Severe droughts and the tuberculosis and HIV care cascade in the Brazilian Amazon: a Bayesian, structural, interrupted time-series analysis of 24 years of surveillance data

IF 20.5 1区医学 Q1 ENVIRONMENTAL SCIENCES

Lancet Planetary Health Pub Date : 2026-05-01 Epub Date: 2026-06-03 DOI:10.1016/j.lanplh.2026.101459

Rodrigo C Menezes MD PhD, Rodrigo M Teles, Isabella B B Ferreira MD, Leandro S Garcia MSc, Jair S Pinheiro MSc, Tiago F Mota PhD, Klauss Villalva-Serra MD, Leandro R Macedo PhD, Moreno M S Rodrigues PhD, Artur T L Queiroz PhD, Marcelo Cordeiro-Santos Prof, Bruno B Andrade Prof

{"title":"Severe droughts and the tuberculosis and HIV care cascade in the Brazilian Amazon: a Bayesian, structural, interrupted time-series analysis of 24 years of surveillance data","authors":"Rodrigo C Menezes MD PhD, Rodrigo M Teles, Isabella B B Ferreira MD, Leandro S Garcia MSc, Jair S Pinheiro MSc, Tiago F Mota PhD, Klauss Villalva-Serra MD, Leandro R Macedo PhD, Moreno M S Rodrigues PhD, Artur T L Queiroz PhD, Marcelo Cordeiro-Santos Prof, Bruno B Andrade Prof","doi":"10.1016/j.lanplh.2026.101459","DOIUrl":null,"url":null,"abstract":"<div><h3>Background</h3><div>In river-dependent regions such as the Brazilian Amazon, severe droughts can disrupt care continuity by isolating municipalities and fragmenting supply chains. We assessed how four major droughts affected tuberculosis and HIV care cascades in Amazonas state.</div></div><div><h3>Methods</h3><div>We analysed monthly data (Jan 1, 2001, to Dec 31, 2024) from all 62 municipalities of the Amazonas state, Brazil, grouped into nine river basins and stratified by sex when available. Droughts were defined by the relative maximum cumulative water deficit value being below −2·0 or river discharge at or below the fifth historical percentile. We prespecified four major drought episodes affecting Amazonas at a monthly resolution: May to June, 2005; May to June, 2010; August, 2015, to April, 2016; and October, 2023, to February, 2024. Outcomes were primary-care consultations, tuberculosis and HIV notifications, treatment interruption, and disease-specific mortality. Additionally, we did a prespecified subgroup analysis stratified by sex for both diseases’ indicators to assess whether drought-associated deviations differed systematically between men and women. We fitted Bayesian hierarchical negative binomial models to estimate counterfactual trajectories and summarised effects as incidence rate ratios (IRRs) and excess events per 100 000 population.</div></div><div><h3>Findings</h3><div>The 2015–16 drought coincided with widespread consultation deficits (statewide peak IRR 0·78 [95% credible interval 0·68–0·89]) and pronounced increases in HIV treatment interruption across all basins (statewide peak IRR 6·01 [2·12–15·33]), particularly among men (Rio Negro peak IRR 5·57 [3·02–9·78]). By contrast, the 2023–24 drought coincided with increased care consultations statewide (peak IRR 1·63 [1·25–2·11]) but persistent treatment interruption and mortality signals, including female-predominant increases in HIV treatment interruption in Rio Negro (peak IRR 2·87 [1·27–5·63]) and female mortality statewide (1·97 [1·05–3·39]), with additional basin-level mortality increases in Alto Solimões, Baixo Amazonas, and Médio Amazonas. Tuberculosis notifications increased mainly in 2023–24 in specific basins, especially in Baixo Amazonas (peak IRR 2·09 [1·54–2·76]), Médio Amazonas (1·74 [1·30–2·28]), and Baixo Solimões (1·60 [1·24–2·07]).</div></div><div><h3>Interpretation</h3><div>Severe droughts were associated with basin-specific disruptions in tuberculosis and HIV care cascades in the Brazilian Amazon, with clearer signals in adherence-dependent and downstream outcomes than in case detection. Strengthening resilience will require anticipatory, continuity-focused strategies that safeguard treatment access during climate-related transport constraints.</div></div><div><h3>Funding</h3><div>Fundação de Amparo à Pesquisa do Estado do Amazonas and Brazilian National Council for Scientific and Technological Development.</div></div>","PeriodicalId":48548,"journal":{"name":"Lancet Planetary Health","volume":"10 5","pages":"Article 101459"},"PeriodicalIF":20.5000,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Lancet Planetary Health","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S254251962600032X","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2026/6/3 0:00:00","PubModel":"Epub","JCR":"Q1","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}

引用次数: 0

Abstract

Background

In river-dependent regions such as the Brazilian Amazon, severe droughts can disrupt care continuity by isolating municipalities and fragmenting supply chains. We assessed how four major droughts affected tuberculosis and HIV care cascades in Amazonas state.

Methods

We analysed monthly data (Jan 1, 2001, to Dec 31, 2024) from all 62 municipalities of the Amazonas state, Brazil, grouped into nine river basins and stratified by sex when available. Droughts were defined by the relative maximum cumulative water deficit value being below −2·0 or river discharge at or below the fifth historical percentile. We prespecified four major drought episodes affecting Amazonas at a monthly resolution: May to June, 2005; May to June, 2010; August, 2015, to April, 2016; and October, 2023, to February, 2024. Outcomes were primary-care consultations, tuberculosis and HIV notifications, treatment interruption, and disease-specific mortality. Additionally, we did a prespecified subgroup analysis stratified by sex for both diseases’ indicators to assess whether drought-associated deviations differed systematically between men and women. We fitted Bayesian hierarchical negative binomial models to estimate counterfactual trajectories and summarised effects as incidence rate ratios (IRRs) and excess events per 100 000 population.

Findings

The 2015–16 drought coincided with widespread consultation deficits (statewide peak IRR 0·78 [95% credible interval 0·68–0·89]) and pronounced increases in HIV treatment interruption across all basins (statewide peak IRR 6·01 [2·12–15·33]), particularly among men (Rio Negro peak IRR 5·57 [3·02–9·78]). By contrast, the 2023–24 drought coincided with increased care consultations statewide (peak IRR 1·63 [1·25–2·11]) but persistent treatment interruption and mortality signals, including female-predominant increases in HIV treatment interruption in Rio Negro (peak IRR 2·87 [1·27–5·63]) and female mortality statewide (1·97 [1·05–3·39]), with additional basin-level mortality increases in Alto Solimões, Baixo Amazonas, and Médio Amazonas. Tuberculosis notifications increased mainly in 2023–24 in specific basins, especially in Baixo Amazonas (peak IRR 2·09 [1·54–2·76]), Médio Amazonas (1·74 [1·30–2·28]), and Baixo Solimões (1·60 [1·24–2·07]).

Interpretation

Severe droughts were associated with basin-specific disruptions in tuberculosis and HIV care cascades in the Brazilian Amazon, with clearer signals in adherence-dependent and downstream outcomes than in case detection. Strengthening resilience will require anticipatory, continuity-focused strategies that safeguard treatment access during climate-related transport constraints.

Funding

Fundação de Amparo à Pesquisa do Estado do Amazonas and Brazilian National Council for Scientific and Technological Development.

查看原文本刊更多论文

巴西亚马逊地区的严重干旱、结核病和艾滋病护理级联：对24年监测数据的贝叶斯、结构、中断时间序列分析。

背景：在巴西亚马逊等依赖河流的地区，严重干旱会使市政当局孤立，供应链断裂，从而破坏医疗服务的连续性。我们评估了亚马逊州四次重大干旱是如何影响结核病和艾滋病护理级联的。方法：我们分析了巴西亚马逊州所有62个城市的月度数据（2001年1月1日至2024年12月31日），将其分为9个河流流域，并按性别分层（如果有的话）。干旱的定义是相对最大累积亏水值低于-2·0或河流流量等于或低于第5个历史百分位数。我们预先确定了影响亚马逊地区的四次主要干旱事件：2005年5月至6月；2010年5月至6月；2015年8月至2016年4月；从2023年10月到2024年2月。结果是初级保健咨询、结核病和艾滋病毒通报、治疗中断和疾病特异性死亡率。此外，我们对两种疾病的指标进行了预先指定的亚组分析，按性别分层，以评估男性和女性之间干旱相关偏差是否存在系统性差异。我们拟合贝叶斯层次负二项模型来估计反事实轨迹，并将影响总结为发病率比（IRRs）和每10万人中的过量事件。研究结果：2015-16年的干旱与广泛的咨询不足（全州最高IRR为0.78[95%可信区间0.68 - 0.89]）和所有流域艾滋病治疗中断的明显增加（全州最高IRR为6.01[2.12 - 15.33]）相吻合，尤其是男性（黑人最高IRR为5.57[3.02 - 9.78]）。相比之下，2023-24年的干旱同时伴随着全州范围内护理咨询的增加（IRR峰值为1.63[1.25 - 2.11]），但持续的治疗中断和死亡率信号，包括里约热内卢Negro （IRR峰值为2.87[1.27 - 5.63]）和全州女性死亡率（1.97[1.05 - 3.39]），以及Alto Solimões、Baixo Amazonas和m dio Amazonas的流域死亡率增加。2023 ~ 2024年，各流域结核病通报数主要呈上升趋势，其中亚马逊河流域（Baixo Amazonas）（峰值IRR为2.09[1.54 ~ 2.76]）、亚马逊河流域（m ~ 2.28]）和亚马逊河流域（Baixo Solimões）（峰值IRR为1.60[1.24 ~ 2.07]）呈上升趋势。解释：严重干旱与巴西亚马逊地区结核病和艾滋病毒护理级联的流域特异性中断有关，与病例检测相比，在依从性依赖和下游结果中具有更清晰的信号。加强抗灾能力需要前瞻性的、以连续性为重点的战略，以确保在气候相关的运输限制期间获得治疗。资助：亚马逊州经济发展基金会和巴西国家科学和技术发展委员会。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Lancet Planetary Health Multiple-

CiteScore

28.40

自引率

2.30%

发文量

272

审稿时长

8 weeks

期刊介绍： The Lancet Planetary Health is a gold Open Access journal dedicated to investigating and addressing the multifaceted determinants of healthy human civilizations and their impact on natural systems. Positioned as a key player in sustainable development, the journal covers a broad, interdisciplinary scope, encompassing areas such as poverty, nutrition, gender equity, water and sanitation, energy, economic growth, industrialization, inequality, urbanization, human consumption and production, climate change, ocean health, land use, peace, and justice. With a commitment to publishing high-quality research, comment, and correspondence, it aims to be the leading journal for sustainable development in the face of unprecedented dangers and threats.