서브메뉴
검색
Heat and Hearts: Computational Phenotyping of Cardiovascular Disease in the Context of Extreme Heat
Heat and Hearts: Computational Phenotyping of Cardiovascular Disease in the Context of Extreme Heat
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202104808
- ISBN
- 9798291588956
- DDC
- 610
- 서명/저자
- Heat and Hearts: Computational Phenotyping of Cardiovascular Disease in the Context of Extreme Heat
- 발행사항
- [Sl] : Northwestern University, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 200 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
- 주기사항
- Advisor: Allen, Norrina B.;Kho, Abel N.
- 학위논문주기
- Thesis (Ph.D.)--Northwestern University, 2025.
- 초록/해제
- 요약Climate change and global warming are considered public health and medical crises in the United States (U.S.) with significant projected morbidity and mortality this century as temperatures rapidly increase. Cardiovascular disease (CVD) is currently the largest burden on our healthcare infrastructure with nearly 50% of U.S. adults living with CVD and more than 70% at an elevated risk of a CV event due to comorbid risk factors. Extreme heat exposure is a known exacerbator of CVD with the risk of cardiac arrest, heart failure, and other heart-related outcomes increasing 10-30% for every degree centigrade above regional thresholds. Certain sociodemographic categories and clinical characteristics promote susceptibility to heat-related CVD (HRCVD). Yet, the identification of populations most vulnerable to HRCVD remains methodologically underdeveloped, particularly at both the individual and community scales.Chicago, the site of one of the deadliest heatwaves in U.S. history in 1995, is projected to experience the greatest increase in temperature-related mortality among major US cities by 2090. A substantial share of future heat-related morbidity and mortality is expected to stem from cardiovascular disease, due to the confluence of physiological vulnerability and environmental stress. While prior studies have examined either community-level or individual-level modifiers of the heat-CVD association, none have integrated both dimensions-limiting our understanding of joint or interacting risk factors. Recent advances in machine learning offer powerful tools for quantifying heat exposure and identifying complex constellations of clinical and ecological predictors of CVD risk. By leveraging these methods to develop computational phenotypes across both community areas and individuals, this work aims to identify those most vulnerable to CVD events during extreme heat, enabling more precise and context-sensitive public health interventions.This dissertation presents a multi-level, computational approach to phenotyping HRCVD, integrating environmental exposure data with clinical, demographic, and spatial information in Chicago from 2010 to 2023. This work begins with an introduction that highlights the importance of heat-health studies and the complexity of the methods therein. Chapter 1 ends with a scoping review of the methodologies and study designs used in heat-health studies from across the world. In Chapter 2, I explore the temperature estimate tools that are most appropriate for characterizing extreme heat exposure in Chicago and demonstrate that suitability of Daymet for integration with health data for phenotype development. In Chapter 3, I employ ecological models at the Chicago community area level using generalized additive modeling and spatial clustering to quantify critical thresholds for excess HRCVD morbidity and mortality across all 77 community areas. These analyses identify geographic heterogeneity in heat vulnerability and demonstrate the modifying role of sociodemographic and socioeconomic factors. At the individual level, in Chapter 4 I leverage a longitudinal cohort of over 100,000 geocoded Chicago patients from a multi-institutional electronic health record (EHR) to define HRCVD phenotypes based on structured EHR data. Using regularized regression, interaction modeling, and phenome-wide association studies, I isolate key features associated with increased heat sensitivity to develop and validate a proof-of-concept HRCVD nomogram for calculating individual HRCVD risk. Together, these findings advance a framework for characterizing heat-related cardiovascular risk using computational phenotyping methods.The findings in these chapters support our original hypotheses that both community-level conditions and individual-level clinical factors independently and jointly modify the relationship between heat exposure and cardiovascular events. Community areas with greater social vulnerability experienced higher excess mortality during heat events, consistent with the hypothesis that neighborhood context plays a critical role in population-level susceptibility. At the same time, individual-level analyses confirmed that prior clinical features contribute significantly to HRCVD risk; some of which is known in the literature and some of which represent novel findings. Together, these results demonstrate the value of integrating ecological and clinical data using computational phenotyping to improve risk stratification and guide targeted prevention efforts. By bridging ecological and patient-level perspectives, this work supports targeted public health strategies and clinical interventions that can better mitigate the cardiovascular impacts of a warming climate.
- 일반주제명
- Medicine
- 일반주제명
- Public health
- 일반주제명
- Environmental health
- 키워드
- Climate change
- 키워드
- Chicago
- 키워드
- Health data
- 기타저자
- Northwestern University Health Sciences Integrated PhD Program
- 기본자료저록
- Dissertations Abstracts International. 87-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017358907
■00520260202104808
■006m o d
■007cr#unu||||||||
■020 ▼a9798291588956
■035 ▼a(MiAaPQ)AAI32166516
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a610
■1001 ▼aGraffy, Peter M. C.▼0(orcid)0000-0002-2156-7175
■24510▼aHeat and Hearts: Computational Phenotyping of Cardiovascular Disease in the Context of Extreme Heat
■260 ▼a[Sl]▼bNorthwestern University▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a200 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-03, Section: B.
■500 ▼aAdvisor: Allen, Norrina B.;Kho, Abel N.
■5021 ▼aThesis (Ph.D.)--Northwestern University, 2025.
■520 ▼aClimate change and global warming are considered public health and medical crises in the United States (U.S.) with significant projected morbidity and mortality this century as temperatures rapidly increase. Cardiovascular disease (CVD) is currently the largest burden on our healthcare infrastructure with nearly 50% of U.S. adults living with CVD and more than 70% at an elevated risk of a CV event due to comorbid risk factors. Extreme heat exposure is a known exacerbator of CVD with the risk of cardiac arrest, heart failure, and other heart-related outcomes increasing 10-30% for every degree centigrade above regional thresholds. Certain sociodemographic categories and clinical characteristics promote susceptibility to heat-related CVD (HRCVD). Yet, the identification of populations most vulnerable to HRCVD remains methodologically underdeveloped, particularly at both the individual and community scales.Chicago, the site of one of the deadliest heatwaves in U.S. history in 1995, is projected to experience the greatest increase in temperature-related mortality among major US cities by 2090. A substantial share of future heat-related morbidity and mortality is expected to stem from cardiovascular disease, due to the confluence of physiological vulnerability and environmental stress. While prior studies have examined either community-level or individual-level modifiers of the heat-CVD association, none have integrated both dimensions-limiting our understanding of joint or interacting risk factors. Recent advances in machine learning offer powerful tools for quantifying heat exposure and identifying complex constellations of clinical and ecological predictors of CVD risk. By leveraging these methods to develop computational phenotypes across both community areas and individuals, this work aims to identify those most vulnerable to CVD events during extreme heat, enabling more precise and context-sensitive public health interventions.This dissertation presents a multi-level, computational approach to phenotyping HRCVD, integrating environmental exposure data with clinical, demographic, and spatial information in Chicago from 2010 to 2023. This work begins with an introduction that highlights the importance of heat-health studies and the complexity of the methods therein. Chapter 1 ends with a scoping review of the methodologies and study designs used in heat-health studies from across the world. In Chapter 2, I explore the temperature estimate tools that are most appropriate for characterizing extreme heat exposure in Chicago and demonstrate that suitability of Daymet for integration with health data for phenotype development. In Chapter 3, I employ ecological models at the Chicago community area level using generalized additive modeling and spatial clustering to quantify critical thresholds for excess HRCVD morbidity and mortality across all 77 community areas. These analyses identify geographic heterogeneity in heat vulnerability and demonstrate the modifying role of sociodemographic and socioeconomic factors. At the individual level, in Chapter 4 I leverage a longitudinal cohort of over 100,000 geocoded Chicago patients from a multi-institutional electronic health record (EHR) to define HRCVD phenotypes based on structured EHR data. Using regularized regression, interaction modeling, and phenome-wide association studies, I isolate key features associated with increased heat sensitivity to develop and validate a proof-of-concept HRCVD nomogram for calculating individual HRCVD risk. Together, these findings advance a framework for characterizing heat-related cardiovascular risk using computational phenotyping methods.The findings in these chapters support our original hypotheses that both community-level conditions and individual-level clinical factors independently and jointly modify the relationship between heat exposure and cardiovascular events. Community areas with greater social vulnerability experienced higher excess mortality during heat events, consistent with the hypothesis that neighborhood context plays a critical role in population-level susceptibility. At the same time, individual-level analyses confirmed that prior clinical features contribute significantly to HRCVD risk; some of which is known in the literature and some of which represent novel findings. Together, these results demonstrate the value of integrating ecological and clinical data using computational phenotyping to improve risk stratification and guide targeted prevention efforts. By bridging ecological and patient-level perspectives, this work supports targeted public health strategies and clinical interventions that can better mitigate the cardiovascular impacts of a warming climate.
■590 ▼aSchool code: 0163.
■650 4▼aMedicine
■650 4▼aPublic health
■650 4▼aEnvironmental health
■653 ▼aCardiovascular disease
■653 ▼aClimate change
■653 ▼aChicago
■653 ▼aHealth data
■653 ▼aComputational phenotyping
■690 ▼a0564
■690 ▼a0470
■690 ▼a0573
■71020▼aNorthwestern University▼bHealth Sciences Integrated PhD Program.
■7730 ▼tDissertations Abstracts International▼g87-03B.
■790 ▼a0163
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17358907▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.


