Objective This study evaluates the applicability of three overseas simplified assessment models–the Treatment Eligibility in Africa for the Hepatitis B Virus (TREAT-B), the Hepatitis B in Africa Collaborative Network (HEPSANET), and the HBeAg, Platelet count, ALT, and Albumin (HePAA) score–for guiding antiviral therapy initiation in chronic hepatitis B virus (HBV) carriers in China. The findings are expected to inform the development of a simplified assessment tool for antiviral therapy initiation tailored to chronic HBV carriers in China.

Methods The clinical data of 870 chronic HBV carriers who underwent regular follow-up by the Shandong Provincial Center for Disease Control and Prevention in Zhangqiu district, Jinan city, in November 2023 were retrospectively collected. Two hepatitis B specialists from a tertiary-level hospital assessed whether these chronic HBV carriers required antiviral therapy based on the Guidelines for the Prevention and Treatment of Chronic Hepatitis B (2022 Edition), and the assessment results served as the actual values. The simplified assessment models REAT-B, HEPSANET, and HePAA were used to determine whether chronic HBV carriers required antiviral therapy, with the results serving as predicted values. The actual values and predicted values were compared, and the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity of the three assessment models were calculated to evaluate their discriminatory ability.

Results Among the 706, 358, and 715 chronic HBV carriers who met the inclusion criteria for the TREAT-B, HEPSANET, and HePAA models, respectively, the number of carriers identified as eligible for antiviral therapy was 562 (79.6%), 282 (78.8%), and 571 (79.9%). The comparison between the predicted values and the actual values showed that REAT-B had the AUROC of 0.640 (95%CI: 0.597–0.683), the sensitivity of 26.3% (95%CI: 22.7%–30.2%), and the specificity of 92.4% (95%CI: 86.7%–96.1%). HEPSANET showed the AUROC of 0.568 (95%CI: 0.520–0.615), the sensitivity of 12.8% (95%CI: 9.1%–17.2%), and the specificity of 94.7% (95%CI: 87.1%–98.5%). HePAA showed the AUROC of 0.590 (95%CI: 0.553–0.625), the sensitivity of 22.8% (95%CI: 19.4%–26.4%), and the specificity of 92.4% (95%CI: 86.7%–96.1%). All the three models exhibited low accuracy in determining antiviral therapy eligibility for the HBV carriers in China.

Conclusions The predictive accuracy of the REAT-B, HEPSANET, and HePAA models is low in China, and thus they are not suitable for use in China at present.