Digital health decade in action: case studies

Oncompass – Hungary (IVSZ)

Oncompass Medicine is a start-up that developed an artificial intelligence-based medical software for targeted cancer therapy. Thanks to artificial intelligence, the software can compute more than 20,000 potential associations between cancer genes and targeted therapies in 20 milliseconds to predict each patient’s response to targeted therapies. It increases cancer patients’ chance to receive the right effective treatment, lowers unnecessary costs, and helps pharmaceutical companies develop therapies faster and safer.

Corti AI – Denmark (IT-Branchen)

Corti is a scaling-up company that uses voice recognition and artificial intelligence and to help medical professionals make life-saving decisions in split seconds. By analysing millions of patient interviews they identify patterns in conversations and use them to predict potentially fatal incidents like cardiac arrest.

Population Health Analytics and predictive modelling

Impact Pro is a browser-based analytics system used for identifying and stratifying population to provide a holistic view of risk drivers and actionable insights.

Improving healthcare:

Impact Pro is an advanced data analytics suite used by health systems to identify, profile and stratify their populations. It helps determine individuals who would benefit from tailored intervention programmes and which intervention programmes are likely to have an impact on the quality of individuals’ health based on over 1800 clinical markers. It supports clinicians and analysts across a health system to gain a better understanding of the health status of their populations and plan care programmes appropriately.

Having been used in the US health and care market for around 30 years, it has now been calibrated for the UK.

How does it work?

Impact Pro processes data from primary and secondary health and care services to create episode treatment groups which provide a structured timeline of care for individual patients e.g. hospitalisations, outpatient visits, prescriptions. It uses over 1800 risk markers which indicate the presence and/or severity of diseases and treatments for each patient, and applies statistical machine learning to predict the probability of future hospital admissions over a given timeframe. Based on episode treatment groups, risk markers and probability models, Impact Pro allows end users to segment patients into cohorts (registries) with similar conditions, levels of risk, and potential actionability. The first UK release includes 21 Population Health Registries grouped under seven headings:  Healthy, Acute Episodic, Chronic – Stable, Chronic – At Risk, Complex, Terminal, Not Yet Defined.

To adapt the solution for the UK, approximately 600,000 clinical and drug codes were mapped from the US to the UK clinical terminology and clinically validated.

Partners: Optum UK (part of UnitedHealth Group) partnered with an Electronic Health Record provider to include Impact Pro into its wider platform. The customers and end users are Integrated Care Systems (partnerships between the organisations that meet health and care needs across an area, including family doctors, hospitals, pharmacies, local councils and others.). The first live implementation is due for release in June 2022.

Main technologies and data: Primary and secondary care data; statistical machine learning regression models; predictive analytics; clinical and drug code mapping.

Member: UnitedHealth Group

Smarter insights to treat complex arrythmias

Improving healthcare: Helping to improve procedure efficiency, lab efficiency and patient outcomes via collaboration between physicians, electrophysiologists and electrophysiology lab administrators, using multicentre studies and AI-based data analysis for analysing treatment approach.

How does it work?

The CARTONET™ digital platform stores all CARTO^® 3 System cases, analyses them using AI-based algorithms and stores the analysis on the cloud. This enables the physician to review the analysis of each case and to perform modifications of the algorithm analysis in a simple way. The system can then generate statistics based on hundreds of cases’ analysis data, and aggregate this information on cases that were identified to be in multicentre studies to present the global statistics which can lead to research in analysing the treatment approach.

Partners: This solution was developed by Biosense Webster, a Johnson & Johnson medical device company, with collaboration Siemens Healthineers’ teamplay platform on top of Microsoft Azure cloud.

Main technologies: “WEB PACS”, Database, AI – deep learning.

Member: Johnson & Johnson

Enabling personalised care along cancer care pathways

Improving healthcare:Improved patient outcome, for instance for lung and prostate cancer patients, with an AI-Pathway Companion that supports physicians in providing better care with complex diagnosis and therapeutic decisions along care continuum.

How does it work?

The AI-Pathway Companion focuses on domain expertise in the field of oncology. The infrastructure layer shows the patient’s relevant clinical history, current patient status, diagnostic images and results from the pathological reports, eliminating the need to access information from multiple databases or files.

Leveraging AI technologies to provide further insights, the AI-Pathway Companion compares the patient’s clinical status against the current guidelines and through a decision diagram supports physicians in their daily decision-making for next steps in terms of diagnosis and therapy.

Through ongoing development and refinement of data interfaces, this Companion aims to increase the number of disease parameters captured throughout the clinical pathway, thereby expanding the relevant data points with the goal of advancing personalised cancer care.

Member: Siemens Healthineers

Novel AI-Powered Solution for Cardiovascular Disease Prevention

Improving healthcare:

The digital-first approach to Cardio Vascular Disease (CVD) prevention provides a holistic set of digital tools and coaching placing participants on a healthier path by personalizing care, gathering data, and providing actionable insights, all in a user-friendly mobile app.

The program seeks to reduce the broader burden of disease for those at risk while lowering costs for employers and health plans. According to the World Health Organization, at least 80% of heart disease, stroke, and type 2 diabetes are preventable.

How does it work?

Medical devices capture blood pressure and weight automatically tracked in the app alongside food, medications, physical activity, glucose, and A1C from thousands of other integrations. The program combines behavioral science principles, data science, and customized educational content to guide members towards positive habits in weight management, physical activity, and diet for a heart-healthy life.

It uses machine learning algorithms and more than 30 billion health data points from millions of members worldwide. Blood pressure insights help members at risk for CVD understand how their blood pressure is trending and offer tactical recommendations and support in real-time.

Partners: Bayer worked together with One Drop to develop this solution.

Main technologies and data: Data Science (including Machine Learning, Artificial Intelligence) across clinical, behavioural, and self-reported data sets, integration of diagnostics data (blood pressure, weight, glucose), self-reported data (food intake, manually and via barcode scanning, physical activity, medication intake, etc.)

Member: Bayer

Remote patient monitoring for respiratory conditions

Improving healthcare:

Connected positive airway pressure (PAP) and ventilation devices can support treatment delivery to patients by monitoring the patient’s treatment remotely and transmitting the data in a secure manner to the HCPs.

Remote monitoring enables HCPs to manage and coordinate care for their patients with chronic diseases, understand how a treatment protocol works, and decide on customised interventions if needed, even possibly remotely, aiming at limiting the burden of transportation to the hospital and risks associated with hospital stays.

Such tools can help an estimated 936 million individuals aged 30–69 years worldwide with their sleep apnoea and an estimated 380 million people worldwide with COPD.[1]

Remote monitoring services drive patient engagement, adherence, and better outcomes in treating sleep apnoea and other chronic conditions.[2]

How does it work?

Connected devices – for chronic respiratory disorders and diseases such as sleep apnoea and chronic obstructive pulmonary disease (COPD) – collect data on how the treatment is being delivered and how the device is being used by the patient. Treatment device data can be transferred remotely and analysed to provide valuable insights to healthcare providers on treatment protocol.

For example, analysis can identify operational issues such as air leakage resulting in disturbance for the patient’s partner or non-effective therapy. Data can also be used to identify patient compliance problems and enables HCPs to make better decisions in order to balance treatment intensity with outcomes.

Partners: This solution enables data to flow between the device, the platform, third-party cloud providers, and care teams.

Main technologies and data: Machine therapy data is recorded on the device; the data is encrypted and transmitted over the internet to a platform where it is dynamically processed often with advanced algorithms. The data is accessed by the healthcare provider to monitor the patient’s use and treatment.
Member: ResMed

Smart watches in the early identification of cardiac arrythmia

Improving healthcare:

It contributes to reducing stroke and other atrial fibrillation (AF)-related complications though a smart device (wristband and wristwatch) running an AF screening app to collect photoplethysmography (PPG) signals.

Technology incorporated in smart watch and associated App run on any smart phone provides early warning of potential cardiac irregularities that warrant further investigation, potentially enabling the early identification of cardiac problems.

How does it work?

Monitoring is via a wearable device that when worn provides continuous home monitoring with smart device–based PPG technology for AF screening. This enables AF detection leading to early interventions that may reduce the likelihood of stroke and other AF-related complications. The use of a wearable Smart Watch addresses the problem of low detection and adherence rates that characterise current management approaches for patients with suspected AF.

Patients with “possible AF” episodes using the PPG algorithm were further confirmed by health providers among the MAFA (mobile AF app) Telecare center and network hospitals, with clinical evaluation, ECG, or 24-h Holter monitoring.

246,541 individuals downloaded the PPG screening app, with 187,912 individuals using the smart watches to monitor their pulse rhythm. Among those with PPG monitoring, 424 (0.23%) received a “suspected AF” notification. Of those, 227 individuals (87%) were confirmed as having AF.

During algorithm development phase and beta phase, data was collected from both people with and without AF problem, with even distribution on gender, and wide coverage of different aging groups (18 to 80).

Partners: Hospitals, universities

Main technologies and data: An app using an algorithm to screen for photoplethysmography signals from smart watch and wrist band

Member: Huawei

Virtual care management for patients living with chronic conditions

Health insurers can provide eligible beneficiaries with remote patient monitoring through Vivify Health platform which pairs them with nurse-led care teams and connected devices, such as a weight scale, blood pressure cuff, glucose monitor, or pulse oximeter, to monitor their daily vitals and inform interventions. This offers a continuous health care experience, keeping individuals connected to their care team and empowering them to manage their health from the convenience of their homes.

Improving healthcare:

The program serves more than 30,000 members and has reduced in-patient admissions by 5 percent and readmissions by 65 per cent, reduced mortality by 3.4 percent this year and resulted in 97 per cent patient satisfaction.

How does it work?

Eligible members can be connected with a Registered Nurse for care management and offered devices that sync to a Vivify app which helps them track their clinical and biometric information, such as glucose levels, blood pressure pulse and weight. This info is reviewed by the nurse on a daily basis and if changes occur the care provider may engage. The set-up enables monitoring at home or on the go.

The solution increases digital engagement and supports more frequent interaction for patients with chronic conditions to help drive positive behaviour and help enable access to timely clinical intervention. It allows users to recognise and better manage their symptoms and comorbidities, transforming episodic care into proactive, ongoing care management.

Partners: health insurance company with its own or outsourced remote care team; digital care management capability provider that enables individuals to communicate and share information with their care management staff.

Main technologies and data: platform of connected devices with smart paths to enable patient management, including in-app video chat, messaging and telephonic capabilities. Patients can either use their own device, or a tablet is provided by the insurer, together with other peripheral devices specific to their chronic condition(s) to help track their biometric data and symptoms. The tablets provide a simplified interface, tailored to the physiology of older users – screen brightness, larger icons, and electro-sensitivity optimized for thinner skin.

Member: UnitedHealth Group

Improving outcomes with a haematology network

Improving healthcare:

Improving outcomes for haematological cancers patients by increasing the knowledge and understanding of this disease through a collaborative platform that allows for the analysis of data on multiple data sets with methodological and statistical possibilities.

It increases the value of the data by enabling their re-use across a wide range of research studies and encourages publishing results so that insights can contribute towards improving patient outcomes. Datasets on haematological malignancies from 23,000 patients have already been transformed so they can be used for analysis.

How does it work?

The Haematology Outcomes Network in Europe (HONEUR) is a secure, collaborative platform run as a federated model, where the data stay at the respective sites and the analysis is executed at the local data sources. It uses the common data model called OMOP (Observational Medical Outcomes Partnership), ensuring participating sites maintain local governance and can initiate their own research questions. No patient-level data are stored on the HONEUR portal – only aggregated results of a research question can be shared securely through firewalls.

Data partners not only benefit from participating in a network that is pioneering data management and analysis, but their efforts are also compensated at Fair Market Value, and contributors can even increase the value of their data via expanded authorship in publications and potential sponsored studies.

Partners: Janssen collaborated with a consortium with partners from 9 countries, including cancer centres, university institutes and hospitals and other local data projects.

Main technologies and data: Federated data model, patient-level data, disease characteristics, patient baseline characteristics, medications, outcomes.

Consortium: HONEUR

Identifying and studying new COVID-19 variants using cloud technology

Improving healthcare:

The Global Pathogen Analysis System (GPAS) facilitates better tracking and management of diseases and pathogens worldwide to ensure a fast and effective response. This includes an early warning system to track the spread of variants in real time.

The GPAS is run by a not-for-profit entity, offering a global turnkey solution that standardises how genomic sequence data is processed. This is a new type of partnership with academia, non-profit, and industry working together. GPAS is part of the Public Health England New Variant Assessment Platform.

How does it work?

GPAS’ services are cloud-based and secure which alleviates any current issues with worldwide genomic sequencing improving equal distribution of infrastructure and operational support. Unlocking genomic insight requires expertise and access to high-performance computing and sharing data across borders needs a highly secure digital environment (and patience, and trust).

The secure, cloud-based system can analyse and translate data to give users an instant global perspective on their findings whilst maintaining full ownership and control by protecting personal identifiable information (PII). It can also integrate with sequencing infrastructures and data repositories that already exist for seamless tracking.

Partners: The Global Health Security Consortium, the Lawrence J. Ellison Institute for Transformative Medicine and the Tony Blair Institute for Global Change have worked in coordination with Oxford University and Oracle to support the development of the platform and to get it in the hands of global researchers.

Main technologies and data: In-built computational infrastructure for rapid – same-day – data analysis, supporting epidemiological research, therapeutics, and diagnostics.

Member: Oracle

Breakthroughs in pancreatic cancer

Improving healthcare:

Developing better treatment for pancreatic cancer patients by increasing understanding of the pancreatic ductal adenocarcinoma (PDAC) biology and patient stratification through combining genomics and imaging phenomics.

How does it work?

The PancAIM platform integrates the entire range of genomics with radiomics and pathomics and applies a data-efficient two-staged AI method relying on four central concepts of AI in healthcare: data providers, clinical expertise, AI developers, and medtech companies to connect to data and introduce AI into healthcare.

Among the main steps of the project, are:

• Develop digital platform integrating genomics and clinical PDAC data.
• Develop and use unimodal AI biomarkers for integrative research.
• Develop and select most promising AI-assisted clinical products integrating omics and medical imaging data.
• Implement and validate clinical products.
• Sustain the platform for further research and clinical applications.

Partners: A consortium of industry and research universities from 9 countries.

Main technologies and data: combines genomics and imaging phenomics, enable development of clinical use AI applications.

Consortium: PancAIM

Boosting trust and uptake of AI in precision oncology

Improving healthcare:

Increased clinical trust and adoption of enhanced AI in oncology. The research findings are available to the public free of charge which supports advancements in care and treatment development.

How does it work?

EuCanImage is building a highly secure, federated and large-scale cancer imaging platform integrating advanced capabilities and new standards to develop and validate integrative decision support systems for precision oncology.

It will determine the optimal facets of ethical data collection, including legal and ethical framework for oncological imaging to ensure a privacy-by-design approach to the platform’s development.

It will offer a user friendly catalogue for cancer imaging and non-imaging data, for which a comprehensive suite of opensource tools and procedures for data anonymisation.

The goal is to build and demonstrate a FAIR and GDPR-compliant and scalable platform for leveraging large-scale, high-quality and interoperable cancer imaging datasets adequately linked to biological and health cancer data.

Partners: It is a multidisciplinary consortium combines the expertise of 20 partners from 11 countries. It includes major universities, research institutes, and industry partners.

Main technologies and data: Tools will be designed for standardising image data cross sites and scanners enabling the large-scale data and machine learning to provide semi-automated capabilities for data curation and annotation, which can be executed in a distributed privacy-preserving manner, utilising also synthetic image generation.

Consortium: EuCanImage

Using big data to for safer pregnancies with diabetes patients

Improving healthcare:

A state-of-the-art analysis provided relevant insights about the use of Remote Patient Monitoring systems for Gestational diabetes Mellitus (GDM) care provision. Gestational diabetes (GDM) develops in women during pregnancy because the mother is not being able to produce enough insulin.

How does it work?

The BigMedilytics pilot group has created an AI solution for GDM care, including an app, that facilitates self-management and remote-monitoring of GDM through the use of a mobile application linked to a secure medical portal.

The primary data source is the collection of daily blood sugar levels (BSLs), at least four times a day (‘Fasting’ and ‘Post meal’) using glucometers equipped with Bluetooth, thereby obtaining a lengthy review of glycaemic control, which ensures a prompt recognition of high sugar levels. Other relevant information and clinical data were collected for the development of a prognostic model.

Real- time data and historical data were collected to gain improved understanding of patterns of BSL in the population of patients with GDM.

All necessary documentation has been produced to ensure full compliance with both EU and National Irish regulation in terms of data privacy and ethical aspects, with a particular emphasis on the GDPR requirements and security measures to avoid any ethical issue or data breach.

Partners: The study performed with industry partners was initiated by collecting and analysing data records from two different hospitals based within the RCSI Hospital group in Ireland – the Rotunda Hospital in Dublin and Our Lady of Lourdes Hospital in Drogheda. In total 131 records were analysed.

Main technologies and data: Use of RPM solutions in combination with AI models for GDM care

Consortium: BigMedilytics

Better therapeutic effectiveness for cancers

Improving healthcare:

Improving cancer care insights for bladder cancer and multiple myeloma by using privacy-preserving machine learning to analyse patient-level data.

How does it work?

The Augmenting Therapeutic Effectiveness through Novel Analytics (ATHENA) data science innovation project, launched in 2020, facilitates the re-use of clinical data for secondary research by using a federated data network model for data analytics. Data remain local, under governance of the data custodian (in this case, the hospital), and the analysis is brought to the data. Only query results will go back to a central location and no patient-level data leave the hospital.

In partnership with the Universities of Leuven and Ghent, IMEC, Robovision, Inovigate and several other hospital partners and small & medium sized companies, ATHENA uses an AI system and supports research organisations and biomedical companies to analyse the data. It was conceived with the aim to advance medical science, for the development of new treatments and to accelerate clinical research. Hospitals around Europe can join the initiative, potentially increasing the volume of data as well as the robustness of the insights.

Partners: Participating hospitals (UZ Leuven, UZ Gent, AZ Groeninge, CHU Liège, OLV Aalst), KU Leuven, Imec, Illumina, Robovision, Inovigate.

Main technologies and data: Advanced machine learning methods, federated networks.

Consortium: ATHENA

Healthcare and technology companies are providing electronic health records solutions

Austria: Electronic health records for faster and more precise physicians

In Austria, the roll out for the nationwide Electronic Health Record (ELGA) started five years ago, in 2015. Today, work is already underway on the implementation of valuable extensions, such as the eVaccination pass or care networks as part of ELGA. As a modern and secure infrastructure, ELGA is available to all citizens and all those who receive care in the Austrian health care system. It facilitates access to health data for patients and authorized ELGA health service providers – attending physicians, hospitals, nursing homes or pharmacies. ELGA is working with different models of opt-out, leading to a general coverage of 97 per cent of the insured population as enrolled ELGA users.

The Netherlands: A secure portal for exchanging COVID-19 patient data

Since its launch in 2020, 95% of Dutch hospitals have already been connected to the portal. It optimises the use of healthcare resources by allowing hospitals to seamlessly share COVID-19 patient information with one another. The COVID-19 portal running on XDS cloud services, available to all Dutch hospitals, is not linked directly to an individual hospital’s EPD (Electronic Patient Dossier), PACS (Picture Archiving and Communication System) or pathology department systems. Instead, specific information, such as a patient’s radiology images, reports and patient summary is shared via the portal.

Australia: A federal level opt-out approach to Electronic Health Records

By the end of 2018, a My Health Record was created for every Australian, unless an individual chosen not to have one. It is now a nation-wide EHR platform, overcoming challenges of multilevel government structures. This is one of the largest National Health Record Platforms worldwide, allowing authorized health professionals access to some 3 billion clinical documents belonging to 23 million citizens. Australia also is an active contributor to the Global Digital Health Partnership (GDHP), a platform for governments to exchange experiences in digitisation of their own health system.