Brief Communication: Survey Based Study on Food Insecurity during COVID-19 for Households with Children on a Prescribed Gluten-Free Diet

Abstract

Gluten-free (GF) foods are generally costlier than their gluten containing counterparts. We conducted an anonymous electronic survey to assess food insecurity in households with a child on a prescribed GFD and how this relates to GFD adherence during the COVID-19 pandemic. The proportion of households who screened positive using the Hunger Vital Sign was similar to national rates (19–24%). About 5% of families who screened food secure reported GF food insecurity. Parent-reported intentional gluten consumption due to limited GF food availability in the household increased during the pandemic (7.5%). Food insecurity should be considered in the management of celiac disease.

Introduction

The United States Department of Agriculture defines food insecurity as a lack of consistent access to enough food for an active, healthy life.¹ Feeding America projected that 1 in 6 children in the United States may experience food insecurity in 2021.² The US began to experience the socioeconomic impact of COVID-19 pandemic around February 2020; however, there is limited data on how the pandemic may have impacted families with medical dietary restrictions. Celiac disease (CD) is a chronic immune disease for which a gluten-free diet(GFD) is the only treatment. Market basket studies have consistently found that gluten-free (GF) foods are more expensive than their gluten-containing counterparts.³ State and federal food assistance programs often do not account for the increased cost of GF food.⁴ We conducted an anonymous survey to better understand food insecurity in households with a child on prescribed GFD and how food insecurity may affect GFD adherence during the pandemic.

Methods

An anonymous electronic survey was distributed through celiac support groups from March to May 2021(Supplemental Methods). Concurrently, up to three messages were posted at 2-week intervals on CD forums on Facebook and Twitter. Participants completed a single survey with items related to current food security and were asked to recall retrospectively their pre-pandemic experiences. Food insecurity risk was evaluated using the validated Hunger Vital Sign Screener and National Center for Health Statistics US Household Six-Item Short Form Food Security Survey Module.^{5, 6} The Hunger Vital Sign Screener was also adapted to assess GF food insecurity risk by incorporating “gluten-free food” in each screening question. Intentional gluten ingestion due to lack of GF food availability was assessed by parent-report.

Odds of gluten ingestion and food insecurity before and after the pandemic were modeled using multivariate logistic regression. Socioeconomic status was estimated using the Economic Innovation Group’s licensed Distressed Communities Index (DCI). This composite zip-code level measure is based upon data from the United States 2014 to 2018 American Community Survey 5-Year Estimates and the Census Bureau’s business pattern datasets from the same period. To compare categorical variables, χ² testing was used with a p< 0.05 for significance.

Results

The 378 households with children with CD that completed the survey primarily self-identified as white (89%, 338). The median age of GF children in these households was 7 years; 61% (250) were female, 80% (329) were the only child with celiac and 68% (282) attended public school. CD was most often diagnosed by intestinal biopsy (82%, 309). Most (67%, 253) had been on GFD for greater than 2 years. Most had a household income > $80,000 (61%, 230), resided in a mortgaged home (83%, 312), and were relatively well-educated (defined as having a bachelor’s degree or higher (81%, 308) (Table 1). Only 48% of respondents provided a zip-code. Available demographic characteristics did not vary according to whether zip-code was provided. All 10 zip-code regions in the US were represented with respondents primarily from zip-codes classified as “comfortable” or “prosperous” based on their DCI score.

Table 1:

Demographics

	Total Survey Participants (378)
Demographics	N	%	95%CI
Female Child	235	62.17%	[0.57, 0.67]
Child Attending Public School	267	70.63%	[0.66, 0.75]
Child(ren) on a GFD in household
1	301	79.63%	[0.75, 0.84]
2	65	17.20%	[0.14, 0.21]
≥3	12	3.17%	[0.02, 0.05]
Adults on a GFD in household
0	174	46.03%	[0.41, 0.51]
1	147	38.89%	[0.34, 0.44]
2	49	12.96%	[0.10, 0.17]
>3	8	2.12%	[0.01, 0.04]
Diagnosed by Intestinal Biopsy	330	87.30%	[0.84, 0.90]
GFD Duration
Less than 1 year	49	12.96%	[0.10, 0.17]
1–2 years	70	18.52%	[0.15, 0.23]
More than 2 years	253	66.93%	[0.62, 0.72]
Household Information
Heads of Household Married	326	86.24%	[0.82, 0.90]
Head of Household Education Level
Some HS-High School diploma	16	4.23%	[0.02, 0.07]
Some College	40	10.58%	[0.08, 0.14]
Undergraduate Degree	147	38.89%	[0.34, 0.44]
Graduate	150	39.68%	[0.35, 0.45]
Professional	11	2.91%	[0.01, 0.05]
Head of Household Employed, working >40 hours	155	41.01%	[0.36, 0.46]
Combined Household Income
<$30,000	20	5.29%	[0.03, 0.08]
$30,000-$80000	59	15.61%	[0.12, 0.20]
More than $80,0000	230	60.85%	[0.56, 0.66]
Mortgaged/Owns Home	312	82.54%	[0.78, 0.86]
Urban vs Rural
Rural	19	5.03%	[0.03, 0.08]
Small Town	25	6.61%	[0.04, 0.10]
Suburban	121	32.01%	[0.27, 0.37]
Urban	30	7.94%	[0.05, 0.11]
Distressed Community Index Score Quintile
Prosperous	97	25.66%	[0.21, 0.30]
Comfortable	46	12.17%	[0.09, 0.16]
Mid-tier	25	6.61%	[0.04, 0.10]
At-risk	21	5.56%	[0.03, 0.08]
Distressed	6	1.59%	[0.01, 0.03]
Missing Zip-code	183	48.41%	[0.43, 0.54]

When retrospectively considering their pre-pandemic situations, 19% (72) households with at least one child following a GFD screened positive for food insecurity. When asked specifically about GF foods, 21% (81) of the households screened positive. Rates of self-reported food insecurity (24%, 90, p=0.09) and GF food insecurity (27%, 101, p=0.11) both appeared to have increased during the pandemic. The proportion of households who were food secure but screened positive for GF food insecurity remained constant (5%,19) (Figure 1).

Figure 1:

Food Insecurity, Intentional Gluten Ingestion and Gluten-Free Food Availability (n=413) (a) Risk of Food Insecurity and Gluten-Free Food Insecurity in Households with Children on a Gluten-Free Diet was measured via the validated and adapted Hunger Vital Sign Screener. There was no significant difference in food insecurity before and during the pandemic. (b) Intentional Gluten Ingestion due to Unavailability of Gluten-Free Foods before and during pandemic was compared with test of proportions showing a significant difference (p value < 0.05) (c) Gluten-Free Food availability before the pandemic and (d) Gluten-Free Food availability during the pandemic showed decreasing perceived availability of gluten free foods with families requiring more visits to different stores or use of online shopping to access products.

Intentional gluten ingestion due to unavailability of GF foods rose during the pandemic [4.2% vs 7.5 %, p=0.06] and was more likely among those who screened positive for food insecurity (aOR=7.6, 95%CI 1.7–33.8) or GF food insecurity (aOR=12.4, 95% CI 2.5–61.3). The odds of eating gluten were further increased if multiple household members followed a GFD (aOR=3.7, 95%CI 0.90, 15.25). (Table 2). All households reported that the availability of GF foods decreased during the pandemic(p<0.001).

Table 2:

Unadjusted and Adjusted Odds Ratios(OR) from logistic regression for factors associated with Intentional Gluten Ingestion

Characteristics	OR	95%CI	AOR*	95%CI
Multiple Household Members on a Gluten Free Diet	2.9	[1.16,7.36]	3.7	[0.90, 15.25]
Positive Screen for Food Insecurity	5.4	[2.45,11.72	7.6	[1.74, 33.8]
Positive Screen for Gluten-Free Food Insecurity	7.3	[3.22, 16.78]	12.4	[2.5, 61.32]

^*Adjusted Odds Ratios(AOR) were adjusted for Distressed Communities Index(DCI), income, employment and housing

Rural communities had the highest rates of food insecurity and GF food insecurity. While food insecurity rates did not change significantly during the pandemic for rural communities, small towns tended to be the most negatively impacted by the pandemic with an increase in rates of both food insecurity and GF food insecurity (Supplemental Table).

Discussion

Through our survey-based study with convenience sampling, households with at least one child on a prescribed GFD reported an increase in food insecurity and GF food insecurity. During the pandemic, GF food availability dropped significantly, necessitating families to go to more stores to acquire GF foods than what was required prior to the pandemic. Intentional gluten ingestion due to limited GF food availability nearly doubled during the pandemic. Members of food insecure and GF food insecure households were more likely to eat gluten. Pre-existing rural-urban disparities in food security persisted during the pandemic, and small-town households showed the greatest increase in food insecurity.

Our study is the first to differentiate GF food insecurity from non-specific food insecurity. We acknowledge that retrospective recall of pre-pandemic food insecurity, convenience non-random sampling with bias to Caucasians and affluent families relative to the general population, and missing zip-codes all limit the generalizability of our study. However, our study begins the discussion on the association of food insecurity and intentional gluten ingestion. We may also be underestimating food insecurity given that our survey was conducted online. This may explain why families living in shelters or in subsidized housing who may have less access to the internet were underrepresented. About half of the participants did not include their zip-codes so this data only relates to the subset with available data.

Children with CD may still experience GF food insecurity in households that are otherwise food secure. GF food insecurity may pose a health risk by increasing the likelihood of intentional gluten ingestion. More research is needed to understand how food insecurity impacts GFD adherence and its impact on health outcomes. Until then, gastroenterologists may wish to consider the potential role of food insecurity in the evaluation of nonresponsive CD.