Trends in Current Smoker Profiles Among US Adults From 2019-2024: An Unsupervised Clustering Analysis
Author(s)
Kyla Finlayson, MS1, Vicky W. Li, MPH2, Nate Way, PhD3.
1Biostatistician, Real-World Evidence, Oracle Life Sciences, Austin, TX, USA, 2Oracle Life Sciences, Austin, TX, USA, 3Oracle Life Sciences, Santa Barbara, CA, USA.
1Biostatistician, Real-World Evidence, Oracle Life Sciences, Austin, TX, USA, 2Oracle Life Sciences, Austin, TX, USA, 3Oracle Life Sciences, Santa Barbara, CA, USA.
OBJECTIVES: This study identified potential user profiles among current smokers in a general US adult population from 2019 to 2024 using unsupervised clustering.
METHODS: Data on 68,922 current smokers were analyzed from the 2019-2024 US National Health and Wellness Survey. Each year of data was analyzed cross-sectionally. K-medoids clustering with t-SNE dimensionality reduction was implemented to create distinct clusters of respondents with similar characteristics among those who reported being current smokers. In total, fifty-one variables including sociodemographic characteristics, cigarette and nicotine product use, intention to quit smoking, and comorbidities including respiratory, cardiovascular, and mental health conditions were included in analysis. Average silhouette width was used to determine the optimal number of clusters.
RESULTS: Among the 10,959 respondents in 2019, two distinctive profiles of smokers were identified: (1) younger, mostly male smokers who used more nicotine products and reported more ER visits; and (2) older, lower income cigarette-only smokers who reported more comorbidities and were less likely to be trying to quit smoking. Among the 32,669 respondents in 2020-2022, four distinctive profiles were identified: (1) younger, mostly male smokers who used more nicotine products; (2) mostly white, less educated cigarette-only smokers who reported lowest number of ER visits; (3) mostly non-white, high comorbidity smokers who tried to quit smoking; and (4) older, lower income dual users of cigarettes and nicotine products who reported more comorbidities. Among the 25,294 respondents in 2023-2024, four distinctive profiles of smokers were identified: (1) younger, mostly male smokers who used more nicotine products and reported highest number of ER visits; (2) mostly non-white, high comorbidity smokers who tried to quit smoking; (3) mostly non-white, uninsured cigarette-only smokers; and (4) older, mostly white cigarette-only smokers.
CONCLUSIONS: In a broadly representative US adult population, K-medoids clustering identified shifting trends in clusters of current smokers during pre-COVID-19, COVID-19, and post-COVID-19 time periods.
METHODS: Data on 68,922 current smokers were analyzed from the 2019-2024 US National Health and Wellness Survey. Each year of data was analyzed cross-sectionally. K-medoids clustering with t-SNE dimensionality reduction was implemented to create distinct clusters of respondents with similar characteristics among those who reported being current smokers. In total, fifty-one variables including sociodemographic characteristics, cigarette and nicotine product use, intention to quit smoking, and comorbidities including respiratory, cardiovascular, and mental health conditions were included in analysis. Average silhouette width was used to determine the optimal number of clusters.
RESULTS: Among the 10,959 respondents in 2019, two distinctive profiles of smokers were identified: (1) younger, mostly male smokers who used more nicotine products and reported more ER visits; and (2) older, lower income cigarette-only smokers who reported more comorbidities and were less likely to be trying to quit smoking. Among the 32,669 respondents in 2020-2022, four distinctive profiles were identified: (1) younger, mostly male smokers who used more nicotine products; (2) mostly white, less educated cigarette-only smokers who reported lowest number of ER visits; (3) mostly non-white, high comorbidity smokers who tried to quit smoking; and (4) older, lower income dual users of cigarettes and nicotine products who reported more comorbidities. Among the 25,294 respondents in 2023-2024, four distinctive profiles of smokers were identified: (1) younger, mostly male smokers who used more nicotine products and reported highest number of ER visits; (2) mostly non-white, high comorbidity smokers who tried to quit smoking; (3) mostly non-white, uninsured cigarette-only smokers; and (4) older, mostly white cigarette-only smokers.
CONCLUSIONS: In a broadly representative US adult population, K-medoids clustering identified shifting trends in clusters of current smokers during pre-COVID-19, COVID-19, and post-COVID-19 time periods.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
MSR116
Topic
Methodological & Statistical Research
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics
Disease
SDC: Respiratory-Related Disorders (Allergy, Asthma, Smoking, Other Respiratory)