Vaping and smoking as risk factors for arthritis: Evidence from double machine learning
- 1The Rinker College of Business, Palm Beach Atlantic University, West Palm Beach, Florida, United States of America.
- 2Department of Agricultural and Applied Economics, Texas Tech University, Lubbock, Texas, Lubbock, United States of America.
Abstract
Smoking is a well-documented risk factor for chronic diseases, including arthritis. Vaping, often marketed as a safer alternative, has risen sharply, yet its implications for musculoskeletal health remain unclear. We examine the associations of vaping and smoking with arthritis using Double Machine Learning (DML), and present illustrative economic scenarios projecting the potential cost implications of arthritis cases associated with these behaviors. Data from the Behavioral Risk Factor Surveillance System (BRFSS) of the Centers for Disease Control and Prevention (CDC), collected in the United States, for adults aged 18–64 years (N = 36,768) is used. Arthritis is the outcome variable; vaping and smoking are the exposure variables, with estimates conditioned on twenty-seven demographic, behavioral, and health-related covariates (See Table 1). Results show that vaping is associated with a 2.12 percentage point higher prevalence of arthritis than among non-vapers, while smoking is associated with a 2.22 percentage points increase compared with non-smokers. When accounting for concurrent nicotine behaviors, the estimated associations remain largely stable, with smoking exhibiting a slightly larger estimate of 2.14 percentage points while vaping attenuated to 1.91 percentage points. Vaping was estimated to be associated with approximately 190,581–254,109 arthritis cases annually, corresponding to a potential economic cost projection of 133.4−−177.9 million. Smoking was statistically associated with approximately 518,886–687,413 cases, corresponding to $363.2 –$481.2 million. Combined, the projected economic burden associated with vaping and smoking was approximately 496.6–659.1 million annually under conservative assumptions. Relative to published estimates of the total U.S. arthritis burden, these projections represent approximately 0.09%–0.22% of arthritis-related expenditures under direct healthcare cost assumptions and up to 9.23% when broader indirect and societal costs are considered. Our findings suggest that vaping is associated with musculoskeletal risk comparable to, or potentially exceeding, that of smoking, and do not support the perception of vaping as a risk-free alternative. The illustrative projections may inform policy discussions on vaping and smoking in relation to the broader arthritis-related health and economic burden.
Related
- Modeling Factors Associated with Smoking Dependence Severity Using Binary Classification-Based Machine Learning: Prediction with Hematological and Demographic Patient Data
- Machine learning–driven identification of whole-food dietary patterns associated with anxiety and rheumatoid arthritis comorbidity patterns: evidence from the NHANES and the USDA MyPyramid Equivalents database
- Individual-ecological triangulation of risk factors for peripheral artery disease: concordant evidence for smoking and diabetes, discordant evidence for kidney disease
- Cardiovascular adverse events of tocilizumab and baricitinib in COVID-19 and rheumatoid arthritis : A real-world disproportionality and machine learning analysis
- Integrating machine learning and artificial neural networks for dynamic analysis of a fractional order smoking-induced lungs cancer model
BibTeX
@article{chike2026vaping,
title = {{Vaping and smoking as risk factors for arthritis: Evidence from double machine learning}},
author = {Onyedikachi Emmanuel Chike and Syed Badruddoza and Conrad Lyford},
journal = {PLOS One},
year = {2026},
month = oct,
doi = {10.1371/journal.pone.0360863},
url = {https://doi.org/10.1371/journal.pone.0360863},
note = {CC BY},
}
RIS
TY - JOUR
TI - Vaping and smoking as risk factors for arthritis: Evidence from double machine learning
AU - Onyedikachi Emmanuel Chike
AU - Syed Badruddoza
AU - Conrad Lyford
PY - 2026
DA - 2026/10/23
JO - PLOS One
DO - 10.1371/journal.pone.0360863
UR - https://doi.org/10.1371/journal.pone.0360863
AB - Smoking is a well-documented risk factor for chronic diseases, including arthritis. Vaping, often marketed as a safer alternative, has risen sharply, yet its implications for musculoskeletal health remain unclear. We examine the associations of vaping and smoking with arthritis using Double Machine Learning (DML), and present illustrative economic scenarios projecting the potential cost implications of arthritis cases associated with these behaviors. Data from the Behavioral Risk Factor Surveillance System (BRFSS) of the Centers for Disease Control and Prevention (CDC), collected in the United States, for adults aged 18–64 years (N = 36,768) is used. Arthritis is the outcome variable; vaping and smoking are the exposure variables, with estimates conditioned on twenty-seven demographic, behavioral, and health-related covariates (See Table 1). Results show that vaping is associated with a 2.12 percentage point higher prevalence of arthritis than among non-vapers, while smoking is associated with a 2.22 percentage points increase compared with non-smokers. When accounting for concurrent nicotine behaviors, the estimated associations remain largely stable, with smoking exhibiting a slightly larger estimate of 2.14 percentage points while vaping attenuated to 1.91 percentage points. Vaping was estimated to be associated with approximately 190,581–254,109 arthritis cases annually, corresponding to a potential economic cost projection of 133.4−−177.9 million. Smoking was statistically associated with approximately 518,886–687,413 cases, corresponding to $363.2 –$481.2 million. Combined, the projected economic burden associated with vaping and smoking was approximately 496.6–659.1 million annually under conservative assumptions. Relative to published estimates of the total U.S. arthritis burden, these projections represent approximately 0.09%–0.22% of arthritis-related expenditures under direct healthcare cost assumptions and up to 9.23% when broader indirect and societal costs are considered. Our findings suggest that vaping is associated with musculoskeletal risk comparable to, or potentially exceeding, that of smoking, and do not support the perception of vaping as a risk-free alternative. The illustrative projections may inform policy discussions on vaping and smoking in relation to the broader arthritis-related health and economic burden.
SN - 1932-6203
ER -