The science behind
We are building a customized system that is able to predict fatigue in people with cancer or multiple sclerosis and correlate it to measurable factors, so that our users will be able to plan their daily activities prioritizing the things they love most, while being proactive in managing their fatigue and the factors that most affect it.
​
Our strategy is to train an Artificial Intelligence (AI) algorithm to recognize emergent patterns in people’s physiological parameters, when they are fatigued. We were able to isolate 5 specific physiological parameters that both strongly correlate with fatigue and are easy to monitor and we have developed our custom infrastructure of tests and sensors, able to quantify these parameters in real people. We then embedded the measurement system in our Fuel Test App and started a first trial round.
With the help of our corporate partner and various patient associations, we have been able to build a growing network of enthusiastic volunteers among multiple sclerosis and oncology patients, who actively supported our project by using our Test App and participating in the user testing of the final user experience of Fuel.
Behind Fuel
"We are training an AI algorithm to recognize emergent patterns in people's physiological parameters."
Physiological parameters
We selected 5 physiological parameters to train our algorithm, based on literature review of studies related to healthcare fields and to work accidents prevention.
We then started collecting this data through the Fuel Test App.
​
01 / Eye-blinking rate
02 / Heartbeats
03 / Sleep
04 / Reaction Time Test
05 / Perceived level of fatigue
First testing round
We run a first testing phase with the help of a group of volunteers among the patients that have helped us throughout the project, also thanks to the support of some of the main patient associations helping people with multiple sclerosis and oncology patients. We then provided a group of patients with a smart band and asked them to use our Test App for some weeks.
​
The data related to the parameters mentioned above was then collected in a database. This gives us possibility to analyze it and look for patterns as a preliminary study: our goal is to find recurrent patterns that correlate the fatigue level perceived by a person with the parameters detected in order to associate a fatigue score to physical data, collected all day long, so that we can obtain a fatigue trend across all the day.
The future of Fuel
The future of Fuel will be the implementation of an artificial intelligence in order to make the correlation between fatigue score and physical data as accurate as possible and as passive as possible for patients: working on a big amount of data, the algorithm will be able to automatically understand patterns between them.
​
This algorithm will allow us not only to obtain a fatigue trend, but also to make a prediction on how this trend will evolve during a time frame: this prediction will be more accurate as the amount of data collected increases and it will give to patients the possibility to choose in which activities spend their energy. Having a prediction of when they will feel tired will enable our users to arrange their daily schedule in order to prioritize the activities that they care more about.
​
In a future, fatigue trends could be used not only to optimize patients’ routine, but also to understand how medicine and drugs affect their feeling of fatigue in an objective and data based way.