Susan Dorsey

Getting Big Data to Get Along

‌Our Researcher: Susan Dorsey, PhD, RN, FAAN

By Dan Mezibov
(from the fall 2018 issue of Nursing For/um magazine)

conceptual illustration of figures working with dataDo researchers all speak the same scientific language? Not necessarily, especially when they collect data differently.

The various ways researchers define demographic data (such as age, ethnicity, and marital status), symptom and self-management data (such as intensity of pain), and other patient characteristics can hinder investigators’ ability to test hypotheses across studies conducted at different sites. Too often, small sample sizes from single sites and other limitations make it difficult to draw meaningful conclusions.

"You’re not comparing equal measures," says Susan Dorsey, PhD ’01, MS ’98, RN, FAAN, professor and chair of UMSON’s Department of Pain and Translational Symptom Science. What’s more, the average large study requires 12-18 months to secure funding and another five years to complete, Dorsey explains. With the use of a common set of measures that have been identified up front combined with biospecimens banked across many studies, "it is possible to get an answer to my question in one year instead of eight, at much less expense," she says.

Inside the Research

When data sharing and analysis cannot take place across research sites, "the impact of our science is weakened," Dorsey explains. Analyzing data for a collaborative study with colleagues in Europe, Dorsey and her team searched for other studies that measured the symptoms of peripheral nerve damage related to chemotherapy in the same way, but, she says, "we were unable to find a replication cohort, and the work stagnated."

To help develop such leverage, Dorsey and 13 co-authors published a landmark study in the Journal of Nursing Scholarship in 2017 describing a collaboration of five schools of nursing – Case Western Reserve University, Duke University, Emory University, the University of Maryland, and the University of Washington – aimed at examining whether data collected at different sites without the use of common data elements (CDEs) could be combined. This study provided further support for the idea that CDEs need to be identified up front and that trying to combine results across studies upon their completion does not work. The study was one of three important articles by Dorsey and colleagues that began a national conversation about the need to collect CDEs.

That discussion was initiated, in part, in 2014 when Patricia A. Grady, PhD, MS ’68, RN, FAAN, one of Dorsey’s co-authors and director of the National Institute for Nursing Research (NINR) from 1995 until her retirement this summer, convened a meeting of the NINR Center directors to identify CDEs to be collected in studies of behavioral symptoms related to sleep disturbance, fatigue, pain, cognition, and depression and for evaluating patients’ self-management of their symptoms. NINR has expanded the effort by establishing its own CDE repository.

Nursing didn’t invent CDEs, Dorsey says, but was "very early to adopt CDEs associated with symptom and self-management science. We’re leading in that effort."

To move the science further, Dorsey, Grady, and their co-authors also have published a "minimum set" of biomarkers to use as CDEs when studying the underlying biological basis for pain and other symptoms. Moreover, in a separate commentary, Dorsey and colleagues outlined how health methodologies such as genomics and other omics disciplines can be targeted to manage chronic pain more effectively, especially in older adults.

In 2009, a $2.4 million NINR Center grant helped establish the interdisciplinary University of Maryland, Baltimore (UMB) Center to Advance Chronic Pain Research, which aims to conduct translational research on a variety of chronic pain conditions using molecular, cellular, physiological, imaging, and omics methods. Dorsey and Joel D. Greenspan, PhD, professor in the University of Maryland School of Dentistry, co-direct the interdisciplinary effort, which joins researchers from the University of Maryland schools of Nursing, Medicine, and Dentistry and the University of Maryland’s Marlene and Stewart Greenebaum Cancer Center. Dorsey is also a multiple principal investigator for the Omics Associated with Self-Management Intervention for Symptoms Center at UMSON, which is using CDEs to determine how individual differences influence the effectiveness of exercise on managing chronic pain.

"We know that exercise can reduce or even ameliorate pain, but we also know it doesn’t work for everybody. And we don’t know the dose, intensity, or timing of what would work for some," Dorsey explains.

Her goal, she says, is to tailor precise interventions to individual patients by developing a "biosignature" that identifies which self-management strategies eliminate or reduce burdensome symptoms from disease or treatment. "Right now," Dorsey points out, "it’s pretty much one size fits most."