Household food insecurity before and during COVID-19 pandemic and its association with perceived stress: population-based studies

Food insecurity remains a global public health problem, encompassing not only hunger but also uncertainty and inadequacy of regular food access. FAO estimated 2.37 billion people lacked adequate food in 2020, a 15% increase from 2019. In Brazil, after a decade of improvements, food insecurity has risen since 2013; by 2018, nearly 40% of households experienced food insecurity, with regional disparities. During the COVID-19 pandemic, national estimates suggested over half of households faced food insecurity, likely due to economic impacts. Mental health disorders, including stress, have also increased during crises such as the pandemic. Evidence links food insecurity with stress and other mental health issues across countries and income levels. The study aimed to assess the prevalence of household food insecurity before and during COVID-19 and its association with perceived stress among adults in Criciúma, Southern Brazil.

Prior reports show substantial increases in food insecurity globally during 2020 (FAO), with Brazil experiencing rising rates since 2013 and sharp increases during the pandemic, albeit with lower prevalence in the South Region relative to other regions. Studies in Brazil and other LMICs documented increased food insecurity during COVID-19 and concurrent rises in mental health problems such as stress, anxiety, and depression. Pre-pandemic literature indicates food insecurity is associated with higher odds of stress and poorer mental health; a multi-country study linked even mild food insecurity to worse mental health and life satisfaction, independent of income and development level. Mechanistically, food insecurity may elevate stress via socioeconomic instability and poorer diet quality (deficits in vitamins and minerals) that are implicated in stress physiology, suggesting bidirectional links between diet, nutritional needs, and stress.

Design and setting: Two cross-sectional, population-based studies in Criciúma, Santa Catarina, Brazil. Study 1 (Health of the Criciúma Population) was conducted in 2019 (pre-COVID-19). Study 2 (Mental COVID) ran from October 2020 to January 2021 (during COVID-19). Population: Adults aged ≥18 years living in the urban area. Sampling: Two-stage cluster sampling in both studies. In 2019, 77 of 306 census tracts were randomly selected; 618 households were systematically sampled within tracts, and all eligible residents invited. In 2020–2021, 60 census tracts were randomly selected; 607 households were systematically sampled. Data collection: Home interviews by trained personnel (with PPE during the pandemic study). Sample size: Calculated for each study with 80% power, 95% CI, assumed 50% prevalence, 5 p.p. error, design effect 1.5, and 15% for losses; required n=662 per study. Measures: - Household food insecurity: 2019 used the full Brazilian Food Insecurity Scale (EBIA), 14 items (3-month recall) with scoring and classification into food security, mild, moderate, and severe food insecurity (age-appropriate scoring ranges). For analyses, categorized as food secure vs. food insecure (mild/moderate/severe). 2020–2021 used the short-form EBIA (5 items; 3-month recall) for screening; any positive response classified as food insecure. - Perceived stress: Perceived Stress Scale (PSS-14), validated for Brazil; Likert 0–4 per item over the past month; total score 0–56; categorized into quintiles; highest quintile defined perceived stress. - Covariables: Sex, age (18–29, 30–39, 40–49, 50–59, ≥60), skin color (white, black, mixed), schooling (0–4, 5–8, 9–11, ≥12 years), monthly income (<BRL 1,000; 1,001–2,000; >2,000), job status (employed: yes/no), marital status (married/not married), household crowding (1–2, 3–4, ≥5 persons), overweight (BMI ≥25 kg/m2 for adults; ≥27 kg/m2 for older adults; BMI from self-report), and diet quality. - Diet quality: Constructed from frequency of healthy (fruits, vegetables, milk, legumes, chicken, red meat) and unhealthy (sweets, soda/processed juice) foods, with 0–4 scoring per item based on frequency; total 0 (best) to 28 (worst); categorized in tertiles. Statistical analysis: Descriptive statistics with relative/absolute frequencies and 95% CIs. Group comparisons with chi-squared tests accounting for complex sampling. Associations between food insecurity (exposure) and perceived stress (outcome) estimated by Poisson regression with robust variance, reporting prevalence ratios (PRs) and 95% CIs. Adjusted analyses used a three-level hierarchical model: Level 1 (sex, age, skin color); Level 2 (schooling, income, employment, marital status, crowding); Level 3 (overweight, diet quality). Variables retained via backward selection if p<0.20. All analyses used Stata 16.1 with svy for complex sampling weights. Ethics: Written informed consent obtained. Approvals: University of South Santa Catarina Ethics Committee (Dec 2018; protocol 04033118.4.0000.0119) and Brazilian National Ethics Research Committee (July 2020; protocol 30955120.0.0000.5324).

- Sample: 1,683 adults (n=820 in 2019; n=863 in 2020–2021); response rates 86.1% (2019) and 75% (2020). - Food insecurity prevalence: 25.8% (95%CI: 22.3–29.6) in 2019 and 21.6% (95%CI: 18.9–24.4) in 2020–2021; no significant overall difference between periods. Higher prevalence among less educated, unmarried, non-white individuals in both periods; younger age and worst diet quality associated with higher food insecurity in 2019 but not in 2020–2021. - Perceived stress prevalence: ~38% in both surveys (38.8% in 2019; 37.5% in 2020–2021). Higher among females and those with poorer diet quality in both periods; higher among younger adults pre-pandemic; higher with greater household crowding pre-pandemic. - Association between food insecurity and perceived stress: Pre-pandemic (2019): Food insecurity associated with higher perceived stress (crude PR=1.41; 95%CI: 1.12–1.76). Association persisted after adjustment (PR=1.29; 95%CI: 1.02–1.63). During pandemic (2020–2021): No significant association in crude (PR=1.11; 95%CI: 0.90–1.35) or adjusted models (PR=1.10; 95%CI: 0.89–1.35).

The study shows that about one quarter of adults in Criciúma experienced household food insecurity both before and during the early pandemic period, and roughly two in five reported high perceived stress at both times. Pre-pandemic, food insecurity was significantly associated with higher prevalence of perceived stress, aligning with prior evidence linking food insecurity to mental health problems. During the pandemic, despite persistently elevated stress levels, the association between food insecurity and stress was not observed. The authors suggest that emergency social protection measures (e.g., cash transfers, school feeding adaptations) and the relatively stronger local economy and labor market in Santa Catarina may have mitigated stress specifically related to food access and availability among food-insecure households, attenuating the association during this period. Socio-demographic gradients persisted: food insecurity was more common among those with lower schooling, lower income, and non-white skin color; stress remained higher among women and those with poorer diet quality. The findings underscore complex, context-dependent pathways linking food insecurity and stress, including socioeconomic instability and diet-related nutritional mechanisms, and highlight the role of policy responses in buffering adverse mental health impacts.

Food insecurity increased the likelihood of perceived stress before, but not during, the COVID-19 pandemic in Criciúma. The pandemic period did not worsen overall prevalence of food insecurity or perceived stress in this setting, nor sustain their association, possibly due to emergency social policies and local economic conditions. Nonetheless, individuals with lower socioeconomic status remain disproportionately affected by both outcomes, reflecting entrenched social inequalities. Sustained and expanded social and health policies, including the Brazilian National Food and Nutrition Policy and protections for the human right to adequate food, are needed. Continued post-pandemic monitoring of food insecurity and mental health is recommended.

Use of different EBIA versions across periods (full 14-item in 2019 vs. short-form 5-item in 2020–2021) limited comparability of food insecurity severity classification in the pandemic study. Sample composition may be biased due to interviews during business hours, leading to overrepresentation of women and older adults (though somewhat reduced in 2020–2021).