Session: Applications for Health Data Management with Solid
Friday, March 31, 2023 (9:00 - 12:30)
In this session we will focus on the technical requirements for implementing Solid-based health data stores. In addition to Solid's native tools and capabilities, there are additional solutions, such as digital identities, verifiable credentials and smart device interlinking, that should be considered for a successful implementation of Solid in healthcare. We will provide some technical background as well as examples of successful implementations of solid-based solutions; we will also discuss the advantage of solid-based data spaces for the formulation of explanatory models with the aim of discovering relevant interrelationships between entities.
Our citizens highly value their data rights granted by the legislation such as GDPR, but they experience difficulties to exercise their rights due to a lack of technology and standards. Ensuring that citizens can make decisions about how their health data can be used and shared will significantly encourage citizens to participate in health-related research. In this talk, Chang will present a Solid-based Citizen-centric data Platform (called TIDAL) that facilitates interactions by citizens and researchers for health research. TIDAL stores personal data in Solid pods as RDF with well-known health-related vocabularies (e.g., SNOMED CT) and controls access to query fine-grained subsets of personal data. Researchers can post a human- and machine-readable digital consent structured by the Data Privacy Vocabulary. Then, TIDAL combines with a federated learning method to analyze personal health data from individuals’ Solid pods. Finally, Chang will discuss how TIDAL and other Solid-based tools could offer a new data paradigm of sharing and using personal data and ultimately increase the availability of personal data for societal relevant uses.
"Our ambition is to help solve societal challenges. In tomorrow's society, half of the solution is about bringing together the right data." Leading digital platforms collect and store much data, but this data is often not accessible to other companies who could use it to create innovative solutions. These solutions could then address current societal challenges or improve the livelihoods of citizens. Citizens expect more transparency and choose to be more in control of their personal data. Addressing this challenge requires the overhaul of legal barriers and the involvement of a trusted third party. The Flemish Data Utility Company is a trusted third party that helps companies to shine in the Data Economy, respecting the privacy of citizens and consumers in terms of personal data.
Use.id can be used to build a consent-driven ecosystem by leveraging W3C Solid & WebID infrastructure to easily build your ecosystem, onboard participants and let them exchange data. Use.id allows your business to get a WebID so you can receive data, share data, or build a customer-centric digital ecosystem. Use.id is currently being implemented in various industries among which more and more in the health ecosystem. An interesting use-case being developed is in the context of the long-term follow-up of patients in the earliest stages of Alzheimer disease. This is achieved by storing brainscan data and novel biomarkers in Solid Pods. The patient can then grant access to medical actors, researchers and members of ecosystems such as WeAre. Topics that will be discussed are a.o.; user experience, interoperability & access control.
Due to the integration of genomics, novel invitro methods, and in-silico methods in the last decades, personalized medicine has seen a rapid development and wide application in several clinical fields. The concept of personalization refers to the adaptation of therapies in accordance with a patient's bio-medical fingerprint, such as his or her genotype, his or her phenotype, as well as specific biomarkers. Using such characteristics, a model can be defined to make an accurate assessment of response, risk, and planning of a therapy. However, such therapy adaption takes place as a relatively slow process. This is because there is both no control in real time of the patient’s condition as well as more accurate accounting of patient’s environment. Furthermore, therapy planning as well as the mode of action (MOA) of pharmacological compounds is performed based on clinical tests obtained from relatively small patient cohorts that do not reflect the complexity of real patient populations. The main problem is that it is detrimental to the patient's quality of life to track therapy inappropriately. For instance, polypharmacy is very difficult, not only because patients must perform a careful administration of the intake of several drugs (for instance, in some cases some patients are consuming more than 16 drugs each day), but also because of their complex interactions between these substances. For this reason, more accurate documentation systems are required, which could help to setup conditions for, for instance, critical risks due to changes in the patient’s environment. It is therefore necessary to create information systems that store critical information like genotype, vital signs, and subjective assessments about the current state of health. But such systems require fundamental requirements for both data interoperability and data safety. They should not be stored either in corporate or institutional silos and should offer the possibility to share data not only to medical personal, but also to research institutes. In a nutshell, this vision requires patient centered solutions aiming to preserve critical clinical data, for instance generated with IoT – Wearables, able to provide data connectivity as well. SOLID represents a logical step for patient-centered solutions, pharmacovigilance, and patient-centered therapies and represents a plausible cornerstone to make this paradigm change real. We report on the concept and first steps in implementing a COVID-App using SOLID as a proof of concept towards this vision.
We'll look into some of the underlying technologies used commonly across real-world deployments of Solid. We'll show the high level architectures from the BBC's Together+ project in the UK (https://www.bbc.co.uk/togetherpod, which is still live until May 2023), and a hospital-care survey project from VITO in Belgium. Specifically we'll look at these real-world deployments of Solid in terms of their Data Modelling (using Linked Data), and their use of the powerful development tools provided by Inrupt, such as SDK tools, sophisticated Authorization (using both Access Control Policies (ACPs) and Access Grants (via W3C Verifiable Credentials)), a detailed ETL Tutorial, and the Artifact Generator.