If you want to understand someone’s overall health, try shaking their hand.
It might seem strange, but research has shown that the strength of a person’s grip is a good indicator of many aspects of their well-being, from bone density to cognition to cardiovascular health and even mortality risk.
But what constitutes a strong grip? How strong should a young woman’s grip be, compared to an older man’s? Is a person’s grip strength impacted by their race or their height?
To answer those questions, you don’t need a doctor. You need a data scientist.
That’s why Cameron Lee has been so interested in this research. She’s a data science major who just graduated from UWM in December, and she also has worked with UWM’s Success Through Aging Research (STAR) research program, which encourages diversity in the field of aging research. With help from her mentor, Professor Inga Wang from UWM’s School of Rehabilitation Sciences & Technology, Lee has been researching grip strength in hopes that she can make some helpful contributions to the field.
Lee’s research project has four goals:
1. Examine what factors can impact a person’s grip strength, such as race, sex, age, and body mass index (BMI).
2. Create a mathematical model that scientists can use as a guideline to determine a person’s grip strength based on age.
3. Follow trends of changes in grip strength over a person’s lifespan to determine what grip strength looks like through childhood, young adulthood, middle age, etc.
4. Make a machine learning algorithm to predict a person’s grip strength.
“The awesome thing about this is we could guide physicians,” said Lee. “We could give clinical people potential areas in which they can look, based on these algorithms, to help do more personalized care management around the idea of their patient being frail.”
Getting a grip on the data
Grip strength is measured by a tool called a dynamometer, which calculates the force exerted by a person’s forearm muscles. The Centers for Disease Control collected data on grip strength from 2011-14 and released it as part of their National Health and Nutrition Examination Surveys published annually. It was from these surveys that Lee drew her data – and it was a lot of data.
That’s where data science comes in. Data scientists often work with data sets that are so big they can’t be parsed by a single person, even with the aid of a computer. Data scientists often use machine learning to help them make sense of the information.
That’s just what Lee did. She built an algorithm to teach the computer what sort of information she was looking for and the patterns she wanted it to recognize.
For example, she designed her algorithm to look at examples of someone’s grip strength based on their sex and race. Given that information, she can tell her program, “Now that you’ve seen this set of people, here are new people. Predict their age groups based on prior people,” she said.
Lee’s work continues: “What I am hoping to do is create an electronic frailty index to predict whether someone is frail or not,” she added.
Health care professionals rely on several indicators to measure a person’s frailty, including walking speed, involuntary weight loss, activity level, and exhaustion, as well as grip strength. Obviously, said Lee, some of those tests can be difficult to perform on patients who are injured or sick.
“The idea of this electronic frailty index is to use past … data like their blood or their urine samples. Can we use those things to predict frailty, and what markers are we looking for?” Lee said. “The issue is, can these laboratory tests actually do that? Can they measure weakness or exhaustion? That’s why grip strength research is preliminary research to this bigger project.”
Lee will be conducting her work as part of the STAR post-baccalaureate program at UWM over the next year, and she’s applied to begin her PhD work at UWM with Wang as her advisor.
Health and research
Lee has presented her work at seven conferences, including the Conference of Undergraduate Research in Long Beach, California; the American Congress of Rehabilitation Medicine Conference in Dallas; and the Annual Biomedical Research Conference for Minoritized Scientists in Pittsburgh.
She hopes to become an epidemiologist with the National Institutes of Health or the World Health Organization in the future, where she could lead public health studies. After all, data science is a useful tool beyond measuring grip strength.
“It’s so interdisciplinary,” said Lee. “We can answer questions like, what kind of populations are more susceptible to (certain diseases) based on the data records? Can we predict that? What are the actual statistics of people becoming sick? … It fits in every single area – aging, cancer, health disparities.”
Lee was always a “math kid,” but during the COVID-19 pandemic, she grew interested in public health. She transferred to UWM from Lakeland University to pursue a greater variety of public health and machine learning classes. She’s grateful that UWM gave her the tools to pursue her passions – especially research.
“Research is for everyone, whether they believe it or not,” Lee said. “It exposes to me things that I just never would have thought about or realized were part of this world, and I get to meet the brightest, smartest people, and we get to share our experiences with each other. I think there’s never-ending opportunity, and I’m really confident in the foundations I’ve built here at UWM.”
That’s data you can rely on.
By Sarah Vickery, College of Letters & Science