Increasing the effectiveness in the diagnosis of dysphagia with Artificial Intelligence

Oropharyngeal dysphagia (OD), or difficulty swallowing, is commonly seen in older adults. The disease causes malnutrition, dehydration, and aspiration pneumonia that, in severe cases, can be fatal. OD is often not recognised as the main cause of these complications because there is no systematic screening to diagnose it. Treatment is only feasible with minimal interventions, such as fluid and texture adaptation, nutritional supplementation, and oral hygiene.

It is estimated that less than 20% of cases in hospitalised patients are diagnosed, treated and corrected. Current systems to identify dysphagia are based on questionnaires and interviews, which are only successful in 55% of cases. In Europe, it is estimated that about 30 million people suffer from this difficulty in swallowing.

Prospect of more efficient diagnosis

The research nurse Alberto Martín, from Hospital de Mataró – in collaboration with the director of research and innovation at the Maresme Health Consortium (CSdM) Pere Clavé – has developed a screening system to predict the risk of suffering from OD in hospitalised patients. The Artificial Intelligence Massive-Screening for Oropharyngeal Dysphagia (AIMS-DO) is a patented software for automatic diagnosis, without the need for intervention by the nursing or medical staff.

Implementation in other hospitals

Currently, the AIMS-OD is in a pilot phase at Hospital de Mataró where patient records from between 2013 and 2017 are being analysed with anonymised and encrypted data from more than 5,000 patients.  The system has shown a success rate of 85% based on the analysis of the more than 17,000 variables that can influence the risk of suffering from OD.

Another advantage is that the system can perform the same number of screenings in eight seconds as would be done in 42 hours with traditional methods. The algorithm has been refined over the last three years based on experience in OD research. Hospital de Mataró has carried out studies describing risk factors, normalised variables and information about drugs.

(Photo: from left to right Pere Clavé and Alberto Martin, collecting the Creatic award).The system is part of the PECT Mataró-Maresme project: innovation ecosystem for caring cities, financed by the European Union Regional Development Fund (ERDF). It also has the support of CaixaImpulse and the Mataró university centre Tecnocampus, which was awarded the Creatic prize worth €12,000. This has been welcomed within the StartHealth programme, which provides expert advice for the launch of new start-ups in the health sector. According to Clavé, the goal is to implement the intelligent system in every hospital in Europe.