Seleziona la lingua

Get in touch with Sleepacta:

Progetti Sleepacta

Insomnia: now diagnosed with artificial intelligence and wearable devices

Home > News > Insomnia: now diagnosed with artificial intelligence and wearable devices

We are pleased to report an article published in AboutPharma.

It took 30 years to update the diagnosis of insomnia to include among the symptoms considered essential the impact the condition has on daily activities. Now thanks to artificial intelligence, machine learning and wearable devices, it could take much less to transform a purely clinical diagnosis into an objective one, based on data collected from simple devices, as is done with blood pressure and hypertension. Explains Ugo Faraguna, a neurophysiologist and professor at the University of Pisa: “Today, one is considered a chronic insomniac who has difficulty falling asleep, maintaining sleep, or waking up early for at least three times a week for at least three months, with symptoms of daytime sleepiness (without the latter criterion, one could simply be dealing with a short sleep). Diagnosis is based on the complaints the patient reports, and there may be under or overestimation. Also because there are no objective instrumental reference values as there are for other conditions.”


A physiological range for sleep as well


Faraguna’s goal is just that: to use consumer-grade wearable devices that are low-cost and convenient to wear (such as bracelets and smart watches) to collect data and one day arrive at objective parameters on sleep quality and duration that can be used for clinical purposes. “Today, the diagnosis of hypertension is made on the basis of blood pressure values collected with easily accessible devices found in pharmacies or in people’s homes,” the physiologist points out, “which has saved many lives. The same is not yet possible with insomnia, a condition linked to numerous disorders, from Alzheimer’s to cardiovascular and mental disorders, but for which the only instrumental test is polysomnography, a complex investigation that few perform. But sleep is a physiological variable like blood pressure or cholesterol, and at wait the ultimate goal is to have reference values exactly as you have for the others.”


Wearable devices


A few years ago so-called trackers appeared, which allow a single wearable device to collect numerous parameters, including sleep. “The community of scientists working on sleep disorders had high hopes for smartbands and smartwatches,” comments Farguna. “We thought we finally had sensors that were accessible to anyone and could collect the regulatory data needed to establish the physiological range I mentioned earlier. But we soon realized that these were not medical devices. The software is not scientifically validated and verified and cannot be used for clinical purposes to perform certifications.”


Artificial intelligence


This is where artificial intelligence comes in. With this in mind, in 2017 Faraguna and other partners founded Sleepacta a spin-off from the University of Pisa that led to the development of a technology (“Sleep”) that can transform raw data from trackers into reliable data using artificial intelligence. Faraguna continues, “We download data from the bracelets and analyze it with our artificial intelligence tools, turning an object that is not medical into a reliable, certified, validated data with a series of publications and certification from the Ministry of Health to support it.”

To get to this point, Faraguna says he and colleagues tested the different devices on the market today by comparing them with polysomnography, which remains the gold standard for diagnosing insomnia. The process made it possible to train and validate the algorithm and the bracelets, which are constantly monitored whenever new ones come on the market. “That of scientific validation is a key step,” Faraguna emphasizes. “Those who deal with insomnia must use these sensors, but they must do so in a validated and certified way, with solid data. And above all, these data should not be given directly to patients but to professionals who know how to read them.”


The Sleepacta model


Sleepacta’s model is just that. They provide the technology to public and contracted hospitals and also to pharmacies on the ground. Those who go to these facilities receive a commercial wearable device on loan, which some already own, to use for about a week and then bring back. The data is downloaded and sent to the startup, which shortly thereafter provides a clinical report. At that point, the doctor evaluates the parameters to see if there are alterations in sleep physiology, if an intervention has worked or if an intervention plan should be applied. “We started selling this service to hospitals in 2018 and recently started distributing it to pharmacies as well. Interest is high from clinicians and has grown so much over the years, partly because more and more specialists are interested in sleep, from cardiologists to dentists,” Faraguna says.


An under-diagnosed disease


The service currently can be paid for by the patient or by the SSN in public or contracted hospitals provided by Sleepacta, where the cost of monitoring varies by region. Concludes Faraguna, “About 500 million smartbands/smartwatches were shipped last year alone. This shows us that measurement systems are already widespread, we just need to use them in the right way. Their application is extremely important, especially for insomnia: it is estimated that more than 90 percent of cases go untreated because they are not recognized.”

Placeholder autore articolo
Redazione Sleepacta