Electronic medical records have been around decades. Their power to help other patients is starting to be unleashed.

For decades, electronic medical records have offered the potential for making one patient's experience useful to the next.

Such real-world experience could offer insights into which type of person does best on which medication, how best to sequence different drugs or how to add up unusual symptoms into a diagnosis.

But unless it was part of a formal research trial, or passed down from one doctor to the next, that type of information hasn't been available.

"It's really frustrating to imagine we're this far along in our technology and we're still asking really basic questions like what works for whom and at what cost," said Kristin Kostka, a computational epidemiologist at Northeastern University.

Now, several groups are making the kinds of advances that could finally ensure learning from one patient isn't lost.

One such system, called Observational Health Data Sciences and Informatics, or OHDSI – pronounced “Odyssey" – includes 300 databases and over 800 million patient records, analyzed by researchers like Kostka from industry, academia and government.

At Stanford University School of Medicine, researchers recently developed a system to mine medical records for insights that won't reveal personal data but will describe the course of treatment and response for a group of similar patients.

Their Clinical Informatics Consult Service, now provided by spinoff company Atropos Health, includes records from Stanford hospital and national collaborators. They can also help a hospital or network set up a similar system at their own institution – amassing thousands if not millions of patient records as well as the possibility of looking through purely local data.

"We wanted to bring this leverage to every single patient encounter," said Dr. Saurabh Gombar, an adjunct professor at the Stanford School of Medicine, who helped develop the new system.

He and the rest of the team wanted their system to work quickly enough to be useful in real-time for doctor and patient facing a crucial decision.

It is helping in a growing number of cases, he said.

For instance, when a pediatrician worried a child with an abnormal constellation of symptoms might be developing multiple sclerosis, the Stanford service showed in a single day the symptoms did not indicate a heightened risk for MS.

"We were able to help a child and their family get a more resolute answer on what at the outset could have been a very scary diagnosis," Gombar said.

An elderly woman with a recurrent rare cancer was able to learn quickly from the Clinical Informatics Consult Service that a radical surgery another doctor had suggested would not likely extend her life, Gombar said.

"It made us feel great that we built a system that really helped individuals make informed decisions," he said.

The service has also been used to change institutional policy. Stanford wondered whether it really needed to give patients with ulcers a four-drug combination of antibiotics, or whether a three-drug regimen would work just as well. Looking at thousands of patients revealed that three was typically enough.

And it's speeding up research studies, allowing doctors to answer questions quickly and with confidence, Gombar said.

Such real-world evidence can add to information from clinical trials, because "it reflects the true variability of patients' lives and experiences," said James Weaver, an epidemiologist who collaborates with the OHDSI community.

The most complicated patients are often excluded from clinical trials, but they're the ones who face the biggest risk from the wrong medication and where doctors need the most guidance, Gombar said. Plus, clinical trials often run for a limited time, while real-world evidence can often provide longer-term results, he said.

Slow, incremental change

No single technological advance is making OHDSI's system and Stanford's service possible, experts said.

These systems are the result of medical and computer scientists eager to get more out of digital medical records, a growing acceptance that such records contain useful information, and a willingness to collaborate to produce a comprehensive understanding of health and disease.

"It's becoming more accepted as the science of real-world data and real-world evidence advances," said Weaver, associate director of observational health data analytics at Janssen Research & Development, a Johnson & Johnson company. "People unfamiliar with the area are beginning to realize its potential and trust the evidence."

Federal funding from National Institutes of Health is helping to spur data sharing and best practices, Kostka said.

"No one alone is going to solve these problems," she said.

Electronic medical records have long been ignored as a source for this kind of data, Gombar said, because they are messy, with coding aimed at obtaining insurance reimbursement rather than tracking patient outcome.

"That inserts noise into the system," he said.

But by compiling data from hundreds of thousands or even tens of millions of patients, clear signals can be separated from that noise, Gombar said, and the information becomes extremely useful.

Atropos Health's Clinical Informatics Consult Service is designed to rapidly clean up data, making it easy to use at other institutions as well, he said.

To confirm the validity of the service, the Stanford team looked in its database for bad drug side effects and showed that their findings were similar to ones from clinical trials.

Training, Gombar said, will be key to using the system to best effect. Users have to be careful to weed out spurious connections.

There should always be a cautionary note about technology, but a well-designed system can protect against human bias, said Dr. Cesar Padilla, an obstetric anesthesiologist at Stanford, who wasn't involved in developing his institution's new service.

At some point, doctors will use systems like Stanford's to predict which pregnant people will fall ill, monitor pregnancies at risk and determine the best approach and care system for preventing a bad outcome.

"It's going to change the way we practice in the future," Padilla said.

Identifiable factors such as asthma, autoimmune disease and high blood pressure can be used to predict people at higher risk, for instance, but are too rarely used today to do that, he said and wrote in a recent article.

The Stanford program is a model of what can become possible, Kostka said: "What they're building is a really nice workbench to create that evidence set."

Patient perspective

Weaver's personal situation proves how useful systems like this can be.

He had been diagnosed with advanced melanoma, a lethal form of skin cancer. Earlier this year, he endured surgery and then radiation to eliminate tumors that had spread to his brain.

With two young children, ages 4 and 9, Weaver wanted the most effective treatment possible, but he also wanted to avoid another brain surgery.

"It was very important to me to not do that again," he said.

Often, a patient with advanced melanoma like his is given a regimen of two immunotherapy drugs, ipilimumab and nivolumab. But some evidence suggests that ipilimumab can aggravate surgical and radiation sites, which would potentially lead him to another round of surgery.

Weaver wanted to figure out whether it would be OK to take just the second drug. Clinical trials focus on people earlier in the course of their disease, not post-surgery and radiation, so they weren't much help.

Instead, he dug into OHDSI databases.

In about 30 hours, he was able to scan 300 million patient records. The search turned up no one exactly like him, but suggested that a few people sort of like him had done well on nivolumab alone.

Combined with information from other sources, Weaver and his oncologist decided to skip ipilimumab. Five months later, he feels good – and good about his choice.

"My energy and strength are increasing," he told a group of colleagues at a recent virtual scientific conference. He is gaining weight and hasn't needed another round of neurosurgery: "I can participate in most of the important areas of my life again."

He's back to work, horsing around with his kids, and has started a long-deferred doctoral program.

The future of medical care, Weaver hopes, will be one where every clinical decision can be made as his was: confidently and "directly informed by all available evidence."

Contact Karen Weintraub at kweintraub@usatoday.com.

Health and patient safety coverage at USA TODAY is made possible in part by a grant from the Masimo Foundation for Ethics, Innovation and Competition in Healthcare. The Masimo Foundation does not provide editorial input.

This article originally appeared on USA TODAY: Electronic medical record systems advances show power to help patients