Faculty of Health Sciences, Curtin University, Perth, Western Australia, Australia.
Wesfarmers Centre of Vaccines and Infectious Diseases, Telethon Kids Institute, Perth, Western Australia, Australia.
PLoS One. 2019 Aug 9;14(8):e0221027. doi: 10.1371/journal.pone.0221027. eCollection 2019.
Tuberculosis (TB) remains a major health problem worldwide and in Ethiopia. This study aimed to investigate the spatial distributions of notified TB over the whole territory of Ethiopia and to quantify the role of health care access, environmental, socio-demographic, and behavioural factors associated with the clustering of TB.
A spatial analysis was conducted using national TB data reported between June 2016 and June 2017 in Ethiopia. Spatial clustering of TB was explored using Moran's I statistic and the local indicator of spatial autocorrelation (LISA). A multivariate Poisson regression model was developed with a conditional autoregressive (CAR) prior structure and with posterior parameters estimated using Bayesian Markov chain Monte Carlo (MCMC) simulation with Gibbs sampling to investigate the drivers of the clustering.
A total of 120,149 TB cases were reported from 745 districts in Ethiopia during the study period; 41,343 (34%) were bacteriologically confirmed new pulmonary TB and 33,997 (28%) were clinically diagnosed, new, smear-negative pulmonary TB patients. The nationwide annual incidence rate of notified TB was 112 per 100,000 population. The highest incidence was observed in three city administrative regions, namely Dire Dewa (348 cases per 100,000 population), Addis Ababa (262 per 100,000 population) and Harari (206 per 100,000 population), and the lowest incidence was observed in Somali region (51 per 100,000 population). High-high spatial clustering of notified TB was detected at Humera, Gog, and Surima district, and low-low clustering was detected in some districts located in the Somali region. Poor health care access (IRR = 0.78; 95%CI: 0.66, 0.90) and good knowledge about TB (IRR = 0.84; 95%CI: 0.73, 0.96) were negatively associated with the incidence of notified TB.
Substantial spatial clustering of notified TB was detected at region, zone and district level in Ethiopia. Health care access and knowledge about TB was associated with incidence of TB. This study may provide policy makers target hotspot areas, where national control programs could be implemented more efficiently for the prevention and control of TB, and to address potential under-reporting in poor access areas.
结核病（TB）仍然是全球和埃塞俄比亚的一个主要健康问题。本研究旨在调查埃塞俄比亚全国范围内报告的结核病的空间分布，并量化与结核病聚集相关的医疗保健获取、环境、社会人口和行为因素的作用。
使用 2016 年 6 月至 2017 年 6 月期间在埃塞俄比亚报告的国家结核病数据进行空间分析。使用 Moran's I 统计量和局部空间自相关指标（LISA）探索结核病的空间聚类。使用具有条件自回归（CAR）先验结构的多变量泊松回归模型，并使用贝叶斯马尔可夫链蒙特卡罗（MCMC）模拟中的 Gibbs 抽样来估计后验参数，以调查聚类的驱动因素。
在研究期间，埃塞俄比亚 745 个区报告了 120149 例结核病病例；其中 41343 例（34%）为细菌学确诊的新肺结 核，33997 例（28%）为临床诊断的新、涂片阴性肺结核患者。全国报告的结核病发病率为每 10 万人 112 例。发病率最高的是三个城市行政区，即 Dire Dewa（每 10 万人 348 例）、亚的斯亚贝巴（每 10 万人 262 例）和哈拉里（每 10 万人 206 例），发病率最低的是索马里地区（每 10 万人 51 例）。在 Humera、Gog 和 Surima 区检测到结核病报告的高-高空间聚类，而在索马里地区的一些区检测到低-低聚类。医疗保健获取不足（IRR = 0.78；95%CI：0.66，0.90）和对结核病知识的了解（IRR = 0.84；95%CI：0.73，0.96）与报告的结核病发病率呈负相关。
在埃塞俄比亚的区域、区和区一级检测到结核病报告的大量空间聚类。医疗保健获取和对结核病的了解与结核病的发病率有关。本研究可能为决策者提供目标热点地区，以便在这些地区更有效地实施国家控制计划，以预防和控制结核病，并解决医疗保健不足地区潜在的报告不足问题。
