Information Infrastructure for Health

It's not an Internet of Things: It's an Internet of People

ATI Scoping Workshop, Edinburgh

Neil D. Lawrence

18th November 2015

Global Health

• Governance for Global Health Response: Lessons from the Ebola Epidemic Royal Society, 15/09/2015
  • What lessons do we learn for future global crisis?
  • Funding, vaccines, responsibility.
  • Little about improving communication.
    • Prof Oyewale Tomori: Why didn't Nigeria Succumb?

What is an Information Infrastructure?

• Talk at ControlPoint "Right First Time" Day.
  • Audience of Utility companies: Gas, Water, Electric.
• What would a modern Water network look like?
  • Would we build such massive infrastructure?
  • It seems we can't even afford to properly maintain existing architecture.

Data Science in Africa

Data Science in Kenya

• IBM Nairobi: Monitoring Small Farms

• Makerere University:

  • Crop Monitoring (Ernest Mbweze)
  • Kudu: Distributed Bid Matching (Kenneth Bwire)

• Collaboration with UN Global Pulse and Makerere

  • Disease Monitoring with Gaussian Processes (Ricardo Andrade Pacheco, Martin Mubangizi, John Quinn & NL)

Farr Institute at HeRC

• Manchester and HeRC
  • Devolution of power in Social and Health to Manchester Region.
• Manchester has long led in Health Informatics (Iain Buchan & Connected Health Cities)
  • Agenda extends acoss N8

UK - Where do we Lack Infrastructure

Case Study: Mental Health

• NewMind Network (led by Chris Taylor)

  • Early Diagnosis of Dementia, Parkinsons.
  • Support for Bipolar disorder
  • What are the implications for Humans?

• Clinicians, patients, device manufacturers.

  • Ideal domain for data & learning

  • London meeting next week

Major Challenge

• "Data Liquidity"
  • If we can't get to the data we can't learn.
• Clinical Interests
  • Hard-won data is hard to obtain.
  • Working better in computational biology than health informatics?
  • Is it patient interests or practitioner interests?

Technical Challenges

• Interconnecting data

  • Data of different types/modalities.
  • Weighting of text, images, clinical tests, genomics, transcriptomics, social media, music listening, loyalty card.

• Scale: When Big Data is Small

• Massively Missing Data

• Steaming nature of data

  • Health vs Advertising

Time horizon: 10 years?

Social Challenges

• Interconnecting data

  • Loss of privacy and trust care.data fiasco

• Models of data ownership

  • Data protection (actually people protection)
  • Carphone Warehouse and TalkTalk.
  • Need for a "Digital Enlightenment"

• Jonathan Price: Doughty Street Chambers

• Co-evolution of Law and Technology

Time horizon: ???

Answer?

• Closer integration of technical and social solutions.

• It is not enough to provide centralised learning solutions working on large data.

  • inference integrated within information infrastructure

• More focus on

  • information leakage,
  • frameworks of trust,
  • user-centric models for learning and data storage

Where does the ATI come in?

• Open meeting September 2014:
  • "Critical mass"
  • But no-one could define what that meant!
• Leadership + Critical Mass can do something
  • Enabling the social revolution requires a technical revolution.
• A Coalition of the Competent
• Corevolution of technology, law and society.
  • This needs multi-million pound institutes to achieve.