Making Data Speak The Same Language:
Alumna Rebecca Kush's Clinical Data Interchange Standards Consortium sets the standard
By Leslie Linthicum
Scientists searching for a better treatment for Parkinson's or a cure for diabetes might spend years and millions of dollars in clinical research studies.
When it comes to reporting their data, which might hold precious clues for other researchers on similar paths, one research team might code the sex of study participants as male or female. Another might choose M or F. Someone else may assign males the number 1 and females the number 2. Someone else may have no data sets for "sex" because they classified that information as "gender."
To a computer trying to search that data for meaningful patterns that might advance medical research, it's as if the studies are speaking different languages—or even talking gibberish. Compound that by different ways to describe patients' race, how they responded to a therapy, whether their side effects were serious, mild or non-existent, and much of the information gleaned from clinical studies is useless to others.
Worse than comparing apples and oranges, the system of reporting clinical trial results throws figs, bananas and persimmons into a confusing fruit basket of data.
Rebecca Kush (BS '75) has devoted two decades of her professional career to advocating for data standards that can be easily interpreted and shared. It's more than just for the sake of convenience, she says; the stakes are enormous.
Locked in all of those separate data sets could be a promising therapy to slow memory loss in Alzheimer's, a better treatment for asthma or even a cure for cancer.
"Every piece of data is precious," says Kush, who is the president and CEO of the Austin-based Clinical Data Interchange Standards Consortium, known as CDISC. "You can't waste data, and you do waste data when it is formatted and reported in ways that are not sharable."
CDISC, a non-profit that Kush founded in 1997, has advocated for a common research reporting language not just for American medical research but globally.
At the end of 2014, CDISC achieved a breakthrough. The U.S. Food and Drug Administration adopted rules to take effect at the beginning of this year requiring that data be submitted in standardized formats that will be available electronically. Last spring, Japan followed suit.
"It is a big step forward," Kush says. And it was, for her, a major milestone in a long battle.
"So many people told me this couldn't be done," she says. "All along the way."
Her next goal is to persuade the National Institutes of Health to follow similar requirements.
Eliminating the hodgepodge of formats among the thousands of clinical trials investigating therapies that are ongoing at any time, Kush says, can help science get closer to effective treatments and cures.
If data doesn't speak the same language, she says, a clinical trial in Hong Kong could be repeating the same mistakes and reaching the same dead end that a clinical trial in Kansas already did. That wastes precious time, she says, and it also puts patients through unnecessary trials.
And miscommunicating data can also fail to connect successes; that researcher in Kansas might find data in a study in Germany that could spark a breakthrough.
"When patients participate in clinical research to help collect data they do it to help themselves and for the good of others. And they expect their data to be shared and used," Kush says. "They don't expect it to get siloed and not used going forward."
For people outside the world of medical research and clinical trials, it might come as a surprise that there have been no standards for collecting, formatting and sharing research results. In fact, it was when Kush was working in pharmaceutics and someone asked that innocent question—"Don't you guys have any kind of standards in your industry?"—that the germ of the idea that would become CDISC started to form.
Kush, who grew up in Los Alamos (she was Rebecca Daniels when she graduated from Los Alamos High School in 1971), comes from a Lobo family. "Four kids and my dad all graduated from UNM," she says.
Kush credits the Honors College for helping her see different ways to learn, how to think outside the box and to think big. She went through UNM on some scholarships and money she made babysitting, and she recently endowed a scholarship to benefit a Native American student attending the Honors College.
Her degree was in biology and chemistry and, although her father, a research chemist at Los Alamos National Laboratory, thought she should go to medical school and become an allergist, her path was different.
She did attend the University of California, San Diego, med school, but for her doctorate, not her M.D. "At that point in time," she says, "life and death was something I wasn't ready for."
She did clinical research in diabetes among the Pima Indians for the NIH in Phoenix and when her husband got a job that took them to Austin, she did a postdoc in immunology at the University of Texas and went to work for a drug testing company. Her husband's company, Schlumberger, transferred him to Japan and then Paris and then back to Austin in 1988.
Meanwhile, Kush ran a lab that did nutritional analysis and completed a second postdoc, but was at loose ends.
By then, she was ready to become a physician, but UT didn't have a medical school at the time.
Answering a newspaper ad, she applied to be a phlebotomist at a clinical research organization, moved up the ranks at the company and began to understand the data collection process and start to see its inefficiencies. After hearing that question—"Don't you guys have any standards in your industry?"—and having to answer, "No, we don't," Kush started organizing what would become CDISC. The organization has grown into a global consortium with coordinating committees in Asia and Europe, 35 full-time employees and a $7 million annual budget.
The consortium's motto is "Strength through collaboration" and Kush, 64, hopes for a future in which science is open, data is shared and data sets all speak to each other in a shared language.
"You have to a have a vision of what you're building and where you can go with it," she says. "Or else you would get frustrated and quit."