Periodic Reporting for period 1 - LETHE (LETHE (λήθη) – A personalized prediction and intervention model for early detection and reduction of risk factors causing dementia, based on AI and distributed Machine Learning)
Reporting period: 2021-01-01 to 2021-12-31
Dementia has long been considered to be neither preventable nor treatable, but while the underlying illnesses may not be curable, today we know that the disease course might be modifiable with adequate preventive interventions at an early timepoint. The Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) showed a positive effect after a 2-year intervention period targeting several lifestyle and vascular risk factors simultaneously. FINGER was the first large randomised controlled trial of its kind demonstrating the possibility of preventing cognitive decline using a multi-domain intervention among older at-risk individuals.
Performing on such a clinical-scientific evidence, new information and communication technology (ICT) and machine learning algorithms can help leveraging daily personalized interventions to prevent dementia and monitoring relevant risk factors. This is the vision of LETHE, an EU project funded by the H2020 programme.
In phase one, the project provides a data-driven risk factor prediction model for older individuals at risk of cognitive decline. This model is based on big data analyses of observational and longitudinal intervention datasets from our four participating European clinical centres.
In the second phase, an ICT-based semi-automated intervention protocol named FINGER 2.0 is provided to people at risk, engaging participants in their individual lifestyle. A prediction model, based on the big data analyses conducted in phase one, provides a machine-learning based approach towards the risk of development of dementia.
Across an intervention trial lasting 18 months, LETHE will establish novel digital biomarkers for early detection of risk factors unobtrusively detected by an ICT-based passive and active monitoring of patient behaviour and lifestyle. The personalized ICT-based preventative lifestyle intervention will support individualised profiling, personalised recommendations, feedback, and support.
Successful conduction of this project will provide the opportunity to demonstrate the effectiveness and efficacy of an ICT based semi-automated and personalized dementia prevention setup in daily life settings of persons at risk of dementia. By including a large study population of at-risk persons, this will result in widely reducing the individual risk profile of participants.
LETHE project aims at contributing to the current knowledge of the disease and the possible influence of different lifestyle factors on dementia progression. The underlying model will help to project and forecast the development of related risk factors. It thereby aims to help address risk early and shift the onset of cognitive decline and / or lower the progression speed. Moreover, the LETHE project is expected to have a long-term effect on the healthcare sector and provide a foundation for health literacy strategies. The project will highlight the necessity for fully design-centred interventions and solutions for the target population at risk. LETHE aims to demonstrate that enhanced risk detection and management tools in the context of cognitive impairment can contribute to increased effectiveness and efficiency in long-term care management of chronic disease affecting 55+ population.
Performing on such a clinical-scientific evidence, new information and communication technology (ICT) and machine learning algorithms can help leveraging daily personalized interventions to prevent dementia and monitoring relevant risk factors. This is the vision of LETHE, an EU project funded by the H2020 programme.
In phase one, the project provides a data-driven risk factor prediction model for older individuals at risk of cognitive decline. This model is based on big data analyses of observational and longitudinal intervention datasets from our four participating European clinical centres.
In the second phase, an ICT-based semi-automated intervention protocol named FINGER 2.0 is provided to people at risk, engaging participants in their individual lifestyle. A prediction model, based on the big data analyses conducted in phase one, provides a machine-learning based approach towards the risk of development of dementia.
Across an intervention trial lasting 18 months, LETHE will establish novel digital biomarkers for early detection of risk factors unobtrusively detected by an ICT-based passive and active monitoring of patient behaviour and lifestyle. The personalized ICT-based preventative lifestyle intervention will support individualised profiling, personalised recommendations, feedback, and support.
Successful conduction of this project will provide the opportunity to demonstrate the effectiveness and efficacy of an ICT based semi-automated and personalized dementia prevention setup in daily life settings of persons at risk of dementia. By including a large study population of at-risk persons, this will result in widely reducing the individual risk profile of participants.
LETHE project aims at contributing to the current knowledge of the disease and the possible influence of different lifestyle factors on dementia progression. The underlying model will help to project and forecast the development of related risk factors. It thereby aims to help address risk early and shift the onset of cognitive decline and / or lower the progression speed. Moreover, the LETHE project is expected to have a long-term effect on the healthcare sector and provide a foundation for health literacy strategies. The project will highlight the necessity for fully design-centred interventions and solutions for the target population at risk. LETHE aims to demonstrate that enhanced risk detection and management tools in the context of cognitive impairment can contribute to increased effectiveness and efficiency in long-term care management of chronic disease affecting 55+ population.
In the first year of the project the following main results could be achieved:
The project successfully started on 01.01.21 and proactive cooperation between all technical and clinical partners launched the first phase of the project by laying the foundation for the development of the initial prediction model. During the first year of the project, clinical partners worked closely to construct a harmonized data set of over 100.000 data points which can be used to conduct a common analysis. This work included harmonization of the data on a feature level, as well as setting up a common data backend with respective data processing as well as data controller agreements to ensure privacy and security of the data. All involved partners agreed on these legal documents. The data set is unique in its size and qualities and carries a lot of potential for the further progressing project.
Furthermore, the project partners did essential work in anticipation of the short integration period at the beginning of 2022. The project published a whole integration architecture including use cases. The technical design is partially based on user requirements tests which have been performed prior. The project has begun the integration of the sensor framework as well as the development of the app, which will be the main touchpoint for LETHE users. Through timely preparation and the joint work on the technical design, all beneficiaries achieved a good understanding of the LETHE setup and the technical implementation.
Within the work package (WP) 7 the project delivered the study protocol in time. This can be considered a major achievement as the study protocol is the basis for the process of obtaining the Ethical Committee approval, which begins in 2022.
With the start of the project, LETHE achieved a comprehensive cooperate design and media appearance addressing several stakeholders. The LETHE consortium has already been represented in several press releases and radio appearances, scientific events as well as workshops with other European projects like RADAR-AD, BRAINTEASER, and FEMaLe, all in the first year alone. Furthermore, the project has already produced several videos, leaflets in several languages, as well as an information brochure.
The project successfully started on 01.01.21 and proactive cooperation between all technical and clinical partners launched the first phase of the project by laying the foundation for the development of the initial prediction model. During the first year of the project, clinical partners worked closely to construct a harmonized data set of over 100.000 data points which can be used to conduct a common analysis. This work included harmonization of the data on a feature level, as well as setting up a common data backend with respective data processing as well as data controller agreements to ensure privacy and security of the data. All involved partners agreed on these legal documents. The data set is unique in its size and qualities and carries a lot of potential for the further progressing project.
Furthermore, the project partners did essential work in anticipation of the short integration period at the beginning of 2022. The project published a whole integration architecture including use cases. The technical design is partially based on user requirements tests which have been performed prior. The project has begun the integration of the sensor framework as well as the development of the app, which will be the main touchpoint for LETHE users. Through timely preparation and the joint work on the technical design, all beneficiaries achieved a good understanding of the LETHE setup and the technical implementation.
Within the work package (WP) 7 the project delivered the study protocol in time. This can be considered a major achievement as the study protocol is the basis for the process of obtaining the Ethical Committee approval, which begins in 2022.
With the start of the project, LETHE achieved a comprehensive cooperate design and media appearance addressing several stakeholders. The LETHE consortium has already been represented in several press releases and radio appearances, scientific events as well as workshops with other European projects like RADAR-AD, BRAINTEASER, and FEMaLe, all in the first year alone. Furthermore, the project has already produced several videos, leaflets in several languages, as well as an information brochure.
The main progress beyond the state of the art is expected in the following areas: Personalised risk prediction of cognitive decline based on artificial intelligence (AI) and biomarkers (digital and conventional); an ICT setup which allows a large-scale personalized interventions leading to prevention of cognitive decline; an implemented automated health behaviour change setup leading to personal empowerment and general disease knowledge regarding influential factors regarding disease and symptom progression.
Although parts of the mentioned areas have already been addressed in RP1, the full potential will be reported in the last project report. In RP1 the project has already achieved the development of an initial prediction model of Mini Mental State Examination (MMSE) parameters based on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, a detailed planning outline for the personalized ICT intervention framework as well as a theoretical model of health behaviour change mechanisms.
As the project concludes its first year at the time of reporting period (RP) 1, only minor impact achievements can be reported at this point. In general, the joint data harmonization of the four participating clinical centers has yielded promising results regarding a joint research effort as well in the context of the development of the joint protocol. The consortium considers this model as a potential blueprint for wider use for further examinations in clinical practice. Additionally, the implemented model on health behaviour change and the adaptation of the intervention model could have a large impact on persons at risk world-wide and could widen the access to dementia prevention tools, which of course would have an extensive impact on general efforts and costs of care.
Although parts of the mentioned areas have already been addressed in RP1, the full potential will be reported in the last project report. In RP1 the project has already achieved the development of an initial prediction model of Mini Mental State Examination (MMSE) parameters based on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, a detailed planning outline for the personalized ICT intervention framework as well as a theoretical model of health behaviour change mechanisms.
As the project concludes its first year at the time of reporting period (RP) 1, only minor impact achievements can be reported at this point. In general, the joint data harmonization of the four participating clinical centers has yielded promising results regarding a joint research effort as well in the context of the development of the joint protocol. The consortium considers this model as a potential blueprint for wider use for further examinations in clinical practice. Additionally, the implemented model on health behaviour change and the adaptation of the intervention model could have a large impact on persons at risk world-wide and could widen the access to dementia prevention tools, which of course would have an extensive impact on general efforts and costs of care.