Artificial intelligence – A gradual revolution in hospitals

Artificial intelligence

A gradual revolution in hospitals

Since 2010, the use of artificial intelligence is booming. Healthcare establishments are beginning to get to grips with this phenomenon, but, for the moment, intelligent software and robots have made only a limited impact in this sector. We take a closer look at the AP-HP Paris hospital trust and Nantes university hospital.

A chatbot capable of exchanging text messages with patients before and after their outpatient hospital visits? The AP-HP intends to equip its 23 outpatient surgery units with just such technology by 2019. In the event of an alarming or incoherent response from patients, the care staff will be notified and can view the messages exchanged via a secure platform, and contact patients by telephone if necessary.

This method, which has been tested at the Saint-Antoine hospital in Paris, has “proven its effectiveness” compared with conventional follow-up by telephone, according to an AP-HP spokesperson. The trust reports that pre-operation instructions are “better followed” and the number of patients late by more than 30 minutes has been “reduced ten-fold”, citing a study published in 2016 in the Praticien en anesthésie-réanimation (anaesthesia and resuscitation practitioners) journal. This leads to better patient management and, potentially, improves the way in which the hospital departments in question are organised.

Numerous applications

In practice, the use of tools based on artificial intelligence in healthcare establishments is still in the very early stages, and there is a long way yet to go. For the moment, such tools are limited to warning systems triggered when two incompatible medications are prescribed to the same person, voice recognition tools, X-ray diagnosis assistance software, and autonomous robots that can transport small items of equipment to the sterilisation room.

However, there are a great many potential applications. Intelligent software has the potential to offer time-savings to technical, administrative and medical staff. “We have plans to optimise the coding of the hospital’s activities,” explains Pierrick Martin, Deputy Director of the digital services department at Nantes university hospital. The aim is to develop a software program capable of automatically suggesting PMSI codes. The Nantes hospital is also interested in the idea of a computer-assisted maintenance management system for its premises and equipment. “Artificial intelligence would enable us to optimise our preventive and curative maintenance efforts,” says Pierrick Martin.

The AP-HP has implemented a “healthcare data warehouse” which enables certain data (e.g. medico-financial, pharmaceutical) to be used “under controlled conditions” to “develop and test artificial intelligence algorithms” for research purposes and to improve care. The use of clinical data using algorithms could also, in time, enable “personalised prevention services” to be deployed. This would more than likely result in the need to recruit staff with new skill sets, such as data scientists.

The digital hospital of the future

To promote the development of sophisticated and reliable tools, the AP-HP and Nantes university hospital jointly launched, in 2017, a call for expressions of interest in the digital hospital of the future, and have selected 15 projects to which they will lend their support. Some of these projects are based on artificial intelligence, such as the creation of a digital copy of the hospital, “a true digital dynamic mock-up of the hospital that serves to simulate the flow of patients within the real hospital, within a single building, or on a floor, in a ward or in a particular room”.

Another project aims to develop a digital companion that is “multi-sensory and multi-skilled, capable of interacting, understanding, anticipating and, to an extent, responding to the needs and expectations of patients as to their physical safety, psycho-physical well-being, questions relating to their stay in hospital, and their state of health”.

The healthcare sector can thus be expected to evolve significantly in the coming years. In particular, numerous companies, from digital giants (such as Google, Facebook, Microsoft and IBM in the USA and Baidu in China) to small start-ups, are looking at applications to identify signs of ocular degeneration through a simple digital “sweep” of the eye, or to detect cases of tuberculosis from X-ray images, and so on.

Visitors to the forthcoming Paris Healthcare Week will get a sneak peek at these and other innovations, and a taste of the opportunities presented by such technologies.