The Intersection of Food Insecurity and Tobacco Use: A Scoping Review

Abstract

Cigarette smoking is increasingly concentrated in socioeconomically disadvantaged groups, and food insecurity also disproportionately affects lower-income groups. Recent studies have suggested that smoking and food insecurity operate as risk factors for one another, but there is limited understanding of their intersection. This scoping review aimed to synthesize the published literature on the association between food insecurity and tobacco use across population groups in the U.S. and Canada. We searched PubMed, Web of Science, and PsycINFO using keywords. Studies included were published in English between 2008–2018, reported empirical findings, measured both tobacco use and food insecurity, and considered either variable as a study outcome. Nineteen articles were identified; 6 examined tobacco use as an outcome variable and 13 examined food insecurity as an outcome variable. Most articles were of studies using cross-sectional designs. Study samples ranged from general populations, clinical samples, and underserved populations. For each article, we extracted information including specific findings related to the association between food insecurity and tobacco use. We synthesized the current research by formulating a model by which food insecurity and tobacco use are bi-directionally associated. This scoping review concludes that the co-occurrence of food insecurity and tobacco use exists across populations in the U.S. and Canada. As the evidence is largely from cross-sectional investigations, there is a need for longer term, comprehensive assessments of relationships between tobacco use and food insecurity. Such investigations can inform policies and interventions aimed towards addressing the inequitable burden of tobacco use and of food insecurity among disadvantaged populations.

Background

Tobacco use, particularly cigarette smoking, is disproportionately concentrated among populations of lower socioeconomic status (Hiscock, Bauld, Amos, Fidler, & Munafò, 2012; Jamal et al., 2018), and largely contributes to social gradients in health (Stringhini et al., 2010). Causes of socioeconomic disparities in smoking are complex, necessitating a social ecological approach to understand and address such disparities (U.S. National Cancer Institute, 2017). Previous reviews have articulated the multitude of factors that are associated with socioeconomic status and cigarette smoking (Hiscock et al., 2012; Pampel, Krueger, & Denney, 2010). This includes the disproportionate burden of chronic life stressors—such as financial stress—that lower-income smokers face (Siahpush, Borland, & Scollo, 2003; Siahpush, Borland, Yong, Cummings, & Fong, 2012; Widome, Joseph, et al., 2015). Financial stress can maintain smoking behavior (Mulder, de Bruin, Schreurs, van Ameijden, & van Woerkum, 2011) and pose barriers to cessation (Kalkhoran, Berkowitz, Rigotti, & Baggett, 2018).

A particularly prominent financial stressor that disproportionately impacts socioeconomically disadvantaged groups is food insecurity. Food insecurity occurs when access to enough food for an active and healthy life is limited by a lack of money or other resources (Coleman-Jensen, Rabbitt, Gregory, & Singh, 2018), or when there are limitations in the ability to acquire personally acceptable foods in socially acceptable ways (Anema, Vogenthaler, Frongillo, Kadiyala, & Weiser, 2009). In 2017, 31% of U.S. households living at or below 185% of the federal poverty level (FPL) experienced any food insecurity at some point during the year, in contrast to 6% of households living above 185% of FPL (Coleman-Jensen et al., 2018). Racial/ethnic minority households and female-headed households are also more likely to experience food insecurity as compared to the national average (12%), affecting 22% of non-Hispanic Black households, 18% of Hispanic households, and 30% of female-headed households in 2017 (Coleman-Jensen et al., 2018). The overall inequitable experience of food insecurity across socioeconomic status, race/ethnicity, and gender is an important health disparity issue. While there is an important understanding that food insecurity exists alongside other intersecting and structural risk factors such as employment and housing insecurities (Laraia, Leak, Tester, & Leung, 2017), food insecurity is nevertheless considered to have an independent effect on a range of physical and mental health conditions (Gundersen & Ziliak, 2015).

The inverse association between socioeconomic status and tobacco use has been extensively described in the literature (Pampel et al., 2010). Since 2008, studies have highlighted the association between food insecurity and cigarette smoking. Similar to studies examining the phenomenon of smoking-induced deprivation—that is, smokers reporting that they spent money towards purchasing cigarettes that should have been spent on household essentials such as food (Siahpush, Borland, & Yong, 2007), earlier studies largely reported that smoking is a risk factor for food insecurity (Armour, Pitts, & Lee, 2008; Cutler-Triggs, Fryer, Miyoshi, & Weitzman, 2008). Yet given that food insecurity is known to be a stressful experience, and that many smokers report smoking for stress relief, more recent studies report that food insecurity may be a risk factor for cigarette smoking (Hosler & Michaels, 2017; Kim & Tsoh, 2016). As cigarette smoking remains a leading cause of preventable disease and death in the U.S. (U.S. Department of Health and Human Services, 2014), understanding the co-occurrence of food insecurity and tobacco use use, including their potential bidirectional association, is crucial for understanding and mitigating health disparities.

Aims

The overarching aim of this scoping review is to examine the current empirical literature on the association between food insecurity and tobacco use. More specifically, we examine the following key questions: In the published literature, is food insecurity considered an explanatory variable for tobacco use, or is tobacco use considered an explanatory variable for food insecurity? What are plausible mechanistic explanations that offered by current studies? By examining these key research questions, this scoping review examines the range of population groups or study samples for whom the association between tobacco use and food insecurity has been documented between 2008 and 2018, and provides critical summaries of study designs, samples characteristics, and measures of food insecurity and tobacco use. By conducting this review, we aim to offer recommendations for future research and practice aimed at addressing health disparities related to tobacco use, to food insecurity, and the combination of the two. As the context of tobacco use and of food insecurity can largely differ across cross-national contexts, this review specifically focuses on study samples from the U.S. and Canada.

Methods

Data Sources

Published articles were identified using three databases—PubMed, Web of Science, and PsycINFO. Search terms were identical across the databases: ((“food insecur*”) OR (“food secur*”) OR (hunger)) AND ((smoking) OR (cigarette*) OR (tobacco)). Inclusion criteria for articles were as follows: (1) published in English; (2) published in a peer-reviewed journal; (3) presented original research; (4) published between January 1, 2008, through June 1, 2018; (5) study sample is of participants from the U.S. or Canada; and (6) measured food security/insecurity and smoking/tobacco use within the study. As scoping reviews can be iterative, such that parameters may be added or refined on the basis of the literature itself (Arksey & O’Malley, 2005), we added an additional inclusion criteria which was that either food security/insecurity or smoking/tobacco use is treated as a dependent variable in the study’s main analysis or any reported sub-analyses.

Article Screening and Data Extraction

As an initial screening, we independently reviewed article titles and abstracts from the search results of each database to determine the article’s eligibility. If the study’s relevance was not apparent from the title and/or abstract, the full-text article was retrieved. A data charting spreadsheet was developed to uniformly capture relevant information across all studies reviewed. The variables extracted included information regarding the purpose of the study and the general study design, the geographic location in which study samples were drawn, a description of the sample characteristics (e.g., sample size, the gender and racial/ethnic composition, and any special features of the sample), and instruments used to measure food insecurity and tobacco use. As part of the data charting, we summarized findings that were specific to the association between food insecurity and smoking (or vice versa). We noted statistical associations as reported by the study’s analysis (e.g., odds ratios or regression coefficients), also noting any covariates that were included.

This article’s second author conducted the initial extraction, which was subsequently reviewed by the first author. Discrepancies regarding the relevance of the articles were resolved through discussion and mutual agreement. This overall process also allowed us to determine which articles did not ultimately meet the review eligibility after we conducted data extraction based on the full-text article. In some of the articles, the association between food insecurity and smoking was not a primary focus, although an association was reported elsewhere within the findings. These types of articles were excluded, and lastly, articles were grouped according to whether the dependent variable was food insecurity or tobacco use/smoking.

Results

Figure 1 depicts the study selection criteria. We note that a manual search yielded two additional articles that we included in our review. These studies were published online during the study timeframe, but did not appear in our initial search.

Figure 1.

Flowchart of Article Selection Process for Scoping Review

Overall Study Characteristics

Nineteen studies reported associations between food insecurity and tobacco use status (Tables 1 and 2). Seventeen were based in the U.S., and two were based in Canada. Among the U.S. studies, six were based on national samples, four were based in California, four in Eastern states (Connecticut, Delaware, New York, and Pennsylvania), two in Midwestern states (Ohio and Wisconsin), and one in Texas. In the two Canada studies, one was based in Ontario, and the other sampled from eight different provinces. Per inclusion criteria, all studies presented quantitative findings; 16 used cross-sectional study designs, and 3 used longitudinal designs, with follow-up periods spanning 6 weeks, 12 months, and 3 years. Five studies examined study participants with lower income, such as tenants living in subsidized housing, women who were homeless or unstably housed, persons using mobile food pantries, and persons below 200% of the federal poverty level. In terms of special characteristics of the study samples, two were focused on veterans, two were focused on persons living with HIV, and one was focused on youth. In general, the study samples were heterogeneous with respect to sex, age, race/ethnicity, income, and specific community-based populations.

Table 1.

Summary of Findings of 6 Studies Examining Smoking as the Dependent Variable

First Author (year)	General Study Design, Purpose, and Location	Sample Characteristics	Measures of Food Insecurity and Smoking	Reported Findings on the Association between Smoking and Food Insecurity
Castro (2015)	Cross sectional study, from 2012/2013 Geographic Research on Wellbeing Study Study purpose: To examine whether psychosocial and environmental factors account for racial differences in smoking status among mothers Location: Alameda, Los Angeles, Orange, Sacramento, San Diego, and Santa Clara Counties, California, USA	542 ever-smoking women who gave birth 43% White, 17% US-born Latina, 17% immigrant Latina, 15% Asian/Pacific Islander, 7% Black	Food insecurity measured by 6-item Food Security Survey Smoking measured by self-report of having smoked 100+ cigarettes in lifetime, and current smoking status	Among food insecure women, 59% had quit smoking (i.e., were former smokers). Among food secure women, 76% had quit smoking. Food insecurity was associated with significantly decreased the odds of quitting smoking (OR=0.55), adjusting for race/ethnicity, age, partner status and educational attainment. However, food insecurity was no longer significantly associated with quitting smoking when additionally controlling for income, perceived neighborhood safety, having friends who smoke, and smoking in the home.
Hosler (2017)	Cross-sectional study, data collected in 2013 Study purpose: To investigate associations between smoking and food distress at the individual and neighborhood levels Location: Schenectady, New York, USA	1,917 adults 60% female 49% non-Hispanic White, 26% non-Hispanic Black, 11% Guyanese, 10% Hispanic, 4% multiracial/other	Food insecurity measured by single-item on whether did not have enough food at home often or sometimes in the past 12 months Smoking measured by self-report of having smoked 100+ cigarettes in lifetime and whether now smokes every day or some days	Among individuals with food insecurity, 57% were smokers. Among individuals without food insecurity, 33% were smokers. Food insecurity was associated with significantly increased the odds of smoking (OR=1.77) adjusting for age, sex, race/ethnicity, educational attainment, and household income, anxiety disorder, alcohol binge drinking, fruit and vegetable consumption, food pantry usage, receipt of SNAP benefits, neighborhood food environment, and food shopping behaviors.
Kim (2016)	Cross-sectional study, from 2011/2012 California Health Interview Survey Study purpose: To examine correlates of smoking among low-income young adults Location: California, USA	1,511 low-income young adults aged 18–30 48% female 66% Hispanic/Latino, 18% non-Hispanic White	Food insecurity measured by 6-item Food Security Survey Smoking measured by self-report of having smoked 100+ cigarettes in lifetime, and whether now smokes every day, some days, or not at all	Among food insecure young adults, smoking prevalence was 27%. Among the food secure young adults, smoking prevalence 16%. Food insecurity was associated with significantly increased odds of current smoking (OR=1.54), controlling for age, sex, race/ethnicity, nativity and English proficiency, education, poverty, usual source of health care, alcohol use, and psychological distress. Food insecurity was also associated with significantly increased the odds of daily smoking (OR=1.91), adjusting for covariates listed above. However, food insecurity was not significantly associated with non-daily smoking.
Kim (2017)	Longitudinal cohort study, data collected from 2009–2012 Study purpose: To longitudinally examine associations between food insecurity severity and smoking among women who are homeless or unstably housed Location: San Francisco, California, USA	247 unstably housed women living with and without HIV 41% Black/African American, 30% White, 18% Multiracial, 5% Latina/Hispanic, 2% Asian/Pacific Islander	Food insecurity measured by 6-item Food Security Survey. Categorized as no/mild, moderate, and severe food insecurity Smoking measured by self-report of smoking 1+ cigarette per month in past 6 months	Among current smokers, 63% were food insecure and 42% were severely food insecure. Among non-smokers, 51% were food insecure and 26% were severely food insecure. Severe food insecurity was longitudinally associated with significantly increased odds of smoking (OR=1.68), adjusting for age, race, education level, marital status, HIV status, mental health, and substance use (alcohol, marijuana, cocaine, methamphetamine, and heroin). Mild/moderate food insecurity was not associated with smoking.
Perkett (2017)	Intervention study on tobacco education using pre- and post-test design, conducted in 2016 Study purpose: To evaluate cardiovascular health status and to determine impact of low-intensity smoking cessation education intervention among mobile food pantry participants Location: Delaware, USA	144 adults using mobile food pantries in the community 69% female 53% African American, 43% White, 5% Other	Food insecurity measured by 2-item screener Smoking measured by whether smoked a cigarette (even one puff) in past 30 days. Assessed time to first cigarette upon awakening, and cigarettes per day	Among smokers, moderate/severe food insecurity prevalence was 48% and 42%, respectively. Among nonsmokers, moderate and severe food insecurity prevalence was 43% and 21%, respectively. Moderate/severe food insecurity was associated with significantly increased odds of smoking (OR=4.98), adjusting for race, job status, gender and education.
Robson (2017)	Cross-sectional study, from 2014–2015 Youth Risk Behavior Survey Study purpose: To examine relationship between food insecurity and cardiometabolic risk factors in adolescents Location: Pennsylvania, USA	4,994 high school students 50% female 70% non-Hispanic White, 16% non-Hispanic Black, 8% Hispanic, 3% Asian, 3% Other	Food insecurity measured by single item on how often went hungry because there was not enough food in home during the past 30 days Smoking measured by self-report of smoking 1+ cigarettes in the past month	Among food insecure adolescents, smoking prevalence was 19%. Among food secure adolescents, smoking prevalence was 11%. Food insecurity was associated with significantly increased odds of smoking (OR=1.81), adjusting for age, sex, race/ethnicity, grade in school, and neighborhood safety

Notes: OR = odds ratio

Table 2.

Summary of Findings of 13 Studies Examining Food Insecurity as the Dependent Variable

First Author (year)	General Study Design and Study’s Stated Purpose	Sample Characteristics	Measures of Food Insecurity and Smoking	Reported Findings on the Association between Smoking and Food Insecurity
Armour (2008)	Cross-sectional study, from 2001 Panel Study of Income Dynamics Study purpose: To quantify association between food insecurity and smoking in low-income families Location: USA	2,099 low-income families 48% Black, 38% White, 34% Other	Food insecurity measured by 18-item Food Security Survey Smoking measured by self-report on whether smokes cigarettes, and how many cigarettes per day	In food insecure families, adult smoking prevalence was 44%. In food secure families, adult smoking prevalence was 32%. Smoking was associated with significantly increased odds of food insecurity (OR=1.44), adjusting for age, race, education, marital status, number of children, family income, current alcohol use, metropolitan area residence, and region of US.
Bekele (2018)	Cross-sectional study, data collected in 2011–2013 Study purpose: To estimate the prevalence of food insecurity among adults living with HIV and to identify individual and household level factors associated with food insecurity	649 people living with HIV/AIDS, recruited from community-based AIDS service organizations 34% female 53% White, 25% Black	Food insecurity measured by 10-item Health Canada Household Food Security Scale Module, based on US Department of Agriculture’s Food Security Survey Smoking measured by self-report of smoking in the past 30 days	Among current smokers, food insecurity prevalence was 76%. Among non-smokers, food insecurity prevalence was 64%. Smoking was associated with significantly increased odds of food insecurity (OR=1.71), adjusting for sex, ethnicity, level of education, household structure (e.g., single, couple), household income, difficulty with housing costs, history of incarceration, substance use, and depressive symptoms.
Brostow (2017)	Cross sectional study, from 2012 Health and Retirement Study and 2013 Health Care and Nutrition Mail Survey Study purpose: To examine prevalence of food insecurity among older veterans Location: Ontario, Canada	2,560 veteran and non-veteran adult men 74% White, 21% African American/Black, 5% Other	Food insecurity measured by a subset of questions from Food Security Survey Smoking measured as current smoker or ever smoker, measure not further specified	Among food insecure male veterans, smoking prevalence was 40%. Among food secure male veterans, smoking prevalence was 13%. Among male veterans aged 65+, smoking was associated with increased odds of food insecurity (OR=5.26), adjusting for age, income, race/ethnicity, marital status, military pension, history of tobacco and alcohol use, body mass index, physical activity, health behaviors, comorbidities, physical mobility, dementia, total word recall summary scores, any difficulty with activities of daily living, diagnosis of emotional/ nervous/ psychiatric problems, and depression. Smoking was not significantly associated with food insecurity when examining all male veterans, and male veterans aged <65 years.
Cutler-Triggs (2008)	Cross sectional study, from 1999–2002 National Health and Nutrition Examination Survey Study purpose: To investigate rates and severity of child and adult food insecurity in households with and without adult smokers Location: USA	8,817 households with children aged 17 years or younger 49% female 59% White, 21% Hispanic, 15% Black, 6% Other	Food insecurity measured by 18-item Food Security Survey. Categorized as full, marginal, low, and very low food security Smoking household measured by single-item assessing whether anyone who lives in the household smokes cigarettes anywhere inside the home	In households with adult smokers, adult food insecurity prevalence was 26% and child food insecurity prevalence was 17%. In households without adult smokers, adult food insecurity prevalence was 12% and child food insecurity prevalence was 9%. Among adults, being a smoking household was associated with significantly increased odds of low (OR=2.2) and very low food security (OR=2.3), adjusting for age, sex, race/ethnicity, and poverty.
Fitzgerald (2011)	Cross-sectional study (data collection period not specified) Study purpose: To examine association of food insecurity with type 2 diabetes among Latinas, and to identify risk factors related to food insecurity in this population Location: Hartford, Connecticut, USA	201 Latinas aged 35–60, with and without Type 2 diabetes	Food insecurity measured by 6-item Food Security Survey Smoking measured by self-report of “yes or no”; measure not further specified	Smoking was associated with significantly increased odds of very low food security (OR=3.74), adjusting for SNAP participation, depressive symptoms, level of nutrition knowledge, and waist circumference. Smoking was not significantly associated with low food security.
Gucciardi (2009)	Cross-sectional study, from 2005 Canadian Community Health Survey Study purpose: To determine household food insecurity prevalence among Canadians with diabetes and its relationship with diabetes management and health Location: Alberta, British Columbia, Northwest Territories, Nova Scotia, Nunavut, Ontario, Prince Edward Island, and Quebec Provinces, Canada	6,237 individuals aged 12 years or older with diabetes	Food insecurity measured by 18-item Food Security Survey Smoking measured by self-report on whether smoke cigarettes	Among individuals with food insecurity, 32% were smokers. Among individuals without food insecurity, 16% were smokers. Smoking was associated with significantly increased odds of food insecurity (OR=1.71), adjusting for age, sex, diabetes duration, insulin status, having a regular medical doctor, having had the effects of a stroke, adjusted income ratio, household education level, First Nations status, smoking status, and physical activity level.
Hernandez (2017)	Cross-sectional study, data collected in 2012–2013 under a parent prospective health behavior study Study purpose: Examine financial stress as a potential mechanism that links smoking and food insecurity Location: Dallas, Texas, USA	238 adults recruited from the community 32% female 61% non-White	Food insecurity measured by 2-item screener Smoking measured by self-report of having smoked 100+ cigarettes in lifetime, and whether now smokes every day, some days, or not at all. Smoking status verified by expired carbon monoxide (CO) values ≥ 8–10 ppm	Among individuals with food insecurity, 42% were smokers. Among individuals without food insecurity, 18% were smokers. Smoking was significantly and positively associated with food insecurity; this association was significantly mediated by financial strain. Mediational model adjusted for age, sex, race, education and health insurance.
Hood (2013)	Cross sectional study, data collected in 2011 Study purpose: Characterize smoking behaviors and cessation-related interests among subsidized housing tenants Location: Ohio, USA	301 tenants in subsidized multi-unit housing 86% female 84% African American	Food insecurity risk measured by 2-item screener Smoking measured by self-report of having smoked 100+ cigarettes in lifetime, days per week smoked now, and number of cigarettes usually smoked on smoking days	Among current smokers, 59% were at risk of food insecurity. Among non-smokers, 41% were at the risk of food insecurity. Smoking increased the odds of food insecurity risk (OR=1.73), controlling for education, employment, health insurance, and physical limitations.
Iglesias-Rios (2015)	Cross-sectional study, from 1999–2008 National Health and Nutrition Examination Survey Study purpose: Examine food insecurity, smoking, and acculturation among Latinos Location: USA	6,681 Latino adults 49% female 37% born in US, 37% born in Mexico, 27% born in other Latin American country	Food insecurity measured by 10-item Food Security Survey; categorized as full, marginal, and low food security Smoking measured by self-report of having smoked 100+ cigarettes in lifetime, and whether now smokes every day, some days, or not at all.	Among individuals with full food security, 20% were current smokers, compared to 26% among marginally secure and 26% among low food secure. Current smoking increased the odds of marginal/low food security, compared to full food security (OR=1.51), adjusting for sex, age, educational attainment, poverty status defined by the poverty index ratio, marital status, survey year, and acculturation indicators (language spoken at home, years in US, nativity).
Jih (2018)	Cross-sectional study, from 2012/2013 Health Care and Nutrition Study, a supplemental survey of Health and Retirement Study Study purpose: To examine whether older adults with higher chronic disease burden are at increased risk of food insecurity Location: USA	3,552 community dwelling adults aged 50 years or older 59% female 69% White, 14% Black, 13% Latino, 4% Other	Food insecurity measured by 6-item Food Security Survey Smoking measured by self-report as a part of sociodemographic and health measures, measure not further specified	Among adults with food insecurity, 30% were current smokers; among adults without food insecurity, 15% were smokers. Current smoking was not associated with significantly higher odds of food insecurity, adjusting for number of chronic conditions, age, sex, race/ethnicity, marital status, highest education level, household size, housing type, employment status, wealth categories in quartiles, health insurance, self-rated health status, body mass index, and cost-related medication non-adherence.
Kim-Mozeleski (2018)	Longitudinal cohort study, data collected in 2011–2012 Study purpose: To prospectively examine the association of cigarette smoking and food insecurity in a cohort of persons living with HIV Location: San Francisco, California, USA	108 adults living with HIV engaged in vocational rehabilitation services 90% male 51% non-Hispanic White, 18% Hispanic/ Latino, 16% Black/African American, 16% Multiracial/Other	Food insecurity measured by 9-item Household Food Insecurity Access Scale, a continuous measure with a range of 0–27 Smoking measured by self-report of how often smoked cigarettes in past 3 months	Current smokers, compared to non-smokers, had significantly higher levels of food insecurity at baseline and at 12-month follow-up. Smoking status at baseline predicted food insecurity severity at 12-month follow-up (Beta=0.31, p=0.01) adjusting for food insecurity at baseline, age, sex, race/ethnicity, years since HIV diagnosis, disability status, employment status, substance use, hazardous drinking, depressive symptoms, mental health quality of life, and physical health quality of life.
Tolzman (2014)	Cross-sectional study Study purpose: To determine prevalence and predictors of food insecurity in a USDA-identified “food desert” Location: La Crosse, Wisconsin, USA	2,068 residents living in a designated food desert 72% female 89% White, 2% Black, 2% Hmong, 1% Native American, 7% Other	Food insecurity measured by US Department of Agriculture Community Food Security Assessment Toolkit Smoking measures not reported but categorized into daily, occasional, former, and never smokers	Prevalence of food insecurity was 30% for daily smokers; 28% for occasional smokers; 11% for former smokers; 7% for never smokers Current/occasional smoking increased odds of severe food insecurity (OR= 3.6), adjusting for age, race, income, education, health insurance status, home ownership, and employment status
Widome (2015)	Cross-sectional study Study purpose: To document prevalence and correlates of food insecurity among US veterans Location: USA	866 US veterans who served in Iraq and Afghanistan since 2001, participating in the NorthStar survey 45% female 90% Non-Hispanic White, 9% Non-White	Food insecurity measured by 6-item Food Security Survey, categorized as high/marginal, low, and very low food security Tobacco use measured via self-report of past 30 day use, categorized as none, some days, and every day	Among veterans with high/marginal food security, 19% were daily smokers and 13% were non-daily smokers. Among veterans with low food security, 34% were daily smokers and 11% were non-daily smokers. Among veterans with very low food security, 42% were daily smokers and 18% were non-daily smokers.a Daily tobacco use was associated with significantly increased odds of very low food security (OR=2.28), adjusting for marital status, income, number of children in household, general health status, and hours of sleep. In adjusted models, daily tobacco use was not significantly associated with low food security. Non-daily tobacco use was neither associated with low nor very low food security.

Notes: OR = odds ratio.

^aThe percentages given here were not reported in the article but were calculated based on the sample sizes given in Table 1 of this article.

Measures of Tobacco Use and of Food Insecurity

As part of the data charting, we examined instruments used to measure tobacco use and food insecurity. For tobacco use measures, 18 studies were specific to assessing cigarette smoking, and one study assessed tobacco use more generally (Widome, Jensen, Bangerter, & Fu, 2015). All 19 studies used self-report instruments; only one study biochemically verified self-reported smoking status via expired carbon monoxide values (Hernandez et al., 2017). There was notable variability in terms of the self-report instruments used to measure smoking status. Six studies used the standard two-item set of assessing whether the individual has smoked 100 or more cigarettes in entire lifetime, followed by an assessment of current smoking status (Castro, Heck, Forster, Widome, & Cubbin, 2015; Hernandez et al., 2017; Hood, Ferketich, Klein, Wewers, & Pirie, 2013; Hosler & Michaels, 2017; Iglesias-Rios, Bromberg, Moser, & Augustson, 2015; Kim & Tsoh, 2016). The remaining studies used various single-item measures assessing current smoking in reference to the past month, 3 months, or 6 months, or did not provide specific information on the measure used. Six studies discerned patterns of smoking, such as daily or non-daily smoking (Gucciardi, Vogt, DeMelo, & Stewart, 2009; Hood et al., 2013; Kim & Tsoh, 2016; Kim-Mozeleski et al., 2018; Tolzman, Rooney, Duquette, & Rees, 2014; Widome, Jensen, et al., 2015). We note that studies generally did not assess levels of nicotine dependence or readiness to quit.

For measures of food insecurity, ten studies used the U.S. Department of Agriculture (USDA) Adult/Household Food Security Survey modules which assesses food insecurity experienced in the past 12 months (Armour et al., 2008; Castro et al., 2015; Cutler-Triggs et al., 2008; Fitzgerald, Hromi-Fiedler, Segura-Pérez, & Pérez-Escamilla, 2011; Gucciardi et al., 2009; Iglesias-Rios et al., 2015; Jih et al., 2018; Kim, Flentje, Tsoh, & Riley, 2017; Kim-Mozeleski et al., 2018; Widome, Jensen, et al., 2015). However, the studies varied in whether they used the 18-item, a 10-item, or 6-item versions of the module. The remaining studies either used the 1- or 2-item food insecurity screeners which are based on the module, with the exception of one study that used the U.S. Agency for International Development’s Household Food Insecurity Access Scale (Kim-Mozeleski et al., 2018). The majority of the studies categorized responses dichotomously into food secure or food insecure, whereas three studies considered the severity of food insecurity (Cutler-Triggs et al., 2008; Kim et al., 2017; Perkett et al., 2016), such as severe food insecurity (or very low food security).

Food Insecurity’s Contribution to Tobacco Use

Table 1 summarizes the study characteristics and findings from six studies in which current smoking status was examined as a dependent variable in the analysis, and food insecurity was examined as an independent or predictor variable. All studies reported in Table 1 conducted analyses using multivariable logistic regression. There was a significant and independent association between food insecurity and odds of current smoking, with one study examining the odds of former smoking (quitting) among ever smoking mothers in California (Castro et al., 2015). In this group of articles, two studies reported more nuanced patterns. For instance, in addition to food insecurity being associated with significantly greater likelihood of current smoking, food insecurity was associated with significantly greater likelihood of daily smoking, but not non-daily smoking, among lower-income young adults living in California (Kim et al., 2017). In a two-year longitudinal analysis of women in San Francisco who were marginally housed, the severity of food insecurity, rather than any food insecurity, was associated with significantly greater likelihood of being a current smoker (Kim et al., 2017). It is notable that most of the studies in Table 1 reported odds ratios in the range of 1.5 to 1.9, whereas one study of adults using mobile food pantries in Delaware reported that food insecurity was associated with a nearly five-fold increase in the odds of current smoking (Perkett et al., 2016).

Tobacco Use’s Contribution to Food Insecurity

Table 2 summarizes the study characteristics and findings from 13 studies in which food insecurity was examined as a dependent variable and smoking was examined as an independent or predictor variable. In this group of studies, the data analysis also comprised largely of multivariable logistic regression, with two exceptions. One study tested a mediational or indirect effects model and reported that among community-dwelling adults in Dallas, Texas, financial strain significantly explained the positive association between smoking and food insecurity (Hernandez et al., 2017). Another study conducted a linear regression analysis that considered food insecurity as a continuous variable which was measured across two time points, and reported that smoking status at study baseline was predictive of food insecurity severity at a 12-month follow-up among persons living with HIV in San Francisco (Kim-Mozeleski et al., 2018).

Similar to the studies reported in Table 1, select studies included in Table 2 also reported more nuanced patterns. In a study of veterans, the association between smoking and food insecurity was only significant for veterans aged 65 years or older, whereas there was no significant association for veterans less than 65 years of age (Brostow, Gunzburger, & Thomas, 2017). In another study of veterans, daily tobacco use was associated with significantly higher odds of very low food security, but there was no significant association for non-daily tobacco use as an independent variable, nor low food security as a dependent variable (Widome, Jensen, et al., 2015). One study assessed smoking at the household level rather than the individual level (i.e., whether any adult in the household smokes anywhere inside the home) among participants in the 1999–2002 National Health and Nutrition Examination Survey, and found that smoking households had significantly greater odds of food insecurity (Cutler-Triggs et al., 2008). Notably, there was one study in this group which reported that there was no significant association between smoking and food insecurity in the multivariable analysis; among adults aged 50 or older in the 2012/2013 Health Care and Nutrition Study, there was a significant bivariate association between smoking and food insecurity that was no longer significant when adjusting for covariates (Jih et al., 2018).

Discussion

This scoping review examined the empirical literature published from 2008–2018 on the association between food insecurity and tobacco use, primarily cigarette smoking. About one-third of articles examined tobacco use as the outcome variable, and the remaining two-thirds examined food insecurity as the outcome variable. Therefore, it has been more common in the literature to examine tobacco use as an explanatory variable for food insecurity. Although most of the studies in this review were cross-sectional studies, and the three longitudinal studies were not necessarily designed to offer conclusions regarding causality, this body of research collectively contributes to knowledge that tobacco use and food insecurity are significantly associated with one another. Many health disparity populations were reflected in the study samples, as were important health care needs such as diabetes. Across study samples, the association between food insecurity and tobacco use were independent of important sociodemographic characteristics such as education and income, as well as factors related to substance use and mental health in the several studies that included such measures. The findings particularly emphasize the disproportionate burden of tobacco use—particularly cigarette smoking—among population groups experiencing food insecurity, and conversely, the disproportionate burden of food insecurity among tobacco users compared to non-users.

An important objective of this review was to gather plausible mechanistic explanations for the reciprocal association, towards informing future directions in this growing area of research. Figure 2 depicts a working conceptual model of the potential bidirectional pathways. In considering the pathways linking food insecurity with cigarette smoking, psychological stress and negative emotions were mentioned across a number of studies covered in this review (Castro et al., 2015; Gucciardi et al., 2009; Iglesias-Rios et al., 2015; Kim & Tsoh, 2016). Whereas over a third of the studies assessed various mental health-related variables such as psychological distress, depressive symptoms, and mental health-related quality of life (Bekele et al., 2018; Brostow et al., 2017; Fitzgerald et al., 2011; Hosler & Michaels, 2017; Kim et al., 2017; Kim & Tsoh, 2016; Kim-Mozeleski et al., 2018), such variables were mostly regarded as statistical controls to conclude that food insecurity’s effect on tobacco use was independent of mental health measures. Worrying about running out of food is an important screening question for food insecurity, and qualitative findings have delineated the emotional hardship of food insecurity (Canales, Coffey, & Moore, 2015), suggesting a direct or causal effect of food insecurity on one’s perceptions of stress, worry, and anxiety. It is also commonly known that smokers report smoking for stress and tension relief and to regulate overall mood (Berlin et al., 2003; Leventhal & Zvolensky, 2015), and therefore it is plausible that stress is a key mechanism that links food insecurity and tobacco use, particularly cigarette use.

Figure 2.

A Working Conceptual Model on the Intersection of Tobacco Use and Food Insecurity

In addition to a psychological pathway, several studies discussed the possibility of a physiological pathway related to nicotine’s effects on hunger and appetite (Iglesias-Rios et al., 2015; Kim et al., 2017; Kim & Tsoh, 2016). Key indicators of food insecurity include cutting or skipping meals, eating less than what would be considered enough, feeling hungry, and losing weight due to lack of money or food-related resources. Prior to 2006, the severity of food insecurity was formally differentiated as occurring with hunger or without hunger; while these terminologies are no longer formally used, food insecurity remains an important risk factor for hunger (National Research Council, 2006). Appetite and hunger suppression is one of the physiological effects of nicotine (Jo, Talmage, & Role, 2002), and nicotine withdrawal is associated with increased eating (Benowitz, 2008). Taken together, this raises the possibility that depending on the severity of food insecurity, smokers experiencing food insecurity use cigarettes to cope with hunger. It is possible that there is also behavioral reinforcement over time of nicotine’s effects on hunger, as suggested by patterns relating to food insecurity and daily smoking (Kim & Tsoh, 2016), and considering that severe food insecurity, but not mild or moderate food insecurity, was associated with smoking (Kim et al., 2017).

In considering the pathways linking tobacco use with food insecurity (Figure 2), numerous studies raised the possibility that tobacco-related expenditures create or exacerbate financial strain by competing with spending on household necessities (Armour et al., 2008; Hernandez et al., 2017; Iglesias-Rios et al., 2015; Kim et al., 2017; Kim-Mozeleski et al., 2018; Widome, Jensen, et al., 2015). In a study of veterans, tobacco use patterns were correlated with the degree of food insecurity, such that daily tobacco use increased the odds of severe food insecurity (Widome, Jensen, et al., 2015). However, non-daily tobacco use had no significant association with food insecurity, whether it was severe food insecurity or any food insecurity. This suggests that more frequent tobacco use worsens food insecurity. One study discussed that dietary behaviors and food acquisition practices may differ among smoking versus non-smoking households (Cutler-Triggs et al., 2008), which parallels the literature on smoking-induced deprivation (Siahpush et al., 2012). Another study raised the possibility that specific to persons living with HIV, for whom tobacco use may be particularly harmful to health, tobacco use contributes to food insecurity by negatively influencing one’s health status, thereby limiting workforce participation (Kim-Mozeleski et al., 2018).

Alongside these individual-level pathways, studies also applied social ecological lenses to acknowledge the role of structural factors that can impact either tobacco use, food insecurity, or both. Gendered experiences were emphasized by studies focusing exclusively on women (Castro et al., 2015; Fitzgerald et al., 2011; Kim et al., 2017). Underlying inequities across race/ethnicity, socioeconomic status, immigration status, and acculturation were highlighted in several studies (Cutler-Triggs et al., 2008; Fitzgerald et al., 2011; Iglesias-Rios et al., 2015), and discussions of built environment and neighborhood-level access to healthy foods were also raised (Hosler & Michaels, 2017). Children in low-income households are exposed to both food insecurity and to secondhand smoke (Cutler-Triggs et al., 2008). As tobacco use becomes increasingly concentrated among populations that are underserved or marginalized, there is a need to consider how multiple health risk factors converge to create a double burden of tobacco use and food insecurity that adversely affects the physical and mental health of affected populations across the developmental spectrum. We emphasize that the conceptual model depicted in Figure 2 is a working model, not meant to be comprehensive of all possible pathways and factors, but that these were candidate pathways and factors raised in the articles that were included in this review.

Limitations

There are several limitations to this review. This review was inclusive of articles that were published in peer-reviewed journals between 2008 to mid-2018, as captured by relevant search terms using three widely used databases. We may not have captured all relevant articles published within the study time frame, nor were we able to include studies that were published since mid-2018. We chose 2008 as the starting point because to our knowledge, the first two studies on the topic of food insecurity and tobacco use were both published in 2008, but it is possible that there were earlier studies that were not included here. Our search terms were specific to the concept of food insecurity/security (including hunger), and therefore we were not able to include studies that would be relevant when considering food insecurity in the broader context of economic insecurity. We also restricted the review to studies in the U.S. and Canada, based on an understanding that the context of tobacco use and of food insecurity can vary widely. As this was a scoping review, we did not evaluate individual studies for the quality of the evidence, evidence of bias, or conduct a meta-analysis of the quantitative findings, which would be relevant to a systematic review. Despite these limitations, to our knowledge, this article provides the most comprehensive review of this area of work to date.

Implications

This review offers several directions and implications for health promotion research and practice. The studies included here largely did not assess important characteristics related to tobacco use, such as readiness to quit and levels of nicotine dependence, nor did they assess characteristics related to food insecurity, such as recency or frequency of the experience (e.g., when or how often in the past 12 months one experienced food insecurity). By the current state of the research, the temporal direction between food insecurity and tobacco use is not well understood, as in whether tobacco use precedes food insecurity or vice versa. It is important to consider that while the majority of smokers initiate smoking during their youth, socioeconomic disparities in smoking tend to emerge during the early adulthood years (Pampel, Mollborn, & Lawrence, 2014). Taken together, applying a life course perspective to investigate the intersection of food insecurity and tobacco use may inform prevention of tobacco-related health disparities. Greater specificity of quantitative measures, and buttressing quantitative studies with qualitative investigations, may also provide important contextual data to identify key next steps for research.

A predominant finding of the current review is that the relationship between tobacco use and food insecurity is robust across study settings and samples. The high prevalence of smoking in persons experiencing food insecurity—and conversely, the high prevalence of food insecurity among smokers—warrants concerted attention given the rising burden of chronic diseases. There are several possibilities for extending health promotion into sectors that are not traditionally focused on health. For instance, the Supplemental Nutrition Assistance Program (SNAP) is the largest federal safety net for food insecurity, and program participation has been demonstrated to reduce levels of household food insecurity (Mabli & Ohls, 2015) as well as psychological distress (Oddo & Mabli, 2015). Alongside supporting existing tobacco control strategies, such as smoke free housing policies and media campaigns, considering partnerships with SNAP, the Special Nutrition Assistance Program for Women, Infants, and Children (WIC), and other social service infrastructures as a means to reach smokers experiencing food insecurity may provide important opportunities to engage with smokers as they seek assistance with food, and offering assistance with cessation.

Community-based food assistance settings (e.g., food pantries, community meal programs, and soup kitchens) may also provide important opportunities for linkages. There are already numerous health promotion programs and interventions being conducted in food pantries, but they largely focus on improving nutrition education and dietary behaviors related to managing chronic diseases (An et al., 2019; Dave, Thompson, Svendsen-Sanchez, McNeill, & Jibaja-Weiss, 2017). Addressing tobacco use is traditionally not a focus of food assistance programs, yet the findings presented in this review illustrate that tobacco use can exacerbate food insecurity. Leveraging food assistance settings to conduct outreach for tobacco cessation may reach low-income smokers who are underserved (Perkett et al., 2016).

More than half of households who use food assistance programs do so on a regular basis (Weinfield et al., 2014), with many relying on emergency food programs as a long-term strategy for procuring food (Holmes et al., 2018). Without undermining the critical need to promote food security through structural and policy level interventions, the current reality presents repeated engagement opportunities with individual smokers to conduct tobacco cessation outreach. For instance, this may involve expanding the Ask-Advise-Connect model—which proactively connects smokers to tobacco quitlines (Vidrine et al., 2013)—to outside of the health care setting and into community-based settings that address food insecurity. The role of lay or community health workers in delivering tobacco cessation has been previously articulated (Lautner et al., 2019; Tsoh et al., 2015), and extending this model to food assistance programs may also be a viable and scalable strategy. In parallel, conducting food insecurity screening and providing food assistance referrals for individuals seeking smoking cessation services would identify individuals who may be contending with food and other unmet needs that can complicate efforts to quit. For example, food insecurity may exacerbate common nicotine withdrawal symptoms (e.g., increased anxiety, greater worries about food and managing appetite) which may be addressed through more tailored methods.

Towards reducing the disproportionate burden of smoking on low-income populations, focusing on the role of food insecurity has implications for developing, implementing, and testing approaches that are scalable and population-based. This review has also identified key knowledge gaps, as much remains to be examined and learned towards informing reductions in tobacco use and promotion of food security. Furthermore, future studies are warranted to investigate the reciprocal association, in terms of the extent to which tobacco cessation decreases food insecurity, and the extent to which tobacco cessation outcomes are improved by addressing food insecurity.

Author bios:

Jin Kim-Mozeleski, PhD, is Assistant Professor in the Department of Health Promotion and Policy at the University of Massachusetts, Amherst, USA. Rajshree Pandey, PhD, MPH, was a graduate student in the Department of Biostatistics and Epidemiology at the University of Massachusetts Amherst, at the time of this writing.