An inspiring online event on 3 February 2021. We are discussing the future of data-driven health solutions that focus on fairness for all stakeholders: people, business and the public sector. We are asking questions such as: What is fairness in health? What role does trust play in data-driven health services? What needs to change and who needs to act? Most of all, we are launching “The Fair Health Data Challenge“.
Event speakers:
- Jaana Sinipuro, Project Director, IHAN – Human-driven data economy, Sitra
- Dipak Kalra, President, The European Institute for Innovation through Health Data (i~HD)
- Pekka Kahri, Technology Officer, HUS Helsinki University Hospital
- Markus Kalliola, Project Director, Health data 2030, Sitra
- Tiina Härkönen, Leading Specialist, Sitra
6. Source: van der Kamp & Plochg: ”The Health System Quartet” in:
Sturmberg J (ed), Embracing Complexity in Healthcare, Springer 2019, pp. 113-123
Cure Care
Heal Deal
Emphasis on
Professional
Agents
Emphasis
on
Personal
Agents
It’s going to get a lot more personal!
9. The part of the economy that focuses on creating
services and data-based products in an ethical
manner.
In the fair data economy data is shared between
different parties with individuals' consent in a
seamless and transparent way.
People use services that improve their well-being and
everyday life, companies achieve growth through
innovation, and the well-being of society increases
Vision is that Europe has a well-functioning
data market, where companies that use data
responsibly and open-mindedly succeed with
smart services
Our response to
The European
strategy for data
10. DATA ECOSYSTEM IS A GROUP OF ENTITIES
THAT WANT TO CREATE NEW BUSINESS BY
SHARING DATA WITH EACH OTHER
Data is shared with the data owner’s permission (including
individual’s permission) and according to the rules set in the data
ecosystem’s rulebook.
An ecosystem that follows fair rules creates value for all participants.
Data can be shared in the ecosystem more
freely, transparently and safely.
11. There must be a balance
between all stakeholders’ needs,
interests and rights
12. Working paper link: https://www.sitra.fi/en/publications/towards-trustworthy-health-
data-ecosystems/
14. Electronic Health Records for Clinical Research
A RECIPE FOR
TRUSTWORTHY DIGITAL
HEALTH
Professor Dipak Kalra
President of i~HD
Trusted!
Quest for data-driven and fair health solutions
SITRA
3rd February 2021
16. The spectrum of data use: from care to research
3
Individual level health data
EHR systems, apps, sensors, genomics,
Clinical Decision Support, AI
Used for:
• Health status monitoring
• Continuity of care (including the patient
and caregivers)
• Care pathway tracking, clinical workflow
management
• Real-time feedback and guidance to
patients and clinicians
• Personalised medicine
• Disease interception, prevention and
wellness
Population level health data
EHR systems, regional & national
eHealth infrastructures
Reused for:
• Quality and safety, care pathway
optimisation
• Medical device and algorithm
refinement
• Pharmacovigilance
• Public health surveillance
• Public health strategy
• Health services and resource
planning
Big health data
national & international research
infrastructures,
federated query platforms
+ cross-sectoral services
Reused for:
• Digital innovation: devices,
sensors, apps
• AI development
• Personalised medicine and bio-
marker research
• Diagnostics development
• Drug development
• Disease understanding and
stratification
17. 4
The Sitra report highlights some
important challenges
• Heterogeneous legislation
• Data incompatibility
• Delays in digitalisation
• Digital literacy
• Support for innovation
• Lack of trust and transparency
• Emerging threats
• Public health crises
18.
19. Demand driven push for interoperability
6
Market response
Citizen demand
e.g. value
propositions and
business models
for SMEs
e.g. interoperability
certification for
citizen-facing
digital solutions
Demand for interoperable EHRs
Demand for reliable and
trusted apps and wearables
20. Example data quality issues from hospitals
7
34% of weight errors led to medication-dosing errors
48% of these patients required additional
monitoring, examination or treatment
Data quality variation across example EU hospitals
- the data most needed for clinical research
(partial table)
• Most EHR data is captured by busy junior staff, using various EHR systems
• Staff have no access to training in data quality
• Patients also have no training! (but their data is becoming increasingly important)
21.
22. The FAIR principles: a commitment by data sources
Data must be Findable
a searchable method to discover resources, with standardised metadata and a
repository identifier
Data must be Accessible
retrievable metadata, and potentially retrievable data via appropriate protocols
and controls
Data must be Interoperable
metadata is standardised, data conforms to relevant published standards
Data must be Reusable
there is transparency about the terms under which the data may be reused
9
23. Data sharing agreements protect all parties
10
Code of
Practice
Data sources
Data users
Federation
data sharing
agreement
Data enrichment
Handling of analysis results
Agreeing recognition and reward
Demonstrating compliance
Measures to protect privacy
Data specification and access
Research purposes and protocol
24. Some “grey” GDPR compliance areas researchers
(and DPOs) struggle with
Is new consent really needed when a large data set is reused? (every time?)
How precisely should purpose be explained, when reuse is intended?
How maximal can a data set for research reuse be?
Can anonymised data retain one way linkage, for updates?
What is acceptable anonymisation for rare disease groups, genomics, images…?
What is required for a data set to be open?
Can we have standards for what is “adequate” or “proportionate”?
What data protection assurances can realistically be required, assessed, enforced?
Can we create a level European playing field for those who wish to reuse data?
11
We should strive for an EU aligned code of practice
The Data Governance Act may help
25. The challenge with gaining public acceptance of health data reuse
12
Individual level health data
EHR systems, apps, sensors, genomics,
Clinical Decision Support, AI guidance
Used for:
• Health status monitoring
• Continuity of care (including the patient and caregivers)
• Care pathway tracking, clinical workflow management
• Real-time feedback and guidance to patients and clinicians
• Personalised medicine
• Disease interception, prevention and wellness
• Healthcare provider reimbursement
Population level health data
EHR systems, regional & national
eHealth infrastructures
Reused for:
• Healthcare provider performance and planning
• Quality and safety, care pathway optimisation
• Medical device and algorithm refinement
• Pharmacovigilance
• Public health surveillance
• Public health strategy
• Health services and resource planning
Big health data
national & international research infrastructures,
federated query research platforms
+ cross-sectoral infrastructures & services
Reused for:
• Disease understanding and stratification
• Digital innovation: devices, sensors, apps
• AI development
• Personalised medicine and bio-marker research
• Diagnostics development
• Drug development
Decreasing public understanding of why and how data are used
Increasingly unfamiliar data users
Increasing distance of data results from the patient
Increasing time from data use to demonstrated value
Perceived lessening choice and greater cybersecurity risk = harder to trust
27. Citizens’ Juries have highlighted…
14
https://onlinelibrary.wiley.com/doi/full/10.1002/lrh2.10200
…that the public are concerned about the use of
health data even if their identify is not exposed
29. How do we reach societal acceptability?
Data protection regulations prioritise the rights of the individual to privacy
Clinical research can bring important benefits to society
Many surveys indicate patients are in favour of their data being re-used
for research
But the GDPR hype sometimes breeds fear (public, DPOs, CEOs…)
The public need greater transparency about why and how health data are
used, safeguarded, and the benefits of that use
We need to find the right balance between the rights of the individual
and the benefits for society
16
30. 17
Showcasing the benefits
of research using health
data
Defining the elements of
trust
Assuring the
trustworthiness of data
custodians and data
users
Transparency in the uses
of health data
Principles,
codes of conduct,
security controls
Public perceptions,
preferences, priorities
Evidence and accountability
to the public and professionals
A holistic approach is needed
Capturing the outcomes
from data use
Consensus on
- kinds of health data
- purposes of use
- types of user
31. TOWARD DATA-DRIVEN
HEALTH SERVICES AND
ECOSYSTEMS – CASE HUS
.
Pekka Kahri, Technology Officer, HUS
Trusted! Quest for data-driven and fair health solutions –event, February 3, 2021
32. HUS OPERATES IN 23 SITES
2
Southern Finland special
catchment area for
tertiary care together with
districts of
Southern-Karelia (Eksote),
Kymenlaakso (Carea) and
Päijät-Häme (PHHYKY)
33. 16,000
childbirths
24.5 million
laboratory tests
2,900,000 PATIENT VISITS
Healthcare in 2019
330,000
elective referrals
82,000
emergency referrals
680,000
PATIENTS
TREATED
860, 000
treatment
days
2,800
hospital beds
92,000
surgeries
1.8 million
imaging examinations
453
organ transplants
34. ARTIFICIAL INTELLIGENCE & HUS
Start with ”Low hanging fruits” e.g.
automation of routine tasks with
RPA / Chatbots / AI
Improve understanding through
pilots – accept that many pilots
will fail
Build strategic understanding of
required competences, tools and
partners
Focus on clinical applications with
imaging & ICU monitor big data –
critically ill patients
Process
management
and
optimization
Clinical
decision
support
Image
analytics
Predictive
clinical
analytics
35. GETTING STARTED WITH AI –
KEY QUESTIONS FROM
HOSPITAL’S PERSPECTIVE
• How to make health data available for AI
research and algorithm development?
• How to develop needed competences and
engage with partners in co-development?
• How to get meaningful and validated AI
tools to clinicians’ use in patient care?
5
36. GETTING STARTED WITH AI –
KEY QUESTIONS FROM
HOSPITAL’S PERSPECTIVE
• How to make health data available for AI
research and algorithm development?
• How to develop needed competences and
engage with partners in co-development?
• How to get meaningful and validated AI
tools to clinicians’ use in patient care?
6
“Substance side
of things to solve”
“Process side of
things to solve”
37. GETTING STARTED WITH AI –
KEY QUESTIONS FROM
HOSPITAL’S PERSPECTIVE
• How to make health data available for AI
research and algorithm development?
• How to develop needed competences and
engage with partners in co-development?
• How to get meaningful and validated AI
tools to clinicians’ use in patient care?
7
38. Decades of clinical information stored
as digital data
1980 1990 2000 2010 2020
Lab and pathology
Disease-specific quality registers
EHR
Operative and financial data
Radiology ERP
PACS
Digital ECG
Surgical ERP
ED & ICU
Obgyn
Genomics
39. Azure ML
Data Lake
Integrations
HUS DATALAKE INFORMATION ARCHITECTURE
Integrations
Azure Data Factory
IoT Data
Laboratory
Imaging
EHR (Apotti/Epic)
Pathology
Genomics
Health Village
Operative/Finance
Raw
Data
Zone
Master
Data
Zone
Distribution
Pseudonymization
Azure Data Factory
Harmonization
Validation
Azure Databricks
Research
Clinical
Operational
Reporting
Azure Databricks
Azure Machine
Learning
Applications
ETL
Azure Data Factory
Permissions
Workflow
Analytics & AI
Reporting & BI
Integration
Power BI
Synapse Analytics
API Management API for FHIR
Azure Databricks Azure ML
Azure Data Lake Storage
40. SERVICES FOR DATA USE
Researchers and innovators need services to navigate
the datalake and the regulatory “shallows”
• Information services with process guidance on data
availability, dataset formation, secure data
processing environments
• Research services with process guidance concerning
data permits, research permits, ethical committees
• Solution oriented legal support services with strategic
backing
41. GDPR, MDR,
Clinical trials regulation,...
REGULATORY ENVIRONMENT FOR INNOVATION,
DEVELOPMENT, VALIDATION AND TESTING
Clinical research
•Involves patients
•Act on medical research
•Ethical committee & institutional
review
•Good clinical practice guidelines
Biobank research
•Biobank act
•Wide consent from sample-
donating participants
•Ethical review
•FinBB collaboration and shared
servies
Registry research
•Secondary use of social and health
data - the new act
•Division of tasks between individual
data controllers and Findata
•Collaboration and shared services Attaching clinical
data from EHR
Feasibility studies,
control groups
Possibility to re-contact
donors / patients
42. GETTING STARTED WITH AI –
KEY QUESTIONS FROM
HOSPITAL’S PERSPECTIVE
• How to make health data available for AI
research and algorithm development?
• How to develop needed competences and
engage with partners in co-development?
• How to get meaningful and validated AI
tools to clinicians’ use in patient care?
12
43. Rare
diseases
Acute
leukemia
Home
dialysis
eMOM
GDM
Health
Village
PROJECTS
• Innovative development models
• New services & products addressing real-time needs
• Agile, standardised collaboration platform
Ecosystem Partners & Output
• Patient data
• Clinical activities
• Understanding patient needs
• Patient reported outcomes
• Product / service needs
AI Head
Analysis
Child with
diabetes /
IHAN
• New global solutions
• Improved patient care
• Provider satisfaction
• Mydata – secure and safe
eCare
for Me
PROJECTS
5.2.2021
13
The world’s fastest track to commercialization for digital health and wellbeing innovations
SPIN
OFF
PROJECTS
NN project
HUS DIGITAL HEALTH ECOSYSTEM
44. YOU MUST WORK WITH PROCUREMENT
• Traditional tenders (or textbook innovative procurement) are not
giving best outcomes
• Well-established large companies rule, SMEs and startups find it difficult to launch
business with large hospitals
• Lengthy processes, AI industry is highly dynamic and it’s highly difficult to evaluate
which company has best competences for the task at hand
5.2.2021
14
Case: AI development partner for TBI prediction algorithm
• Formal tendering process – agile team with clinical, legal and IT competence
• Assignment for shortlisted companies which includes interaction with clinicians
45. GETTING STARTED WITH AI –
KEY QUESTIONS FROM
HOSPITAL’S PERSPECTIVE
• How to make health data available for AI
research and algorithm development?
• How to develop needed competences and
engage with partners in co-development?
• How to get meaningful and validated AI
tools to clinicians’ use in patient care?
15
46. PHYSICIAN ROLE AND SKILLS
• AI will be part of routine tools for physicians -> basic
understanding of AI for all
• Physician role is needed in every stage of AI
development
• Problem identification and definition
• Information architecture and data management
• Analytics and IT
• Development and validation processes
• Governance and ethics
5.2.2021
16
47. 12.9.2018
17
Applications and systems for social and healthcare that utilize algorithms,
artificial intelligence, machine learning and data analytics
CERTIFIED QUALITY SYSTEM (ISO 9001, ISO 13485)
48. EFFICIENT DEPLOYMENT AND INTEGRATION TO
OPERATIVE SYSTEMS IS A MUST
Apotti (EPIC) EMR system & Maisa patient portal Digital Health Village
54. TEHDAS Joint Action
Towards the European Health Data Space
Joint Action Towards the European Health Data Space,
TEHDAS
Markus Kalliola
Project Director, TEHDAS Coordinator
The Finnish Innovation Fund Sitra
55. TEHDAS Joint Action
Building momentum for the European Health Data Space
07/12/2020
• ”We need common data spaces - for example, in the energy or healthcare sectors. This
will support innovation ecosystems in which universities, companies and researchers
can access and collaborate on data.”
Commission President von der
Leyen, State of the Union address,
16/09/2020
• ”The European Council welcomes the creation of common European data spaces in
strategic sectors, and in particular invites the Commission to give priority to the health
data space, which should be set up by the end of 2021.”
European Council,
Summit Conclusions,
02/10/2020
• Plan: Legislative proposal for the European health data space, incl. impact
assessment, Articles 114 and 168 TFEU, Q4 2021
Commission work programme
2021, 19/20/2020
• ”I had five messages for the Leaders today: The first one is on data sharing: We have to
share comprehensive and accurate data in real time with the ECDC platform.”
President von der Leyen,
Statement after Covid-discussions,
29/10/2020
• “The Commission and the German Council Presidency announced their intention to
work closely together on a secure European Health Data Space for better healthcare,
research and health policymaking.”
German EU Presidency’s
Conference on Digital Health 2020
– EU on the Move, 11/11/2020
• “The set-up of the European Health Data Space will be an integral part of building a
European Health Union…”
Commission Communication on
building a European Health Union,
11/11/2020
• COM(2020) 767 final
Commission proposal for a
European Data Governance Act,
25/11/2020
Council Conclusions on COVID-19 lesson learned in
health,
09/12/2020:
“CALLS upon the European Commission, the
Member States, and all relevant public and private
stakeholders to jointly collaborate in order to
deliver a functioning European Health Data Space,
…”
“WELCOMES the close cooperation between
Member States and the Commission in preparing
the Joint Action for the European Health Data
Space…”
56. TEHDAS Joint Action
2021 - Commission & Presidency joint press release on activities
paving the way for better data-driven health care in Europe
• proposed by the Commission in Q4 of 2021
Legal text on the European
Health Data Space
• to propose options on governance, infrastructure, data quality and data solidarity and empowering
citizens with regards to secondary health data use in the EU
TEHDAS Joint Action
• under the EU4Health programme, as well as common data spaces and digital health related innovation
under Horizon Europe and Digital Europe programmes
Investments to support the
EHDS
• for secondary health data use developed through engagement with relevant actors
Targeted Codes of Conduct
• to demonstrate the feasibility of cross border analysis for healthcare improvement, regulation and
innovation
A pilot project
• for digital transformation of health and care available for Member States as of 2021 under Recovery and
Resilience Facility, European Regional Development Fund, European Social Fund+, InvestEU.
Other EU funding opportunities
Adapted from the joint press release 11/11/2020
57. TEHDAS Joint Action
Joint Action Towards European Health Data Space
Project Coordinator: Finnish Innovation Fund Sitra, Finland
Project Acronym: TEHDAS
Starting Date: 01/02/2021
Project Duration: 30 months
Participants: 26 (22 Member States, 4 other countries)
Co-funding: 3rd EU Health Programme (2014-2020)
Budget: Revised €4.16 mill.
EU €2.5 mill. + MS €1.66 mill.
58. TEHDAS Joint Action
Mission statement
The Joint Action helps the
Members States and the
Commission in developing
and promoting concepts for
sharing of data in secondary
use for purposes for citizens’
health, public health, as well
as health research &
innovation in Europe.
In the future Europe, citizens,
communities and companies
benefit from a protected and
secure access to health data
regardless where it is stored.
59. TEHDAS Joint Action
Towards the European Health Data Space
Participant
Countries
18 June 2020
Member States interested in participating (22)
Other countries interested in participating (4)
Eligible, not participating
MT
60. TEHDAS Joint Action
Joint Action is expected to produce as its outputs options for
• Data Governance for the European Health Data Space (EHDS), complementing the
horizontal European Data Spaces framework, including functions and
responsibilities of the relevant actors.
• Guidelines on using health data for research and policy making
• Guidelines on Ethical, Legal and Social issues in the European Health Data Space
• Data Quality framework encompassing semantic interoperability and FAIR
principles, as well as anonymization and pseudonymisation techniques.
• Infrastructure architecture and technical interoperability guidelines to enable
European Health Data Space services
• Economics models focused on the sustainability of European Health Data Space
Adapted from CHAFEA Ares(2020)1379455 -
05/03/2020: Policy expectation for the joint actions
2020
61. TEHDAS Joint Action
Towards the European Health Data Space
Contact information
Markus Kalliola
Project Director, TEHDAS Coordinator
Markus.kalliola@sitra.fi
+358503591499
@markuskalliola
62. The Fair Health Data Challenge
The challenge is a unique fair
data-driven solution contest
aspiring to identify existing or soon-
to-be-released European health and
wellbeing solutions that are based
on data sharing and fair values.
Applications deadline 9th of April
Apply here
sitra.fi/healthchallenge
What is fairness in health?
What role does trust play in
data-driven health services?
What needs to change and
who needs to act?
63. Name Title Organisation
Emmanuel Bacry Chief Scientific Officer Health Data Hub
Artur Olesch Journalist aboutdigitalhealth.com
Cátia Pinto MD, Public Health Specialist SPMS - Shared Services for Ministry of Health
Thomas Plochg Director Federation for Health
Stephan Schug Chief Digital Health Strategist IQmed Healthcare
Jaana Sinipuro Project Director Sitra
Ilkka Räsänen (secretary) Leading Specialist Sitra
Board of Domain Experts
64. The Challenge Rules
The candidates are evaluated against the
following factors
1. Managing one’s personal data
2. User-friendliness & accessibility
3. Data-sharing model
4. Fairness factor
5. Security & confidentiality
6. Innovativeness
7. Trust factor
Open until
9th of April