It’s a common sight to see Professor Joe Cavanaugh sitting around a table with UI IPRC faculty, students and collaborators as he consults with them on their study designs, analytic approaches and methodological interests. As UI IPRC’s longstanding biostatistician, he has worked on projects and data involving numerous injury issues—from childhood trauma and intimate partner violence to teen driving and farm equipment crashes.
Cavanaugh is also Head of the UI Department of Biostatistics at the College of Public Health. He said, “The majority of research questions in public health must be addressed using data, and conclusions must be defended based on empirical evidence. Data is becoming increasingly prevalent, yet data sources are frequently large, multi-faceted, and complex. The appropriate analysis of such data often requires advanced training in biostatistics.”
The role of the biostatistician, Cavanaugh said, is to understand both the data and the context in which it arises, and to determine how the data should be appropriately analyzed and/or modeled. The biostatistician must also recognize what conclusions are supported by the data and what conclusions are not.
“In other words, the role of the biostatistician is to serve as an interpreter of the story told by the data,” he said. “If the biostatistician fails in his or her role, the study conclusions may be either erroneous, misleading, or devoid of important nuance.”
One area Cavanaugh has helped advance the injury field is through methodological discoveries. Several years ago he and his former doctoral advisee Patrick Ten Eyck worked on UI IPRC’s project to evaluate the effectiveness of Iowa’s anti-bullying law using data from the Iowa Youth Survey. They found that the variable selection criteria typically used in a ubiquitous modeling framework were fundamentally flawed. They designed an algorithm to address this problem.
“This methodological innovation served as part of the research for Patrick’s thesis, and led to a 2017 methodological publication in the journal Sankhya. The Iowa Youth Survey modeling application appears in the paper, and we used it to illustrate the methodology,” he said.
UI IPRC Director Corinne Peek-Asa said the work Cavanaugh does with the UI IPRC is critical.
She said, “The field of injury and violence compared with, for example, public health research in areas such as cancer and infectious disease, is newer, and we don’t yet have the sophistication in our statistical approaches. Joe is helping us advance methodology for the field. And he is a fabulous collaborator – he’s deeply interested in teasing out what our data can tell us while at the same time exploring innovative strategies to work with data.”
For Cavanaugh’s first decade of his academic career, he served on the faculty in the Department of Statistics at the University of Missouri, focusing on methodological and theoretical innovations. It was here, Cavanaugh said, where he began to engage in more collaborative work. His first substantive applied interdisciplinary project was a collaboration with a trauma surgeon to investigate the efficacy of the trauma system in the state of Missouri using two large state trauma databases.
“This project reminded me of what had originally inspired me to become a statistician: to use my quantitative skills to help address practical scientific problems. My quest to become more engaged in collaborative work eventually led me to the Department of Biostatistics at the University of Iowa, and my collaborative endeavor at the University of Missouri served as the impetus for my affiliation with the UI IPRC,” he said.
Cavanaugh said UI IPRC provides opportunities for a biostatistician that are both challenging and worthwhile.
He said, “The studies conducted by the UI IPRC are vast, varied, and compelling, and always address critical problems that directly impact public health. The UI IPRC is comprised of experts representing a multitude of disciplines, and the leadership is exemplary. It is a wonderful environment in which to work.”
Cavanaugh teaches courses in basic statistical methodology, categorical data analysis, and model selection. He often utilizes datasets from injury-related research projects applications to illustrate the analytical and modeling techniques covered in these courses. Thus, Cavanaugh assists with one of the UI IPRC’s major goals: to integrate its research, teaching, and outreach missions, and in particular, to expand the presence of injury topics in all types of coursework.
For example, Cavanaugh uses the example of the 2011 rhabdomyolysis (rhabdo) outbreak that resulted in the hospitalization of 13 Hawkeye football players. Cavanaugh participated in the internal investigation to examine the possible causes.
“I analyzed data provided by the football program to try and characterize risk factors associated with the outbreak. I use this data for several applications that I present to my analysis of categorical data classes,” he said.
Cavanaugh said injuries affect all individuals, regardless of age, gender, ethnicity or socio-economic status, and can result in both physical and psychological harm.
He said, “Most individuals view the injury field through a fairly myopic lens, perhaps first thinking of injuries resulting from car crashes and next considering injuries that result from violence. Obviously, this type of trauma is pervasive, and warrants attention. However, the injury field is incredibly broad.”