Changing Smoke to Data
Second only to heart disease, smoking is among the leading causes of death in the United States, accounting for 455,000 deaths per year. This slow, silent killer leads to more deaths by itself than car accidents, suicide, drug overdose, and gun violence combined; yet it’s not something we generally hear much about.
Unless, of course, you’re Christopher Cambron, assistant professor at the University of Utah College of Social Work. His work is grounded in early professional experiences that highlighted the difficulties of substance abuse treatment, and the important role that access to social and economic resources can play in trying to improve health behaviors—something now backed by both his professional peers and his own research. “Tobacco is one of the major killers of Americans today, but it flies under the radar because many of the people dying from it are from a lower socioeconomic status and don’t have a lot of voice,” Dr. Cambron explains. Working from a health equity perspective, he hopes his research can improve our understanding of how to reduce the clear socioeconomic disparities in tobacco use and smoking cessation.
The most traditional method of data collection in observational research studies on substance use is retrospective surveys—administered once a year, or once every few years. Using these data, researchers are able to see trends in self-reported use and abuse over time. This kind of retrospective survey data is important for showing broad trends of behavior over longer periods of time.
“We have clear indications that substance abuse and addiction are problematic for health across the life course, employment prospects, general wellbeing, quality of family life—it’s clear across many dimensions,” said Dr. Cambron. “But in order to do something about that, you need to understand how factors that influence behavior are working in real time. If you’re trying to connect broad life-course perspectives to treatment perspectives, traditional surveillance data can only get you so far.” He continued, “You need to understand lived, in-the moment-experiences to enhance the interventions we currently have, and to build new ones.” He explained that what people tell you in retrospective surveys can be different from what they feel in real time. “Memories are fallible. People misremember. We tend to live through our stories, which may or may not be accurate.”
This is why the innovative method of data collection known as experience sampling or ecological momentary assessment (EMA) is so important for addiction research. Instead of collecting retrospective data every year, Dr. Cambron’s recent work utilizes data gathered every day from study participants’ smartphones. Each day, participants get randomly pinged on their phones with survey questions about their momentary experience with trying to quit smoking: How are you feeling? Where are you? Are you around anyone who is smoking? When was the last time you smoked a cigarette? How stressed are you? Answering these questions multiple times a day for a few weeks generates a very rich and dense description of lived experience. Whereas a traditional survey approach might end up with 500 participants followed for a few years and leading to 1500 data points, EMA studies regularly produce datasets with tens of thousands of data points.
“Human behavior is multi-faceted and complex so these very dense datasets allow us to detect small changes we might not see with more traditional survey methods. But small changes can be important as they accrue over time,” explained Dr. Cambron. “Stress is a really good example of this. We’ve known that it’s an important factor in addiction for many years. But, the scope of how important stress is, because of its broad and consistent impact on many aspects of experience, becomes much clearer using EMA approaches.”
Dr. Cambron is referencing two recent studies he conducted that demonstrated the impact of stress on multiple factors known to be important for smoking cessation. In an academic sense, previous research has shown that stress has an effect on a multitude of psychological factors: affect, craving, abstinence self-efficacy (the belief in one’s ability to quit), and coping and smoking expectancies (beliefs about how smoking will affect one’s mood in the moment, and beliefs about whether the coping skills you have will be effective in the moment). A momentary instance of stress impacts all these things in ways that can make smoking relapse more likely during a quit attempt.
Because Dr. Cambron’s recent studies used these very dense EMA datasets, he was able to detect the impact of stress on these factors simultaneously. “With the amount of data we have, we can start to statistically test more fine-grained aspects of theoretical models of addiction—something researchers have struggled with previously outside of laboratory settings. EMA studies hopefully can help better align our interventions with the lived experiences of those participating in the interventions.”
Dr. Cambron is hopeful about where further use of these technologies could lead. “One of the next big questions is: Can we use technology to intervene in real time? Can we design strategies to bolster treatment or provide additional supports in the moment?” He continued, “Smoking is a huge killer in America and disproportionately kills lower socioeconomic status folks. Over 75% of adults have some sort of smartphone these days, so the hope is we can provide low cost, technology-based treatment to people without consistent access to health care.” For him, that’s reason enough to keep going.