Privacy In The Era Of Personalised Medicine Dr Kelvin Ross, Dr - - PowerPoint PPT Presentation

Privacy In The Era Of Personalised Medicine

Dr Kelvin Ross, Dr Brent Richards, Dr Zoran Milosevic, Dr Mark Pedersen

Dr Kelvin Ross Director, IntelliHQ kelvin@intelliHQ.com.au

Machine Intelligence Progress

intelli

• AI Innovation Centre for

Healthcare

• NFP partnership between

GCUH, research and industry

Innovation In Innovation In

• Attract Innovators
• Readiness Assessment
• Customer Validation & Trials
• Benefit Evaluation

Innovation Ou Innovation Out

• Idea Development
• Venture Establishment
• Team Development
• Funding Advisory
• Commercialisation Pathways

GCUH

Customer Innovation Partner

Delivery Model

Universities

• Local: Griffith, Bond
• Regional: QUT, UQ, SCU
• National: Curtin, …
• International: NUS, MIT, Tapei,

… Ventures Ventures Ventures Ventures Ventures Advisory Board Investment Fund Investment Fund Investment Fund Commercialisation Pathway Professional Associations

• College of Intensivists,

RACS, … Heath Technology Assessment Facility Precision Medicine Data Platform Other

• Data 61, AEHRC
• Health & Knowledge Precinct

Government

• Qld Govt, Advance Queensland
• Federal Government

Industry Partners

• Platform: AWS
• Devices: GE, Philips. …
• EMR: IMDSoft, …
• SMEs: TechConnect, KJR,

Veriluma

• Other: Verily, CIMIT

Selection Board Research & Education Health Economics Advisory Global Innovation Hubs

• Startup Health, TMCx, CIMIT, NIA, …

Mentoring / Advisory

PMDP

a patient physiology data platform supporting AI-based decision systems intrinsic to emerging Health Technologies.

• 1. Improve Health Outcomes
• 2. Create local opportunities for skills, jobs and venture development
• 3. Attract Entrepreneurs & Investment
• 4. Achieve Global Recognition for Research, Innovation & Commercialisation

Time-Series Physiologic Data Lab Results Orders Interventions Pre-Admission Data Demographics Outcomes

Treatment Outcomes Interventions

PREDICT OUTCOME INTERVENTION ANALYSIS

AI and machine learning techniques provide prescriptive/prescriptive analytics and decision support to optimise health outcomes. Comprehensive predictive and prescriptive models of disease and intervention can be realised from deep data sets. Prediction algorithms can pre-empt onset of sepsis in advance of organ failure, allowing for early intervention. Reinforcement learning can be used to understand optimal policy for interventions for individualised patient conditions and diagnoses. This can augment clinical decision processes to advise on optimal intervention, such as drug dosage.

Machine Learning

DAT DATA

Interventions Bio-signals Outcomes

Precision Medicine Precision Medicine Journey Journey

Data Data Actions Actions Go Goals als

Public Data Sets

https://www.kaggle.com/tags/healthcare

Other resources:

• https://physionet.org
• http://www.ukbiobank.ac.uk
• https://healthcare.ai
• https://www.datasciencecentral.com/profil

es/blogs/10-great-healthcare-data-sets

• https://cloud.google.com/blog/big-

data/2017/05/new-healthcare-and- population-datasets-now-available-in- google-bigquery

Source: https://ai.googleblog.com/2018/05/deep-learning-for-electronic-health.html

Big Data Growth

ICU Rooms

• 25 Beds, 250m

rows

Inpatient Patches

• 250 patients

Outpatient Patches

• 2,500 patients

Community Care

• 12,500 patients

IOT devices will grow demand for health data analytics Additional Units, Hospitals, Health Regions

Platform Architecture

Data Sources Big Data Store Data Compute

(map/reduce)

Toolset: Analytics Visualisation

Modelling Tools Systems Integration Tools

ETL ETL Heart Rate Analysis Prediction Models Data Science Platform Mortality Analysis Septic Shock

Deeper ICU Data Sets

ANZICS ~ 1,400k (A+NZ) eICU-CRD ~ 200k (US) MIMIC III ~ 60k (US) GCICU ~ 20k (GC)

PMDP

Other hospitals and HHSs

# patients

200+ research publications

Deeper Data

First 24 hrs

HR_HI = 118 HR_LO = 44

Information Loss

Hourly Minutely

Deep Medical Datasets

Clinicians and data scientists need access to deep de-identified data. Deep data creates new tiers

• f predictive power, and

enables new front end applications. Data sources from many hospitals need resources that will scale.

Need Data Scale

250k

2k

Data Is The New Oil

00101100101001010111001010010110010100101100101100101100101001011001010010110010100101100 10110100111010010101100101000010110010100101100110100101100101001011001010010110010100101 10010100101101110010100101100101001011001010010110010100101100111010010110010100101100101 10010101110010100101100101010101100101001011001010010110010100101100101001011001010010101

Flatiron

https://www.forbes.com/sites/davidshaywitz/2018/02/18/the-deeply-human-core-of-roches-2-1b-tech-acquisition-and-why-they-did-it/

Reduced Control Groups via Real World Evidence http://cancerres.aacrjournals.org/content/78/4_Supplement/P3-17-03 Flatiron Health makes software to improve the workflow in cancer clinics and aggregates anonymized data from that software to share with pharmaceutical companies and research institutes.

"relentless curation of data--an army of "data janitors" transforming EHR data into analyzable, actionable information."

Injest Refine Models Products

Data Is The New Oil

Data is ingested from

• hospitals. Hospital

contribute to a data consortia, joining their data with many others to provide a richer data set. Consolidating, cleaning and de- identifying data is the first step. Value adding to the data sets, such as labelling, augmentation, and quality assurance enhances data sets for optimised model creation. Machine learning models are constructed from the data sets. Rich data sets provide the underlying basis for powerful models. Machine learning models are incorporated within products to deliver value, translating decisions into actions, capturing more data into the machine learning feedback loop.

UK House Of Lords AI in the UK: ready, willing and able?

https://publications.parliament.uk/pa/ld201719/ldselect/ldai/100/100.pdf

• Government (Public) data including health data is of

considerable value. Data is the infrastructure for artificial intelligence - providing a fundamental framework for AI systems to function

• Could provide considerable economic benefit to the community
• Concerned that data could lack equitable access and be

monopolized by large technology companies

• Commercial use of public health data raises community

concerns

• Community expects societal benefit from data sharing
• But data on its own is raw, siloed and of limited value.

Significant work is required to extract it, merge and add value.

• Public sector lack guidance on how to negotiate agreements

Stages of Data Readiness

Source: Ready…set…AI — preparing NHS medical imaging data for the future, Dr Hugh Harvey, https://towardsdatascience.com/ready-set-ai-preparing-nhs-medical-imaging-data-for-the-future-8e85ed5a2824

Adapted from: https://arxiv.org/abs/1705.02245

Democratisation Of Data

Public Private

How to Leverage Private Investment? How to grant data access equity? What is considered acceptable use?

‘Identifiability spectrum’ Credit: Understanding Patient Data

De-Identification Process

Hosp A PRIVATE LIMITED OPEN

• PHI under tight access

constraints

• Data linkage between

data sets HIPAA Safe Harbor: Limited Data Sets HIPPA Safe Harbor: incl timestamp de-identification

• PKI used to provide confidential

mapping between patient/provider IDs, and de-identified IDs

• 18 safe harbor rules

less timestamp de-identification Identifiable patients

• 1. Names
• 2. Geographic subdivisions smaller than a state
• 3. All elements of dates (except year) for dates

that are directly related to an individual, and all ages over 89 and all elements of dates (including year) indicative of such age

• 4. Telephone numbers
• 5. Vehicle identifiers and serial numbers, including

license plate numbers 14.Health plan beneficiary numbers 15.Full-face photographs and any comparable images 16.Account numbers 17.Any other unique identifying number, characteristic, or code, except as permitted by paragraph (c) of this section [Paragraph (c) is presented below in the section “Re- identification”]; and 18.Certificate/license numbers

• 6. Fax numbers
• 7. Device identifiers and serial numbers
• 8. Email addresses
• 9. Web Universal Resource Locators (URLs)

10.Social security numbers 11.Internet Protocol (IP) addresses 12.Medical record numbers 13.Biometric identifiers, including finger and voice prints

HIPAA “Safe Harbor” De-identification Rules

Sepsis Survivability Improvement Over 10 Years

Scrambling admission date loses predictive power

Source: Kaukonen et al, Mortality Related to Severe Sepsis and Septic Shock Among Critically Ill Patients in Australia and New Zealand, 2000-2012, JAMA, 2 April 2014

Hosp A PRIVATE Hosp B Private Datasets Identified Patients, Providers LIMITED Hosp C Merged Limited Data Sets Partially De-identified

ML Requires Scale Via Merged Datasets

Different EMRs, different Devices, Different Configurations

Data Governance

HHS Source PMDP

• Who owns the data?
• How is access evaluated?
• What processes are required?
• What legislation, regulation and policies apply?
• How does data custodian, ethics approval get re-assigned?
• What are the technical restrictions and limitations in the data as a

result?

• How is data access monitored & surveilled?

Roles: Data Owner Data Custodian Data Provider Assurance Ethics Committee

Jurisdiction

Patchwork of legislation and regulation

Security Access

ML Target X Y Z Data Source A

✓

B

✓ ✓ ✓

C

✓ ✓

• Source data tagging basis for

access model

• Security privileges allocate

access to designated items for specific target models

• Record of access maintained

in log

• Verifiable Data Audit
• https://deepmind.com/blo

g/trust-confidence- verifiable-data-audit/

Data Operator

Data Trust

Data Provider Data Trust Data User Assurer

• Custodian for de-identified data
• Manage Policy and Terms of Use
• Coordinate Ethics Approval
• Democratised funding models

for access

• Investment model for partners
• Operate data platform
• Value add data
• Provide de-identified

data

• Assign custodianship to

Data Trust

Thankyou

Dr Kelvin Ross Director, IntelliHQ kelvin@intelliHQ.com.au