Presentation at AI morning in April 13th at Tampere University of Technology Kampusklubi. "AI Morning in April 13th experiments with a new distinctive concept and remixes together machine learning and analytics in the two verticals of healthcare and industry! There is a huge common ground in diagnostics of people and machines, and the same algorithms can be used in both. The presenters from healthcare and industry keynote a conversational networking forum in theme: 'Health: analytics'." See http://www.aiaamu.fi/

1. © 2017 Nokia1
Nokia Technologies
Cloud Platform for Remote Patient Monitoring. Case:
Stroke Remote Care
AI Morning on Health: analytics (http://www.aiaamu.fi/)
“Machine learning and analytics in the two verticals of healthcare and industry “
Friday 13th, April 2018, Tampere University of Technology / Kampusareena
Dr. Petri Selonen, Cloud SW Lead @ Nokia TECH DH Patient Care Engineering, Tampere
petri.selonen@nokia.com
Public

2. © 2017 Nokia2
Problem
• Stroke is Global Problem
• number #2 cause of death in the world
• number #1 cause of adult disability
• every 6th of us will suffer stroke
• Impact of Stroke
• USA: 800,000 strokes and130,000 deaths
• Finland, 15,000 strokes and 3000 deaths
• 7% of national healthcareexpenditure
• Preventability
• Risk of recurrentstroke ~3-20% within 3 months
• Atrial fibrillation major indicator
Public

3. © 2017 Nokia3
Solution
Public
INTERNATIONAL STROKE CONFERENCE Abstract WMP62: Remote Home Monitoring of Risk Factors After Stroke
or TIA to Improve Secondary Prevention - A Pilot Study. Stroke. 2018;49:AWMP62

4. © 2017 Nokia4
Implementation
Public
Analytics

5. © 2017 Nokia5
Analytics as a (part of a) medical device product
• Define Intended Use of the Product: is it a Medical Device?
Public

6. © 2017 Nokia6
EU Medical Device definition
For the purposes of this Directive, the following definitions shall apply:
(a) ►M5 ‘medical device’ means any instrument, apparatus, appliance, software, material or other article, whether used alone or in
combination, including the software intended by its manufacturer to be used specifically for diagnostic and/or therapeutic purposes and
necessary for its proper application, intended by the manufacturer to be used for human beings for the purpose of: ◄
— diagnosis, prevention, monitoring, treatment or alleviation of disease,
— diagnosis, monitoring, treatment, alleviation of or compensation for an injury or handicap,
— investigation, replacement or modification of the anatomy or of a physiological process,
— control of conception,
and which does not achieve its principal intended action in or on the human body by pharmacological, immunological or metabolic means,
but which may be assisted in its function by such means;
Public

7. © 2017 Nokia7
Analytics as a (part of a) medical device product
• Define Intended Use of the Product: is it a Medical Device?
• Find right classification for your Product
• risk analysis, hazards, harms, mitigation, residual risk, risk/benefit analysis
• Provide clinical validation of the algorithm
• algorithm complexity (or type) does not determine its classification
• Implement appropriate medical Quality Management System
• ISO 13485, IEC 62304, ISO 14971 risk analysis, verification, validation, change management
• classification determines the extent to which QMS is applied and how the system is declared
• algorithm implementationneeds to follow the QMS based on its classification, too
• Support Data Security and Privacy (GDPR, HIPAA)
• Health data, even pseudonymized, belongs to GDPR special categories of personal data and may only be processed in specific circumstances
• Automated decision making requires explicit and specific consent
Public

8. © 2017 Nokia8
Things are Evolving
However, managing Patient Risk will continue being at the core of SaMD creation. Take-away: When creating a medical product, embrace
medical regulation from get-go, instead of trying to avoid it or hoping to deal with it later.
Public
February 13, 2018: FDA permitted marketing of clinical decision support software for
alerting providers of a potential stroke in patients (ML based analytics of CT images for
stroke, alerting neurology specialist for validation).
April 11, 2018: FDA permitted marketing a ML based diagnostic device for
ophthalmology through retinopathy without clinician intervention.
March 29, 2018. FDA announced Digital Health Software Precertification Pilot Program
allowing participating technology companies to market SaMD products without premarket
review (low-risk) or with streamlined pre-market review to support faster iteration

9. © 2017 Nokia9
Nokia Technologies
Thank You!
Public

10. © 2017 Nokia10
Patient Care Cloud Platform Asset : SaaS for Remote Patient Monitoring
• Medical ready
• EU Class I Medical Device compliancy with technical file for self-declaration
• US Class I Medical Device Data System (MDDS) compliancy
• US HIPAA Privacy and Security Rule compliancy audited, EU GDPR
compliancy in progress
• Documentation, processes, procedures, QA
• Validated in pilots
• HUS dept. of Neurology pilot for RPM of stroke patients focusing on BP and
ECG arrhythmia analytics, leading to patient interventions
• >10000 hours of ECG / 9.6 Billion samples analysed
• clinician satisfaction 4.6/5 and willingness to commercialize
• TEKES BRAVE project
• Enterprise ready
• native FHIR data models
• audited technical, physical, and administrative safeguardsfor security and privacy
• SAML 2.0 SSO, 2FA, KMS, ABAC/RBAC,audit logging, threatmonitoring, etc.
• design for scalability, high availability, fault tolerance,resilience
• Versatilityvia leading edge technologies
• enablers for near real-time stream processing and analytics of continuoustime-series
data, as well as point measurements
• software defined infrastructure,fully automateddeployment,ability to launch
customer specific single-tenantstacks
• microservices architectureand containerizationfor customization
Confidential
A cloud platform for building remote patient monitoring solutions in medical environments.
Support for ingestion and analytics of ECG, blood pressure, weight, activity, and sleep data.

11. © 2017 Nokia11
Architecture

12. © 2017 Nokia12
Analytics as a (part of a) medical device product
• Health data, even pseudonymized, belongs to GDPR special categories of personal data and may only be processed in specific circumstances, most
common ones being:
• Data subject has given explicit and specific consent
• Processing is done for medical purposes by a medical professional
• Processing is done for scientific research purposes
• Automated decision making which significantly affects the data subject (e.g. determines treatment)requires explicit and specific consent, and data
subject has the right to obtain human intervention
• In practice you should always seek consent, however since consent is specific, you need to seek it again if there is a new type of processing
• Need to maintain contact with user, or you cannot introduce new types of processing
• It is unlikely that your consents are in sync over the entire dataset – maintain provenance metadata which tells you where the data came from and what kind of consents have
been given
• Typically you would create a separate anonymized (not pseudonymized!) dataset for R&D purposes
• Able to freely experiment on the data without GDPR limitations while building the model
• In production, when the model is applied to personal data, you are subject to GDPR and processing is based on consent (or medical purposes)
Public

13. © 2017 Nokia13
Development and operations processes defined for medical compliance
Culture for medical product development successfully created
Confidential
Designchoice
Marketentryofthesolutionwithspecific
intendeduse
Regulatory AffairsQuality ManagementDevelopment Work
Set requirements
Fulfill requirements
EU directives
US FDA
Technical stds
IEC 60601
ISO 13485 ISO
14971
US CFR 820
IEC 62304
IEC 62366
Processes,
Design control
Documents &
Records
Risk management
Other reqs Privacy

14. © 2017 Nokia14
FDA Medical Device definition
A medical device is an instrument, apparatus, implement, machine, contrivance, implant, in vitro reagent, or other similar or related article,
including a component part, or accessory which is:
• recognized in the official National Formulary, or the United States Pharmacopoeia, or any supplement to them,
• intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treatment, or prevention of disease, in man
or other animals, or
• intended to affect the structure or any function of the body of man or other animals, and which does not achieve its primary intended
purposes through chemical action within or on the body of man or other animals and which is not dependent upon being metabolized
for the achievement of any of its primary intended purposes.
Public

15. © 2017 Nokia15
Take-away
When creating a medical product, embrace medical regulation from get-go, instead of trying
to avoid it or hoping to deal with it later.
Public