Improving Predictive Asthma Algorithms with Modeled Environment Data for Scotland

Scotland has the highest prevalence of asthma in the world and some of the poorest health outcomes from asthma. Epidemiological studies have demonstrated the impact of pollution, pollen and adverse weather conditions on poor asthma outcomes, e.g. asthma attacks and hospitalisations. Asthma has been estimated to cost Scotland £92million per year, including £10 million per year on unscheduled care, e.g. A&E attendances. Efforts to identify cost-effective approaches to better predict poor outcomes in people with long-term conditions, such as asthma, is a key healthcare policy.

Current approaches to help patients and clinical staff to predict future asthma attacks and intervene to prevent these are less than optimal. This has been due to a lack of an accurate risk score to predict which people are likely to have an asthma attack. Reasons for the poor quality of current risk scores include the lack of exposure data for known pollutants, weather and pollen which is a factor in triggering asthma attacks.

In the past, studies have tended to estimate these pollution, weather and pollen exposures from a few widely dispersed monitoring stations. This, along with difficulties in measuring where people are and how they are exposed, has meant that the pollution/weather/pollen exposures are not as accurate as they could be.

In this application our aim is to bring together academic communities and their datasets from health, environmental and geospatial sciences, demographic and social sciences. Datasets include individual's information from general practice, hospitals, census (for place of work and travel), Scottish education (for school location) and environmental data. We also plan to improve our environmental data so that we can improve resolution (e.g. to smaller geographical areas). Using a trusted research environment which allows the study researchers to safely link and analyse information from these data sources, we will carry out a statistical analysis and develop a risk score for asthma attacks. We will also work to refine and further develop methods for better forecasting weather, pollen and pollution data to feed into this risk score. We will then trial the accuracy of this risk score's potential to reduce asthma attacks, hospitalisations and deaths in a future project.

Keywords:
Asthma, meteorology, atmospheric composition, pollution, pollen, attack, primary care, secondary care, data linkage, unscheduled care, forecasting, risk scores.

Stakeholders:
Stakeholders involved in our collaboration for this project include The University of Edinburgh's Usher Institute and the Administrative Data Research Centre - Scotland, Strathclyde University's Department of Mathematics and Statistics, The Center for Ecology and Hydrology and NHS National Service Scotland, Information Services Division. Key Stakeholders also include public health experts and health policy/regulatory experts, in particular NHS Health Scotland and the Scottish Environment Protection Agency, who have expressed a strong interest in the findings of this pilot project. Other stakeholders are scientific researchers, data scientists and statisticians, public health planners, medical staff and regulatory/policy officers.

The proposed research will primarily impact a wide range of groups:

The scientific community will benefit from the improved integration of medical and health experts, data scientists and environmental modellers. The researchers involved in the project benefit from the opportunity for interdisciplinary research and the design of an application which will serve as a demonstrator for a conceptual approach towards better integration of environment and health datasets. The direct involvement of the Asthma UK Centre for Applied Research and the Administrative Data Research Centre will ensure that the project results are directly feeding into strategic research objectives in the relevant units concerned with health and health care planning (e.g. the Scottish Government). Finally, the use of environmental modelling data for the assessment of health impacts will provide feedback on the relevance and robustness of such datasets, validated by comparison with observations from monitoring networks, and further advance the better integration between observed and modelled data. This will in particular offer relevant insights for environmental regulators and policy makers (e.g. SEPA), who have a remit to collect data and report on improvements in air quality and other environmental parameters.

General practices , secondary care respiratory services, hospital A&E departments will benefit from the results of this proposed research as it will provide valuable insights into the mechanisms how environmental pollution events may affect the influx of (high) patient numbers. This will provide a proof-of-concept for the development of more advanced information systems and establish the basis for a wider research activity aiming to build an operational forecasting system for improved health care planning.

Public health planners, as well as organisations directing public health research and resource use such as NHS, Public Health Intelligence Scotland and charities such as Asthma UK will benefit from the research due to its contribution to the longer term capability development for advance warning of both patients and frontline care services, in addition to gaining a better, quantitative insight into the relationships between multiple drivers of episodic health effects.

The general public, specifically population groups suffering from diseases i.e. the Asthma UK Centre for Applied Research the patient advisory group, will benefit from the research by increasing the resilience of front line medical services and enabling better planning, resulting in faster, more efficient treatment. This will extend as well to those population groups which are more vulnerable, e.g. children and OAPs.

The proposed research thus has a high potential to produce outcomes that are directly beneficial to public health, as well as serving as a pilot study to build the quantitative and qualitative expertise for the development of a full scale research grant. To maximise these benefits, the project team will communicate results in the following ways:
- Traditional scientific communication via a peer-reviewed article at the end of the project, as well as the organisation of a focused scientific workshop at a relevant conference or science meeting.
- Direct communication to public health experts and health policy/regulatory experts, in particular NHS Health Protection Scotland and Scottish Environment Protection Agency (see letters of support), who have expressed a strong interest in the findings of this pilot project, will ensure that the development is informed by existing and emerging stakeholder needs.
- A workshop with stakeholders at the end of the project will be organised to discuss a wider funding application, involving scientific researchers, data scientists and statisticians, public health planners, medical staff and regulatory/policy officers.