AIR SAFETY. A programme to prevent aerosol borne respiratory diseases using Artificial Intelligence

Scientific Abstract (499 words, max allowed 500 words) Research question How to enable simple user-introduced mitigations to reduce hospital airborne disease transmission by implementing an Artificial Intelligence driven air-Safety Tool (AISaT) platform? Background Inadequate hospital room airflow drove nosocomial disease transmission during the COVID-19 pandemic. Safeguarding against future respiratory viruses with a rapidly-deployable technology is a critical NHS need. Aims and Objectives We aim to implement our AISaT software. The software will recommend an optimal mitigation solution and a list of possible ranked alternatives, based on equipment already available. Users will thus create bespoke mitigation-device layouts to reduce aerosol-droplet levels in clinical-spaces. Objectives: Understand airborne disease risks and mitigation options Enhance the AISaT software to evaluate increasingly complex hospital environments at dramatically faster speeds to recommend best low-cost mitigation strategies. Assess software effectiveness and implementation in clinical environments Assess cost efficiency Develop an NHS-wide roll-out strategy, informed by national policy. Methods We will develop a programme theory using the MRC Complex Evaluation Framework, combining implementation-science, quality-improvement, and social-science theory. Two systematic reviews will inform the programme about factors influencing nosocomial disease transmission and solutions. We will embed key stakeholder perspectives including staff and patients. Our AISaT software currently takes time to make predictions. Using experiments and computational simulations, we will dramatically enhance the AI software s speed and accuracy. Using an RNA bacteriophage virus in an environmental chamber, we will confirm reduction in infectivity with mitigations deployed according to AISaT recommendations. Next, we will develop alternative techniques using oligonucleotides and aerosol-droplets counting for easy clinical deployment. We will also perform usability testing. Clinical trials in 2 hospitals will assess effectiveness of AISaT recommendations in outpatient, day-case rooms, and wards to demonstrate reduced infection risks. In parallel, mixed methods process-evaluation will generate a convergence coding-matrix to develop an implementation strategy. Using data-envelope analysis, we will examine technical efficiency of AISaT software recommendations. We will then evaluate roll-out at 10 NHS hospitals across England representing healthcare inequalities using observations and interviews to identify barriers and enablers, informed by the Consolidated-Framework for Implementation-Research. We will engage national policy-makers, translating research findings into policy-relevant outputs. We will form an independent programme steering committee. Timelines for delivery Within 12 months, we will identify NHS safe-air needs and affordable retrofit approaches. Testing[TM1] effectiveness in clinical locations will occur during years 2-3, while cost-effectiveness demonstration will happen years 2-5. Roll-out will occur in the last 3 years. Policy engagement will continue throughout the programme. Anticipated Impact and Dissemination Up to 1 million NHS patients visits at 10 study sites will directly benefit from safer air. Low energy costs will help the NHS towards achieving Net Zero goals. If another COVID-19 surge occurs, using AISaT could save 1,500 lives nationally per mont­h. Our programme is embedded with a strong PPIE programme including professional stakeholder panels, Patient Advisory Group, taster weeks, summer schools, and student science festivals. In collaboration with British Science Association CREST Awards, we will engage a new generation of STEM scientists for a lasting educational legacy.