Periodic Reporting for period 1 - RE-SAMPLE (RE-SAMPLE)
Período documentado: 2021-03-01 hasta 2022-08-31
Patients with Chronic Obstructive Pulmonary Disease (COPD) often also have other Complex Chronic Conditions (CCCs), like diabetes, depression or chronic heart failure. As these conditions have common risk factors and overlapping symptoms, this easily leads to delay of appropriate treatment. Flare ups (called ‘exacerbations’) of COPD and CCCs are detrimental to patient, health and social outcomes. Therefore it is important that patients receive timely care, which can prevent worsening of their disease(s), and that fits their personal needs and characteristics.
RE-SAMPLE will answer upon the challenge of the increasing number of patients with COPD and CCCs who are in need of such integrated, personalised, and holistic approach to manage care. We do this by monitoring disease progression and other parameters in daily life (called real-world data; RWD) and combine these with other relevant information, like clinical data from the hospital, research data and data from the environment. With this data, we use artificial intelligence to predict upcoming CCC exacerbations and to coach patients in their self-management in an adaptive and tailored manner. These are integrated in an eHealth programme, called the virtual companionship programme. This programme is offered to patients (i.e. virtual companion) and healthcare providers (i.e. active support programme) and supported by shared-care facilities for additional monitoring. These together serve shared-decision making and cross care paths towards daily life. RE-SAMPLE’s architecture strictly follows privacy and security-by-design principles and uses highly-innovative techniques (i.e. secure multiparty computation). The integrated disease management fits values and needs by continuous involvement of stakeholders and our Pan-European service model will ensure deployment in different systems throughout Europe.
RE-SAMPLE will answer upon the challenge of the increasing number of patients with COPD and CCCs who are in need of such integrated, personalised, and holistic approach to manage care. We do this by monitoring disease progression and other parameters in daily life (called real-world data; RWD) and combine these with other relevant information, like clinical data from the hospital, research data and data from the environment. With this data, we use artificial intelligence to predict upcoming CCC exacerbations and to coach patients in their self-management in an adaptive and tailored manner. These are integrated in an eHealth programme, called the virtual companionship programme. This programme is offered to patients (i.e. virtual companion) and healthcare providers (i.e. active support programme) and supported by shared-care facilities for additional monitoring. These together serve shared-decision making and cross care paths towards daily life. RE-SAMPLE’s architecture strictly follows privacy and security-by-design principles and uses highly-innovative techniques (i.e. secure multiparty computation). The integrated disease management fits values and needs by continuous involvement of stakeholders and our Pan-European service model will ensure deployment in different systems throughout Europe.
In the first 18 months of the project, we focused on building a strong foundation for the RE-SAMPLE platform and services in terms of a collective knowledge base and active end-user involvement, leading to definition of our functional specifications, platform architecture, release of our first functional prototype and the start of an observational cohort study for RWD gathering. More specifically, we have accomplished the following main achievements:
Continuous end-user involvement and requirements
The successful design and implementation of the RE-SAMPLE platform and services necessitates a good understanding of the end-users, including their context of use, values and needs. Therefore, in RE-SAMPLE we followed an human-centred approach where stakeholders are engaged throughout the project. Together with these end users, we identified parameters that patients and professionals find important for monitoring. Also we elicited user needs in terms of privacy and data collection, self-management, decision making, data visualisation, communication, and coaching. Low-fidelity prototypes, scenarios and data-flow diagrams were developed, and prototypes were evaluated. This work led to the specification of user requirements for the virtual companionship programme. Next, the first prototypes and data visualization were further specified, developed and tested in end-user studies. In addition first iterations of the RE-SAMPLE service model were developed together with the stakeholders and consortium members.
Collective knowledge base
Knowledge and parameters that could be important for predicting exacerbations of COPD or comorbid CCCs are currently unknown or distributed among different (heterogenous and multimodal) data sources and representations. Therefore we have established a collective knowledge base, by identifying and bringing together retrospective datasets, as well as clinical-scientific knowledge and end-user preferences and expertise (see ‘continuous end-user involvement’ above). We performed state-of-the art literature (review) studies about clinical predictors, as well as diagnostic tools to differentiate acute heart failure from acute COPD. Furthermore, we identified and gathered retrospective data from the pilot sites including information from existing clinical studies and Electronic Health Record data, and from environmental databases for e.g. weather. Also a common methodology for data homogenisation is established along with standardization requirements.
Platform architecture and first functional prototype
Taking into account the sensitive type of (health) data to be handled, the RE-SAMPLE platform needs to be have a secure and privacy-securing data management in all its components and communications.
Therefore the RE-SAMPLE platform architecture - i.e. the name of the set of components and subsystems that make up the technological solution - has been defined from an early phase of the project. The development followed privacy- and security-by-design principles to ensure all functional and non-functional (security and privacy) related requirements imposed by the General Data Protection Regulation, ethical rules and the users. Also, data processing processes (pseudonymization, integration of heterogeneous multimodal datasets) were specified and ecosystems, roles, and key components identified. The first predictive models trained on the retrospective datasets collected by the three pilot sites are available to predict exacerbations of COPD. Furthermore the RWD monitoring application has been developed to use in the observational cohort study of RE-SAMPLE, which will later evolve to the virtual companion application.
Observational Cohort study
Current CCC disease management relies heavily on information acquired during scheduled visits when patients are usually stable, whereas actual symptoms and changes during daily life are not quantified. Therefore, in the early phase of the project the research protocol for our observational cohort study was designed. It has received ethical approvals in the three pilot sites (Netherlands, Italy and Estonia), and patient inclusion has started. We identified a set of parameters (e.g. daily symptom diary, physical activity) to be included in the cohort from the beginning and which is iteratively updated. This is assessed with the RWD monitoring application by the patient in daily life.
Continuous end-user involvement and requirements
The successful design and implementation of the RE-SAMPLE platform and services necessitates a good understanding of the end-users, including their context of use, values and needs. Therefore, in RE-SAMPLE we followed an human-centred approach where stakeholders are engaged throughout the project. Together with these end users, we identified parameters that patients and professionals find important for monitoring. Also we elicited user needs in terms of privacy and data collection, self-management, decision making, data visualisation, communication, and coaching. Low-fidelity prototypes, scenarios and data-flow diagrams were developed, and prototypes were evaluated. This work led to the specification of user requirements for the virtual companionship programme. Next, the first prototypes and data visualization were further specified, developed and tested in end-user studies. In addition first iterations of the RE-SAMPLE service model were developed together with the stakeholders and consortium members.
Collective knowledge base
Knowledge and parameters that could be important for predicting exacerbations of COPD or comorbid CCCs are currently unknown or distributed among different (heterogenous and multimodal) data sources and representations. Therefore we have established a collective knowledge base, by identifying and bringing together retrospective datasets, as well as clinical-scientific knowledge and end-user preferences and expertise (see ‘continuous end-user involvement’ above). We performed state-of-the art literature (review) studies about clinical predictors, as well as diagnostic tools to differentiate acute heart failure from acute COPD. Furthermore, we identified and gathered retrospective data from the pilot sites including information from existing clinical studies and Electronic Health Record data, and from environmental databases for e.g. weather. Also a common methodology for data homogenisation is established along with standardization requirements.
Platform architecture and first functional prototype
Taking into account the sensitive type of (health) data to be handled, the RE-SAMPLE platform needs to be have a secure and privacy-securing data management in all its components and communications.
Therefore the RE-SAMPLE platform architecture - i.e. the name of the set of components and subsystems that make up the technological solution - has been defined from an early phase of the project. The development followed privacy- and security-by-design principles to ensure all functional and non-functional (security and privacy) related requirements imposed by the General Data Protection Regulation, ethical rules and the users. Also, data processing processes (pseudonymization, integration of heterogeneous multimodal datasets) were specified and ecosystems, roles, and key components identified. The first predictive models trained on the retrospective datasets collected by the three pilot sites are available to predict exacerbations of COPD. Furthermore the RWD monitoring application has been developed to use in the observational cohort study of RE-SAMPLE, which will later evolve to the virtual companion application.
Observational Cohort study
Current CCC disease management relies heavily on information acquired during scheduled visits when patients are usually stable, whereas actual symptoms and changes during daily life are not quantified. Therefore, in the early phase of the project the research protocol for our observational cohort study was designed. It has received ethical approvals in the three pilot sites (Netherlands, Italy and Estonia), and patient inclusion has started. We identified a set of parameters (e.g. daily symptom diary, physical activity) to be included in the cohort from the beginning and which is iteratively updated. This is assessed with the RWD monitoring application by the patient in daily life.
With RE-SAMPLE we aim to progress beyond the state of the art by delivering 1) new, unique insights in the course and interrelationships between (bio)markers, lifestyle characteristics, and clinical and well-being parameters creating holistic models for the onset and progression of CCC exacerbations; 2) a breakthrough in knowledge in timely detecting exacerbations to start preventive treatment effectively; 3) definition of clinical endpoints to identify and monitor individual patients with CCCs that can be used in a personalised eHealth application in daily care; 4) highly-innovative secure and trustworthy handing of data and machine learning methodology using multiparty computation.
RE-SAMPLE will disrupt guidelines on how to tailor and implement adaptive predictive CCC care in which individuals benefit optimally by tailored CCC disease management. RE-SAMPLE significantly impacts towards this direction not only through the development of the platform and companionship programme, but also through its uptake in the privacy-sensitive domain of healthcare, by healthcare organizations throughout Europe, to alleviate the overall societal and economic burden of CCCs.
RE-SAMPLE will disrupt guidelines on how to tailor and implement adaptive predictive CCC care in which individuals benefit optimally by tailored CCC disease management. RE-SAMPLE significantly impacts towards this direction not only through the development of the platform and companionship programme, but also through its uptake in the privacy-sensitive domain of healthcare, by healthcare organizations throughout Europe, to alleviate the overall societal and economic burden of CCCs.