New Technologies Detect Kidney Irregularities

Two companies are offering products that use artificial intelligence (AI) to help improve kidney conditions.

AI start-up Medial EarlySign (Kfar Malal, Israel) has shown how the combination of AI and electronic health record (EHR) data can facilitate early detection and treatment of kidney problems and can help slow or even help prevent progression to end stage renal disease. AI refers to the concept of machines being able to carry out tasks in a manner considered to be “smart, while Machine Learning is an application of AI based on the idea that we should give machines access to data and let them learn for themselves,” notes Forbes magazine contributor Bernard Marr.

Medial EarlySign’s machine learning–based model analyzes information available in EHRs to predict which patients may be at high risk of experiencing renal dysfunction in the coming year. The technology assesses a combination of laboratory test results, demographics, medications, diagnostic codes, and other factors to predict the risk of renal dysfunction within the next year.

By isolating less than 5% of the 400,000 patients with diabetes selected among the company’s database of 15 million patients, the algorithm was able to identify 45% of patients who would progress to significant kidney damage within a year, prior to their becoming symptomatic. This represents 25% more patients than would have been identified by commonly used clinical tools and judgment, the company says.

The AI firm DeepMind (London, UK) has been expanding into healthcare. The Google-owned company’s app, Streams, is being used in the Royal Free, a London teaching hospital. One of Streams’ first uses is to rapidly alert clinicians to potential cases of acute kidney injury (AKI) in patients, reports ZDNet.The company noted that the Streams app is not strictly AI. “The more time we spent with the clinicians at the Royal Free, the more it became obvious that … their core challenge was in how you actually implement an algorithm to change the way care is delivered in practice,” not necessarily the most perfected algorithm.

Streams allows AKI to be detected in several hours, rather than a day or two, ZDNet wrote. A low hemoglobin and elevated urea might point to blood loss, while an elevated white cell count might result from an infection.

DeepMind hopes that in the future the Streams app could be used to study the performance of clinical teams—recording how long it takes to respond to an AKI alert, for example—and patient outcomes related to certain clinical activities.

April 2018 (Vol. 10, Number 4)