VISION & METHODOLOGY
Supporting the quality of life of patients undergoing cancer treatment.
FAITH aims to provide an Artificial Intelligence application that identifies and analyses depression markers in people that have undergone cancer treatment, with the goal of increasing awareness of their mental health situation, giving them the possibility to improve their quality of life and providing intelligent post-cancer support.
In order to do so, FAITH will collect and monitor a range of health indicators. These will allow FAITH to analyse and infer information about the mental status and quality of life of a person in a non-intrusive way.
CHALLENGES
FAITH’s ‘AI Angel’ analyses depression markers to predict negative trends in patients’ mental health.
The selected markers refer to several lifestyle categories, including activity, voice patterns, sleep, and nutrition.
FAITH analyses potentially sensitive data, data that needs to be protected. To safeguard user’s privacy, we record data on the user’s mobile phone only, and nowhere else.
This is then analysed using a Federated Machine Learning approach, which does not require the sensitive data to be sent to the Cloud, or leave the user’s device otherwise, for processing. Moreover, FAITH will deliver personalised AI models directly to each patient’s device.
FAITH monitors markers that could suggest negative trends potentially leading to depression: it does not make automatic diagnoses of depression, neither aims to replace clinicians in doing so.
Rather, it means to works in support of clinicians, providing them an additional tool for their practice. That is why we validate its concept through trials involving clinicians. We want to ensure the ‘AI Angel’ is as useful as possible to them.
To ensure Artificial Intelligence is ethical and trustable, it must be transparent: it should provide not only an output, but also an understandable explanation of its reasoning.
Explainable AI is key in healthcare: our ‘AI Angel’ will provide doctors with a reason for its output, allowing them to interpret the results in order to make a diagnosis and take a clinical decision.
FAITH collects part of the data in a “passive mode”, namely with no explicit interaction between the system and the user. However, other information require an active interaction, high engagement, and motivation from the user’s side.
To elicit this kind of information, an ‘Angel App’ on the patient’s phone will prompt them with specific questions for them to answer. The App is designed with user experience in mind, to foster proactive and regular users’ engagement.
TESTING
FAITH will have trial sites in two hospitals through the European Union (in Spain and Portugal). They will carry out trial pilots involving real end users (both clinicians and patients) to collect the data necessary to train the AI algorithms behind the FAITH solution.
After an initial requirements gathering, the concept will be prototyped for a real-life situation trial to use at the hospital pilot sites. The trials, specifically related to each use case, will allow FAITH to be TESTED by doctors and patients according to their specific needs.
This input will feed back further round of requirements gathering and concept refinement.
All pilots will be subject to the highest ethical standards to assure the privacy and protection of data.
EXPECTED RESULTS
FAITH will provide an Artificial Intelligence application that remotely identifies and analyses depression markers to predict negative trends in people that have undergone cancer treatment.
This concept would present healthcare providers with advanced warnings to allow timely intervention, and increase awareness of cancer patients’ mental health situation, improving their quality of life and allowing them to receive intelligent post-cancer support.
TIMELINE
VISION & METHODOLOGY
Supporting the quality of life of patients undergoing cancer treatment.
FAITH aims to provide an Artificial Intelligence application that identifies and analyses depression markers in people that have undergone cancer treatment, with the goal of increasing awareness of their mental health situation, giving them the possibility to improve their quality of life and providing intelligent post-cancer support.
In order to do so, FAITH will collect and monitor a range of health indicators. These will allow FAITH to analyse and infer information about the mental status and quality of life of a person in a non-intrusive way.
CHALLENGES
FAITH’s ‘AI Angel’ analyses depression markers to predict negative trends in patients’ mental health.
The selected markers refer to several lifestyle categories, including activity, voice patterns, sleep, and nutrition.
FAITH analyses potentially sensitive data, data that needs to be protected. To safeguard user’s privacy, we record data on the user’s mobile phone only, and nowhere else.
This is then analysed using a Federated Machine Learning approach, which does not require the sensitive data to be sent to the Cloud, or leave the user’s device otherwise, for processing. Moreover, FAITH will deliver personalised AI models directly to each patient’s device.
FAITH monitors markers that could suggest negative trends potentially leading to depression: it does not make automatic diagnoses of depression, neither aims to replace clinicians in doing so.
Rather, it means to works in support of clinicians, providing them an additional tool for their practice. That is why we validate its concept through trials involving clinicians. We want to ensure the ‘AI Angel’ is as useful as possible to them.
To ensure Artificial Intelligence is ethical and trustable, it must be transparent: it should provide not only an output, but also an understandable explanation of its reasoning.
Explainable AI is key in healthcare: our ‘AI Angel’ will provide doctors with a reason for its output, allowing them to interpret the results in order to make a diagnosis and take a clinical decision.
FAITH collects part of the data in a “passive mode”, namely with no explicit interaction between the system and the user. However, other information require an active interaction, high engagement, and motivation from the user’s side.
To elicit this kind of information, an ‘Angel App’ on the patient’s phone will prompt them with specific questions for them to answer. The App is designed with user experience in mind, to foster proactive and regular users’ engagement.
TESTING
FAITH will have trial sites in two hospitals through the European Union (in Spain and Portugal). They will carry out trial pilots involving real end users (both clinicians and patients) to collect the data necessary to train the AI algorithms behind the FAITH solution.
After an initial requirements gathering, the concept will be prototyped for a real-life situation trial to use at the hospital pilot sites. The trials, specifically related to each use case, will allow FAITH to be tested by doctors and patients according to their specific needs.
This input will feed back further round of requirements gathering and concept refinement.
All pilots will be subject to the highest ethical standards to assure the privacy and protection of data.
EXPECTED RESULTS
FAITH will provide an Artificial Intelligence application that remotely identifies and analyses depression markers to predict negative trends in people that have undergone cancer treatment.
This concept would present healthcare providers with advanced warnings to allow timely intervention, and increase awareness of cancer patients’ mental health situation, improving their quality of life and allowing them to receive intelligent post-cancer support.