The global burden of chronic hepatitis B virus infection: comparison of country-level prevalence estimates from four research groups

Abstract

Background: Progress towards viral hepatitis elimination goals relies on accurate estimates of chronic hepatitis B virus (HBV)-infection prevalence. We compared existing sources of country-level estimates from 2013 to 2017 to investigate the extent and underlying drivers of differences between them.

Methods: The four commonly cited sources of global-prevalence estimates, i.e. the Institute for Health Metrics and Evaluation, Schweitzer et al., the World Health Organization (WHO) and the CDA Foundation, were compared by calculating pairwise differences between sets of estimates and assessing their within-country variation. Differences in underlying empirical data and modelling methods were investigated as contributors to differences in sub-Saharan African estimates.

Results: The four sets of estimates across all ages were comparable overall and agreed on the global distribution of HBV burden. The WHO and the CDA produced the most similar estimates, differing by a median of 0.8 percentage points. Larger discrepancies were seen in estimates of prevalence in children under 5 years of age and in sub-Saharan African countries, where the median pairwise differences were 2.7 percentage and 2.4 percentage points for all-age prevalence and in children, respectively. Recency and representativeness of included data, and different modelling assumptions of the age distribution of HBV burden, seemed to contribute to these differences.

Conclusion: Current prevalence estimates, particularly those from the WHO and the CDA based on more recent empirical data, provide a useful resource to assess the population-level burden of chronic HBV-infection. However, further seroprevalence data in young children are needed particularly in sub-Saharan Africa. This is a priority, as monitoring progress towards elimination depends on improved knowledge of prevalence in this age group.

Keywords: Hepatitis B; disease burden; indicator; infectious diseases; modelling; monitoring; prevalence; sub-Saharan Africa; viral-hepatitis elimination.

Figure 1.

Distribution of country-level estimates of chronic HBV infection prevalence (A) across all ages and (B) in children under 5 years of age, from the Institute for Health Metrics and Evaluation (IHME), Schweitzer et al, the World Health Organization (WHO) and the CDA Foundation (CDA). Dots represent country-specific estimates spread according to the density distribution of the data, for the 112 (A) and 72 (B) countries covered by the four groups.

Figure 2.

Within-country variation in estimates from different groups (A) for chronic HBV infection prevalence across all ages and (B) for chronic HBV infection prevalence in children under 5 years of age. Within-country variation is represented by the mean absolute deviation (MAD), and categories show the 25th, 50th, 75th and 90th percentile of MAD values. White shading represents countries where the MAD could not be calculated because less than two groups provided estimates.

Figure 3.

Factors contributing to differences in estimates of chronic HBV infection prevalence in sub-Saharan Africa. (A) shows the distribution of pairwise relative differences between estimates across all ages for different characteristics of the underlying empirical data, for (left) comparisons of estimates from the World Health Organization (WHO) with the Institute for Health Metrics and Evaluation (IHME), and (right) of WHO with Schweitzer estimates. (B) shows the prevalence ratio of estimates across all ages to estimates in children under 5 years of age by research group, which reflects the modelled age distribution of prevalence. The number of country-specific estimates represented in each category are: (A) 21 with no recent empirical data and 19 with recent empirical data for each comparison, 9 with no empirical data for WHO-IHME and 2 for WHO-Schweitzer, (B) 49 for IHME and WHO, 25 for the CDA Foundation (CDA). The 9 countries with no empirical data underlying the WHO estimate are Botswana, Chad, Comoros, Djibouti, Guinea-Bissau, Lesotho, Mauritius, Sao Tome and Principe, Swaziland.