HEALTH (HarnEssing Artificial intelligence to Lead Transformative Healthcare)

Background/rationale Poor air quality is the largest environmental risk to public health, with up to 36,000 UK deaths occurring every year.

Exposure to air pollutants can lead to exacerbations of ill health, increased healthcare utilisation (primary care consultations and inpatient hospital admissions), multi-morbidity and mortality. Air pollutants can exert disproportionately large health effects on disadvantaged and vulnerable groups.

The UK Health Security Agency (UKHSA), formerly Public Health England, estimated that total costs to the NHS and social care (2017-25) will be £1.6 billion.

The Department for Environment Food and Rural Affairs have highlighted the need to communicate with the public during air pollution episodes, with UKHSA acknowledging that the use of behavioural interventions, alongside other local and national strategies, can have a synergistic effect.

We want to co-produce a personalised intervention that will notify individuals when air pollution levels are likely to increase, thus helping them to make informed decisions on how to reduce their exposure and/or better manage the consequences of exposure.

However, this information must be communicated to people in the right way, at the right time, and with the right content.

Aim and Objectives To co-produce our communication strategy with disadvantaged and vulnerable groups, tailoring it to their needs and available local infrastructure, and mitigating against digital exclusion.

To develop and internally validate a proof of concept risk score to identify those at higher risk and needing intervention.

Development work plan We will use existing links with organisations such as NHS North Tyneside Future Case Working Group, Connected Voice, Healthwatch, Age UK, VODA and local community groups to reach target groups.

We will specifically include those perceived at high risk of digital exclusion, such as people with disabilities, ethnic minority groups and/or lower income and deprived groups, those living in rural areas and older people. Participants will have the choice of participating in either a semi-structured interview or focus group.

The topic guide will include questions on what information individuals want to receive about elevated air pollution levels, when, where and how they would like to receive it, and potential benefits and barriers of each, including how we mitigate against digital exclusion.

We will co-produce the content of these messages, the range of actionable recommendations (including how local initiatives and infrastructure could help support desired change), and explore any potential unintended consequences.

Building on our previous work looking at how different multi-morbidity clusters utilise healthcare, we will also develop a proof of concept risk score to investigate the risk of increased healthcare utilisation and/or acute unplanned admission across different multi-morbidity clusters when air quality is poor.

We will use multivariable logistic regression and identify the predictors for high-risk individuals and their associated weights to transform the prediction model into a risk score.

Anticipated Impact and Dissemination We enjoy excellent working relationships with relevant leaders in NHS England, DEFRA, WHO Global Air Pollution group, UK charities, UK Faculty of Clinical Informatics, and Integrated Care Systems, and are ideally placed to ensure that our findings reach senior decision-makers.