© 2023 WRVO Public Media
Your Source for NPR News
Play Live Radio
Next Up:
Available On Air Stations

Latest in health: algorithms for diagnosing disease, the health lab of the future


For today's latest in health segment, we look at solutions for two different kinds of patients -- one for those with genetic diseases and another for healthy people looking to stay that way.

First, we start with a new algorithm that could improve the diagnosis of rare diseases. Second, we look at a New York City lab that educates patients on making healthy lifestyle changes.

Using computers to find important mutations

Rare diseases can be marked by a genetic mutation, but when it comes to diagnosis, finding that gene and ruling out others could take clinicians hours -- potentially an entire week of work. But a team of researchers at Stanford is trying to change that.

Gill Bejerano is an associate professor at Stanford University School of Medicine who teaches in the fields of developmental biology, computer science, biomedical data science and pediatric genetics. Earlier this year, in a paper published in Genetics in Medicine, his team revealed an algorithm that could improve our ability to diagnose rare diseases.

Bejerano described how, currently, the process is long and arduous to send DNA out to a lab and analyze the results.

“You send them out to sequencing. The sequencing comes back. You analyze it really carefully, and then, you’re faced with, let’s say, 200 genes, any one of which could have caused the person’s disease,” Bejerano said. “There are medical professionals -- clinicians -- who sit down and essentially go gene by gene over these 200 in the hope of finding the one gene that can explain the patient’s disease.”

"Our basic observation behind this was: clinician time is expensive. Computer time is cheap once we get the computer to do absolutely as much as it can before the clinician comes on the scene."

That is where the algorithm can help the most, as well as making the process cheaper, Bejerano said.

“Our basic observation behind this was…clinician time is expensive,” Bejerano said. “Computer time is cheap once we get the computer to do absolutely as much as it can before the clinician comes on the scene.”

One key thing to note, Bejerano said, is that they are not interested in replacing the clinician, but setting them up on better footing to make decisions that only a human can -- does this person have this particular rare disease and is this the gene mutation that caused it?

“The idea is that when the clinician shows up on the scene, they go down the list the way the computer’s organized the list for them and hopefully immediately, at the first gene, they just look at it and say, ‘Yep, everything looks right. This is the diagnosis,’” Bejerano said. “The whole premise there is make clinicians work more with patients by doing as much as possible with the computer before the clinician starts to look at the case.”

The algorithm combs through a lot of information to make these lists: The patient’s genes, phenotypes -- essentially the signs and symptoms the patient is experiencing -- and finally, a body of peer-reviewed papers.

“You can essentially go to a body of all peer-reviewed papers that were ever published…and you try to assemble a database of all genes and which phenotypes they may be causing to different diseases,” Bejerano said.

Their database includes information about which genes cause which diseases and what the symptoms are. That is the information the algorithm cycles through, trying to compare them to the patient’s actual concerns are.

The algorithm -- which is called Phrank, with a p-h for phenotype, combined with the word rank -- is a work in progress, as are many algorithms, Bejerano said.

“We want to see as many patients as possible that have already been diagnosed,” Bejerano said. “It’s the cases [doctors] diagnose that I need to put in my system. I want my system to put the causal gene that [they] found at the top.”

The 5,000 rare diseases in Phrank’s database are all caused by mutations to a single gene and have no environmental component. If a patient has that mutation in their genome, they will get the disease, but Bejerano sees an expansion from there.

“We’re going to start with rare diseases, build this wonderful, semi-automated system that help the clinicians deal with the flood of data and then expand from rare diseases into the more common ones… because right now, common diseases are somewhat mysterious in their underlying genome,” Bejerano said. “But we and the community at large are making headway in those respects, as well.”

Teaching healthy people about successful habits

The Icahn School of Medicine at Mount Sinai in New York City is home to an experimental clinic that focuses on providing patients with the health information they need to make healthy lifestyle choices.

Katharine Schwab, an associate editor at Fast Company who covers technology, design and culture, visited Mount Sinai’s Lab100, a collaboration between Mount Sinai and a design studio known for mixing digital tech with physical architecture.

Lab100 is a really interesting space, Schwab said, where patients feel “like [they are] in the future.” The clinic is meant to redesign healthcare by showing patients how their health decisions affect them in the long run.

“The goal of the lab is to help healthy people understand what the baseline of their health is…to try and connect lifestyle choices to actual health metrics that can be tracked over time,” Schwab said.

The clinic conducts many different tests that are far more extensive than those of a typical doctor visit, Schwab said, like a grip test using a virtual reality headset. In addition, physicians explain the process and the purpose to the patient before and during each procedure.

One of the more surprising results healthy, middle-age individuals who walk into the clinic may receive is that they have some inefficiencies in some areas, like sub-par balance, for which they were not tested at their primary care doctor office.

“It’s something that we don’t test for at all,” Schwab said. “It’s not something most of us probably even pay attention to.”

Schwab said that by informing patients of these issues, patients can then make the necessary lifestyle choices to improve their overall health.

“By giving people those interventions -- uncovering some of these hidden things that, because we don’t test for them, you wouldn’t know that you have a deficiency in that area -- you can hopefully keep people out of the hospital and help them stay healthier for longer,” Schwab said.

This approach, Schwab said, is rather unique because it challenges patients to think about health differently, analyzing the decisions they make every day for possible areas of improvement.

That data can help the patients as well as future studies. Schwab consented for Mount Sinai to use her anonymous lab results in other studies -- an option all patients receive at the clinic. If patients consent, their data, without their identity attached, will be used in long-term studies that allow scientists to track how people’s lifestyle changes impact their overall health. Though the program is too new to produce results yet, physicians there are hoping to understand trends in health care improvement.

“They’ve got a long way to go before they get to that part, but that is the goal,” Schwab said.