InsideTraCS — with Marcy Boynton

Marcella Boynton, PhD, is an assistant professor in the UNC School of Medicine and a biostatistician at NC TraCS. In this role, she supports biomedical and behavioral researchers as they conduct studies focused on improving healthcare delivery and patient outcomes—providing guidance and statistics services to translational scientists as they work through the complicated process of designing, implementing, and statistically analyzing their studies. Boynton also supports the educational efforts at NC TraCS by teaching seminars and helping to develop educational programming geared toward the translational science workforce.

Marla Broadfoot, NC TraCS science writer, recently spoke with Boynton about how psychometrics can inform clinical and translational science, her research on stressors that negatively affect health outcomes for marginalized groups, and how she mentors junior faculty in innovative research and quantitative methods.

How did you get involved in research, and what drew you to biostatistics?

I had some tremendous science teachers in high school, and one encouraged me to apply to a National Science Foundation-funded summer program at Purdue University. I worked in an entomology lab for seven weeks, and still remember the title of my project: The effect of Beauveria bassiana (Naturalis-T) on beneficial organisms in turf grass. Although I didn't ultimately join the agricultural sciences, it was a wonderful entrée into the world of research. It was very hands-on work with direct implications for mitigating the environmental impacts of traditional chemical pesticides.

Fast forward to my college days, where I became a psychology major conducting research in an animal research lab, worked as an oral surgery assistant during the summers, and studied abroad in France during my junior year. These experiences were incredibly formative but needed a focus. After a lot of thought about what topics I was most passionate about, I decided to go to a doctoral program in social psychology at the University of Connecticut with a specialization in health behavior research.

Ironically, I struggled in my first graduate statistics class. I found the material impenetrable and disconnected from the "real world" science that I wanted to do. Over time, and with a lot of hard work, I slowly mastered the topic. It took a lot of missteps and repetition, but to do the science I was most interested in, I needed to learn several advanced statistical methods, so I did. This training opened a door to a postdoctoral fellowship in advanced quantitative methods at Duke University, and my career as a methodologist and statistician took off from there!

One of your focus areas is on psychometrics, a term I've never heard of. What is psychometrics, and how can it inform clinical and translational science?

Psychometics is focused on the measurement of psychological and cognitive processes. Every research study measures one or more variables, many of them having to do with human thoughts, feelings, and behaviors. This type of measurement can be surprisingly difficult to do well. Most of us have completed numerous surveys over the course of our lives. Sometimes the measures in these surveys are fully tested and validated; however, most are simply created ad hoc by the research team. The result is that many measures designed to assess the same psychosocial or cognitive processes vary substantially between research studies in terms of their psychometric properties and validity.

Much of behavioral health science is therefore inconsistent in its methods and findings, negatively impacting scientific rigor and reproducibility. Because translational science is largely concerned with human health, rigorous, well-tested measures are essential. As one example, patient-reported outcomes (PROs) are becoming increasingly central to evaluating the impact of medical interventions. Measures in this area focus on all sorts of health predictors and outcomes, including mental health, quality of life, and disease symptoms. Without well-tested, reliable measures, we cannot evaluate whether what we are doing is improving patient health.

As part of your research, you've examined how group-specific stressors can negatively affect health outcomes, especially for historically marginalized groups. What have you found?

It is well-documented that our identities and lived experiences have direct impacts on our health and well-being. Having both social and financial capital greatly improve our chances at living a long and healthy life. Conversely, high levels of stress, deprivation, and lack of access to healthcare shortens lives and decreases quality of life. I have always been interested in understanding how social marginalization and stress—such as discrimination, social isolation, and maladaptive coping processes—impact people's health decision-making and outcomes.

"Giving people access to healthy, supportive healthcare environments that considers the whole human being can be a powerful tool to improve patient outcomes"

 Boynton

What I have consistently found in my research across multiple groups, including people of color, immigrant communities, and gender and sexual minorities, is that people living in largely the same social conditions can have substantially different responses to stress; however, as you increase the pressures that people are under and remove social supports that people can turn to, you see a marked decline in overall community health.

Some individuals can thrive under even the harshest conditions, but the reality is that putting people in a pressure cooker, with no way to escape, tends to be incredibly harmful. People are more likely to use maladaptive coping strategies, such as risk-taking, substance use, and patterns of unhealthy eating, when under high stress and lacking the ability to turn to healthier coping strategies.

One of the best ways we can use policy to support human health is to provide resources and other forms of support in an accessible, non-judgmental way. I realize that this seems like an incredibly obvious thing to say, but research consistently shows that care and compassion can go a long way in enhancing human health. Also, giving people access to healthy, supportive healthcare environments that considers the whole human being can be a powerful tool to improve patient outcomes.

As part of your work at NC TraCS, you serve as the biostatistics faculty mentor for the CTSA K12 Program and the UNC BIRCWH Program. Could you tell me more about that role and the type of advice you give to your mentees?

It has been one of the greatest honors of my life to serve as a mentor to numerous junior faculty pursuing research in translational science. Currently, I am a mentor for the TraCS K12 Junior Faculty Career Development Program, and I also mentor several other individual junior faculty with similar types of funding support. In this role, I typically provide guidance on research design and methods, which takes the form of consultative meetings and reviewing lots of grant proposals and manuscripts.

It turns out that my early struggles in statistics have served as a wellspring that allows me to more clearly explain complex statistical concepts. This ability allows me to provide actionable guidance to junior faculty on how to use more innovative and rigorous research and quantitative methods. This work includes conducting regular professional development sessions related to how structural factors, as well as our own implicit biases, can contribute to problematic scientific methodologies and health inequities.

I also advise faculty as they go through the process of responding to reviewer feedback on their grants and papers. It can be very painful to see one's work get picked apart by others. I regularly remind my mentees that reviewers are people too, and they don't always get things right. That said, reviewers have provided the gift of their time and experience, with little to no payment in return, and so we need to take the time to carefully think through the feedback in reviews and wring every helpful suggestion from them. Their science will be much improved as a result! In other words, welcome feedback from your colleagues, but you are the final arbiter as to what suggestions to take.

What do you think is the biggest challenge to the wider field of clinical and translational science?

For many decades, researchers did not consider the perspectives of the people they were studying. This contributed to biased results, ineffective interventions, and, at times, unethical research studies. Slowly, things are starting to change, but we still have a long way to go when it comes to respectfully and fully incorporating community member perspectives into translational science. TraCS is doing exciting work with respect to community engagement, recruitment, and inclusion. I am hopeful that we will greatly advance this field in the coming years. Without the voice of the community in translational science, we will never be able to further the patient centered care we aspire to promote!