Sugar-sweetened beverage intake and convenience store shopping as mediators of the food insecurity–Tooth decay relationship among low-income children in Washington state

Abstract

Introduction

There are oral health disparities in the U.S. and children in food-insecure households have a higher burden of tooth decay. Identifying the mechanisms underlying the food insecurity–tooth decay relationship could inform public health interventions. This study examined how sugar-sweetened beverage (SSB) intake and frequent convenience store shopping mediated the food insecurity–tooth decay relationship for lower-income children.

Materials and methods

Cross-sectional study data included a household survey, beverage questionnaire, and dental examination. The sample included 452 lower-income, racially-diverse, child-caregiver dyads in 2018 from King County in Washington state. The exposure was household food insecurity, the outcome was untreated decayed tooth surfaces, and the proposed mediators were SSB intake and frequent convenience store shopping (≥2 times/week). Causal mediation analyses via the potential outcomes framework was used to estimate natural indirect and direct effects.

Results

Fifty-five percent of participants were in food-insecure households, the mean number of decayed tooth surfaces among children was 0.87 (standard deviation [SD] = 1.99), the mean SSB intake was 17 fluid ounces (fl/oz)/day (SD = 35), and 18% of households frequently shopped at a convenience store. After adjusting for confounders, household food insecurity and log-transformed SSB intake (fluid ounces/day) were positively associated with decayed tooth surfaces, but not at the a α = 0.05 level (mean ratio [MR] 1.60; 95% confidence interval [CI] 0.89, 2.88; p = .12 and MR 1.16; 95% CI 0.93, 1.46; p = .19, respectively). Frequent convenience store shopping was associated with 2.75 times more decayed tooth surfaces (95% CI 1.61, 4.67; p < .001). SSB intake mediated 10% of the food insecurity–tooth decay relationship (p = .35) and frequent convenience store shopping mediated 22% (p = .33).

Conclusions

Interventions aimed at addressing oral health disparities in children in food-insecure households could potentially focus on reducing intake of SSBs and improving access to healthful foods in lower-income communities.

Introduction

Tooth decay is highly prevalent among U.S. children and disproportionately impacts socioeconomically-disadvantaged children [1]. Hispanic children, Black children, and children in low-income households are at particularly high-risk for tooth decay [2,3]. The high burden of tooth decay in these subgroups stem from underlying structural and systemic inequalities such as poor access to dental insurance, quality dental care, and nutritious foods [4], inequalities in wealth and income [5], and racism [6]. Consequences of childhood tooth decay include poor overall health, pain, missed school days, lower quality of life, and costly health care needs [7–9]. Previous studies have reported that household food insecurity is an important risk factor for tooth decay in children [10–15]. According to the United States Department of Agriculture, food is a household-level economic and social condition of limited or uncertain access to adequate food [16]. Understanding the mechanisms underlying the household food insecurity–tooth decay relationship could identify the most relevant pathways to target when developing public health interventions aimed at preventing tooth decay and improving the oral health and overall health of low-income children.

Few studies have considered the pathway for the household food insecurity–tooth decay relationship. The one study aimed to examine diet quality as a potential mediator using a large nationally-representative sample of U.S. children but did not conduct mediation analysis because the association between overall diet quality and tooth decay was not statistically significant, in accordance with the traditional Baron and Kenny approach to mediation [14]. Other potential mediators include specific components of diet (e.g., sugar-sweetened beverage [SSB] intake) and behaviors associated with diet quality (e.g., food store shopping), although these have not been considered before in the literature. Excess intake of added sugars, including those from SSBs, is causally associated with the development of tooth decay, because sugar acts as a substrate for oral bacteria that cause tooth decay [17,18]. Children with low socioeconomic status (SES) and children in food-insecure households tend to have higher levels of SSB intake compared to children with higher SES and children in food-secure households [17,19–23].

Frequent convenience store shopping is a potential mediator of the relationship between food insecurity and tooth decay because convenience stores are an important setting where excess added sugars and SSBs are purchased. Research has shown that the most common convenience store purchase is soda [24,25], convenience store purchases frequently exceed recommendations for added sugar intake [25], and convenience stores are the most common setting that low SES populations purchase SSBs [26,27]. In addition, food-insecure households are more likely to shop at convenience stores [28–30]. This results from both limited access to other food store options and from food assistance program cycling where shopping behaviors shift to affordable food options when food money runs out [31]. If analyses show that convenience store shopping is a relevant mediator, public health interventions could modify the food environment by improving access to affordable and healthy foods in convenience stores and by changing shopping and purchasing behaviors in convenience stores.

In this study, we leverage causal mediation methods to better understand the mechanisms underlying the household food insecurity–tooth decay relationship. The aim was to determine whether SSB intake and frequent convenience store shopping mediate the food insecurity–tooth decay relationship for lower-income children. Mediation analysis was used in this analysis because it is a tool used to discover and test possible causal relationships by teasing apart direct effects of an exposure on an outcome from indirect effects that occur through a third mediator variable. Findings can inform public health efforts aimed at eliminating oral health disparities among children with low SES.

Materials and methods

Study design and data source

This study used cross-sectional data from a racially-diverse sample of children and caregivers from lower-income households in Seattle and South King County in Washington state. The sample was based on the SeaSAW study (a cohort study designed to evaluate the Seattle SSB tax) [32]. Data collection for SeaSAW including the household survey and beverage consumption questionnaire began in October 2017 and was completed in January 2018 with child dental examinations occurring between April and June 2018. Data were collected in a period for which the enactment of the SSB tax would not influence the variables of interest (the tax went into effect January 2018). Written informed consent was obtained from the caregivers of study participants. The study was approved by Seattle Children’s Hospital Institutional Review Board. Authors had access to information that identify individual participants during and after data collection.

Study population

Children-parent dyads who participated in the SeaSAW study who met study eligibility criteria for the dental examination were invited to participate in this additional oral health study. We also recruited friends and acquaintances of SeaSAW participants, and other members of Spanish and Somali communities with the help of community organizers. There were five eligibility criteria for participation in the additional study: (1) the child was in 1^st to 9^th grade (5 to 16 years old); (2) the child consumed SSBs based on parent report; (3) the annual household income was <312% of the Federal Poverty Level (FPL); (4) the caregiver was ≥18 years old; and (5) the caregiver spoke English, Spanish, Somali or Vietnamese. The screening item about SSBs was, “Does your child ever drink sugary beverages like: regular soda/pop (such as Coke or Sprite), fruit-flavored drinks (like Sunny Delight), coffee or tea drinks with added sugar (like Starbucks Frappucinnos, Arizona Iced Tea, Chai Tea, bubble tea), or regular sports drinks or energy drinks (such as Gatorade or Red Bull)?”. This screening item was different than the main study measure of SSB intake which only captures intake in the past 30 days so it possible that some participants had an affirmative response to the screening item but did not report SSB in the past 30 days in later study measured. Defining low-income as <312% of the Federal Poverty Level corresponds to cut-off points used by Washington state public health insurance programs. The SeaSAW study had slightly different child age criterion; namely, children were required to be either 7–10 or 12–17 years old, but otherwise had the same inclusion criteria [33]. We recruited and limited the sample to be from communities with fluoridated water to eliminate the possibility that community water fluoridation affected study results because water fluoridation is known to weaken the positive associations between added sugar intake and tooth decay [34–36].

Study variables

Exposure.

The study exposure, food insecurity, was assessed with a 2-item screener using a “last 12-months” reference period [37]. Compared to the U.S. Household Food Security Survey Module: Six-Item Short Form [38], the 2-item screener has 95.4% sensitivity and 83.5% specificity [39]. Households with an affirmative response to either item were considered food-insecure.

Mediators.

Potential mediators of interest were child SSB intake and frequent convenience store shopping (Fig 1). SSB intake was measured with a validated beverage intake questionnaire (BEV-Q) that was modified to differentiate between sugary beverages that were soon to be taxed within Seattle (e.g., bottled sports drinks with sugar versus sports drink powder with sugar used to constitute the beverage) [40–42]. The BEV-Q asks about intake (typical frequency [ranging from never or less than 1 time/week to 3+ times/day] and volume [ranging from less than 6 fluid ounces to more than 20 fluid ounces] in the past 30 days) of different types of sugary and non-sugary beverages. The caregiver completed the BEV-Q regarding their child’s beverage consumption for children <11 years old. Children ≥ 11 years old completed their own BEV-Q regarding their consumption. The BEV-Q assessment of SSB intake was tested in adolescents and shown to have high validity against multiple 24-hour dietary recalls, high reliability across repeat measurements, and readability scores appropriate for individuals with a fourth grade education or higher [40–42]. To score the BEV-Q, frequency of intake is converted to the unit of times/day, then multiplied by the typical amount consumed to provide average daily beverage intake in fluid ounces (fl oz) [41]. Intake of each SSB item (including SSBs that were and were not subject to the Seattle SSB tax) was summed to create one variable for total SSB intake in units of fl oz/day. SSB intake was then assessed two ways: log-transformed SSB intake (natural log(x+1)) to account for a right skew and to include zero values, and as a categorical variable (0 fl oz/day, >0 fl oz/day to 8 fl oz/day, >8 fl oz/day to 16 fl oz/day, and >16 fl oz/day). SSB intake was also examined as a binary variable in sensitivity analyses (>0 fl oz/day vs. 0 fl oz/day).

[Figure omitted. See PDF.]

Frequency of convenience store shopping was measured with five possible responses (never, <1 time/month, 1 time/month, 1 time every other week, 1 time/week, ≥2 times/week). It was categorized into three groups (never or <1 time/month, 1 time/month to 1 time/week, and ≥2 times/week. Frequent convenience store shopping was defined as ≥2 times/week.

Outcome.

The study outcome was tooth decay operationalized as a count variable based on the number of surfaces with untreated decay on primary or permanent teeth. Untreated tooth decay was the main outcome because it reflects current disease status; including teeth missing due to disease and filled surfaces may capture disease that occurred before the exposures and mediators of interest in this study. However, as a sensitivity analyses, the number of decayed, missing, and filled tooth surfaces (DMFS) was also assessed as an outcome. The International Caries Detection and Assessment System (ICDAS, version 2 or ‘ICDAS’) was followed for dental examinations [43,44]. A lesion code of ≥3 was the cut off to identify decayed tooth surfaces.

All dental examiners (dentist or dental hygienist) were calibrated in the ICDAS during an intensive 3-day training class which included didactic materials and practice examinations. Duplicate examinations on a sample of study participants were conducted to estimate inter-rater reliability. About 55% of the sample received an examination by both a gold standard examiner and another examiner. Inter-rater agreement was high, >98% surfaces were correctly coded. Cohen’s kappa for all carious surfaces was 0.85 (0.87 for primary and 0.73 for permanent surfaces). These values are classified as substantial to almost perfect [45] and acceptable and like other studies using ICDAS [44,46].

Confounders.

Confounders were variables hypothesized to be associated with the exposure and outcome that do not lie on the causal pathway between food insecurity and tooth decay [47]. We also identified exposure-mediator confounders and mediator-outcome confounders. The minimum set of confounders to adjust for were identified using directed acyclic graphs (DAGs) (S1 File). We adjusted models for the following potential confounders: child age, child race, child ethnicity, child insurance type, caregiver highest level of education, annual household income, and food assistance use (defined as participation in either Supplemental Nutrition Assistance Program [SNAP] or Special Supplemental Nutrition Program for Women, Infants, and Children [WIC].). These variables were based on the caregiver report on the household survey. Although race and ethnicity were used as confounders in this study, we acknowledge that it is racism and discrimination, which were not measured, and not race and ethnicity, that shape an individual’s exposures and opportunities [48,49]. In this study, exposure to racism and discrimination are likely determinants of food environment, income, food insecurity, education, and employment opportunities [48,49].

Statistical analyses

We generated descriptive statistics (mean, standard deviation [SD] or n, %) for participant characteristics. Because the distributions of both tooth decay and SSB intake were skewed, we also report median and interquartile range (IQR). In separate Poisson regression models, we examined the associations between 1) household food insecurity and number of decayed tooth surfaces, 2) SSB intake and number of decayed tooth surfaces, and 3) convenience store shopping number of decayed tooth surfaces. Each model was adjusted for confounders and confidence intervals (CI) were generated with robust standard errors. Because of the age range of the study population, there was a combination of primary, permanent, and mixed dentitions, so all multivariable analyses modeling tooth decay as an outcome included adjustment for the number of tooth surfaces as a covariate in the regression model.

We used causal mediation analysis based on the potential outcomes framework [50,51] to decompose indirect and direct pathways between household food insecurity and tooth decay. We estimated two mediation effects: 1) the natural indirect effect which measures the effect of the exposure (e.g., food security) that operates through the mediator (e.g., SSB intake or frequent convenience store shopping); and 2) the natural direct effect, which measures the impact of the exposure if it did not cause the mediator [52]. The proportion mediated, which can be interpreted as the overall impact of food insecurity on tooth decay attributable to the impact of food insecurity on the mediator, was estimated and expressed as a percentage.

All mediation estimates were generated using the quasi-Bayesian approach developed by Imai et al. that uses Monte Carlo simulations based on normal approximations [50,51]. In the quasi-Bayesian approach, the natural indirect effect and natural direct effect are calculated by estimating the difference in the observed and counterfactual outcomes across many simulations. We used 1,000 simulations to generate the estimates and robust standard errors were used to generate 95% CI. The mediation estimates for are reported on the additive scale (mean differences in the number of decayed tooth surfaces) at the mean level for quantitative confounders or the most frequent level for categorical confounders. For the mediation analysis, we fit a linear regression model for natural log-transformed SSB intake and a Poisson regression model for frequent convenience store shopping conditional on household food insecurity and confounders. We then fit a Poisson regression model for number of untreated decayed surfaces conditional on household food insecurity, the mediator of interest, and confounders (S2 File).

There was minor missingness in the analytical variables (<6%); observations with missing data for the analytical variables were removed for a complete-case analysis. All analysis was conducted in R version 4.03 using the mediation package [50,53].

Statistical assumptions and sensitivity analyses.

Causal mediation analysis requires a number of assumptions for which sensitivity analyses have been developed. First, causal mediation analysis allows for an exposure-mediator interaction, so as a sensitivity analysis, we included an exposure-mediator interaction term in the outcome models to examine whether the natural effect estimates changed [54]. The results of this analysis informed the decision of whether to include exposure-mediator interaction term in all regression models in the main analysis. Second, mediation analysis depends on the assumption of no unmeasured confounding between the exposure and mediator, exposure and outcome, and mediator and outcome (sequential ignorability). Imai et al. have proposed sensitivity analyses to examine this assumption [50,51] but they are not currently supported for count outcome models (i.e., Poisson regression). The main analysis was repeated with DMFS as an outcome. Because we have two proposed mediators in this study, we assumed they were independent for simplicity of the analysis. However, this assumption of independence may not be valid given that convenience stores shopping and SSB intake are likely associated. Violations in this assumption would result in unaddressed confounding in the mediation estimates, which cannot be corrected for using basic mediation methods due to collider stratification bias [55]. Based on descriptive results that suggested that SSB intake may not have a linear relationship with tooth decay, we ran an additional analysis examining the SSB intake mediator as a binary variable.

Results

Sample characteristics

Of the 452 study participants, 27 were excluded for missing data on the exposure variable (food insecurity). The final sample size was 425. The mean age of children in the study was 10 years (SD = 3), 53% were female, 58% were Black or African-American, 33% were Hispanic, 84% were publicly insured, and 13% did not have any dental care in the past 12 months (Table 1). About 18% of caregivers did not complete high school, 24% completed high school/received a GED, 32% completed some college or vocational training, and 25% completed college or a higher degree. Annual household income was less than $24,000 for 40% of the sample, $24,000 to $48,000 for 38% of the sample, and more than $48,000 for 25% of the sample. Nearly one-half (44%) of caregivers used food assistance.

[Figure omitted. See PDF.]

About one-half (55%) of the children lived in a food-insecure household. The mean number of decayed tooth surfaces was 0.87 (SD = 1.99; median = 0; IQR = 0,1) and 29% had at least one decayed tooth surface. The mean intake of SSB intake was 17 fl oz/day (SD = 35; median = 7; IQR = 3, 17. About 10% reported 0 fl oz of SSB intake per day, 42% had >0 to 8 fl oz/day, 21% had >8 to 16 fl oz/day, and 26% had >16 fl oz/day. Almost one-half of the sample (45%) reported that they shopped at convenience store 0 to <1 times/month and 37% reported that they shopped at a convenience store 1 time/month to 1 time/week. About 18% of participants reported that they frequently shopped at a convenience store (≥2 times/week).

Children in food-insecure households tended to have higher SSB intake and had a higher prevalence of shopping at a convenience ≥2 times/week than children in food-secure households (23 fl oz/day vs. 11 fl oz/day and 24% vs. 11%, respectively). In addition, living in a food-insecure household was more common among children whose caregivers identified them as Black or African-American, among households with lower annual incomes, and among households who used food assistance.

Association of food insecurity, SSB intake, and frequent convenience store shopping with tooth decay

After adjusting for confounders, children in food-insecure households had 1.60 times more decayed tooth surfaces than children in food-secure households (95% CI 0.89, 2.88; p = .12) (Table 2). Log-transformed SSB intake was also associated with decayed tooth surfaces (mean ratio 1.16; 95% 0.93, 1.46; p = .19). Compared to children who reported 0 fl/oz of SSB intake per day, those who reported >16 fl oz/day had 1.70 more decayed tooth surfaces (95% CI 0.70, 4.15; p = .24). Children in households that frequently shopped at a convenience store (≥2 times/week) had 2.75 times more decayed tooth surfaces than children in households that shopped <2 times/week at a convenience stores (95% CI 1.61, 4.67; p = < .001).

[Figure omitted. See PDF.]

Mediation analyses

The natural indirect effect of the impact of household food insecurity on tooth decay mediated through log-transformed SSB intake was 0.05 decayed tooth surfaces (95% -0.06, 0.22; p = .35), which indicates that about 9.9% of the influence of food insecurity on tooth decay occurred through SSB intake (Table 3). The natural indirect effect of the impact of household food insecurity on tooth decay mediated through frequent convenience store shopping was 0.64 decayed tooth surfaces (95% CI -0.29, 1.68; p = .33), which indicates that about 22.4% of the food insecurity–tooth decay relationship is mediated through frequent convenience store shopping.

[Figure omitted. See PDF.]

Sensitivity analyses

Including an interaction between exposure and mediators in the mediation analyses did not meaningfully influence the mediation estimates, so an interaction term was omitted from all main analyses (S1 Table). Food insecurity, SSB intake, and frequent convenience stores shopping were not associated with DMFS (S2 and S3 Tables).

Discussion

In this study, we examined whether the household food insecurity–tooth decay relationship was mediated by SSB intake and frequent convenience store shopping for children in lower-income households in King County, Washington state. We found that household food insecurity, SSB intake, and frequent convenience store shopping were associated with higher levels of mean decayed tooth surfaces. Specifically, children in lower-income households that reported shopping at a convenience store ≥2 times/week had 2.4 times more decayed tooth surfaces than children in low-income households that shopped at a convenience store less frequently. According to mediation analyses, SSB intake mediated 10% of the impact of food insecurity on tooth decay and frequent convenience store shopping mediated about 22%, although estimates were not statistically significant. Children with low SES are at a high risk for tooth decay and evidence-based public health interventions are needed to address disparities. Potential interventions include food insecurity prevention, SSB taxes, and community education programs that improve food and beverage purchasing decisions in tandem with structural changes that improve access to healthy and affordable food and beverage options in low-resource communities.

Consistent with the literature that has reported that household food insecurity is a risk factor for childhood tooth decay [10–15], we found that household food-insecurity was associated with 1.6 times more decayed tooth surfaces among the low-income children in this study. We also found that higher levels of SSB intake tended to be associated with higher levels of tooth decay, which aligns with the biological understanding of how sugar acts as a substrate for bacteria that cause tooth decay [17]. One new finding in this study was that frequent convenience store shopping was strongly associated with tooth decay. Past research has suggested that convenience store shopping may have a modest but positive association with other adverse chronic health outcomes in children, such as obesity [56–59]. Convenience stores may be a particularly relevant setting to target for interventions aimed at addressing childhood oral health disparities because it is a high-sugar environment and a common source of not only SSBs [24,25], but also of nutrient-poor foods high in added sugars [60]. Targeting behaviors alone is not adequate as structural factors also shape the distribution of convenience stores and supermarkets and the price of healthy foods in communities. Policies should aim to improve access to a variety of healthy and affordable foods options in all stores in low-income communities.

Although estimated mediation effects were not statistically significant in this study, findings suggested that SSB intake and frequent convenience store shopping transmitted about 10% and 22% of the impact of food insecurity onto tooth decay, respectively. The one known study that examined mediators in the household food insecurity–tooth decay relationship tested diet quality but did not complete mediation analyses because the association between diet quality and tooth decay was not statistically significant [14]. According to previous research, food-insecure households have higher SSB intake [17,19–23] and shop at convenience stores more frequently [28–30], which provides support that each may be a relevant mediator in the food insecurity–tooth decay relationship. However, the lack of statistically significant mediation estimates in this study could mean that SSB intake and frequent convenience store shopping are not truly mediators and the effect of food insecurity on tooth decay occurs via other mechanisms. For example, the impact of food security on tooth decay could be attributed to socioeconomic pathways, given that food security is a marker of SES. It is also likely that the lack of statistically significant mediation effects in this study could be explained by the relatively small study sample size that had insufficient power to detect precise mediation estimates given that confidence intervals were wide but mostly covered effect sizes that would indicate a true relationship in the hypothesized direction. Given the lack of certainty in the study findings, follow-up studies using sufficient sample sizes and a longitudinal study design are needed to verify whether SSB intake and frequent convenience store shopping are truly mediators in the food insecurity–tooth decay relationship.

Regardless of whether SSB intake and frequent convenience store shopping are relevant mediators of the food insecurity–tooth decay relationship, they were each associated with tooth decay which indicates there are opportunities to address oral health disparities through policy and program interventions that impact SSB intake and access to healthy foods at convenience stores. One policy intervention is SSB taxes which not only aim to reduce SSB intake but also divert tax revenue to fund programs for nutrition improvement in low-income populations [61]. According to a meta-analysis of SSB taxes across 62 studies in cities and national governments, SSB taxes were associated with reductions in SSB sales [62]. In Seattle, Washington, which is a region captured in the current study, a SSB tax in January 2018 led to a 22% decline in the volume of SSBs purchased [63]. However, there was no corresponding decrease in SSB intake directly associated with the tax [64]. Regardless, in the two-year period after the tax was enacted, support for the tax increased among low-income residents [65] and the perceived negative health impacts of SSBs increased among low-income residents [66], which suggests that SSB taxes could have positive spillover effects on the incidence of tooth decay among Seattle children. While no experimental studies have examined whether SSB taxes improve oral health, numerous simulation and model-based studies have predicted the potential public health benefits of a SSB tax on oral health outcomes [67]. Predictions in high-income countries like the U.S. show that SSB taxes could lead to reductions in the incidence of tooth decay and reduced spending on treatment costs [67]. An experimental study could identify whether SSB taxes have tangible impacts on tooth decay and can inform evidence-based policy decisions.

There is also opportunity to intervene on oral health outcomes through the food environment [68,69]. While the association of neighborhood access to convenience stores and child SSB intake is equivocal [70–72] and many families travel outside of their neighborhoods to obtain food [73], low-resource neighborhoods still offer the fewest food-purchasing establishments and convenience stores are common [26,74]. In addition, food-insecure households are more likely than food-secure households to not own their own vehicle to use for shopping (they rely on others for transportation or public transportation) and travel shorter distances when shopping [28–30], which means that the neighborhood food environment may be particularly relevant for food-insecure families. At least one study in Pittsburgh, Pennsylvania, has shown added sugar intake dropped after a supermarket was opened in a community with limited access to healthy foods [75]. Ensuring the availability of a variety of food-purchasing establishments that stock healthy and affordable food and beverage options is a step to support healthy food purchasing in low-resource communities. However, additional efforts are also necessary. These efforts include enforcement of living wages, regulations on food advertising, and population-based education programs that publicize the potential negative effects of excess added sugar intake [76], provide education about healthier food and beverage choices, meal preparation resources, and information about how to introduce new foods to picky children.

While not statistically significant, our study findings suggested that about 11% and 20% of the impact of food insecurity on tooth decay was transmitted through SSB intake and frequent convenience store shopping, respectively. Therefore, interventions that reduce SSB intake and promote intake of healthy foods could be a particularly effective way to address oral health disparities in food-insecure communities. For instance, SNAP participants can purchase SSBs with benefits and as of 2016, SSBs are among the ten most commonly-purchased items among SNAP participants [77]. Simulation research suggests that SSB-purchasing restrictions in SNAP could lead to an 8% decrease in the number of decayed teeth among children in SNAP-participating households [78]. Furthermore, policies that implement incentives for healthy food in addition to SSB restrictions may prove to be the most in fruitful in making public health improvements [79]. However, SSB-purchasing restrictions in SNAP remain controversial and there is strong corporate lobbying against such policy changes [80]. Improving food purchasing habits in food-insecure households could also improve oral health disparities. For example, among food-insecure households participating in SNAP, point-of-sale incentives have been linked to higher vegetable and fruit expenditures [81–83].

Strengths and limitations

There are strengths to this study. First, the study focused on a lower-income racially-diverse sample of children, a population that experiences some of the greatest oral health disparities and stands to gain the most benefit from public health interventions. Second, we incorporated recently-developed causal mediation techniques. These approaches improve on traditional mediation techniques by allowing for an exposure-mediator interaction, analysis of a broader variety of exposures, mediators and outcomes, and moves away from a reliance on statistical significance which is required in the traditional Baron Kenny approach [84]. We also add to the growing literature that has identified food insecurity as a relevant oral health risk factor and attempt to understand underlying mechanisms which is key to developing successful evidence-based public health interventions.

Limitations should also be considered when interpreting the findings of this study. First, because our data are cross-sectional we cannot directly infer the temporality of our study measures. All published research reporting the food insecurity–tooth decay relationship has relied on cross-sectional data which emphasizes the need for a well-designed longitudinal study that can more definitively discern temporality. Second, the relatively small study sample size likely contributed to insufficient power for the mediation analyses and may be a primary explanation for the lack of statistically significant mediation findings. According to simulations conducted by Fritz and Mackinnon, at least 500 observations are needed to detect significant direct and indirect effects when studying small effects [85], such as those that would be expected when studying decayed tooth surfaces as an outcome. Future longitudinal and intervention studies should ensure that an adequate sample size is met which can be achieved using newly-developed power calculations for mediation analyses [86,87]. Third, it is possible there was measurement error in SSB intake due to social desirability bias, especially for adolescents who self-reported their intake and the measure of convenience store given that adults reporting shopping behaviors for the household may have not captured adolescent’s independent shopping behaviors. While the BEV-Q used in this study was found to be valid and reliable among adolescents [40–42], this potential measurement error should be considered when interpreting the study findings. Lastly, we were unable to conduct the sensitivity analyses that have been developed for causal mediation analyses because they are not currently supported for models with count outcomes. While we cannot examine how robust the present findings are to the assumptions of causal mediation analysis, future work should aim to incorporate sensitivity analyses as they are developed.

Conclusion

This study examined the extent to which SSB intake and frequent convenience store shopping mediated the household food insecurity–tooth decay relationship among a racially-diverse lower-income sample of children in King County, Washington state. Consistent with previous literature, household food insecurity and SSB intake were associated with higher levels of tooth decay. We also found that frequent convenience store shopping was associated with 2.4 times more decayed tooth surfaces. Although mediation estimates were not statistically significant which may have resulted from the relatively small study sample size, SSB intake and frequent convenience store mediated about 10% and 22% of the impact of household food insecurity on tooth decay. Concerted policy and public health efforts are needed to address tooth decay disparities among children with low SES, including enforcement of living wages, improvements in the access and affordability of healthy foods in low-income neighborhoods, SSB taxes, and community education efforts that improve food and beverage purchasing decisions in low-resource communities.

Supporting information

S1 Checklist. Table 1.

STROBE-nut: An extension of the STROBE statement for nutritional epidemiology.

https://doi.org/10.1371/journal.pone.0290287.s001

(DOCX)

S1 File. Confounder selection.

https://doi.org/10.1371/journal.pone.0290287.s002

(DOCX)

S2 File. Mediator models.

https://doi.org/10.1371/journal.pone.0290287.s003

(DOCX)

S1 Table. Mediating effects of SSB intake and frequent convenience store shopping in the household food insecurity–tooth decay relationship allowing for an interaction between food insecurity and mediator, for children, aged 5 to 16 years, in Seattle and South King County, 2018.

https://doi.org/10.1371/journal.pone.0290287.s004

(DOCX)

S2 Table. Number of DMFS on household food insecurity, SSB intake, and frequent convenience store shopping among children, aged 5 to 16 years, in Seattle and South King County, 2018.

https://doi.org/10.1371/journal.pone.0290287.s005

(DOCX)

S3 Table. Mediating effects of SSB intake and frequent convenience store shopping in the household food insecurity–DMFS relationship for children, aged 5 to 16 years, in Seattle and South King County, 2018.

https://doi.org/10.1371/journal.pone.0290287.s006

(DOCX)

S1 Data. SeaSAW_OH_data_restricted.

https://doi.org/10.1371/journal.pone.0290287.s007

(XLSX)

S2 Data. SeaSAW_OH_data_restricted_codebook.

https://doi.org/10.1371/journal.pone.0290287.s008

(XLSX)

Citation: Hill CM, Chi DL, Mancl LA, Jones-Smith JC, Chan N, Saelens BE, et al. (2023) Sugar-sweetened beverage intake and convenience store shopping as mediators of the food insecurity–Tooth decay relationship among low-income children in Washington state. PLoS ONE 18(9): e0290287. https://doi.org/10.1371/journal.pone.0290287

About the Authors:

Courtney M. Hill

Roles: Conceptualization, Formal analysis, Methodology, Writing – original draft

E-mail: [email protected]

Affiliations: Department of Oral Health Sciences, University of Washington, Seattle, WA, United States of America, Department of Epidemiology, University of Washington, Seattle, WA, United States of America

ORICD: https://orcid.org/0000-0002-6487-5258

Donald L. Chi

Roles: Conceptualization, Investigation, Writing – review & editing

Affiliations: Department of Oral Health Sciences, University of Washington, Seattle, WA, United States of America, Department of Health Systems and Population Health, University of Washington, Seattle, WA, United States of America

ORICD: https://orcid.org/0000-0003-2289-7747

Lloyd A. Mancl

Roles: Methodology, Writing – review & editing

Affiliation: Department of Oral Health Sciences, University of Washington, Seattle, WA, United States of America

Jessica C. Jones-Smith

Roles: Conceptualization, Writing – review & editing

Affiliations: Department of Epidemiology, University of Washington, Seattle, WA, United States of America, Department of Health Systems and Population Health, University of Washington, Seattle, WA, United States of America

Nadine Chan

Roles: Conceptualization, Writing – review & editing

Affiliations: Department of Epidemiology, University of Washington, Seattle, WA, United States of America, Public Health-Seattle & King County, Assessment, Policy, Development and Evaluation Division, Seattle, WA, United States of America

Brian E. Saelens

Roles: Conceptualization, Funding acquisition, Writing – review & editing

Affiliations: Seattle Children’s Research Institute, Seattle, WA, United States of America, Department of Pediatrics, University of Washington, Seattle, WA, United States of America

Christy M. McKinney

Roles: Conceptualization, Funding acquisition, Writing – review & editing

Affiliations: Department of Oral Health Sciences, University of Washington, Seattle, WA, United States of America, Seattle Children’s Research Institute, Seattle, WA, United States of America, Department of Pediatrics, University of Washington, Seattle, WA, United States of America

References

1. Kassebaum NJ, Smith AGC, Bernabé E, Fleming TD, Reynolds AE, Vos T, et al. Global, Regional, and National Prevalence, Incidence, and Disability-Adjusted Life Years for Oral Conditions for 195 Countries, 1990–2015: A Systematic Analysis for the Global Burden of Diseases, Injuries, and Risk Factors. J Dent Res 2017;96:380–7. pmid:28792274

2. Gupta N, Vujicic M, Yarbrough C, Harrison B. Disparities in untreated caries among children and adults in the U.S., 2011–2014. BMC Oral Health 2018;18. https://doi.org/10.1186/S12903-018-0493-7.

3. Centers for Disease Control and Prevention. Oral Health Surveillance Report: Trends in Dental Caries and Sealants, Tooth Retention, and Edentulism, United States, 1999–2004 to 2011–2016. Atlanta, GA: 2019.

4. Northridge ME, Kumar A, Kaur R. Disparities in Access to Oral Health Care. Annu Rev Public Health 2020;41:513. pmid:31900100

5. Singh A, Peres MA, Watt RG. The Relationship between Income and Oral Health: A Critical Review. J Dent Res 2019;98:853–60. pmid:31091113

6. Bastos JL, Celeste RK, Paradies YC. Racial Inequalities in Oral Health. J Dent Res 2018;97:878–86. pmid:29634429

7. Çolak H, Dülgergil Ç, Dalli M, Hamidi M. Early childhood caries update: A review of causes, diagnoses, and treatments. J Nat Sci Biol Med 2013;4:29. pmid:23633832

8. BaniHani A, Deery C, Toumba J, Munyombwe T, Duggal M. The impact of dental caries and its treatment by conventional or biological approaches on the oral health-related quality of life of children and carers. Int J Paediatr Dent 2018;28:266–76. pmid:29288546

9. Mota-Veloso I, Soares MEC, Alencar BM, Marques LS, Ramos-Jorge ML, Ramos-Jorge J. Impact of untreated dental caries and its clinical consequences on the oral health-related quality of life of schoolchildren aged 8–10 years. Qual Life Res 2016;25:193–9. https://doi.org/10.1007/S11136-015-1059-7.

10. Chi DL, Masterson EE, Carle AC, Mancl LA, Coldwell SE. Socioeconomic status, food security, and dental caries in US children: mediation analyses of data from the National Health and Nutrition Examination Survey, 2007–2008. Am J Public Health 2014;104:860–4. pmid:24625141

11. Chi DL, Dinh MA, da Fonseca MA, Scott JAM, Carle AC. Dietary Research to Reduce Children’s Oral Health Disparities: An Exploratory Cross-Sectional Analysis of Socioeconomic Status, Food Insecurity, and Fast-Food Consumption. J Acad Nutr Diet 2015;115:1599–604. pmid:25840937

12. Angelopoulou M V., Shanti SD, Gonzalez CD, Love A, Chaffin J. Association of food insecurity with early childhood caries. J Public Health Dent 2019;79:102–8. pmid:30485438

13. Hill B. Evaluating the association between food insecurity and dental caries in US children 1–19 years: results from the National Health and Nutrition Examination Survey (NHANES) 2013–2014. J Public Health Dent 2020;80:14–7. https://doi.org/10.1111/JPHD.12345.

14. Bahanan L, Singhal A, Zhao Y, Scott T, Kaye E. The association between food insecurity, diet quality, and untreated caries among US children. Journal of the American Dental Association 2021;152:613–21. pmid:34147241

15. Tsai WT, Lawrence HP. Association between psychosocial determinants of adverse childhood experiences and severe early childhood caries among First Nations children. Int J Paediatr Dent 2022;32:352–66. pmid:34358378

16. Economic Research Service. Definitions of Food Security. US Department of Agriculture 2021. https://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-us/definitions-of-food-security.aspx#characteristics (accessed April 15, 2021).

17. Chi DL, Scott JAM. Added Sugar and Dental Caries in Children: A Scientific Update and Future Steps. Dent Clin North Am 2019;63:17. pmid:30447790

18. Valenzuela MJ, Waterhouse B, Aggarwal VR, Bloor K, Doran T. Effect of sugar-sweetened beverages on oral health: a systematic review and meta-analysis. Eur J Public Health 2021;31. pmid:32830237

19. Mendez MA, Miles DR, Poti JM, Sotres-Alvarez D, Popkin BM. Persistent disparities over time in the distribution of sugar-sweetened beverage intake among children in the United States. Am J Clin Nutr 2019;109:79. pmid:30535176

20. Tasevska N, DeLia D, Lorts C, Yedidia M, Ohri-Vachaspati P. Determinants of sugar-sweetened beverage consumption among low-income children: are there differences by race/ethnicity, age, and gender? J Acad Nutr Diet 2017;117:1900. https://doi.org/10.1016/J.JAND.2017.03.013.

21. Lee J, Kubik MY, Fulkerson JA. Diet Quality and Fruit, Vegetable, and Sugar-Sweetened Beverage Consumption by Household Food Insecurity among 8- to 12-Year-Old Children during Summer Months. J Acad Nutr Diet 2019;119:1695–702. pmid:31056369

22. Fernández CR, Chen L, Cheng ER, Charles N, Meyer D, Monk C, et al. Food Insecurity and Sugar-Sweetened Beverage Consumption Among WIC-Enrolled Families in the First 1,000 Days. J Nutr Educ Behav 2020;52:796. pmid:32444189

23. Tomayko EJ, Mosso KL, Cronin KA, Carmichael L, Kim K, Parker T, et al. Household food insecurity and dietary patterns in rural and urban American Indian families with young children. BMC Public Health 2017;17. pmid:28666476

24. Lent MR, vander Veur S, Mallya G, McCoy TA, Sanders TA, Colby L, et al. Purchasing patterns of adults, adolescents and children in urban corner stores: Quantity, spending and nutritional characteristics. Public Health Nutr 2015;18:1706. https://doi.org/10.1017/S1368980014001670.

25. Caspi CE, Lenk K, Pelletier JE, Barnes TL, Harnack L, Erickson DJ, et al. Food and beverage purchases in corner stores, gas-marts, pharmacies and dollar stores. Public Health Nutr 2017;20:2587. pmid:27641618

26. D’Angelo H, Suratkar S, Song HJ, Stauffer E, Gittelsohn J. Access to food source and food source use are associated with healthy and unhealthy food-purchasing behaviours among low-income African-American adults in Baltimore City. Public Health Nutr 2011;14:1632–9. pmid:21450140

27. Madsen KA, Falbe J, Olgin G, Ibarra-Castro A, Rojas N. Purchasing patterns in low-income neighborhoods: Implications for studying sugar-sweetened beverage taxes. Public Health Nutr 2019;22:1807. https://doi.org/10.1017/S1368980019000375.

28. Ma X, Liese AD, Hibbert J, Bell BA, Wilcox S, Sharpe PA. The association between food security and store-specific and overall food shopping behaviors. J Acad Nutr Diet 2017;117:1931. pmid:28366811

29. Lyonnais MJ, Rafferty AP, Pitts SJ, Blanchard RJ, Kaur AP. Examining Shopping Patterns, Use of Food-Related Resources, and Proposed Solutions to Improve Healthy Food Access Among Food Insecure and Food Secure Eastern North Carolina Residents. Int J Environ Res Public Health 2020;17. https://doi.org/10.3390/IJERPH17103361.

30. Lenk KM, Winkler MR, Caspi CE, Laska MN. Food shopping, home food availability, and food insecurity among customers in small food stores: an exploratory study. Transl Behav Med 2020;10:1358–66. pmid:33421081

31. Kharmats AY, Jones-Smith JC, Cheah YS, Budd N, Flamm L, Cuccia A, et al. Relation between the Supplemental Nutritional Assistance Program cycle and dietary quality in low-income African Americans in Baltimore, Maryland. Am J Clin Nutr 2014;99:1006–14. pmid:24622807

32. City of Seattle. Sweetened Beverage Tax Evaluation Reports. City of Seattle 2022. http://www.seattle.gov/sweetened-beverage-tax-community-advisory-board/evaluation-reports (accessed July 27, 2022).

33. Chan N, Chen R, Jones-Smith J, Knox M, Krieger J, Oddo VM, et al. The Evaluation of Seattle’s Sweetened Beverage Tax Baseline Report. Seattle, WA: 2018.

34. Armfield JM, Spencer AJ, Roberts-Thomson KF, Plastow K. Water fluoridation and the association of sugar-sweetened beverage consumption and dental caries in Australian children. Am J Public Health 2013;103:494–500. pmid:23327241

35. Peres MA, Sheiham A, Liu P, Demarco FF, Silva AER, Assunção MC, et al. Sugar Consumption and Changes in Dental Caries from Childhood to Adolescence. J Dent Res 2016;95:388–94. pmid:26758380

36. Melough MM, Sathyanarayana S, Zohoori F v, Gustafsson HC, Sullivan EL, Chi DL, et al. Impact of Fluoride on Associations between Free Sugars Intake and Dental Caries in US Children. JDR Clin Trans Res 2022. pmid:35446163

37. Research Food & Center Action. Screen and Intervene: A Toolkit for Pediatricians to Address Food Insecurity—Food Research & Action Center n.d. https://frac.org/aaptoolkit (accessed July 27, 2022).

38. Coleman-Jensen A, Rabbitt MP, Hashad RN, Hales L, Gregory CA. Six-Item Short Form of the Food Security Survey Module. United States Department of Agriculture, Economic Research Service 2022. https://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-u-s/survey-tools#six (accessed July 27, 2022).

39. Radandt NE, Johnson DB, Scott JM, Corbridge T, Kim AS, Coldwell SE. Validation of a 2-Item Food Security Screening Tool in a Dental Setting. J Dent Child (Chic) 2018;85:114.

40. Hill CE, MacDougall CR, Riebl SK, Savla J, Hedrick VE, Davy BM. Evaluation of the Relative Validity and Test-Retest Reliability of a 15-Item Beverage Intake Questionnaire in Children and Adolescents. J Acad Nutr Diet 2017;117:1757–1766.e5. pmid:28673606

41. Hedrick VE, Comber DL, Estabrooks PA, Savla J, Davy BM. The beverage intake questionnaire: determining initial validity and reliability. J Am Diet Assoc 2010;110:1227–32. https://doi.org/10.1016/J.JADA.2010.05.005.

42. Hedrick VE, Savla J, Comber DL, Flack KD, Estabrooks PA, Nsiah-Kumi PA, et al. Development of a brief questionnaire to assess habitual beverage intake (BEVQ-15): sugar-sweetened beverages and total beverage energy intake. J Acad Nutr Diet 2012;112:840–9. pmid:22709811

43. Ismail AI, Sohn W, Tellez M, Willem JM, Betz J, Lepkowski J. Risk indicators for dental caries using the International Caries Detection and Assessment System (ICDAS). Community Dent Oral Epidemiol 2008;36. pmid:18205641

44. Ismail AI, Sohn W, Tellez M, Amaya A, Sen A, Hasson H, et al. The International Caries Detection and Assessment System (ICDAS): An integrated system for measuring dental caries: Methods. Community Dent Oral Epidemiol 2007;35. https://doi.org/10.1111/j.1600-0528.2007.00347.x.

45. Landis JR, Koch GG. The Measurement of Observer Agreement for Categorical Data. Biometrics 1977;33. https://doi.org/10.2307/2529310. pmid:843571

46. Ekstrand KR, Gimenez T, Ferreira FR, Mendes FM, Braga MM. The International Caries Detection and Assessment System—ICDAS: A Systematic Review. Caries Res 2018;52:406–19. pmid:29518788

47. VanderWeele TJ. Principles of confounder selection. Eur J Epidemiol 2019;34:211. pmid:30840181

48. Odoms-Young A, Bruce MA. Examining the Impact of Structural Racism on Food Insecurity: Implications for Addressing Racial/Ethnic Disparities. Fam Community Health 2018;41:S3. pmid:29461310

49. Duggan CP, Kurpad A, Stanford FC, Sunguya B, Wells JC. Race, ethnicity, and racism in the nutrition literature: an update for 2020. Am J Clin Nutr 2020;112:1409. pmid:33274358

50. Imai K, Keele L, Tingley D. A General Approach to Causal Mediation Analysis. Psychol Methods 2010;15. pmid:20954780

51. Imai K, Keele L, Yamamoto T. Identification, inference and sensitivity analysis for causal mediation effects. Statistical Science 2010;25. https://doi.org/10.1214/10-STS321.

52. Hafeman DM, Schwartz S. Opening the Black Box: a motivation for the assessment of mediation. Int J Epidemiol 2009;38:838–45. pmid:19261660

53. Tingley D, Yamamoto HT, Hirose K, Keele L, Princeton KI. mediation: R Package for Causal Mediation Analysis n.d.

54. Liu SH, Ulbricht CM, Chrysanthopoulou SA, Lapane KL. Implementation and reporting of causal mediation analysis in 2015: A systematic review in epidemiological studies. BMC Res Notes 2016;9:1–21. https://doi.org/10.1186/S13104-016-2163-7/TABLES/6.

55. Vanderweele TJ. Bias formulas for sensitivity analysis for direct and indirect effects. Epidemiology 2010;21:540. pmid:20479643

56. Cobb LK, Appel LJ, Franco M, Jones-Smith JC, Nur A, Anderson CA. The relationship of the local food environment with obesity: A systematic review of methods, study quality and results. Obesity (Silver Spring) 2015;23:1331. pmid:26096983

57. Gorski Findling MT, Wolfson JA, Rimm EB, Bleich SN. Differences in the Neighborhood Retail Food Environment and Obesity Among US Children and Adolescents by SNAP Participation. Obesity (Silver Spring) 2018;26:1063. pmid:29797558

58. Xin J, Zhao L, Wu T, Zhang L, Li Y, Xue H, et al. Association between access to convenience stores and childhood obesity: A systematic review. Obes Rev 2021;22 Suppl 1. pmid:31274248

59. Kraft AN, Thatcher EJ, Zenk SN. Neighborhood Food Environment and Health Outcomes in U.S. Low-Socioeconomic Status, Racial/Ethnic Minority, and Rural Populations: A Systematic Review. J Health Care Poor Underserved 2020;31:1078–114. pmid:33416683

60. Chrisinger BW, Kallan MJ, Whiteman ED, Hillier A. Where do U.S. households purchase healthy foods? An analysis of food-at-home purchases across different types of retailers in a nationally representative dataset. Prev Med (Baltim) 2018;112:15. https://doi.org/10.1016/J.YPMED.2018.03.015.

61. Jones-Smith JC, Knox MA, Coe NB, Walkinshaw LP, Schoof J, Hamilton D, et al. Sweetened beverage taxes: Economic benefits and costs according to household income. Food Policy 2022;110:102277. https://doi.org/10.1016/J.FOODPOL.2022.102277.

62. Andreyeva T, Marple K, Marinello S, Moore TE, Powell LM. Outcomes Following Taxation of Sugar-Sweetened Beverages: A Systematic Review and Meta-analysis. JAMA Netw Open 2022;5:e2215276. pmid:35648398

63. Powell LM, Leider J. Impact of a sugar-sweetened beverage tax two-year post-tax implementation in Seattle, Washington, United States. J Public Health Policy 2021;42:574–88. pmid:34732842

64. Saelens B, Rowland M, Qu P, Walkinshaw L, Oddo V, Knox M, et al. Twelve Month Report: Store Audits and Child Cohort—The Evaluation of Seattle’s Sweetened Beverage Tax. Seattle, WA: 2020.

65. Oddo VM, Knox MA, Pinero Walkinshaw L, Saelens BE, Chan N, Jones-Smith JC. Evaluation of Seattle’s sweetened beverage tax on tax support and perceived economic and health impacts. Prev Med Rep 2022;27. pmid:35656219

66. Sawyer L, Oddo VM, Fretts A, Knox MA, Chan N, Saelens BE, et al. Impacts of the Seattle Sweetened Beverage Tax on the Perceived Healthfulness of Sweetened Beverages. Nutrients 2022;14. https://doi.org/10.3390/NU14050993.

67. Alhareky M. Taxation of Sugar-Sweetened Beverages and its Impact on Dental Caries: A Narrative Review. Saudi J Med Med Sci 2021;9:113. pmid:34084101

68. Mobley C, Marshall TA, Milgrom P, Coldwell SE. The contribution of dietary factors to dental caries and disparities in caries. Acad Pediatr 2009;9:410–4. pmid:19945075

69. Mackenbach JD, Ibouanga EL, van der Veen MH, Ziesemer KA, Pinho MGM. Relation between the food environment and oral health—systematic review. Eur J Public Health 2022:1–11. https://doi.org/10.1093/EURPUB/CKAC086.

70. He M, Tucker P, Gilliland J, Irwin JD, Larsen K, Hess P. The Influence of Local Food Environments on Adolescents’ Food Purchasing Behaviors. Int J Environ Res Public Health 2012;9:1458. pmid:22690205

71. Spoer BR, Cantor JH, Rummo PE, Elbel BD. Food environment does not predict self-reported SSB consumption in New York City: A cross sectional study. PLoS One 2018;13:e0196689. pmid:30356232

72. Shareck M, Lewis D, Smith NR, Clary C, Cummins S. Associations between home and school neighbourhood food environments and adolescents’ fast-food and sugar-sweetened beverage intakes: findings from the Olympic Regeneration in East London (ORiEL) Study. Public Health Nutr 2018;21:2842–51. pmid:29962364

73. Dubowitz T, Zenk SN, Ghosh-Dastidar B, Cohen D, Beckman R, Hunter G, et al. Healthy food access for urban food desert residents: examination of the food environment, food purchasing practices, diet, and body mass index. Public Health Nutr 2015;18:2220. https://doi.org/10.1017/S1368980014002742.

74. Gittelsohn J, Rowan M, Gadhoke P. Interventions in small food stores to change the food environment, improve diet, and reduce risk of chronic disease. Prev Chronic Dis 2012;9. https://doi.org/10.5888/pcd9.110015. pmid:22338599

75. Dubowitz T, Ghosh-Dastidar M, Cohen DA, Beckman R, Steiner ED, Hunter GP, et al. Changes in Diet after Introduction of a Full Service Supermarket in a Food Desert. Health Aff (Millwood) 2015;34:1858. https://doi.org/10.1377/HLTHAFF.2015.0667.

76. Block JP, Subramanian S v. Moving Beyond “Food Deserts”: Reorienting United States Policies to Reduce Disparities in Diet Quality. PLoS Med 2015;12. pmid:26645285

77. Garasky S, Mbwana K, Romualdo A, Tenaglio A, Roy M. Foods Typically Purchased by Supplemental Nutrition Assistance Program (SNAP) Households. Alexandria, VA: 2016.

78. Choi SE, Wright DR, Bleich SN. Impact of Restricting Sugar-Sweetened Beverages From the Supplemental Nutrition Assistance Program on Children’s Health. Am J Prev Med 2021;60:276–84. pmid:33349472

79. Krieger J, Bleich SN, Scarmo S, Ng SW. Sugar-Sweetened Beverage Reduction Policies: Progress and Promise. Annu Rev Public Health 2021;42:439–61. pmid:33256536

80. Paarlberg R, Mozaffarian D, Micha R, Chelius C. Keeping Soda in SNAP: Understanding the Other Iron Triangle. Society 2018;55:308. pmid:35095130

81. Vericker T, Dixit-Joshi S, Taylor J, May L, Baier K, Williams ES. Impact of Food Insecurity Nutrition Incentives on Household Fruit and Vegetable Expenditures. J Nutr Educ Behav 2021;53:418–27. pmid:33526387

82. Parks CA, Han P, Fricke HE, Parker HA, Hesterman OB, Yaroch AL. Reducing food insecurity and improving fruit and vegetable intake through a nutrition incentive program in Michigan, USA. SSM Popul Health 2021;15. pmid:34458551

83. Hulbrock E, Otten JJ, Quinn E, Johnson DB, Lerman S. Exploring the Use of Seattle’s Farmers’ Market Incentive Program (“Fresh Bucks”) by Household Food Security Levels. J Hunger Environ Nutr 2017;12:362. pmid:34354790

84. Kenny DA. Mediation. DavidakennyNet 2021. http://davidakenny.net/cm/mediate.htm#DI (accessed September 20, 2022).

85. Fritz MS, MacKinnon DP. Required Sample Size to Detect the Mediated Effect. Psychol Sci 2007;18:233. pmid:17444920

86. Cheng J, Cheng NF, Guo Z, Gregorich S, Ismail AI, Gansky SA. Mediation analysis for count and zero-inflated count data. Stat Methods Med Res 2018;27:2756–74. pmid:28067122

87. Guo Z, Small DS, Gansky SA, Cheng J. Mediation analysis for count and zero-inflated count data without sequential ignorability and its application in dental studies. J R Stat Soc Ser C Appl Stat 2018;67:371–94. pmid:30983638