In April, 100-year-old Emily Crocker received a package from her doctor in the post at her home in Portsmouth.

Inside was a kidney test, and instructions for her carer on how to download an app, dip a reactive strip into a urine sample and take a photo with a smartphone.

The app, from the Israeli start-up, uses machine learning algorithms, a subset of artificial intelligence that analyses data to find patterns, to diagnose kidney function from the photo. It then instantly shares the results with doctors.

In England, Crocker is one of more than half-a-million diabetes patients who will be getting one of these kits over the next year-and-a-half.

After more than a year of diverting huge resources to battling coronavirus, England’s NHS is now rolling out new artificial intelligence tools as it battles to cope with the enormous backlog of patients whose medical care has been sidelined by the crisis.

Roughly 4.7m people are currently awaiting treatment, including more than 387,000 who have been waiting for more than a year, for a range of serious conditions from cancers to cardiovascular illness and chronic kidney disease.

“Our first goal is to offer [our] service to everybody with diabetes who hasn’t done their regular test in the past 12 months,” said Katherine Ward, chief commercial officer of She estimated that number currently stands at 2.6m people.

Algorithms can offer big savings for overburdened health systems like the NHS: they are cheap to deploy at scale, can help prioritise and triage patients for under-resourced healthcare staff, and their results are instant.

“There has been a massive change in how patients are thinking about digital and how the NHS is thinking about it. We have got the scaling potential . . . at a time when NHS is more perceptive to this need,” said Ward.

At the heart of the NHS’s efforts to roll out AI is the one-and-a-half year old NHSX AI lab and the Accelerated Access Collaborative, the new umbrella organisation for UK health innovation. Together, they hope to be a one-stop shop for any AI start-ups trying to sell their clinical tools across the health service.

Before their creation, start-ups had to pilot, fund and test their technology with individual hospitals and NHS trusts, the bodies that are responsible for providing healthcare across a region, or which have a specialised function. Now, NHSX offers them the chance to have larger trials, running across dozens of NHS trusts, systematically funded.

“We want to find the best technology internationally, so we cast the net wider than the UK. We’d say you can come to the NHS, trial at sufficient scale . . . so you can say it works across the UK population,” said Indra Joshi, Director of Artificial Intelligence for NHSX, who leads the AI lab.

“Our offer to those latter stage companies is . . . we will help that journey to widespread adoption. That’s a lot further than ‘here’s some money and some intros’.”, which had been piloting its kidney tests in a handful of NHS trusts since 2017, has now signed 37 new contracts since September last year, when it was picked as one of 10 companies to share a £140m AI award, run by NHSX and the AAC.

Another AI award winner, Kheiron Medical, a UK start-up whose breast screening algorithm Mia helps radiologists decide if a woman should have further tests for breast cancer, had only managed to work with three locations within the NHS since 2016. Since September, it has signed contracts with 15 trusts, and Mia will screen more than half a million women over the next three years.

“The AI lab is an incredibly effective accelerator for working with the NHS. I think it surpassed all of our expectations on how effective it would be,” said Peter Kecskemethy, chief executive of Kheiron Medical. “Previously, we focused on large-scale studies across the globe, with the US as a major market . . . but with the AI Lab, the UK came into the picture as a really global leader in this space.”

The 45-person AI lab, which was set up in late 2019 by Matt Hancock, secretary of state for health, has narrowed its focus since the pandemic struck to finding solutions to urgent problems exacerbated by the coronavirus.

This has included setting up a Covid-19 chest image database, to train and validate diagnostic algorithms, for instance. Layla McCay, director of policy at the NHS Confederation which represents healthcare organisations across the UK, said the database had allowed patients to be assessed more quickly in A&E, saving radiologists’ time and increasing “the safety and consistency of care across the country”, as well as saving lives.

But it is also taking a longer term view, hoping to make the NHS a global exemplar for an ethical, regulated AI-infused national health system. To this end, it is building relationships with regulators like the Medicines and Healthcare products Regulatory Agency, to help clarify rules for AI software products, and working with independent researchers like the Ada Lovelace Institute on studying ways AI could inadvertently deepen healthcare inequities.

Critics said the NHS remains at the start of a long process. Rachel Coldicutt, the former chief executive of Doteveryone, a think-tank for responsible technology, said that while the NHS has put money into developing AI tech, it has spent less on adapting hospitals so AI can be used effectively.

“When we come out of the pandemic, is there any funding there to help teams in [NHS] trusts . . . think about how they are integrating new tools into business as usual?” she asked.

Ultimately, the promise of AI in healthcare is that it will help doctors to cope with increasing burdens of care in an ageing population and help to offset the spiralling shortages of specialists such as radiologists.

Finbarr Cotter, a haemato-oncology consultant at Barts Health, has been trialling an algorithm called EarlySign Colon Flag, which predicts which individuals have the highest risk of developing colon cancer.

The algorithms have been used on all patients who have been referred to Barts with suspected colon cancer in the past year, roughly 6,000 to 10,000 people, and have been a helpful tool to triage cases, he said. “It’s fascinating, we are learning, this is modern medicine, and it may be the only way to afford our health services . . . as a nation.”

Additional reporting by Sarah Neville