Exploring the Bidirectional Relationship between Food Insecurity and Smoking: A Cross-Lagged Panel Analysis

Background and Purpose: Food insecurity and smoking are interrelated factors that disproportionately affect individuals experiencing poverty. Recent studies suggest a bidirectional relationship between these factors, however, longitudinal research is needed to better understand their interaction over time. This study utilized data from the Future of Families and Child Wellbeing Study (FFCWS) to explore the following research question: Is there a bidirectional relationship between smoking and food insecurity?

Methods: This study used data from Waves 3 and 4 (N = 2,394) of the FFCWS, a longitudinal study examining the effects of social and economic factors on families in U.S. cities. Food insecurity was measured with a binary variable (yes or no) based off respondents’ responses to the U.S. Household Food Security Survey Module. Smoking was measured with a binary variable (yes or no) based off respondents’ answer to the question: “In the past month, did you smoke cigarettes?” Covariates included poverty, race, and depression/anxiety. A cross-lagged panel model with food insecurity and smoking measured at baseline (Wave 3) and follow-up (Wave 4) was fitted using R to explore their relationship over time. Data analysis was conducted without (i.e., unadjusted model) and with the covariates (i.e., adjusted model). Three exploratory models were also examined—each with one of the three covariates—to examine how each covariate(s) may influence the relationship between food insecurity and smoking.

Results: The findings showed a significant cross-lag relationship between food insecurity and smoking in the unadjusted model: Baseline food insecurity was associated with significantly higher odds of smoking at follow-up (β = 0.125, p = 0.021), and baseline smoking was associated with significantly higher odds of food insecurity at follow-up (β = 0.169, p = 0.010). In the adjusted model, however, these relationships became non-significant. Additionally, poverty, race, and depression/anxiety were each significantly and positively associated with future food insecurity and future smoking in the adjusted model. In the exploratory models, the cross-lag relationships between baseline smoking and food insecurity and future food insecurity and smoking remained significant when race was included as a covariate, but were non-significant when poverty or depression/anxiety were included as a covariate.

Conclusions and Implications: There may not be a clear or simple relationship between food insecurity and smoking. Rather, these findings indicate that poverty and mental health are confounding factors that affect whether food insecurity predicts smoking, and vice versa. Relatedly, the results suggest that poverty and mental health may be more influential than current food insecurity and smoking when predicting future food insecurity and smoking. Policy and practice interventions should target poverty and mental health to address food insecurity and smoking. Future collaborative research by social workers, policymakers, and individuals impacted by these issues should explore the relationship between all these factors (i.e., food insecurity, smoking, poverty, and mental health) over time to develop a more nuanced understanding of how to address food insecurity and smoking. For example, future studies should explore the potential mediating roles of food insecurity and mental health in the relationship between poverty and smoking.