Table IV.

Results (for studies investigating association of mental health parameters with diet quality)

								Hypothesis outcome			Effect sizea
Author, year	Diet quality tool	Depression instrument	Dietary assessment	Model	Adjustment	Result	OR, HR or RR, β coefficients, or other statistics	1	2	3	Small	Medium	Large
Mental health parameter: depression
Hall et al., 2017 [84]	Macronutrient and micronutrient scores (based on dietary guideline)	GDAAS	24-h dietary recalls	Binary logistic regression	University, diet perception, breakfast consumption, energy intake, soda consumption, weekly vigorous exercise	Depression had no effect on diet quality scores	N/A			X
Keck et al., 2020 [83]	HEI	PHQ-9	ASA24	Multigroup path analysis	Race, marital status, college status, GAD-7 severity, PHQ-9 severity	There was a significant association of adverse PHQ-9 score with decreased total caloric intake and increased sugar intake. There was no effect on total HEI score	Total caloric intake: b = −27.44, SE 10.67, P < 0.01 Sugar HEI component: b = −0.17, SE = 0.05, P < 0.001	X				X
Laza-revich et al., 2018 [82]	N/A	CES-D	69-item FFQ	Logistic regression analysis	Sex, age, BMI	In women, the fourth quartile of depression score was positively associated with frequent consumption of fast food (i), fried food (ii) and sugary food (iii). There were no associations for men	(i) OR = 2.08 (95% CI 1.14 −3.82, P = 0.018) (ii) OR = 1.92, (95% CI 1.17–3.15, P = 0.010) (iii) OR = 2.16, 95% CI 1.37–3.43, P < 0.001)	X			X
Dalton and Hammen, 2018 [91]	N/A	BDI	Standard measures of daily eating habits	Hierarchical generalized linear modelling (Poisson)	Gender	There was no association of depression with daily maladaptive behaviours (including diet habits)	N/A			X
Mental health parameter: anxiety
								Hypothesis outcome			Effect sizea
Author, year	Diet quality tool	Anxiety tool	Dietary assessment	Model	Adjustment	Result	OR, HR or RR, β coefficients, or other statistics	1	2	3	Small	Medium	Large
Hall et al., 2017 [84]	Macronutrient and micronutrient scores, based on US dietary guidelines	Anxiety: Goldberg depression and anxiety scales	24-h dietary recall	Binary logistic regression	University, diet perception, breakfast consumption, energy intake, soda consumption and weekly vigorous exercise	Anxiety was associated with greater risk of low macronutrient quality	OR 2.35 (95% CI 1.27, 4.38)	X			X
Carlos et al., 2020 [97]	KIDMED	STAI	N/A	Multiple regression analysis	Adhesion to the MD, alcohol consumption, level of emotional eating	Adhesion to the MD was not predicted by state anxiety	N/A			X
Keck et al., 2020 [83]	HEI	GAD-7	ASA24	Multigroup path analysis	Race, marital status, college status, GAD-7 severity, PHQ-9 severity	There was a significant association of adverse GAD-7 score with decreased total caloric intake and increased sugar intake. There was no effect on total HEI score	Total caloric intake: b = −30.16, SE = 10.67, P < 0.01 Sugar HEI component: b = −0.16, SE = 0.05, P < 0.001	X				X
Mental health parameter: stress
								Hypothesis outcome			Effect sizea
Author, year	Diet quality tool	Stress tool	Dietary assessment	Model	Adjustment	Result	OR, HR or RR, β coefficients or other statistics	1	2	3	Small	Medium	Large
Kotecki et al., 2019 [86]	Online questionnaire assessing diet quality	Online questionnaire assessing perceived stress	N/A	Regression, ANOVA, ANCOVA	N/A	Perceived stress was negatively associated with diet quality	Mean diet quality scores were 51.32 for low perceived stress versus 50.17 for high perceived stress (P < 0.05)	X
El Ansari and Berg- Beckhoff, 2015 [85]	Dietary Guideline Adherence Index	PSS	FFQ	Multiple linear regression models	Age, sex, living situation (accommodation during term time), economic situation, BMI, physical activity, faculty	Higher perceived stress score was significantly associated with less frequent food intake of fruit and vegetables. There was no significant association between unhealthy foods and stress	b = −0.12	X			X
Daigle Leblanc and Villalon, 2008 [124]	N/A	PSS	Three-day food record and a FFQ	Pearson’s correlations, Student’s t-tests	N/A	Increased stress was associated with an increased consumption of milk and milk products for 1st-year students at the beginning of the first trimester (i) and of breads and cereals for 4th-year students at the end of the first trimester (ii)	(i) P = 0.05 (ii) P = 0.02	X
Nataskin and Fiocco, 2015 [95]	N/A	PSS	Eating habits confidence scale and Block fat and sodium screener	Linear regression analyses	Perceived stress, diet self-efficacy, age, race and sex	Low levels of perceived stress were associated with the lowest levels of fat and sodium intake	b = −1.07 (P = 0.04)	X			X
Peker and Bermek, 2011 [87]	N/A	PSS	Nutrition section of HPLP II	Pearson’s product moment correlation and stepwise multiple linear regression analysis	Age, place of residence, monthly family income, perceived social support and perceived stress	Perceived stress was negatively associated with healthy diet	r = −0.36 (P < 0.01)	X				X
Cheng and Mohd Kamil, 2020 [125]	N/A	PSS	FFQ and 3-day dietary record	Independent samples t-tests and chi-square tests	N/A	There was no significant difference for all food categories between the non-stressed and stressed groups	N/A			X
Ahmed et al., 2014 [88]	N/A	DASS-21	7-day FFQ	Logistic regression analysis	Age, year of study, family income, parents’ education level, marital status, smoking status	Stressed female students were more likely to eat fast foods (i) snacks and beverages (ii) than unstressed female students. No associations were found for males	(i) OR 1.75 (95% CI: 1.02–3:00) (ii) OR 2.28 (95% CI: 1.30–3.98)	X			X
Almogbel et al. 2019 [89]	N/A	DASS-21	FFQ	Chi-square tests	N/A	Stressed participants consumed more junk foods. Non-stressed participants preferred healthier foods	P < 0.05	X
Papier et al., 2015 [90]	N/A	DASS-21	CSIRO FFQ	Logistic regression analysis	Marital status, study status, living situation, working hours, frequency of exercise, BMI, whether participants were trying to lose weight, smoking status	Stress was negatively associated with consumption of meat alternatives, vegetables and fruits. Stress was positively associated with the consumption of highly processed food	OR 2–3, P < 0.05	X			X
Dalton and Hammen, 2018 [91]	N/A	LSI and Daily Stress Measure	Standard measures of daily eating habits	Poisson linear modelling	Gender	Daily stress (i) and chronic stress (ii) were significantly associated with daily maladaptive behaviours (including unhealthy diet)	(i) b = 0.01 (P = 0.02) (ii) b = 0.02 (P = 0.03)	X			X
Errisuriz et al., 2016 [92]	N/A	Stress: measured by single item Stress management: measured by single item	FFQ	Multiple hierarchical linear regressions	Gender, BMI and race	Perceived stress was positively associated with past week soda (i), coffee (ii), energy drink (iii), salty snack (IV), frozen food (v) and fast food consumption (vi)	(i) b = 0.09 (ii) b = 0.15 (iii) b = 0.14 (iv) b = 0.12 (v) b = 0.15 (vi) b = 0.09 (all P < 0.05)	X			X
Kandiah et al., 2006 [93]	N/A	45-itemized stress-eating survey	45-itemized stress-eating survey	ANOVA	N/A	Only 33% ate healthy when stressed (compared to 80% when not stressed). When stressed, sweet foods were chosen	N/A	X
Oliver and Wardle, 1999 [94]	N/A	Stress- induced eating Questionnaire	Stress- induced eating Questionnaire	Descriptive statistics, chi-squared test	Dieting status, gender	Intake of ‘snack-type’ foods increased during periods of stress; females were more likely to consume sweets and chocolate (I). Intake of ‘meal-type’ foods (fruit and vegetables, meat and fish) decreased during stressful periods	(i) Chi-squared = 10.9 (P < 0.01)	X
Mental health parameter: test anxiety
								Hypothesis outcome			Effect sizea
Author, year	Diet quality tool	Test anxiety tool	Dietary assessment	Model	Adjustment	Result	OR, HR or RR, β coefficients, or other statistics	1	2	3	Small	Medium	Large
Pollard et al., 1995	N/A	State anxiety scale	24-h dietary recall	Repeated measures analysis of covariance	Group (examination stress, control, gender, time, year of study, trait anxiety, social support, dietary restraint)	Students with high trait increased their consumption of total fat, saturated fat and total energy intake between baseline and examination sessions	P < 0.05	X
Trigueros et al., 2020 [98]	KIDMED	The Test Anxiety Inventory	N/A	Structural equation model	Exam anxiety, academic stress, emotional intelligence, resilience	Exam anxiety negatively predicted adherence to the MD	Coefficient = 0.37 (P < 0.001)	X			X
Mental health parameter: academic stress
								Hypothesis outcome			Effect sizea
Author, year	Diet quality tool	Academic stress tool	Dietary assessment	Model	Adjustment	Result	OR, HR or RR, β coefficients, or other statistics	1	2	3	Small	Medium	Large
Trigueros et al., 2020 [98]	KIDMED	Student Stress Inventory Stress Manifestation	N/A	Structural equation model	Exam anxiety, academic stress, emotional intelligence, resilience	Academic stress negatively predicted adherence to the MD	Coefficient = 0.49, (P < 0.01)	X			X
Aljaber et al., 2019 [101]	CES	ASS	N/A	Authors used 12 statements to test the hypothesis that students with high stress levels would eat more unhealthy foods	N/A	11 statements proved the hypothesis, only one statement disproved; the authors accepted the hypothesis that students who have a high stress level eat more unhealthy foods	N/A	X
Mansoury et al. [2015, 100]	N/A	PSS	24-h recall food diary analysed using Diet Plan	t-tests	N/A	Participants experiencing academic stress at T2 demonstrated significantly lower frequency of healthy food intake at T2 compared to T1	P = 0.001	X
Mental health parameter: menstrual distress
Author, year	Diet quality tool	Menstrual distress tool	Dietary assessment	Model	Adjustment	Result	OR, HR or RR, β coefficients, or other statistics	Hypothesis outcome			Effect sizea
								1	2	3	Small	Medium	Large
Bu et al. 2019 [102]	N/A	MDQ	15-item FFQ	Multiple logistic regression analyses	N/A	Negative mood was positively associated with tea, coffee and carbonated beverage intake during the menstrual phase (I). Negative mood was positively associated with banana and dates intake during the premenstrual phase (ii)	(i) b = 0.21, P = 0.0453, OR = 1.23 (ii) b = 0.59, P = 0.0172, OR = 1.81	X			X

Note: Studies ordered according to diet quality tool; if no diet quality tool used, studies were ordered according to depression instrument.

Dietary scores: Healthy Eating Index (HEI), Automated Self-Administered 24-h recall (ASA24), Food frequency questionnaire (FFQ), Mediterranean diet (MD), Health-promoting lifestyle II (HPLP II), Commonwealth Scientific and Industrial Research Organization food frequency questionnaire (CSIRO FFQ), Test of Adherence to Mediterranean Diet (KIDMED), Compulsive eating scale (CES).

Mental health scores: Goldberg depression and anxiety scales (GDAAS), Patient Health Questionnaire 9-Item (PHQ-9), Generalized Anxiety Disorder 7-items (GAD-7), Centre for Epidemiologic Studies scale (CES-D), Beck’s depression inventory (BDI), Depression, anxiety and stress scale (DASS-21), UCLA Life Stress Interview (LSI), Academic Stress Scale (ASS), Cohen’s perceived stress scale (PSS), Menstrual Distress Questionnaire (MDQ), State-Trait Anxiety Inventory (STAI).

Statistics: Odds Ratio (OR), Hazard Ratio (HR), Relative Risk (RR), M (mean), Analysis of variance/covariance (ANOVA/ANCOVA), Pearson’s coefficient (r), Beta coefficient (b), Standard Error (SE).

Not applicable (N/A).

Hypothesis: Good mental health will have a beneficial effect on diet quality, and/or bad mental health will have a detrimental effect on diet quality.

Hypothesis outcomes:

(i) Hypothesis accepted.

(ii) Hypothesis rejected—good mental health had an adverse effect on diet quality.

(iii) Hypothesis rejected—no association between mental health and diet quality.

^aIf applicable.