Title: Moving medically certifiable AI algorithms from the Cloud and onto the Medi-OS Operating System of Medical Devices to automate, make robust and increase uptake of AI in healthcare: Use-case will be community-based spirometry.

Spirometry is a simple medical test used to help diagnose and monitor lung conditions by measuring how much air a patient can breathe out in one forced breath. The medical monitor is called a spirometer.

Spirometry is performed by a trained nurse or physiologist in a GP surgery or during a visit to a hospital clinic. There is a Worldwide shortage of trained and certified staff to perform these tests. Spirometry testing was stopped during the pandemic and it is proving difficult to restart in the NHS because of the shortage of experts available.

**WHY IS SPIROMETRY CARRIED OUT?**

Spirometry is used to diagnose or screen for lung conditions. Lung health conditions cost the NHS £11Billion a year (0.6% of GDP: £10 Billion direct cost to the NHS, £1Billion in lost work) according to the British Lung Foundation (2023).

Conditions diagnosed and monitored using spirometry include asthma, chronic obstructive pulmonary disease (COPD), bronchiectasis, cystic fibrosis and pulmonary fibrosis. Spirometry may also be carried out to check the severity of the condition or to see how a patient is responding to treatment.

**HOW IS SPIROMETRY PERFORMED**

Patients breathe forcefully into a machine three-to-eight times to get reliable results. It requires a skilled clinician to ensure results are of good quality, so these results are not available immediately. After review, results are sent to the doctor who referred the patient for the test for discussion with the patient at a later date. This takes time and is labour-intensive.

**USING ARTIFICIAL INTELLIGENCE TO SUPPORT IMMEDIATE CLINICAL DECISIONS**

There are already web-based AI programmes to accurately "review" breath manoeuvres (data) which are sent to it. These services send back a report in a fully automated process. But this still involves sending patient information away (usually outside of the UK), and there is still a considerable delay before the printable report is returned and reviewed by the doctor.

**INNOVATION**

Using 15,700 existing recordings from patients, the consortium will create new software which will operate on a medical device that can perform expert analysis using AI to give results immediately for use in a GP surgery or community-based diagnostic hub. This will require a strong consortium team to deliver this innovation.

This also opens the way to use this new medical device software, how it learns and how it can be certified as a trustworthy medical device, for a range of other medical monitoring devices.