Artificial intelligence lands in the hospital: helps in the emergency room and reduces waiting lists

Artificial intelligence lands in the emergency room - where seconds can make all the difference - to help doctors decide early and well what to do when there is a suspicion of a fracture or worse, a stroke, but AI can also help hospitals to better manage visits and examinations by preventing patients from 'blowing off' and not showing up for their appointments and eventually suggesting to white coats whether or not to prescribe a CT scan or an MRI. In this way, waiting lists are shortened and places are freed up for those who really need that service. These are some of the most innovative experiences among the first forty included in the Observatory on Artificial Intelligence in Health Care, presented at the Logos & Téchne' Forum, organised in Syracuse a few days ago by Fiaso, the federation that brings together the managers of local health authorities and hospitals.

At the Asl 2 Sgavonese, for example, artificial intelligence has entered emergency radiology and has become an additional 'second reading' tool for triaging patients, with the acceptance of radiologist physicians growing over time: here in particular, an Ia solution that helps detect fractures on all X-ray examinations will be used from 2021. In another hospital in Liguria - the Asl 4 Chiavarese - another Ia system for reading X-rays that helps detect bone fractures has been tested in the emergency room 'Boneview'. The latter technology has also entered the emergency department of the Asst Santi Paolo e Carlo in Milan. In the emergency department of the University Hospital of Perugia, on the other hand, a virtual assistant has entered for early stroke recognition thanks to voice, facial and biometric analysis: this assistant has been well received by doctors for its usability and efficiency during triage, improving diagnosis, intervention and survival times. Artificial intelligence is also used at the Asst Santi Paolo e Carlo in Milan for the automatic analysis of CT images, again in cases of ischemic stroke.

So far the experiences in emergency rooms. But artificial intelligence could also help to better govern the number one problem in healthcare: waiting lists. According to a study using an AI algorithm, in the Apulia region at least four out of ten prescriptions for CT and MRI scans are unnecessary. Carried out by Aress Puglia, the project is based on an analysis of over 17,000 prescriptions: at its heart is a generative artificial intelligence system based on the LLaMA 3.1 model, capable of assessing the clinical appropriateness of a diagnostic prescription. The algorithm was trained to associate each requested examination with the patient's condition, comparing it with official guidelines. The results revealed that only 39% of the requests fully complied with the appropriateness criteria, 43% were inappropriate (i.e. not in fact necessary), and the remainder were only partially useful. At the Asl Napoli 3 Robo-Cup, an automated system using an algorithm applied to the data of the Single Booking Centre, has instead allowed the empty slots between one appointment and the next to be halved, with benefits also on waiting lists. The system made it possible to reduce the percentage of the so-called 'no-show' phenomenon from 40% to 20% by optimising the gaps in appointments left free. Using booking data, Robo-Cup sends text message reminders, allowing patients to confirm or cancel their appointments. Then, again using artificial intelligence, it reassigns available appointments, including through recall services, controlled overbooking and pre-listing to optimise the use of free slots. In the event of a non-response, the intervention is handled by the operator.

'Citizens are ready: they have understood the potential of artificial intelligence in healthcare and are demanding concrete solutions,' explains Giovanni Migliore, President of Fiaso. 'Now it is up to us, as a public healthcare system, to respond with vision and competence. Innovation is no longer a promise: it is already reality, and it is transforming the way we take care of people and organise services'. With the birth of the Faso Ia Observatory, 'a concrete path of innovation has begun in Syracuse,' adds Migliore, 'which will take us to Rome, next 29 January, for a major event at which we will present the results and award prizes for the best experiences gained through the collaboration between healthcare companies, research and technology companies.