Technology company DeepMind claims that Artificial Intelligence (AI) can predict acute kidney injury (AKI) two days before it does any harm. And it has also developed a phone app that can inform clinicians within minutes.
Writing in the journal Nature* the DeepMind team worked with the United States Department of Veterans Affairs (VA), to apply AI technology to de-identified electronic health records collected from a network of over a hundred VA sites.
The research showed that the AI could accurately predict AKI in patients up to 48 hours earlier than it is currently diagnosed. Importantly, it says the model correctly predicted nine out of 10 patients whose condition deteriorated so severely that they then required dialysis.
DeepMind says this could provide a window in the future for earlier preventative treatment and avoid the need for more invasive procedures like kidney dialysis. The model has also been designed so that it might, in the future, generalise to other major causes of diseases and deterioration such as sepsis, a life-threatening infection.
And DeepMind has also developed a new phone app alerting system, known as Streams, to get this information to clinicians in a timely manner in the form of easy to read graphs and results.
During a trial at London's Royal Free Hospital, doctors and nurses received warning signals via a mobile phone app in an average of 14 minutes, when patients' blood tests indicated the condition. Normally, this would have taken several hours.
One of the blood tests looks for high levels of creatinine, which is normally filtered out by the kidneys. Information on other blood markers which can help treat patients is also made available quickly to specialists via the app.
Hospital managers said there had been a knock-on reduction in the cost of treatment. Mary Emerson, lead nurse specialist at the Royal Free, told the BBC the system had made a big difference to her job, and was the first that fits in with the way clinicians work. "It's a huge change to be able to receive alerts about patients anywhere in the hospital," she said. "Healthcare is mobile and real time, and this is the first device that has enabled me to see results in a mobile real-time way."
Peer review of Streams, by doctors from University College London, in the Journal of Medical Internet Research** concluded: "Digital technologies allow early detection of adverse events and of patients at risk of deterioration, with the potential to improve outcomes. They may also increase the efficiency of health care professionals’ working practices. However, when planning and implementing digital information innovations in health care, the following factors should also be considered: the provision of clinical training to effectively manage early detection, resources to cope with additional workload, support to manage perceived information overload, and the optimization of algorithms to minimize unnecessary alerts."
*Tomašev N, Glorot X, Mohamed S. A clinically applicable approach to continuous prediction of future acute kidney injury. Nature 572, 116-119, 31 July 2019. DOI:10.1038/s41586-019-1390-1
**Connell A, Black G, Montgomery H, et al. Implementation of a Digitally Enabled Care Pathway (Part 2): Qualitative Analysis of Experiences of Healthcare Professionals. J Med Internet Res 2019;21(7):e13143. DOI: 10.2196/13143