Scholarlink

<p>Respiratory diseases remain a major public health challenge in tropical coastal cities, where persistent heat-humidity interactions and climate variability shape population vulnerability. This study quantified associations between atmospheric conditions and respiratory hospitalizations in Maceió, Brazil, using a 20-years time series (2000–2019). Weekly hospitalization rates stratified by age (children 0–4&#xa0;years, adults 5–59&#xa0;years, elderly ≥ 60&#xa0;years), were modeled against meteorological variables—including temperature, relative humidity, precipitation, atmospheric pressure, and solar radiation—considering lag structure of 0, 1, and 2&#xa0;weeks. Random Forest regression models were applied to capture nonlinear relationships and forecast hospitalization rates. Minimum temperature was the dominant predictor, exhibiting strong inverse associations across all age groups (ρ = − 0.65, p &lt; 0.001), with effects persisting up to two weeks. Age-specific patterns were observed: children showed immediate sensitivity to thermal and precipitation variables, whereas elderly populations exhibited delayed responses to barometric pressure and evaporation. Model performance was high for children and adults (R<sup>2</sup> = 0.83–0.90) and moderate for the elderly, with Symmetric Mean Absolute Percentage Error ranging from 13 to 25% across groups. Long-term trends revealed declining hospitalization rates among children and adults, contrasted by stabilization and subsequent increases in the elderly after 2010, consistent with demographic aging and increased climate sensitivity. These findings demonstrate the value of machine learning approaches for modeling complex climate-health relationships and provide a transferable framework for climate-informed respiratory risk assessment and early warning systems in tropical coastal environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Age-specific climate sensitivity of respiratory hospitalizations in a tropical coastal city: 20-year Random Forest forecasts

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月3日

查看全文
  • Marcos Paulo Santos Pereira,
  • Rafaela Lisboa Costa,
  • Fabricio Daniel dos Santos Silva,
  • Glauber Lopes Mariano,
  • Heliofabio Barros Gomes,
  • Helber Barros Gomes,
  • Djane Fonseca da Silva,
  • João Otávio Alves Accioly,
  • Jean Souza dos Reis,
  • Eva Maria Mollinedo Veneros,
  • Bruno Coelho Bulcão,
  • Marcelo Félix Alonso,
  • Flavio Manoel Rodrigues da Silva Júnior,
  • Richard James Ladle,
  • Fabiana Rita do Couto Santos Pereira

摘要

Respiratory diseases remain a major public health challenge in tropical coastal cities, where persistent heat-humidity interactions and climate variability shape population vulnerability. This study quantified associations between atmospheric conditions and respiratory hospitalizations in Maceió, Brazil, using a 20-years time series (2000–2019). Weekly hospitalization rates stratified by age (children 0–4 years, adults 5–59 years, elderly ≥ 60 years), were modeled against meteorological variables—including temperature, relative humidity, precipitation, atmospheric pressure, and solar radiation—considering lag structure of 0, 1, and 2 weeks. Random Forest regression models were applied to capture nonlinear relationships and forecast hospitalization rates. Minimum temperature was the dominant predictor, exhibiting strong inverse associations across all age groups (ρ = − 0.65, p < 0.001), with effects persisting up to two weeks. Age-specific patterns were observed: children showed immediate sensitivity to thermal and precipitation variables, whereas elderly populations exhibited delayed responses to barometric pressure and evaporation. Model performance was high for children and adults (R2 = 0.83–0.90) and moderate for the elderly, with Symmetric Mean Absolute Percentage Error ranging from 13 to 25% across groups. Long-term trends revealed declining hospitalization rates among children and adults, contrasted by stabilization and subsequent increases in the elderly after 2010, consistent with demographic aging and increased climate sensitivity. These findings demonstrate the value of machine learning approaches for modeling complex climate-health relationships and provide a transferable framework for climate-informed respiratory risk assessment and early warning systems in tropical coastal environments.