Government publishes NHS AI and data ‘code of conduct’

‘A set of principles that set out what we expect from suppliers of data-driven technologies, and a set of commitments that outline what the government will do to support and encourage innovators in health and care’

Posted 12 September 2018 at 8:55am by

The Department of Health & Social Care has issued guidelines on what it says it will do to encourage better take-up of AI and data-driven approaches in the NHS – and perhaps more interestingly, what it expects suppliers of such technologies to do in return.

Hence its key proviso: “We expect developers of data-driven technologies to work closely with health and care providers to ensure the safe implementation of technology and manage risks to the safety and quality of care in accordance with the expectations set out by the health and care system regulators. Health and care data must be used in accordance with the common law duty of confidence.”

The aim of the code is to make it easier for its suppliers to understand what it needs from them, and to help health and care providers choose “safe, effective, secure technology to improve the services they provide”.

That’s because it notes that AI is being used on this data to develop novel insights, tools to help improve operational efficiency and machine learning driven algorithms, and clinical decision support tools to provide better and safer care – which “presents a great opportunity”, but add that “these techniques are reliant on the use of data that the NHS and central government have strong duties to steward responsibly”.

Ten principles are outlined for AI and data-driven health IT applications, which are:

  • Define the user
  • Define the value proposition
  • Be fair, transparent and accountable about what data you are using
  • Use data that is proportionate to the identified user need (data minimisation principle of GDPR)
  • Make use of open standards
  • Be transparent to the limitations of the data used and algorithms deployed
  • Make security integral to the design
  • Define the commercial strategy
  • Show evidence of effectiveness for the intended use
  • Show what type of algorithm you are building, the evidence base for choosing that algorithm, how you plan to monitor its performance on an ongoing basis and how you are validating performance of the algorithm

It also outlines the 5 ‘commitments’ the government will make in turn back to the market:

  • Simplifying the regulatory and funding landscape
  • Creating an environment that enables experimentation
  • Encouraging the system to adopt innovation
  • Improving interoperability and openness
  • Listening to our users

The Department wants feedback from the market to improve the suggestions, with a new version promised in December – so you may want to fill in its questionnaire (here) if you are such a supplier.