to the assignment of risk score does not mean the effectiveness of screening overrepresents high risk groups, as different sampling fractions were reweighted. The chance of having inclusion criteria with priority is because of our unique community‐based integrated screening (the biochemical tests were routinely checked for health check‐up rather than only for mass screening for HCC) before doing AUS. This does not mean mass screening for HCC with AUS in other places with a high incidence of HCC as in Taiwan should follow suit with our inclusion criteria (priority with risk score) and costs on biochemical tests would not be required if they offer mass screening for HCC with AUS directly. Second, we think an 80.6% attendance rate was high in a community‐based screening program in comparison with other breast and colorectal screening programs in Asian countries, as generally speaking the screening rate is low. Therefore, selection bias would not be serious. However, self‐selection bias due to the 20% nonparticipants has been adjusted in our evaluation method (see e‐Appendix formula). The reported 31% mortality reduction in the conclusion based on the comparison of cumulative mortality between the invited group and the uninvited group took into account selection bias. The point that the invited group included a majority of the high‐risk group but a minority of intermediate and low‐risk subjects was dealt with by weighting different proportions in the proposed methodology, as mentioned above. Third, regarding the technical aspects of AUS, the time interval between the time of US examination and the time of screening was within 3 months. This would not affect the detection rate in our program. Due to the outreach service, we used a portable US machine (GE LOGIQ). As far as operator reliability is concerned, although a sample size of 35 is too few for HCC and liver cirrhosis, the kappa for other related abnormalities was still acceptable. More important, the operators of performing an AUS in a local community setting were well trained by the two senior experts in gastroenterology (Dr. Tsung‐Hui Hu, the second author, and Dr. Chao‐Sheng Liao, the last author), both of whom have practiced AUS for over 20 years. We think that quality control for screening with AUS would not be a serious problem. Indeed, large samples would be suggested for better operator reliability in the future before conducting such a mass screening. Indeed, in our risk score assessment, we did not include nonalcoholic fatty liver disease or alcoholic steatohepatitis alcoholic, as we did not have such information. We understand this plays an important role in HCC cases resulting from those free of hepatitis viral infection. This may be considered in our Group B (free of hepatitis viral infection). In the current scenario, not including this factor may have little effect, as our HCC cases in Taiwan so far are still dominated by hepatitis viral infection. Finally, on the basis of evidence‐based medicine (EBM), we do agree with Cui et al. in pointing out that a randomized controlled trial (RCT) with long‐term follow‐up, evaluation of the sensitivity and specificity of US (several previous studies have been provided), and the aspect of an economic appraisal should be required before HCC screening with AUS is recommended. Our observational cohort study cannot replace the RCT but it provides a new insight into the possibility of considering mass screening for HCC with AUS in places with a high incidence HCC.
SDGs
Publisher
John Wiley and Sons Inc
Type
letter