Silver Circle Profiles
A Look at Pat Cerrito
posted: August 2006Pat Cerrito's original plan was to make a career out of studying theoretical probability. But an introduction to statistics and SAS software sent her down a different path, and today the math professor is helping the University of Louisville's Medical School solve real-world health problems.
Cerrito first began working with SAS in 1974 after a professor mentor introduced her to the software. Several years later, she was teaching at the University of Louisville when she got a request to help some medical college researchers.
"There was no statistical support down at the medical school," Cerrito says. This was 1989 and she was using an early version of SAS, SAS/Stat™. Cerrito's help quickly made her the go-to person for other projects. With the introduction of the SAS data mining tool Enterprise Miner™, she could do even more. In the past 17 years, she has helped college researchers study everything from drug interactions to physician decision making. Her focus now is cost and quality as it relates to physician decision making.
SAS® hits home
One recent study examined the amount of the anti-clotting drug heparin that should be administered during angioplasty procedures. Physicians were deciding independently how much of the drug to give – the guidelines allowed for considerable variability. Cerrito helped researchers discover that the higher the dose, the more likely a patient would experience complications and require an extended hospital stay. She is currently helping researchers study antibiotic usage and its connection to a drug-resistant form of bacteria that frequently requires amputation. This last problem is particularly close to Cerrito's heart – her husband contracted the bacteria during a hospital stay and has struggled to regain his health. Her work has attracted the attention of the National Science Foundation and the National Institutes of Health, which has provided some funding.
Cerrito innovates, introducing marketing and retail methods to clinical researchers. One recent adaptation involved applying to healthcare the "market basket" studies that help online book sellers and movie rental firms suggest new products to customers based on what they purchased or rented in the past. She wants to help physicians use customer relationship management tools not to sell patients more medicine (like a retailer would sell books) but to manage their health better. "This is an area that is almost totally unexplored in the medical literature," Cerrito says.
What she is doing is looking for connections. Like the bookseller that suggests new titles in historical fiction to buyers who have bought other books of that genre, Cerrito wants to see the connections between different medications and what they suggest about a patient's diagnosis, possible drug interactions and future medical problems. Diabetics often have heart trouble. But are certain diabetes drugs more likely to trigger heart problems? Or is heart trouble more often a result of blood sugar levels being out of control for long periods of time? If you mined volumes of data could you see the point where diabetic patients develop heart problems and find the point to intervene to forestall greater damage? All of these questions require the ability to manipulate a set of data in a complicated way – one Cerrito says she couldn't do without SAS.
Cerrito gets better, more reliable results with SAS
SAS can crunch numbers in a manner that other statistical programs can't. For instance, Cerrito notes that some statistical analysis programs will churn out different results depending on what order the information was entered. SAS is order-neutral, leading to better, more reliable results.
What Cerrito has enjoyed most about working with SAS over the past 26 years is the improvements that have allowed her to share her knowledge more readily with students. Starting with the introduction of SAS 6 and continuing through the debut of SAS Enterprise Guide, "the improvements have allowed me to radically change the content of my courses as well as the manner in which they're taught. Students are introduced to SAS on day one, with the focus on learning concepts and interpreting results." Tests are a thing of the past. Instead, students work with data sets and write reports. This year, for the first time, the university is offering a data mining certificate program online. Students have published their work, derived with numbers crunched using SAS.
Cerrito often helps students with statistics-based projects, whether or not the students study directly with her. Robert Topp, a professor in the nursing school, sent Cerrito a student who wanted to study the impact of resistance training on breast cancer recovery. She sent Topp a student who looked at peak usage times at the emergency room in an effort to use staff more efficiently and reduce wait times. The student discovered that ER usage peaked at 3 p.m., which happened to coincide with shift change. As part of a dissertation, the student suggested moving employees to 12-hour shifts – an idea that hospital administrators are considering. "The only way administrators listen is if you've got numbers to back up your idea," Topp says. Using SAS, Cerrito can provide all the support her colleagues and students need.