Deliver intelligence to integrate care: the Connecare way

摘要

Background : According to the European Patient Forum, patient empowerment is the process helping people gain control over their own lives, by increasing their ability to act on issues that they themselves regard as important. Methods based solely on clinical criteria (i.e. professional training and experience) or combined with rules-based approaches (i.e. thresholds for parameters triggering pre-established decisions) constitute most of current practice. In contrast, regular usage of predictive modelling tools for clinical decision support (establishing relationships between sets of variables and outcomes, through statistical or machine learning methods) and/or patient’s support through suitable recommendations or nudges is in its infancy, despite seemingly a natural step towards customization of care. Methods : In the CONNECARE project (ID: 689802), intelligence is delivered to patients and clinicians in the form of automated, contextual, and personalized data analysis and decision making, by dedicated ICT tools: the Recommender System (RS) and three Clinical Decision Support Systems (CDSS). The RS has two goals: (i) enabling patients’ awareness regarding their compliance to prescriptions, thus of the evolution of their condition (i.e. worsening adherence to treatment), and (ii) pursue engagement of patients during treatment, thus a more active role in the process (by sending personalized messages). The goal of the CDSS is threefold: (i) deliver to clinicians risk assessment and stratification services (CDSS-risk), so as to let them focus on the most critical cases, (ii) provide geolocation services (CDSS-map) to color patients on a map depending on severity of condition (e.g., according to Charlson risk index), and existence of socio-cognitive and economic barriers to treatment (i.e., anxiety, living alone, low incomes), while suggesting optimal routes for home visits and nearby, relevant medical facilities; and (iii) assist clinicians in designing personalized clinical pathways (CDSS-path), by suggesting “templates” of the process (in the form of a workflow) to be further specialized manually. Both tools operate by constantly monitoring patients’ data (i.e., latest clinical measurements, current prescriptions) – while respecting privacy concerns – so as to feed algorithms specifically conceived so as to balance the tradeoff between (i) scientific relevance (i.e. state-of-art techniques) and (ii) clinical acceptance (i.e. validation by domain experts). Results : The CONNECARE system is a federation of subsystems each providing goal-oriented functionalities, whose main components are the Self-Management System and the Smart Adaptive Case Management system. To be smart, adaptive, and personalized, they rely on the RS and CDSS, which are now being integrated in the next release of the system, already adopted in implementation studies started in May 2018 in Lleida, Israel, and Groningen. Conclusions : Approaches to increase patient empowerment vary from patient self-management programs, to promoting patient involvement in treatment decision-making, to facilitating the physician-patient cooperation. Under the umbrella of the CONNECARE project, we defined intelligent systems aimed at providing support to patients through a recommender system and to clinicians through a combination of three clinical decision support systems.