Skip navigation
AI in health care .jpg Getty Images

AI in Health Care: It’s Coming, One Disease at a Time

A number of startups are focused on AI in health care, working on technology to detect conditions and diseases and predict their correlation with environmental conditions.

This month at the giant HIMSS 2020 conference, hundreds of technology vendors would have been in Florida discussing new ways to use AI in health care to help patients, providers and payers.

That conference, along with just about everything else in daily life, was canceled due to the coronavirus. But ITPro Today was able to catch up with a few of those vendors in interviews.

The common thread emerging from those discussions is that startups are creating AI and machine learning models to enable doctors to do early detection of many conditions and diseases, including cardiac arrythmia, colon cancer, diabetes and the flu, and also predict the correlations between environmental conditions and certain health issues.

The hope is that early detection of at-risk patients can lead to intervention by health care representatives and ultimately cut back or eliminate costly hospitalizations or readmissions.

Social Networking Factors

Jvion is an Atlanta-based startup vendor that is working on prescriptive analytics for hospital or clinical settings to identify factors that would contribute to patients being readmitted or patients coming down with sepsis or blood poisoning.

It is now attacking the emerging field of “social determinants of health” (SDOH), which include factors such as education, housing, ethnicity, gender, occupation, air quality, food security and many others. The company applies AI in health care to create models to identify the socioeconomic factors that lead to specific health risks.

“Our core offering [Jvion Clinical AI] is the clinical AI solution focused on preventable harm,” said Dr. John Showalter, Jvion’s chief product officer. “We take claims data, EHR data, census tract [and] government data and provide insights.”

The new SDOH product enables health care professionals to identify groups of at-risk patients, even if their underlying medical conditions are unknown. Working out of the Microsoft cloud and using Azure Maps, Jvion is able to create visualizations of social determinants on a community level (see screen shot). 


Jvion uses predictive analytics to identify factors that contribute to the need for health care services, creating visualizations of social determinants on a community level.

“By focusing on just the social determinants of health in this particular solution, we’re able to understand who is at highest risk because of those characteristics, as opposed to who is at highest risk because they have the worst disease,” said Showalter.

Getting the Early Signs

Medical EarlySign, based in Israel and Aurora, Colorado, already has had success in early detection of colorectal cancers by employing AI in health care. The company recently announced successful results of a study assessing the effectiveness of another product, its Pre2D-Flag model to identify individuals at high risk of progression from prediabetes to diabetes.

The study represents just one of several diabetes-related models the company is working on to alleviate the human and financial burden of diabetes over the long term by identifying high-risk patients who could benefit from early intervention strategies, said Dr. Jeremy Orr, CEO of Medical EarlySign.

“There is a whole suite of diabetes predictors. It starts with the prediabetes,” explained Orr. “The next one is looking at complications of patients that already have diabetes, such as chronic kidney disease and heart disease.”

EarlySign works with leading U.S. health systems, including Kaiser Permanente and Geisinger Health, to develop their models. EarlySign recently announced a partnership with Maccabi Healthcare Services of Israel on a flu vaccination campaign that targets patients who are at risk of complications from the flu.

“We are predicting compliance [for getting a flu shot],” Orr said, and working to identify patients who are unlikely to get a flu shot even though they are at high risk for complications. “For those people, they might actually send a nurse to their home and give them a shot. No doubt everybody should get a flu shot, but we’re giving them also extra ammunition to convince the patient, and it's working.”

Cardio Logging

One more startup using AI in health care is Paris- and Boston-based Cardiologs, another Azure customer that is applying deep learning neural networks to electrocardiograms (ECGs) to help cardiologists speed diagnoses.

The company, which just announced a new $15 million funding round, started detecting atrial fibrillation while using a Holter monitoring device and now can examine 15 types of arrythmia, said CEO and Co-Founder Yann Fleureau.

While Jvion and EarlySign are used to predict risk or complications, the main benefits of Cardiologs’ tech are accuracy and speed — detection that is up to five times faster compared with traditional methods, Fleureau said, along with fewer false positives.

Analysis can be done in five minutes instead of 25, a difference that amounts to huge time savings when doctors have to go through hundreds of scans. “The real point is to save all the time of the expert to really focus on treating the patient and removing the mundane part of the work,” he said.

Going With the Workflow

Projects incorporating AI in health care are popping up all over and showing some success. Yet, for every disease that can be spotted more easily, it’s one more piece of software that has to be integrated into clinical workflows. That is just the nature of the business around developing machine learning models, said Fleureau.

“The technology is super vertical, which means that you train the model to a specific task, with specific data sets [and] specific domain expertise,” he said. “The product that you create is very specialized, clinically speaking. And then you build a sales force, which is also specialized on a certain type of customer.”

Providers and patients shouldn’t soon expect any all-in-one AI solution that can take care of every aspect of health care, though Fleureau is looking to electronic medical record vendors to help with the integration piece.

“EMR integration is key. … It's fine for the hospital to have different vendors. But they want to have a unified workflow so they don't have to think about the login password … for every single application.” 

Scot Petersen is a technology analyst at Ziff Brothers Investments, a private investment firm. He has an extensive background in the technology field. Prior to joining Ziff Brothers, Scot was the editorial director, Business Applications & Architecture, at TechTarget. Before that, he was the director, Editorial Operations, at Ziff Davis Enterprise. While at Ziff Davis Media, he was a writer and editor at eWEEK. No investment advice is offered in his blog. All duties are disclaimed. Scot works for a private investment firm, which may at any time invest in companies whose products are discussed in this blog, and no disclosure of securities transactions will be made.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.