MUSC-led study shows factors that may predict substance use in adolescents

May 06, 2024
Woman with her hair pulled back smiles in front of flowers. She is wearing a blue blazer and green blouse.
Dr. ReJoyce Green and colleagues used data from a large study to try to determine what factors may predict which young people are more likely to use substances early in life. Photo by Sarah Pack

ReJoyce Green, an assistant professor of Psychiatry and Behavioral Sciences at the Medical University of South Carolina, led a study that’s getting national attention for its findings on young people and substance use. Her team found that factors ranging from religion to race to a family’s income level appear to play a role. The results are featured in the May 6 online edition of the American Journal of Psychiatry.

Green said the study used a large dataset to tease out those factors. It came from the Adolescent Brain Cognitive Development Study, or ABCD. “It’s a multisite study, and one of the sites where the data is collected is here at MUSC.

ABCD follows young people ages 9 through adulthood. “The ABCD study collects a variety of different measures, from mental and physical health to neurocognitive assessments and neuroimaging. One of the things I’m most interested in is what we can learn about substance use.”

The analysis

Specifically, she and her fellow researchers wanted to know the top predictors for adolescent substance use. So essentially, we broke this large-scale data set into a smaller set of those who were substance naive at baseline. They’d never tried a single sip of alcohol, a single puff of any type of nicotine or cannabis. And then we had data available through a three-year follow-up period. And so, essentially, we’re saying over those three years, ‘How many youths initiated substance use, and what were some of the top predictors of initiation?’

They found those top predictors were in what she called the self-report domain. 

“Sociodemographic characteristics, mental health, physical health aspects, those ended up actually being some of our top predictors for substance use initiation compared with things that are more resource or time-intensive to collect,” Green said.

“So, for instance, hormone data, any type of neurocognitive task, any type of neuroimaging, all things that just take inevitably more and more time and resources and money to collect – those actually did not improve our ability to predict substance use initiation.”

Starting early

The researchers focused on children who began to use substances by age 12 – earlier than many parents would expect them to. “They’re very, very young. So we are really capturing the early stages,” Green said.

“I think that’s really interesting because when I think of a clinical setting, what are some things that I think would be helpful for providers to be aware of when they’re trying to gauge what things might be potential risk or protective factors for initiation?”

Religion as a factor

One of the top predictors: religious preference. “We included numerous religious preference categories and found mixed results, as some religious preferences were associated with a lower likelihood of initiation while others were associated with a greater likelihood of initiation,” Green said.

Green said that’s an area that needs more study. “Future studies might be able to tease apart a little bit more whether there are specific norms or practices within different religious groups that might lead one to be more or less inclined to try something such as a substance.” 

Other factors

Other factors the study suggests predict substance use in adolescents were sociodemographic. 

“Race was a top predictor. So we found that compared to white youth, Black youth were less likely to initiate substance use. We also had a few different income categories, and so we did find that one of the lower-income categories was actually less likely to initiate than one of the higher-income categories.”

The researchers also looked at how available substances were to adolescents. “We did find that was one of our top predictors as well and in the direction you would kind of expect. If they had more access to substances, they were more likely to initiate. This may be helpful for parents to consider how factors such as substance use availability may contribute to their child’s early experiences with a substance,” Green said.

“Other things we also had a history of detention or suspension was more likely to be associated with initiating. Other ones we also had were prenatal exposure to substance use was also among our top predictors. It’s really religion, race, income history, detention, suspension, prenatal exposure and substance use availability were all among our top 10.”

Those were all what she called self-report factors, which her team focused on first. Then, they considered other factors. 

Surprising findings

“In our second model, we added hormones and neurocognitive factors. In our third model, we added over 300 neuroimaging variables for both structural gray and white matter in the brain. And so I think what surprised us was that adding the hormones, neurocognitive factors and neuroimaging variables did not improve our ability to accurately predict whether or not a youth initiated over the follow-up period.”

Size of the group studied

Green’s conclusions in her current study came from analyzing data from about 7,000 adolescents. “In that sample, we had roughly about 14% in our initiation group, and the remaining 85, 86% were in our non-initiation group.”

Her team defined initiation very simply. “Like a single sip of alcohol, a single puff, really trying to get at just that initial starting point. And that includes initiation of any substance. So we include alcohol, nicotine, cannabis, misuse of any prescribed substances from physicians. We really tried to capture it very broadly, knowing that at this age range, it’s so young that we knew that, hopefully, we wouldn’t get a lot of initiation. But just we’re trying to capture as much as we could.”

What’s next

Green said work in this area is far from over. “I’ve only highlighted some of the top predictors. But other predictors that were just below, I think will be helpful in considering what we do next.”

Unlike factors such as race or religion, some of those other predictors may be more malleable. They include mental and behavioral health. 

“So, for instance, greater sensation-seeking, different aspects of impulsivity, greater rule-breaking, those were all things that were associated with greater likelihood of initiation. And I would argue that those are all things that I think could be viewed as modifiable risk factors that we could incorporate in either some type of prevention or early intervention programs.”

Green, who has a bachelor’s degree in psychology and a doctorate in clinical psychology from the University of California-Los Angeles, has authored multiple publications on substance use. She worked with adults struggling with substance use before coming to MUSC, where she shifted to younger adults and adolescents. 

“It’s interesting to go from people who had alcohol use disorders at ages 30 to 40 and saying, “OK, how can we track it back to the earliest touchpoint – and see what can we do there?' So there’s definitely a rewarding aspect to studies like this, finding things we can track – and maybe help prevent later problems.”

The team

Other MUSC researchers on the study included Anna Kirkland, Ph.D.; Brittney Browning; Brittany Bryant, DSW; Rachel Tomko, Ph.D.; Kevin Gray, M.D.; Lindsay Squeglia, Ph.D.; Bethany Wolf, Ph.D.; Andrew Chen, Ph.D.; and Pamela Ferguson, Ph.D. Louise Mewton, Ph.D., from the University of Sydney in Australia was also involved. 

Get the Latest MUSC News

Get more stories about what's happening at MUSC, delivered straight to your inbox.