Digital Life Science

How a Database May Demystify a Deadly Children’s Disease

PALISI meeting in session.

The Nathaniel Adamczyk Foundation’s first PALISI (Pediatric Acute Lung Injury & Sepsis Investigators) meeting in 2010.

Iris Melendez was pregnant with her second child in 2004 when her healthy two-year old son Nathaniel developed a mild cold and a fever. A doctor diagnosed mononucleosis, but 48 hours later the boy had pneumonia, a collapsed lung, and septic shock. Two weeks later he was dead. The cause? Complications of Acute Respiratory Distress Syndrome (ARDS), a sudden condition that typically begins with a lung infection and affects 10,000 or more children a year in the U.S. At least 30 percent of the time, it’s fatal.

In agonizing grief, but with another baby due in 8 weeks, Melendez and her husband Hank Adamczyk began researching ARDS. They wanted to understand why Nathaniel had contracted it, and were desperate to know how to prevent their next child from falling victim to it.

Meanwhile, the mourning couple found themselves in the position of being ARDS fundraisers. “We were overwhelmed by an outpouring of support from friends and family. Donations started pouring in, and we raised a tremendous amount of money without even trying,” says Melendez, a Wharton business school graduate who is a senior manager at a Fortune 50 company. She became president of The Nathaniel Adamczyk Foundation and began exploring how the money could spare other families from the same nightmare.

Researchers believe pediatric ARDS is caused by a combination of factors including genetics, lifestyle, and even environmental exposure during pregnancy. But few clinical tools can control or prevent it, and no effective drug exists. And, as Melendez discovered, “there is very little written about what predisposes you to this.” All of the research she found focused on treating ARDS patients already in the ICU.

“Given how fast it happens, we decided the biggest bang for the buck would be in figuring out what the early warning signs are so children don’t become critically ill,” Melendez says.

She and her board first thought they would fund medical researchers investigating ARDS risk factors. But the more Melendez learned about the competitive nature of medical research, the less interested she became in funding it. The way she saw it, most scientists researching ARDS were hoarding data and resources while competing to get published. “They’re not sitting down and saying, ‘Let’s put our collective energy into the things that are the most promising,’” Melendez says.

She and the board came to believe that there was a more fundamental need they could address: data collection. Melendez credits foundation board member Michael Liebman, a computational biologist and medical data-modeling expert, with the idea to establish a national ARDS repository for data and tissue for use by researchers.

ARDS, Liebman notes, is characterized by as many as 10 unique symptoms. A patient with any 5 of them could be diagnosed with ARDS. “Two patients can have no matching characteristics but have the same diagnosis,” he explains. Liebman says that one key to unraveling the risk factors for ARDS is to stratify, or define the subcategories for, every variety of patient based on specific symptoms and other characteristics such as age, gender, and secondary conditions. What’s presently known as ARDS might, in fact, be several distinct diseases with different causes.

So the Nathaniel Adamczyk Foundation initially undertook a massive data-gathering effort. In 5 years, it has developed access for more than a half million patients by working with 110 pediatric ICUs across the U.S. “There are all kinds of measurements done on these patients to come up with the diagnosis,” says Liebman, who spent several hours in an intensive care unit to observe the process. They tracked as many of them as possible. Already, through analyzing various subgroups in the database, the foundation has shown that data as simple as gender or age can influence a patient’s response to treatment.

Liebman says the project now seeks to collect additional background information that isn’t typically requested in the pediatric ICU. For instance, he says, “Some of these patients appeared in neonatal ICUs before they showed up in the pediatric ICU, but those two don’t share data.”

The foundation aims to raise another $250,000 to build a registry of information from parents about the pregnancy, drugs, exposures, infections, diet, and health care experiences to investigate whether any of these pose a potential risk factor.

“Instead of promising, like most disease foundations, that in 10 years we’ll have a diagnostic or therapeutic, this year we will help patient management and manage survival characteristics, and that will help lead to what we want: new diagnostics and therapeutics,” Liebman says.

What’s more, he says, a similar data collection and analysis approach could be useful to treating many other diseases with diverse sets of symptoms. Cross-tabulating such massive banks of data to track the symptoms and response to treatment of specific subgroups of patients could extract valuable clues about diseases. “Autism, ADHD, and diabetes are all conditions that represent a complex set of diseases that are lumped together, so I’m interested in how we can use data to develop the right approaches to improve the way we diagnose and manage patients, which then can lead to better drug development,” Liebman says.

Note: Adrienne Burke was awarded first place in the Connecticut Press Awards in the Science/Technology Specialty Articles category for this article and its companion piece, “Startup’s Data Helps Women Succeed With In Vitro Fertilization.”

Tags: , , , , , ,