Geospatial Processing System for Social and Environmental Health Research
Lead Research Organisation:
Imperial College London
Department Name: School of Public Health
Abstract
Our current and future health is strongly influenced by our social and physical environment. Changes in income, housing, and environmental exposures are linked to health inequalities but are potentially changeable and preventable through evidence-based interventions. Identifying and understanding the social and environmental factors which impact on the nation's health is, therefore, essential to help reducing health inequalities. This requires highly accurate modelling of environmental exposure and social factors for each resident in the UK, and the linkage of these to patient records, cohort participants and administrative health systems, via the residential address. To enable this, state of the art capacity and capabilities to securely store, process, analyse and share large amounts of data are needed, but are currently not available at Imperial College London.
We propose establishing a Geospatial Processing System for Social and Environmental Health Research (GeoPS) at Imperial College London, a stand-alone, high-performance solution to model address-level social and environmental data and securely link these to individual health records. The GeoPS will provide the storage capacity, memory, and processing power to model exposure to air pollution, noise, housing quality and greenspace for each address in the UK using novel methods and data, including from satellites and Google Street View images. Such models already exist or are currently under development at the MRC Centre for Environment and Health at Imperial College, but the capacity to scale these up to the national level is missing. The GeoPS will follow information government requirements on the save and secure handling of personal information to maintain patient confidentiality. This allows us to link the constructed exposure profiles to the ~300 million individual records that are stored in health data bases at the UK Small Area Health Statistics Unit, Imperial College, and contain information on deaths and birth from the Office for National Statistics and Hospital Episode Statistics from NHD Digital. Linkage is also planned to established Imperial College hosted longitudinal population studies.
These linked data resources will enable advanced epidemiological analyses to guide policy, public health and services to improve the nation's health. Establishing such equipment and infrastructure in house, will enable us to act at pace to emerging social and environmental health issues and provide timely epidemiological evidence to Directors of Public Health, the UK Health Security Agency and policy makers in local and national governments.
We propose establishing a Geospatial Processing System for Social and Environmental Health Research (GeoPS) at Imperial College London, a stand-alone, high-performance solution to model address-level social and environmental data and securely link these to individual health records. The GeoPS will provide the storage capacity, memory, and processing power to model exposure to air pollution, noise, housing quality and greenspace for each address in the UK using novel methods and data, including from satellites and Google Street View images. Such models already exist or are currently under development at the MRC Centre for Environment and Health at Imperial College, but the capacity to scale these up to the national level is missing. The GeoPS will follow information government requirements on the save and secure handling of personal information to maintain patient confidentiality. This allows us to link the constructed exposure profiles to the ~300 million individual records that are stored in health data bases at the UK Small Area Health Statistics Unit, Imperial College, and contain information on deaths and birth from the Office for National Statistics and Hospital Episode Statistics from NHD Digital. Linkage is also planned to established Imperial College hosted longitudinal population studies.
These linked data resources will enable advanced epidemiological analyses to guide policy, public health and services to improve the nation's health. Establishing such equipment and infrastructure in house, will enable us to act at pace to emerging social and environmental health issues and provide timely epidemiological evidence to Directors of Public Health, the UK Health Security Agency and policy makers in local and national governments.
Technical Summary
The MRC Centre for Environment and Health at Imperial College London aims to achieve a step change in determining and understanding the social and environmental drivers of the nation's health and consequently help reducing health inequalities. The advancements in geospatial analytical methods and increased availability of spatially high-resolution data from national data providers (e.g., Ordnance Survey MasterMap), satellites and imaging (e.g., Google Street View) allows for cutting-edge research into spatial and temporal high precision exposure assessment. Linking advanced environmental, exposure and geospatial information to administrative health systems and recruited cohorts for epidemiological and biostatistical analyses, however, requires high-spec solutions to overcome substantial computational and information governance challenges.
We propose establishing a Geospatial Processing System for Social and Environmental Health Research (GeoPS) at Imperial College London, a stand-alone, high-performance solution for secure linkage of health, social and environmental data. The proposed GeoPS consists of fast storage to allow rapid access of large data sets during modelling processes (e.g., linking geospatial and environmental information to geographic location of UK residential addresses); graphics processing units (GPU) for processing of large geospatial environmental and exposure data including from satellites and imaging data; and a backup server to securely store and archive data.
The GeoPS will provide a step change in ICL's capacity and capability to study health and inequalities across the life course by generating powerful, linked data resources to guide policy, public health and services to improve the nation's health. This will strengthen the UK's capacity to deliver UK-wide interoperable and harmonised, FAIR health data assets at scale and pace.
We propose establishing a Geospatial Processing System for Social and Environmental Health Research (GeoPS) at Imperial College London, a stand-alone, high-performance solution for secure linkage of health, social and environmental data. The proposed GeoPS consists of fast storage to allow rapid access of large data sets during modelling processes (e.g., linking geospatial and environmental information to geographic location of UK residential addresses); graphics processing units (GPU) for processing of large geospatial environmental and exposure data including from satellites and imaging data; and a backup server to securely store and archive data.
The GeoPS will provide a step change in ICL's capacity and capability to study health and inequalities across the life course by generating powerful, linked data resources to guide policy, public health and services to improve the nation's health. This will strengthen the UK's capacity to deliver UK-wide interoperable and harmonised, FAIR health data assets at scale and pace.
Description | Due the equipment funded under this scheme not being in place for very long, key outputs related to the research process rather than research outputs. We have established an ongoing dialog with other funders and institutions around the country (including NERC and the MET Office) to identify best practice and develop a data pipelpine and methodologies around linking environmental and geospatial data to health records. This includes work on geoprivacy and the linkage of unique property reference number for linkage under the HDR UK Driver Programme on Social and Environmental Determinants of Health. |
Exploitation Route | Too early to say |
Sectors | Environment Healthcare |
Description | Written evidence to the Environmental Audit Committee's 'Heat resilience and sustainable cooling' inquiry. Sep 2023. D Fecht et al |
Geographic Reach | National |
Policy Influence Type | Participation in a guidance/advisory committee |
URL | https://committees.parliament.uk/work/7811/heat-resilience-and-sustainable-cooling/publications/ |
Description | Collaboration with MET Office on linkage to environmental data to health records |
Organisation | Meteorological Office UK |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Provision of expertise and methodologies on linking environmental and geosptial data to health records. |
Collaborator Contribution | Collaboration is still in developing stage and we are currently scoping out via various workshops on how we can best work together. The contribution will most likely be in the form of environmental data assessts being made available to us for integration to health records. |
Impact | Collaboration is still in developing stage and we are currently scoping out via various workshops on how we can best work together. |
Start Year | 2023 |
Description | Discussion with UKRI on use of JASMIN computing infrastructure |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Third sector organisations |
Results and Impact | Discussion with MRC lessons learned from our study which could feed into the future of use/direction of UKRI-funded computer facility JASMIN. This could lead to future collaborations with other UKRI-funded infrastructure. |
Year(s) Of Engagement Activity | 2024 |
Description | Presentation to Scientific Advisory Board |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Industry/Business |
Results and Impact | Presentation to the Scientific Advisory Board of Our Future Health on 'Geolocation for cohort studies' - 26 Feb 2024 by Daniela Fecht |
Year(s) Of Engagement Activity | 2024 |