NHS Digital Data Release Register - reformatted

University Of Exeter

Project 1 — DARS-NIC-374485-Y2X9C

Opt outs honoured: N

Sensitive: Non Sensitive

When: 2017/03 — 2017/05.

Repeats: One-Off

Legal basis: Health and Social Care Act 2012

Categories: Anonymised - ICO code compliant


  • Hospital Episode Statistics Accident and Emergency
  • Hospital Episode Statistics Admitted Patient Care


The aim of processing HES A&E Admissions and HES Inpatient Admissions data is to model relationships between exposure to natural environments and health outcomes, and to develop methods for economic valuation of public health benefits and dis-benefits of these environments. Data processing will be in connection with three sub-projects of the NIHR-funded Health Protection Research Unit (HPRU) on Environmental Change and Health, which is conducted in collaboration with the Met Office, London School of Hygiene and Tropical Medicine and University College London (partner organisations in the HPRU). The projects are linked to Theme 3 of this HPRU – ‘Health and the Natural Environment’ - and are all led by staff at the European Centre for Environment and Human Health (www.ecehh.org) at University of Exeter Medical School. Specifically, the data processing aims to: 1) improve understanding of the relationships between exposure to land cover types and to allergenic pollen and other air pollutants, and allergic disease and respiratory health outcomes (Project 3.3); 2) improve understanding of the relationships between residential area exposure to green space and other natural environments, and mental and physical health and well-being outcome measures (Project 3.5); and 3) improve understanding of the potential human health impacts of harmful algal bloom (HAB) occurrences, including on respiratory, digestive and neurologic illnesses (Project 3.7). In all three cases, the objective will be to develop an understanding of the health implications of environmental factors, especially in relation to climate change. The objectives of Project 3.3 are linked to another Theme 3 project of this HPRU (Project 3.4) led by staff at the Met Office which aims to produce spatial-temporal models of the distribution of allergenic pollens across the UK. The objectives of Project 3.5 build on earlier work carried out by the current team and financed by the ESRC under the Secondary Data Analysis Initiative (Award No. ES/K002872/1) in which environmental exposure variables were derived for Lower-level Super Output Areas using GIS and relationships to health and wellbeing measures were examined in Census 2011, the British Household Panel Survey and the UK Household Longitudinal Survey. The objectives for processing HES data in relation to Project 3.5 also relate to data analysis which will be carried out using data from UK Biobank as part of this same HPRU project. The objectives of Project 3.5 also link to the HPRU Theme 2 projects (Healthy Sustainable Cities). The objectives of Project 3.7 are linked to a larger collaboration between the University of Exeter Medical School; the UK Met Office; Public Health England; Centre for Environment, Fisheries and Aquaculture Science (CEFAS); and the Scottish Association for Marine Sciences (SAMS). The wider collaborations use climate projections to explore likely changes in occurrences of HABs driven by future marine environmental conditions and is funded by MRC, NERC and MEDMI in addition to NIHR HPRU on Environmental Change and Health. Project 3.7 also builds on over 2 decades of work by Professor Fleming on the exposures and health effects of HABs in the US and globally.

Expected Benefits:

The work will provide high quality scientific evidence to help Public Health England implement effective public health interventions to reduce the burden of ill health associated with unsustainable development, more specifically the health protection activities outlined in UK’s National Adaptation Programme (2013) and the Sustainable Development Strategy for the Health and Social Care System (2014-2020). The work will also produce research of relevance to other government departments, including DEFRA, EA and CEFAS regarding the health co-benefits of environmental policies (particularly adaptation to and mitigation of climate change) and the protection of the natural environment. Analyses related to the HPRU Project 3.3 are conducted in collaboration with the Met Office (a partner organisation in the HPRU), and aim to contribute to the knowledge base required for genera-specific allergenic pollen forecasts which, for example, would potentially help the UK's c.5 million asthma sufferers in the management of their condition. Project 3.5 aims to contribute to the knowledge base concerning the potential health and wellbeing effects for both physical and mental health benefits of interacting with the natural environment, research that might decrease the impacts of obesity, cardiovascular disease, diabetes and mental health issues on the NHS and the economy. Work in this area already conducted by the current team was discussed as being a key strand of evidence in the Parliamentary Office of Science and Technology POSTnote ‘Urban Green Infrastructure’ (#448) and was presented orally at the POSTnote launch in the House of Commons in May 2014. Project 3.7 aims to explore if HABs (which have been shown to exacerbate asthma and other respiratory diseases for example in the US) are having any direct impacts on respiratory, digestive and/or neurologic health in the UK. The project aims to contribute to the knowledge base required to identify potentially vulnerable human populations for targeted interventions; and to model potential provision of early HAB warning systems.


Journal papers and associated conference presentations will be the primary outputs. Results will concern samples in large geographical areas, such as admissions for clusters of ICD10 codes in England and in Greater London; data published will be sample means etc., and no data on specific observations of individuals or small groups of specific observations of individuals will be published. An initial paper on pollen and A&E hospital admissions (related to Project 3.3) will be submitted by May 2016; an initial paper on land cover and physical health (related to Project 3.5) will be submitted by December 2016; an initial paper on valuation of land cover for health co-benefits (related to Project 3.5) will be submitted by December 2016; an initial paper on health impacts of harmful algal blooms (related to Project 3.7) will be submitted by December 2017. Any data included in the outputs will be aggregated with small numbers suppressed in line with the HES Analysis Guidelines. Further papers as well as presentations will follow from all 3 projects.


The initial processing activity is the same for all three projects and requires data at the individual level. Individual level hospital episode data will be supplied by the HSCIC to the University of Exeter and it will then be linked to residential area environmental factors through the non-sensitive Lower-level Super Output Area (field soal). Area level data to be linked on soal include: Generalised Land Use Database; Land Cover Map 2007; rural-urban classification; tree cover density; coastal proximity; distance to inland waterways; daily, monthly and annual mean exposure to air pollutant concentrations, temperature and rainfall; smoking rates estimated at Postcode District and distributed to soal; genera-specific pollen concentrations; oceanographic data; HAB data, both organisms and toxin levels. This linkage activity will involve no data transfer flows (no data will be supplied to HSCIC for linkage) since the area level environmental datasets are stored on the same secure server that the HES data will be stored on. These environmental datasets are publicly available, or have been derived by the University of Exeter from publicly available sources, or are available for this research under a user agreement with the Met Office; or are made available for this research by CEFAS and SAMS. The data are physically stored on a secure server located in a University of Exeter secure datacentre. The server, which operates Active Directory Kerberos authentication, will be accessed for the linkage activity and for all other processing activities from the European Centre for Environment and Human Health, University of Exeter Medical School, Knowledge Spa, Royal Cornwall Hospital, Truro. No data supplied by the HSCIC will be shared with or accessible to third parties or used for any purposes outside of the 3 projects outlined above. Post-linkage processing activities in relation to Project 3.3 (see above) will follow two methodologies which respectively address the health impacts of relatively dynamic and relatively time-invariant (or 'slowly changing') environmental exposures. In the first case, time series regression will be used to analyse daily counts of episodes of potentially allergic diseases against daily variations in environmental airborne exposures, adjusting for seasonal and long-term trend. For example, daily A&E admission counts for respiratory conditions and for allergy, and daily In-patient admission for asthma, will be regressed against current daily (and lagged daily) pollen, weather and air pollutant measures. In the second case, area level standardised period prevalence rates will be calculated by aggregating daily episode data to period, and rates will be regressed against single time point measures of environmental exposures, and against period mean measures of time varying factors. For example, area A&E admission rates 2007/8-2013/4 for respiratory conditions will be regressed against area green space as measured in 2005, social deprivation as measured in 2011, and area mean pollutant concentrations over the period. Processing activities in the analysis phase in relation to Project 3.5 (see above) will mainly examine relationships between relatively time-invariant environmental exposures and area episode rates using the same approach described above. Research hypotheses on environment/health relationships will assume mechanisms mediated by both the engagement in physical activity (which is stimulated by environmental settings), and cognitive attention restoration (which is stimulated by environmental settings). For example, area level standardised period prevalence rates for in-patient admissions for diseases related to circulatory health will be regressed against single time point measures of environmental exposures, and against period mean measures of time varying factors. Processing activities in the analysis phase in relation to Project 3.7 (see above) will use time series and cross-sectional regression models to explore how time varying marine environmental factors (such as nutrient loading, temperature and salinity) as well as identified HAB blooms in marine and freshwaters relate to diseases possibly associated with HAB occurrences, including respiratory, digestive and neurologic illnesses, after controlling for exposures to non-time varying environmental factors. For clarity, all data analyses in connection with the three projects described will either be time series, using days, or aggregations of days, as the unit of analysis, or will be ecological, using geographical areas as the unit of analysis, and in both cases results will be estimations of rates of events associated with environmental conditions. However, data is required at the individual level, rather than at any aggregated level, in order to make possible the linkage of spatial-temporal environmental data to hospital episodes. Published outputs will not specify any information about individuals, nor about groups of individuals, but will be limited to abstracted statistical associations between environmental factors and rates of hospital episodes, and to whole sample descriptive statistics without geographic disaggregation.