Disability-adjusted Life Years of Hepatitis B in Iran during 2009–2019: An Analysis Based on the Global Burden of Disease Study 2019

2.1. Study Design and Setting

The present study is a type of ecological study that was conducted using the Exploratory mixed design method. Exploratory means that these types of studies investigate and discover health-related phenomena, and mixed means that in these studies, the factor of location (multiple-group study) and time (time-trend study) are both evaluated and measured together [13]. In the present study, the amount of DALYs attributed to hepatitis B has been evaluated over time and in different places in Iran.

2.2. Sources of Data Collection and Calculation of DALYs

The Burden of diseases study is the most comprehensive and accurate global epidemiological study. it is the result of examining 359 diseases and injuries and 84 health risk factors in 195 countries and regions of the world (including developed and developing countries). Information about the burden of disease index, years of life lost due to premature death, and years of life lost due to disability are freely available to researchers on the Global Burden of Disease website.

2.3. Disability-adjusted Life Years (DALY)

DALY is a type of health gap index that calculates the years of life lost, either due to premature death or due to non-fatal diseases. This index was defined and used in the Global Burden of Diseases and Injuries (GBD) study to calculate the burden of diseases [14, 15].

Location and description of the study site: Iran is a country in Southwest Asia with an area of 1,648,195 km2 located in the Middle East region. According to the latest World Meter statistics, the population of Iran in June 2020 is estimated at 8,618,234 people, which is the 18th most populous country in the world. Iran lies between 25° 3′ and 39° 47′ N in latitude and between 44° 5′ and 63° 18′ E in longitude. The geographical location of Iran is shown in Fig. (1).

2.4. Analysis

The statistical analysis consisted of three main parts: 1- descriptive analysis, 2- analysis of the time trend of DALYs attributed to hepatitis B from 2009 to 2019, 3- spatial pattern analysis and identification of high-risk and low-risk clusters of DALYs attributed to hepatitis B in the provinces of Iran.

2.5. Time Trend Analysis

To check the time trend, first, the total and average DALYs attributed to hepatitis B during each of the years 2009 to 2019 were obtained, then the changes in the time trend were obtained using Joinpoint regression analysis and the latest version of Joinpoint Regression Program version 4.7.0.0 0 - February 26, 2019, was evaluated. In the joint point regression analysis model, the total and average DALYs (per 100,000 people) were considered as the dependent variable and the year as the independent variable. Finally, considering that the dependent variable followed the Poisson distribution, the natural log-linear model was used to obtain the values of Annual Percent Change (APC) and Average Annual Percentage Change (AAPC). Also, the Grid Search method was used to identify join points, and the Monte Carlo method was used to estimate the p-value with 4499 repetitions. Bayesian information criterion (BIC) and Sum of Squared Errors (SEE) were used to select the best model (model selection) that had goodness of fit [16, 17].

2.6. Spatial Pattern Analysis and Cluster Identification

In the next step, to evaluate the spatial pattern, the number of DALYs attributed to hepatitis B was obtained in each of the provinces during the years 2009 to 2019, and then the zoning of DALYs in Iran was drawn using ArcGIS 10.3 environment. To evaluate the degree of spatial autocorrelation (Spatial Autocorrelation) of the disease, the Global Moran's I index was used. The range of this statistic is from +1 to -1, and the higher the value of this statistic goes to +1, it indicates that it is clustered and Anything that moves towards -1 indicates dispersion of the desired phenomenon [18].

Kuldorff's purely spatial scan statistic embedded in the SaTScan software was used to identify the exact location of primary and secondary high-risk and low-risk clusters and also to estimate the relative risk of DALYs in each of Iran's provinces. Due to the numerical and discrete nature of the number of DALYs in each province, discrete Poisson distribution and a circular window were used to detect clusters. The window whose Log-likelihood ratio (LLR) was maximized was reported as the primary or most likely cluster (a cluster that has a low probability of occurring by chance) and other windows with lower LLR were reported as the Secondary cluster. (a cluster that is more likely to have occurred by chance) [19, 20]. The statistical significance of clusters was obtained by Monte Carlo simulation with a maximum number of 999 repetitions. The relative risk in Kuldorff's spatial scan statistic is estimated using the following equation.

where c is the number of DALYs observed inside the window, E[c] is the expected number of DALYs inside the window, C-c is the number of DALYs observed outside the window, and E[C]-E[c] is the expected number of DALYs outside the cluster. In this study, for Determination of the optimal radius window for spatial cluster analysis, we used the Gini Optimized Cluster Collection method, which has recently been shown that the Gini coefficient can identify several non-overlapping clusters in a more pure, local and more realistic [21, 22]. The optimal radius window in this study was 10%. Finally, the relative risk of DALYs attributed to hepatitis B in all provinces of Iran was estimated and its zoning was drawn in the GIS environment.

3.1. Results of Spatial Autocorrelation and Cluster Identification

Global Moran's I index value of 0.298 was obtained, which indicates high autocorrelation and a high tendency of DALYs to cluster in Iran. (P-value = 0.024). The results of detecting high-risk and low-risk clusters by the Gini-optimized method are shown in Fig. (3). As can be seen, a total of 5 important spatial clusters were discovered, which included 2 high-risk clusters and 3 low-risk clusters.

The full epidemiological, statistical, demographic, and geographic characteristics of the clusters discovered by Gini Optimized are shown in Table 2. According to the results of this table, it can be said that the most likely high-risk cluster with LLR=238.59 is located in the west of Iran (centered at 33.5 N & 48.8 E), which has a radius of 192.03 km and includes the provinces of Lorestan, Kermanshah, Hamadan, Markazi, Khuzestan. It was Ilam (cluster 1 in Fig. 3). In this cluster, the relative risk of incidence is estimated to be 2.31, which means that the amount of DALYs inside this cluster is 2.31 times that of the areas outside the cluster. Also, a secondary high-risk cluster with LLR=167.63 was identified in northeastern Iran, which included North Khorasan and Golestan provinces. On the other, the most probable low-risk cluster with LLR=218.03 was discovered in the northwestern part of Iran (centered at 36.5 N & 48.4 E), which had a radius of 157.8 km and included the provinces of Zanjan, Gilan, and Kurdistan (cluster 3 in Fig. 3). The relative risk in this cluster was estimated to be 0.23, which means that the amount of DALYs inside the cluster is lower than outside the cluster and 0.23 times. Two secondary low-risk clusters were identified in the central and northern regions of Iran (clusters 4 and 5 in Fig. 3), in which the relative risk was estimated at 0.31 and 0.45, respectively.

The relative risk of DALYs in all provinces was estimated using Kuldorff's spatial scan statistic, and finally, to illustrate the estimated RR values, a dotted map according to the estimated RR at the level of different provinces was produced in GIS (Fig. 4). As can be seen, these results were completely consistent with the results of the spatial distribution of DALYs in the provinces of Iran, such that high RR was observed in provinces with high DALYs, and low RR was observed in provinces with low DALYs. The three provinces with the highest relative risk include the provinces of Lorestan (RR=2.22), Sistan and Baluchistan (RR=1.73), Golestan (RR=1.46), and the three provinces with the lowest relative risk include the provinces of Gilan (RR=0.41), Mazandaran (CIR= 0.48) and Kurdistan (CIR=0.51).

3.2. Time Trend Results

The results related to the evaluation of the time trend of total and average DALYs during the years 2009 to 2019 by Joinpoint regression are shown in Figs. (5 and 6). Fig. (5) shows the time trend of total DALYs by year. According to the obtained results, the time trend of total DALYs has decreased significantly between 2009 and 2015 and decreased by 2.08% annually (p<0.05). But in the continuation of this time trend, a join point was observed in 2015, and the total DALYs at this time point have undergone a noticeable change. So, the total number of DALYs decreased between 2015 and 2019 with a lower slope and intensity (APC=-0.61%). In general, during the years 2009 to 2019, the total DALYs decreased by 1.5% annually (p<0.05). More information about this time trend is shown in Table 3.

Fig. (6) also shows the time trend of average DALYs by year. As can be seen, the time trend of the average DALYs has decreased significantly between 2009 and 2015 and decreased by 2.08% per year (p<0.05). But in the continuation of this time trend, a join point was observed in 2015, and the average DALYs between 2015 and 2019 decreased with a lower slope and intensity. (APC = -0.62%) In general, during the years 2009 to 2019, the average DALYs Annual average (AAPC) had a significant decrease of 1.4% (p<0.05). More information about this time trend is shown in Table 2.

4.1. Limitations

This study suffers from the general limitations of GBD studies. The availability and quality of primary data, which are the foundation of the GBD analysis, is the main limitations of GBD estimates, particularly in regions with countries that have poor completeness. Ecological studies have substantial limitations and, although valid for hypothesis generation, are not considered useful for hypothesis testing because of unmeasured and uncontrolled confounding.

4.2. Future Direction

Quantifying indicators related to the health of society requires indicators that are the losses caused by premature deaths. For any reason, put together the burden caused by the disabilities that arise during the course of the disease and its various consequences. Data and express them based on a common unit, and on the other hand, the level of health and its distribution in the society in the form of an index. Therefore, studies of the burden of diseases are one of the most important needs to determine the current situation. The effectiveness of policies and health programs and the prioritization of interventions are considered.