Dr. Indra Joshi, introduced herself as NHSx’ Head of Digital Health and AI (Artificial Intelligence) on Day 2 of AIMed Europe 2019 but admitted she is unsure of what her job title means at this stage. Like the way we are unsure whether there will be real adoption of AI in medicine and healthcare space and how to scale up these technologies in view of the present boom.
In fact, Dr. Joshi’s role does not officially exist before April this year, when the Secretary of State Matt Hancock, soft-launched NHSx, the new digital branch and a joint unit between the Department of Health and NHS (National Health Service) England and Improvement. One thing for sure, most of us seemed to know why AI is being developed. “We need to create efficiency… We aimed to reduce costs, we also need to make a better experience and we need to just make the world a little bit faster,” Dr. Joshi said.
The rules of game
The purpose of NHSx is to build a framework and guidelines for something highly complex like AI. These “rules of game”, according to Dr. Joshi, include a code of conduct published last year, which questioned innovators if they have their respective ethical implications in place. If they are creating an evidence-based and validated product, and ultimately, who is going to pay for these creations, and more. Dr. Joshi said her unit had worked with a PhD student from University College London to compile some of these principles into an online self-assured workbook, for individuals to assess their creations and from there, build a community of similar-mind.
NHSx is also engaged in active conversations with regulators from the region and other parts of the world; partnering working group such as the International Telecommunication Union (ITU), the technical arm of United Nation, to address some of the wider ethical concerns. Closer to home, NHSx works closely with the Information Commissioner’s Office (ICO) and Care Quality Commission (CQC) to uphold public information rights and quality of care.
Bringing good evidence space products into the NHS
At the moment, the two pressing challenges are data access and deployment of AI-driven tools. For the former, a body called Health Data Research UK (HDR-UK) was set up 18 months ago to build pockets of data across the nation with digital innovation hubs to aid in the access of these data. For the later, NHSx partnered with the Accelerated Access Collaborative, a joint effort between NHS England and the Office of Life Sciences, in hope to introduce several products that have got good evidence space into the NHS.
“A good evidence space is something evaluated, shows clinical efficacy but also cost-impact efficacy and we are going to scale up the adoption of those technologies, across the health and care system in England,” Dr. Joshi explained. For now, this is likely to be the approach NHSx deals with AI. Unlike drugs, there is still no nice and clear framework guiding digital health. Over the next 6-12 months, Dr. Joshi believed the public will begin to see some samples of AI-driven tools within the NHS.
No fantastic slides, just foundation building
Meanwhile, the negotiation is on establishing a safe space to continuously test and validate these solutions, as well as to work with the finance team, to sort out the cost of deploying AI within the NHS in a long run.
“If some of you are developing products, which arrives at 5% efficiency gain, how does that relate to running a hospital? You still have to pay clinicians full-time but you only optimize the time by 5%, how does that translate into costing?” Dr. Joshi asked. “We are going to detail everything in a report called ‘AI, how to get it right’ to be released in October… I have no fantastic slides. The stuff we are doing is to create a foundation. If we can create the foundation, if we can enable products to have a clear pipeline, I think we will be able to really see some of the great stuff coming out over the next year or so”.
A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.