Disordered eating in a population‐based sample of young adults during the <scp>COVID</scp>‐19 outbreak

This was not an intervention study. The researchers tested whether specific psychological and socioeconomic factors during the COVID-19 pandemic were associated with disordered eating behaviors in young adults. They examined:

**Primary correlates:** Psychological distress (depressive symptoms), perceived stress, stress management ability, financial difficulties, and food insecurity.

**Outcomes:** Three categories of disordered eating:

1. **Extreme unhealthy weight control behaviors (UWCBs):** Vomiting, laxative use, diet pills, fasting (≥8 hours), skipping meals.

2. **Less extreme UWCBs:** Eating very little, using food substitutes, smoking more to lose weight, skipping meals.

3. **Binge eating:** Loss-of-control eating with a large amount of food, at least weekly.

They also collected open-ended qualitative responses about how the pandemic affected eating behaviors.

**Sample size:** 720 respondents (out of 1,562 invited from the original EAT 2010–2018 cohort, a 46.1% response rate).

**Population:** Young adults from the Minneapolis/St. Paul metropolitan area (Minnesota, USA) who had participated in the EAT 2010–2018 longitudinal study.

**Age:** Mean 24.7 years (SD = 2.0 years), range approximately 22–28 years.

**Sex:** 57.6% female, 42.4% male.

**Race/ethnicity:** 60.3% White, 14.0% Asian, 12.5% Black/African American, 7.5% Hispanic/Latino, 5.7% other/mixed.

**Socioeconomic status:** Diverse—28.1% had low or low-middle socioeconomic status (SES), 24.7% middle SES, 47.2% upper-middle or high SES.

**Setting:** Population-based sample drawn from public schools in the Minneapolis/St. Paul area, originally recruited in 2009–2010.

All measures were self-reported via an online survey (the C-EAT survey) administered in April–May 2020, approximately 4–8 weeks after the WHO declared COVID-19 a pandemic.

**Depressive symptoms:** Modified version of the Center for Epidemiologic Studies Depression Scale (CES-D), 10 items, scored 0–30. Higher scores = more depressive symptoms. Cronbach's alpha = 0.83 (good internal consistency).

**Perceived stress:** Perceived Stress Scale (PSS-4), 4 items, scored 0–16. Higher scores = more stress. Alpha = 0.72.

**Stress management:** Single item: "During the past 30 days, how often have you felt that you were unable to control the important things in your life?" (reverse-scored). Higher scores = better stress management. Range 1–5.

**Financial difficulties:** Single item: "How difficult is it for you to pay for the very basics like food, housing, medical care, and heating?" Scored 1 (not at all difficult) to 4 (very difficult).

**Food insecurity:** Six-item USDA Household Food Security Survey Module (short form). Participants were classified as food secure, low food security, or very low food security.

**Extreme UWCBs:** Count of behaviors (vomiting, laxatives, diet pills, fasting ≥8 hours, skipping meals) in the past 12 months. Range 0–5.

**Less extreme UWCBs:** Count of behaviors (eating very little, food substitutes, smoking more, skipping meals) in the past 12 months. Range 0–4.

**Binge eating:** Single item from the Project EAT survey: "In the past year, have you ever eaten a very large amount of food (e.g., an entire pizza or a whole cake) and felt like your eating was out of control?" Response options: no, yes—less than once per month, yes—about once per month, yes—a few times per month, yes—about once per week, yes—several times per week. Binge eating was defined as at least weekly.

**Open-ended questions:** "How has the COVID-19 pandemic affected your eating behaviors, if at all?" and "Is there anything else you would like to share about how the COVID-19 pandemic has affected your eating behaviors, weight, or health?"

**Study design:** Cross-sectional observational study embedded within a longitudinal cohort (EAT 2010–2018). Participants completed a one-time online survey during the early pandemic period. The researchers examined associations between current psychological/socioeconomic factors and self-reported disordered eating behaviors (past 12 months, which overlapped with pre-pandemic and pandemic periods).

**Why this design matters:**

**No randomisation or blinding:** This is purely observational. There is no intervention, no control group, and no manipulation. The design can identify associations but cannot prove causation.

**Temporal ambiguity:** The disordered eating measures asked about the past 12 months (which included ~4 months of pandemic and ~8 months pre-pandemic). The psychological and socioeconomic measures asked about the current pandemic period. So the "cause" (pandemic stress) and "effect" (disordered eating) overlap in time—you cannot determine which came first.

**Single time point:** This is a snapshot. It cannot track how eating behaviors changed over the course of the pandemic or whether pre-existing disordered eating patterns influenced responses.

**Self-report bias:** All measures are self-reported, which is standard for eating disorder research but subject to recall bias, social desirability bias, and misinterpretation of questions (e.g., "fasting" might mean different things to different people).

**Selection bias:** Only 46% of the original cohort responded. Non-responders may have been more or less affected by the pandemic, potentially biasing results.

**Statistical approach:** The researchers used:

Negative binomial regression for count outcomes (extreme and less extreme UWCBs), adjusting for sex, race/ethnicity, SES, and pre-pandemic disordered eating (from the 2018 wave).

Logistic regression for binge eating (binary outcome: yes/no weekly binge eating), with same covariates.

Qualitative thematic analysis of open-ended responses using a deductive-inductive approach (themes derived from existing literature plus new themes that emerged from the data).

**What this design can prove:**

That certain factors (low stress management, food insecurity, depressive symptoms, financial difficulties) are statistically associated with disordered eating during a pandemic.

The strength and direction of those associations (e.g., each unit increase in depressive symptoms is linked to X% higher count of UWCBs).

**What this design cannot prove:**

That pandemic stress *caused* disordered eating (people with pre-existing eating disorders may have been more stressed).

That improving stress management would reduce disordered eating (that would require an intervention trial).

That these associations are specific to the pandemic (they might exist in normal times too—no pre-pandemic comparison group for the same outcomes).

**Major methodological weakness:** The disordered eating measures cover the past 12 months, but the pandemic only started ~2–4 months before the survey. So the outcomes include pre-pandemic behavior. The authors partially address this by controlling for pre-pandemic disordered eating (from 2018 data), but the 12-month recall window still mixes pre- and pandemic periods.

**Primary outcomes (quantitative):**

**Extreme unhealthy weight control behaviors (UWCBs):**

- Low stress management was significantly associated with a higher count of extreme UWCBs (incidence rate ratio [IRR] = 1.28, 95% CI: 1.07–1.53, p = .007). This means that for each 1-point decrease in stress management (on a 1–5 scale), the expected count of extreme UWCBs increased by 28%.

- Food insecurity (very low food security vs. food secure) was associated with a higher count of extreme UWCBs (IRR = 2.02, 95% CI: 1.11–3.68, p = .021). That's a 102% increase in expected count.

- Depressive symptoms (IRR = 1.04, 95% CI: 1.00–1.08, p = .043) and financial difficulties (IRR = 1.23, 95% CI: 1.01–1.49, p = .036) were also associated with higher counts, but these were smaller effects.

- Perceived stress was not significantly associated with extreme UWCBs (IRR = 1.04, 95% CI: 0.97–1.12, p = .28).

**Less extreme unhealthy weight control behaviors:**

- Food insecurity (very low food security vs. food secure) was associated with a higher count (IRR = 1.67, 95% CI: 1.18–2.37, p = .004).

- Depressive symptoms (IRR = 1.03, 95% CI: 1.01–1.06, p = .009) and financial difficulties (IRR = 1.17, 95% CI: 1.04–1.31, p = .007) were also significant.

- Stress management (IRR = 0.97, 95% CI: 0.87–1.09, p = .64) and perceived stress (IRR = 1.03, 95% CI: 0.98–1.08, p = .24) were not significant.

**Binge eating (weekly or more):**

- Higher perceived stress was associated with 1.27 times higher odds of binge eating (OR = 1.27, 95% CI: 1.09–1.47, p = .002) for each 1-point increase on the PSS-4 (range 0–16).

- Higher depressive symptoms were associated with 1.09 times higher odds (OR = 1.09, 95% CI: 1.03–1.16, p = .003) per 1-point increase on the CES-D-10 (range 0–30).

- Financial difficulties (OR = 1.26, 95% CI: 0.93–1.72, p = .14) and food insecurity (very low vs. secure: OR = 1.62, 95% CI: 0.73–3.58, p = .23) were not statistically significant for binge eating, though the direction was consistent.

**Secondary outcomes (qualitative):**

Six themes emerged from open-ended responses about how the pandemic affected eating:

1. **Mindless eating and snacking** (most common): "I find myself eating more often because I'm bored and I'm at home."

2. **Increased food consumption:** "I am eating more than I normally would because I am home all day."

3. **Generalized decrease in appetite or dietary intake:** "I have been eating less because I am not as active."

4. **Eating to cope:** "I have been stress eating a lot more than usual."

5. **Pandemic-related reductions in dietary intake:** "I am eating less because I am worried about going to the grocery store."

6. **Re-emergence or marked increase in eating disorder symptoms:** "My eating disorder has come back full force."

**Stress management and extreme UWCBs:** A 1-point drop in stress management (e.g., from "sometimes" to "often" feeling unable to control important things) was linked to a 28% increase in the expected number of extreme weight control behaviors. If the average person reported 0.5 extreme behaviors, this would translate to about 0.64 behaviors—a modest but meaningful shift.

**Food insecurity and extreme UWCBs:** People with very low food security reported roughly twice as many extreme UWCBs as food-secure individuals. This is a large effect—comparable to the difference between someone with no eating disorder symptoms and someone with clinically significant symptoms.

**Perceived stress and binge eating:** Each 1-point increase on the 0–16 stress scale (roughly equivalent to moving from "sometimes" to "fairly often" stressed) was associated with 27% higher odds of weekly binge eating. For someone with a baseline 10% probability of binge eating, this would increase to about 12.7%—a small absolute increase but a large relative one.

**Depressive symptoms and binge eating:** Each 1-point increase on the 0–30 depression scale was linked to 9% higher odds of binge eating. A 5-point increase (e.g., from mild to moderate depression) would correspond to about 54% higher odds.

**Acknowledged by authors:**

Cross-sectional design prevents causal inference.

Disordered eating measures assessed past 12 months, which includes pre-pandemic period.

Single-item measures for stress management and financial difficulties (low reliability).

Sample is from one geographic region (Minnesota) and may not generalize nationally or globally.

Response rate was 46.1%, with potential non-response bias (respondents may have been more or less affected by the pandemic).

Self-report measures are subject to recall and social desirability bias.

**Additional critical notes:**

**No pre-pandemic baseline for disordered eating during the pandemic period:** The authors controlled for 2018 disordered eating, but the 2020 measure still mixes pre- and pandemic behavior. A cleaner design would have asked specifically about "since the pandemic began."

**Multiple comparisons:** The authors tested many associations (4 predictors × 3 outcomes = 12 primary tests) without correcting for multiple comparisons. Some significant results may be false positives.

**Small cell sizes for extreme UWCBs:** Only a small proportion of the sample reported vomiting (2.6%), laxative use (1.4%), or diet pills (3.5%). Analyses of extreme behaviors are based on very few cases.

**No objective measures:** Height, weight, and eating behaviors are all self-reported. People tend to underreport weight and overreport healthy eating.

**Pandemic timing:** Data were collected in April–May 2020, very early in the pandemic. Later waves (e.g., winter 2020–2021) might show different patterns as lockdown fatigue, economic hardship, and mental health effects accumulated.

**No assessment of pre-existing eating disorders:** The study didn't distinguish between people with clinical eating disorders and those with subclinical disordered eating. The "re-emergence" theme suggests that for some, this was a relapse, not a new onset.

For someone running their own n=1 experiment to understand how stress affects their eating behaviors:

**What to test:**

**Stress management techniques** (e.g., daily 10-minute mindfulness meditation, structured scheduling, or a "worry time" journaling practice) vs. your usual coping.

**Financial stress reduction** (e.g., a 30-day spending freeze or a budgeting app) if money worries are relevant.

**Routine stabilization** (e.g., fixed meal times, consistent sleep/wake schedule) vs. unstructured days.

**Minimum meaningful duration:**

**At least 2–3 weeks per condition.** The qualitative data suggest that eating patterns shift within days to weeks of stress exposure. A 2-week baseline + 2-week intervention + 2-week washout (if testing multiple conditions) would be a reasonable minimum.

For binge eating specifically, track for at least 4 weeks, as binge episodes may be weekly or less frequent.

**What to measure (specific metrics):**

**Daily log:** Number of meals/snacks, any episodes of loss-of-control eating (yes/no), any extreme weight control behaviors (vomiting, laxatives, fasting, skipping meals).

**Weekly:** Perceived Stress Scale (P