埃塞俄比亚疫苗覆盖的地理和社会经济决定因素:地区层面的空间和多阶段分析
本研究整合埃塞俄比亚地区层面的常规、调查和普查数据,分析麻疹疫苗第一剂(MCV1)和五联疫苗第三剂(Penta3)覆盖率的关联因素,并建立预测模型。分析识别出到卫生中心的通行时间、电气化水平和地区面积为主要关联因素,预测模型的平均绝对误差为11-12%。
要点
- 到卫生中心的通行时间、电气化水平和地区面积是与地区免疫覆盖相关的主要因素。
- 预测模型估计地区层面MCV1和Penta3覆盖率的平均绝对误差为11-12%。
- 地理空间模型对中低收入国家的公共卫生政策和规划具有重要潜力。
核心结论
- 本研究凸显了地理空间模型在中低收入国家公共卫生政策和规划中的显著潜力。
- 通过整合多样数据源并聚焦地区层面,我们提供了识别免疫覆盖缺口的定量框架。
- 使用地理和社会经济数据的方法可有效应用于广泛的公共卫生干预。
研究设计与样本
- 研究整合了地区(woreda)层面的免疫覆盖估计(来自DHIS-2)、人口与健康调查、人口普查以及电力、行政边界和卫生设施地理位置等公共数据库。
- 采用变量选择构建简约的beta回归模型以识别高预测力的协变量。
- 拟合并内部验证广义可加模型以预测MCV1和Penta3覆盖率。
主要发现
- 分析识别出到卫生中心的通行时间、电气化水平和地区面积是与地区层面免疫相关的主要因素。
- 预测模型估计地区层面MCV1和Penta3覆盖率的平均绝对误差为11-12%。
临床怎么用
- 原文未据此给出推荐。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Despite substantial progress over the past decades, many Ethiopian children still lack the full WHO-recommended immunization schedule. Notably, diphtheria-pertussis-tetanus-Hib-HepB and measles vaccines present large coverage disparities in Ethiopia. This study integrated routine, survey and census data from health, geographic and socioeconomic sources at the district level. We then explored associations between extracted covariates and coverage of measles (1st dose, MCV1) and diphtheria-pertussis-tetanus-Hib-HepB (3rd dose, Penta3). Lastly, we developed prediction models of immunization coverage. METHODS: We utilized multiple data sources, including district (known as woreda) immunization coverage estimates from the District Health Information Software (DHIS-2), Demographic and Health Surveys, demographic census, and public databases on electricity, administrative boundaries and health facility geolocations. We sought to develop parsimonious beta-regression models of immunization coverage using variable selection, so as to identify covariates with high predictive power. We then fitted and internally validated generalized additive models to predict MCV1 and Penta3 coverage. RESULTS: Our analysis identified access time to health centers, electrification levels, and woreda sizes as major factors associated with district-level immunization. Our prediction models estimated district-level MCV1 and Penta3 coverage with mean absolute errors of 11-12 %. CONCLUSIONS: This study highlights the significant potential of geospatial models for public health policy and planning in low- and middle-income countries. By integrating diverse data sources and focusing on the district level, we provide a quantitative framework for identifying gaps in immunization coverage. The approach, using geographic and socio-economic data, can be effectively applied to a wide range of public health interventions.
引用
- Forzy T, Tesfaye L, Getnet F, et al. Assessing the geographic and socioeconomic determinants of vaccine coverage in Ethiopia: A spatial and multistage analysis at the district level. Vaccine. 2025. PMID: 40056895. doi:10.1016/j.vaccine.2025.126834
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。