As the ASCAPE project is moving forward, an essential work has been done in order to determine the design of the different pilots, establish the data determinants in terms of quality of life (QoL) issues to be predicted, as well as the variables of interest in retrospective and prospective datasets that will be used for providing predictions and intervention suggestions, and agree on a common framework within the pilots to evaluate the ASCAPE.

The first step of this work was to reach a consensus among the clinical partners on the QoL issues to be predicted and intervene with in the ASCAPE. For breast cancer, the consensus process (clinical experts from Barcelona, CareAcross, Örebro) identified 15 QoL issues to be predicted through ASCAPE-platform: anxiety, body changes, body image, cognitive impairment, depression, dry vagina, emotional symptoms (loneliness), fatigue, hot flushes, insomnia, joint pain, local symptoms after surgery, lymphedema, neurotoxicity, and sexual dysfunction. A separate selection process was then applied to identify the QoL issues suitable for proposing interventions through the ASCAPE-platform: anxiety, depression, joint pain, fatigue, neurotoxicity, hot flushes, and weight gain.

Using similar consensus process as in breast cancer, clinical experts from Athens, CareAcross, and Örebro identified 12 QoL issues to be predicted through ASCAPE for prostate cancer patients: anxiety, bowel dysfunction, cognitive impairment, depression, erectile dysfunction, fatigue, hot flushes, incontinence, low urinary tract symptoms, loss of libido, musculoskeletal pain, and weight changes. The selection process for QoL issues suitable for interventions through ASCAPE for prostate cancer revealed the following issues: anxiety, depression, fatigue, incontinence, weight changes, sexual dysfunction, and hot flushes.

The initial process to build and train artificial intelligence (AI)-based predictive models for the above mentioned QoL issues will be performed through retrospective datasets. Specifically, five retrospective datasets for breast cancer and two for prostate cancer will be used. After consensus among clinical partners on data determinants of interest for prediction models, a mapping process to align the data among retrospective datasets has begun and will be completed during the work that will performed within the context of the project. Furthermore, additional data determinants not available in retrospective datasets have also been identified in order to prepare the collection of these variables during the prospective phase of the pilots. Three prospective pilots for breast cancer (Barcelona, CareAcross, Örebro) and three for prostate cancer (Athens, CareAcross, Örebro) are planned to collect prospective data of interest and further train and optimize the AI-based predictive models.

An important aspect regarding prospectively collected data is the ability to use active monitoring data. The clinical partners along with the technical ones have reached a consensus on the use of wearables to gather active monitoring data from eligible patients. The same wearable device will be used in three of the four pilots whereas the fourth pilot will use a mobile app to gather some of these data. The active monitoring data to be collected and incorporated into the models are: steps, activity time, activity type, calories burned, heart rate, sleep quality, and sleep quantity.