Here we take a look at all of the entries for the Health Tech Awards 2023 category of ‘Best Use of AI and Automation Tools’.
Polygeist and Gloucestershire Hospitals NHS FT
Overview: Real Time Patient Risk and Long Stay Prediction: AI in the hospital at Gloucester Hospitals NHS Foundation Trust.
Why? Reducing long stays in hospital has rightly been a national priority for NHS England over the past few years, and forms a key plank in the NHS Recovery Plan.
What happened? This second phase project demonstrates real time prediction of patients’ length of stay in hospital, personalised risk factor combinations, and likelihood of mortality, using an AI algorithm. Clinicians receive up to date, real-time estimated date of discharge, accurate to within 12 hours, in EPR. This means that every single observation captured by nursing staff and prescribing clinicians, along with all diagnostic test results, are all then processed in real time, surfacing key insight to the right person, at the right time. This technology has been privately trialled on recent historical data, and can predict over 99 percent of long stayer patients. Average error was within just 12 hours of the actual patient discharge time, over a 30 day period. Using population data, key risk factors contributing to that patient’s stay (both medical and logistical) can be flagged to clinicians; this enables them to plan for potential blockers to discharge and onward care. Analytical assessment, matched with recently published, peer reviewed evidence, suggests that clinicians could safely reduce the average stay between one and five days per long staying patient. Over 7,000 lives per year, nationally, can be saved simply by identifying those high risk patients and discharging five days earlier; based on statistical analysis of the GHNHSFT data, and national mortality statistics. Moreover, early discharge reduces the readmission rates by 50 percent, projecting 3.15 million bed days per year freed nationally.
Looking ahead. The team hope to help this technology reach more patients in other trusts, promoting safer and more effective care for patients and helping reduce long stays in hospital.
Answer Digital
Overview: Answer Digital built and deployed the software behind an innovative deployment and federated learning platform that transforms AI implementation and value delivery in healthcare.
Why? The process for deploying AI apps at NHS trusts currently runs on the months-to-years’ time scale because of the bespoke connectivity and hardware issues that need to be solved each time a new product gets deployed.
What happened? The AI Deployment Engine (AIDE), allows healthcare providers to deploy AI models safely, effectively, and efficiently by enabling the integration of AI models into clinical workflows. AIDE is now deployed in 10 NHS trusts. Through connection to a live stream of clinical data, clinicians can access real-time AI analysis deployed within hours. AIDE’s unified infrastructure allows multiple algorithms to run simultaneously through bespoke APIs. This provides NHS trusts with the capability to run multiple AI solutions on a single platform. AIDE lowers the barrier to deploying clinical AI allowing individual trusts to speed up the AI transformation of patient pathways. The Federated Learning Interoperability Platform (FLIP) helps to link data from multiple NHS Trusts to enable AI at scale. Our federated learning approach brings algorithms to the data within each NHS trust’s secure enclave, without the need to share information outside the organisation or break local governance rules. The Federated Learning Interoperability Platform (FLIP) ensures a high level of fidelity in AI output models compared to traditional aggregative data strategies because the data it trains on does not need to be anonymised before use, ensuring adherence to each trust’s governance and data privacy regulations whilst ensuring that AI models are scalable.
Looking ahead. Answer Digital are excited about the prospects that a platform-based approach to AI deployment will bring to the NHS.
PATCHS
Overview: PATCHS is a next generation online consultation system powered by artificial intelligence triage and workflows, helping practices manage rising demand whilst improving patient access.
Why? PATCHS helps tackle key issues in primary care, including workload management, health inequalities, and long waiting times.
What happened? PATCHS is the UK’s fastest-growing online consultation solution, supporting 1,000+ GP practices and 10m+ patients. It enables easy two-way communication between patients and their GP practice. It has been implemented at-scale across multiple regions including North West London (2.2m patients), Cheshire and Merseyside (2.7m patients), West Yorkshire (2.7m patients) and North Central London (400k patients). Results have shown improvements in workload management including 85 percent of all requests being dealt with fully online, a 20 percent reduction in self-care patients submitting requests, and only 22 percent of health requests requiring a face-to-face appointment. Telephone calls reduced by 50 percent, clinical safety improved thanks to PATCHS AI which is estimated to detect 94 percent of all urgent and emergency patient requests, and waiting times were reduced with an average resolution time for a high usage PATCHS region of 5.6 hours. On an inclusivity basis, UoM research found that many older patients found PATCHS easier to navigate than expected and it was their preferred method for contacting the GP practice, whilst patients who struggle with telephone or in-person appointments find it easier to use PATCHS.
Looking ahead. The PATCHS team plans on continuing to develop the system and to roll it out across more regions to enable even more benefits for primary care.
