Patients with underlying chronic conditions are among the most vulnerable in the COVID-19 pandemic and in new potential pandemics. They have a higher risk of developing life-threatening pneumonia when infected with COVID-19 and acute complications of their chronic condition due to other viral and bacterial infections.
Background and context
Protecting the most vulnerable in a pandemic and beyond
Evidence has shown that patients who have reported no pre-existing conditions had a case fatality rate of 0.9% while those with a pre-existing illness have had a 5-10 times higher chance of a negative outcome. Respiratory and cardiovascular diseases are amongst the chronic conditions that represent higher vulnerability to negative outcomes.
This is particularly relevant for nursing homes and home care (in some cases with large travel distances), responsible for a large share of such risk-group patients.
-
We believe that PyXy is capable to substantially reduce the health-economic impact on healthcare and elderly care services by preventing hospitalization in emergent-care situations.
-
In addition, as PyXy can be used by lesser skilled individuals, it will potentially bridge the gap in care settings that face extreme shortage of skilled workers, and can even be used by informal caregivers at home reducing the overall hospitalization rate even further.
Need and opportunity
An improved screening and monitoring service
Being able to differentiate between the patient’s base line and onset of early changes requires continuous tracking of disease-relevant parameters in the risk-cohort who are based at home or nursing homes. Such health surveillance setup will generate enormous amounts of constantly inflowing data, hence it would be physically implausible to closely monitor all those parameters using only health staff’s observations and doctors are hardly available. With use of machine learning methodology, all the data could be analyzed and monitored in real-time, so that non-urgent cases could be under consistent medical technical surveillance while health professionals’ attention is guided to the more urgent cases.
"We aim to create a benchmark in health monitoring using machine learning for personalized clinical decision support.“
Objectives and challenges
Using H2020 funds, the consortium plans to introduce the PyXy to the market within 22 months: a Point-of-Care, nursing-homes, and home use device allowing anyone to perform diagnostic (auscultation) at an accuracy not normally accessible to the medical staff.
In this project the consortium will collect COVID-19 patient data as well as reference data (healthy and other chronic patients). Following the build-up of this unique database, we will adopt the existing AI algorithm to deal with COVID-19 and take the PyXy to production readiness.
The primary objective of the project is to mature, optimise and demonstrate an AI-powered cloud-based system for patient’s lung health assessment and monitoring, which integrates an intelligent sensor device – PyXy – and machine learning to increase the efficiency and the quality of the care delivery to patients from risk groups, especially under a pandemic outbreak.
The methods should have clinically valuable sensitivity and specificity for early detection of deterioration in individual health status by monitoring cardio-respiratory parameters using combined clinical indicators. The system should fit with routine workflows at the health institution as well as fit with established infrastructure of e-health / IT systems.
We have identified several main challenges that shape the project and set the objectives for the project partners to achieve through the proposed activities, leading to the innovation and outputs.
Workpackages and objectives of the project
WP1
Autumn/winter 2021 data collection phase
The objective is to build a unique clinical database, recorded from risk group subjects over time and conclude with a clinical validation of the results. This will help to build up unique knowledge on the development of health parameters of risk-group patients under COVID-19 pandemic outbreak
WP2
Clinical Validation and finalizing the hardware product design
Take the PyXy into high volume production readiness including meeting regulatory requirements, reliability, as well as maturation and design optimization at pre-defined cost.
WP3
Training AI models
Optimize and adapt current AI algorithms to detect Covid-19 relevant parameters based on the unique collected database. The models will be evaluated against clinical outcomes: e.g. SpO2 change, hospitalization, X-ray or CT scan, need for ventilator, medications etc. The aim is detecting abnormalities early and reducing missed cases with onset of alarming symptoms.
WP4
Implementing integration of the innovation into the existing care delivery infrastructure
The related objective is to integrate the system, comprising PyXy, a phone/ tablet app and web interface, and the platform itself, with the healthcare service supplier’s IT infrastructure. The second objective is to demonstrate value proposition and validate usability with customers and end users representing different care settings/ systems (including private and public service providers.
WP5
Exploitation and communication
Generating scientific and clinical data to support the innovation as well as building channels for efficient communication of the new knowledge and dissemination of the project results. Also, raising awareness to prepare market rollout and commercialization. Establishment of business and exploitation plans using verified data from the demonstration activities
WP6
Overall project management, IPR management and preparation for CE mark and FDA
Coordinate development and validation activities to meet regulatory requirements. Efficient project coordination and management to meet the tight project timeline, within the budget, and the risks.
Concept and methodology
The innovative interdisciplinary concept of this project is to supply patients from risk groups with a health surveillance “kit” and use machine learning / AI to continuously analyze their COVID-19 relevant cardio-respiratory parameters obtained from an intelligent sensor device and spot potential deterioration in their lung health during the pandemic situation.
The proposed solution will comprise the following elements:
- PyXy device
- AI / machine learning models
- An app for Android or iOS device
- Web interface
- Cloud platform: Both the app and web clients are connected to the cloud platform. This platform enables analysis using AI / machine learning models and provides connection and synchronization between the mentioned elements. It includes an open API for further integration with external e-health/IT systems and patient’s record.
The solutions' architecture: Paired PxYy device (© Sanolla) with Smartphone or Tablet App presenting the AI analysis. The desktop computer shows the web-interface and the integration with e-Health providers. (App and web-interface © Medsensio)
