Please consider downloading the latest version of Internet Explorer
to experience this site as intended.

Promising solutions to tough medical problems win University research competitions

October 15, 2019
Jessica Goodman and Alec Salminen each present at University research competitions.Jessica Goodman, an integrated behavioral health clinician, researcher and educator in psychiatry and medicine, and Alec Salminen, a PhD student in biomedical engineering, took first place at two recent University of Rochester research competitions. (University photo / Bob Marcotte)

The rate of emergency department visits in the United States is increasing so fast that the Institute of Medicine declares the system is “at a breaking point.”

At the same time, difficulties in the early detection of sepsis make it one of the most expensive diseases to treat globally and the leading cause of death in intensive care units.

So, how should we deal with these two problems?

Jessica Goodman, an integrated behavioral health clinician, researcher, and educator in psychiatry and medicine, and Alec Salminen, a PhD student in biomedical engineering, took first place at two recent University of Rochester research competitions with their answers. Both competitions limit contestants to brief presentations and reward those who can best explain their research to a non-scientific audience.

Predicting emergency visits

“We have been defining ED use based on number of visits alone for over 20 years,” said Goodman, who won the University’s 2019 Steadman Family Postdoctoral Associate Prize in Interdisciplinary Research, worth $1,000.

She described how her team uses data science to also take into account the unique characteristics of emergency department (ED) patients.

The team applied machine learning algorithms to a population health data set of all emergency department visits for the State of Florida for 2009 through 2015. Though the algorithms could not identify patients based on number of ED visits alone, patient groups that were similar within-group and dissimilar between groups could be created using a combination of payer type (e.g., type of insurance or no insurance at all), diagnostic category (e.g., mental health or organ disorders), and number of ED visits.

“So, the message is that the number of ED visits does matter, but only in the context of a broader, systemic look at patients in order to understand who these people are and what they need,” Goodman said. “We can now proceed to see if we can use these algorithms to actually predict patient use so that at the point of care we will be able to identify what group this patient would fall into.”

Bedside sepsis detection

Salminen, who works in the lab of James McGrath, professor of biomedical engineering, took first place and a prize of $500 at the University’s Falling Walls competition. He will represent the University next month in the final Falling Walls Lab competition in Berlin, against 99 other presenters from across the globe.

Salminen described how the blood-brain barrier, a strict regulator of transport into and out of the brain, is disrupted early in the onset of sepsis, leading to delirium or drastic changes in cognitive function.

“Unfortunately, diagnosing sepsis-associated delirium is subjective and often not viable in many clinical settings,” he said.

He is working on a device that recreates the complex brain microenvironment on an inexpensive microfluidic chip. Combined with batch processing and high content imaging, this system would allow early sepsis detection with a simple blood draw at the bedside “permitting corrective patient intervention before the disease progresses into the devastating multiorgan failure associated with the disease,” Salminen said.

Salminen is the second member of the McGrath lab to win this competition. Kilean Lucas took first place two years ago.

Tags: , , , , ,

Category: Science & Technology