The data that is going to be collected during the LifeChamps pilots is of utmost importance as it will serve as the basis for the LifeChamps quality of life and frailty support monitoring and modelling. In July 2021, a series of workshops were conducted among the LifeChamps technical partners with the following goals: First, to define the list of variables that are going to be collected through the sensing devices, the LifeChamps mobile app and the electronic health records. This list includes, for example, the measurement of the heartbeat per minute, the daily number of steps, the self-assessment of quality of life and age. Second, to define any additional variables that might be computed from the original data, i.e. data collected from the several sources. For example, the trend of increasing or decreasing number of steps taken daily may be extracted from the measurement of the number of daily steps during a week. Furthermore, more complex variables, such as a loneliness index, may be inferred from data collected using home localization sensors. Third, to define even more complex analyses, where complex and usually subjective concepts, such as depression or cognitive decline, are assessed using the collected multisource data. This was a very important exercise to ultimately define the LifeChamps pilot requirements, regarding data collection and how the data is going to be organized in the LifeChamps platform (i.e., the schema of the data).
Another crucial exercise was conducted by the LifeChamps clinical partners regarding the potential clinical prediction models that are going to be implemented in the LifeChamps platform. They revisited the clinical questions, which they may be interested to answer within the scope of this project, and better defined the clinically relevant outcomes. Moreover, they grouped the data that is going to be collected during the LifeChamps pilots into very relevant, maybe relevant, and irrelevant predictors of the clinical outcomes of interest from a clinical perspective. LifeChamps data scientists will take this clinical expertise into account when defining the LifeChamps clinical predictive models.
