How medical databases could improve outcomes for childhood asthma and other diseases

Over the past 10 years, rates of childhood asthma in Indiana have decreased, but death rates have increased, according to a new study from the Indiana University School of Medicine.

Researchers studied childhood asthma trends by examining the health records of more than 50,000 patients from 2010 to 2019. The peer-reviewed study, published in the Journal of Asthma, did not look at the reasons behind the trends in the data — but the findings could help inform future research and public health responses that could improve outcomes for patients with asthma and other conditions.

The study was conducted using a statewide medical records database, the Indiana Network for Patient Care, which aggregates records from 95 percent of Indiana’s health care providers.

Historically, health data on issues like asthma has been collected and analyzed by federal agencies. The data can sometimes take months or years to analyze and publish, but this new state database provides information in near real time, said Dr. Colin Rogerson, research scientist at the Regenstrief Institute and assistant professor of pediatrics at the IU School of Medicine and the study’s lead author.

“One of my research goals is then to use this kind of data to test asthma treatments,” he said.

The data the researchers analyzed includes information about medications and respiratory support devices that patients used.

“Having such a large number of patients and data, we can actually test, retrospectively, treatments and see which patients respond better to different medications than others,” Rogerson said.

It’s also important to combine information from medical records with other data, he said.

”Hospital systems and record systems are talking to each other and exchanging information,” Rogerson said. “And that's allowing more and more research to be done.”

To put the health exchange to the test, researchers chose to examine trends in childhood asthma because it is a common disease and provides ample data points.

The researchers found a decrease in hospital admission rates coincided with an increase in the death rate —  5 to 7 deaths per 1,000 hospital encounters from 2016 to 2019, up from 1 to 3 per 1,000 from 2010 to 2014.

The study authors say possible contributors include worsening environmental conditions and an uptick in vaping among teens. But more research is needed to figure out how to improve health outcomes for children at highest risk of asthma-related death.

The Indiana Network for Patient Care makes it easier for medical records to follow patients around if they switch health providers. Patients do not need to take any action for their information to be included. The network was developed by the Regenstrief Institute and is managed by the Indiana Health Information Exchange. It is not available to the general public but gives researchers access to secure, aggregated, de-identified medical information from across the state, all in one place. This helps track disease trends in a timely way, to flag early signs of public health problems and inform prompt public health interventions.

Rogerson said the same method can be applied to other diseases.

“If the computer systems can share information with each other, we could potentially see that different hospital systems are seeing this pattern earlier, and then we could have public health measures to address it,” he said.

Typically, health authorities like the U.S. Census Bureau and the U.S. Centers for Disease Control and Prevention call people on the phone to collect information about prevalent health-related issues and diseases such as obesity, childhood vaccination and heart disease. But this method has been riddled with problems. For example, some people may not pick up the call, Rogerson said, and others would hang up, reluctant to share any personal information and worried it’s a phone scam.

While a medical records database like this is valuable for researchers, Rogerson said randomized clinical trials, which compare a group receiving a particular treatment to a control group, remain “the best way to prove a therapy or treatment is effective or not.”

But mining databases is both faster and cheaper than a randomized control trial.

“If there was a treatment that had been used for a year or two but hadn't had a randomized control trial for it, we could [potentially] use this kind of approach to test it retrospectively, and see if we think it's associated with a better outcome,” Rogerson said.

CORRECTION: A previous version of this story said the health information network is managed by the Regenstrief Institute. That was incorrect. The network was developed by the Regenstrief Institute and is managed by the Indiana Health Information Exchange.

This story comes from a reporting collaboration that includes the Indianapolis Recorder and Side Effects Public Media — a public health news initiative based at WFYI. Follow Farah on Twitter: @Farah_Yousrym.