A systematic review of the association of diet quality with the mental health of university students: implications in health education practice

Abstract University students are at risk of experiencing mental health problems during the transition from home to university. This transition can also adversely affect their diet quality. This review aims to examine bidirectional associations from observational studies regarding the influence of diet quality on the mental health of university students, and vice versa. The databases PubMed, CINAHL, EMBASE, PsycINFO, The Cochrane Library and Web of Science were searched using relevant search terms. The searches were last updated on 15 July 2022. Majority of studies (36 out of 45) found that good diet quality of students was associated with better mental health in terms of depression, anxiety, stress and overall general mental well-being. Moreover, majority of studies (19 out of 23) found that stress and anxiety of students were associated with poorer diet quality. The effect sizes observed were generally small–moderate. Healthy diets of students have been associated with better mental health in terms of depression, anxiety, stress or other mental health issues. Stress experienced by university students has been associated with unhealthy diets. There are implications for health education research, as interventions to improve diet quality at the university level could reduce mental health issues; additionally, interventions to support students under stress may lead to healthier dietary habits when living on campuses. Randomized controlled trials and intervention studies are needed to further investigate these implications.


Introduction
The transition to university from home can have both positive and negative effects. Positive effects include developing independence, lifelong friendships and networking. However, when the outcome of the life transition is positive (such as gaining new skills and paid employment), this experience can still be unpleasant and associated with life dissatisfaction [1]. Some examples of challenges refer to academic, social, personal-emotional and institutional adjustments [2]. Change is often a cause of uncertainty, which in turn can induce higher levels of stress and anxiety [3].
The transition to university from home may affect mood and overall mental health. This is a period of increased risk of onset of mental health problems [4], particularly for the onset of depression and anxiety [5].
In the United States, it is estimated that up to 50% of the students living on university campuses can be affected by mental health problems [6]. This observation appears to be an international issue. A meta-analysis of 34 international studies with university students of various years between 1990 and 2010 showed an average prevalence of The association of diet quality with the mental health of university students depression of 30.6% [7]. Rates were substantially higher than those found in the general population (∼11%) [8]. However, there are studies that have not detected differences in the mental health of students and non-students [9]. The differences in the findings examining mental health prevalence statistics could be attributed to the various methods used by studies, as ways in which symptoms have been assessed are not always appropriate for establishing prevalence per se.
Definitions of mental health may also vary. The World Health Organization defines mental health as 'our emotional, psychological, and social wellbeing' [10]. The terms mental health and mental illness may be used interchangeably, however a person may experience poor mental health even without being diagnosed with a mental illness [11]. For this reason, in order to better understand the various dimensions of mental health, it may be appropriate to look for a broad range of studies that use constructs assessing for 'mental health'; these may vary from specific symptom measures, to well-being measures, to positive self-concept, as well as include constructs such as resilience and self-concept.
The transition to university from home also appears to affect the diet quality of students, as this period is characterized by students adopting poor quality diets [12].
A meta-analysis of studies conducted in the United States, Canada, United Kingdom and Belgium showed a student weight gain of 1.4 kg over two terms [12]. This increment in weight is five times higher than the weight gain expected in the general population over a period of 1 year. In the United States, this has been observed in at least two-thirds of students during their 1st year of university [13]. These changes in diet quality are characterized by increments in the consumption of fast food, relying more on take-out food and less on fresh food [14,15].
Organizations, such as the Food Standards Agency in the United Kingdom, have issued guidelines as to what a good quality diet should consist of [16]. For example, these guidelines include advice about the intake of fruit (≥2 servings/day), vegetables (≥3 servings/day), oily fish (≥200 g/week), fat (≤85 g/day) and non-milk extrinsic sugars (≤60 g/day) [17]. There are also specific types of diets that are in line with the above recommendations, such as the Mediterranean diet which consists of fruits, vegetables, whole grains, seafood, beans and nuts, and whose health benefits have been reported in previous literature [18]. Various diet quality measures have been devised to capture the quality of a diet according to guidelines such as the ones mentioned earlier, although studies often use food frequency questionnaires without considering a diet quality instrument.
Although not the only factor, poor quality diet has been considered a risk factor of mental health problems [19], and some argue that mental health issues could impact diet quality too [20]. Understanding this relationship could have important implications in health education practice. Current practices for addressing the mental health of university students include counselling [21], as well as cognitive, behavioural and mindfulness interventions [22]. More recently, the above interventions are being available via the internet in addition to sessions in person by university counsellors [22,23]. Educating students about mental health has also been shown to be effective [24]. In some cases, students utilize pharmacological options, such as antidepressants prescribed by general practitioners or psychiatrists [25]. However, pharmacological and psychological interventions may not always be able to prevent or resolve mental health issues; hence, diet could be a potential target for the prevention and adjunct treatment of anxiety and depression of students [26].
Even though there is a scarcity of studies investigating the association of diet quality with mood in university students, relevant studies involving the general population are more abundant. These studies have mainly focussed on the effects of diet on depression. Reviews of cross-sectional studies have shown inverse associations of smallmoderate effect size between diet quality scores and depressive symptoms [19,27]. Moreover, a recent meta-analysis of randomized controlled trials examined the efficacy of dietary interventions for symptoms of depression in both clinical and non-clinical populations [26]. This review showed evidence that dietary interventions had a smallmoderate effect on improvement of depressive symptoms. Examples of dietary interventions that were used included individualized dietary counselling, group dietary classes and standardized dietary prescription. In view of the above evidence, there is scope to understand the influence of diet on mental health of students and vice versa by performing a systematic literature review of relevant observational studies. Moreover, establishing a link between diet quality and mental health may be used for practical support involving interventions that could improve both diet quality and mental health of students.

Review aims
To better understand the associations between diet and mental health, this review aimed to interpret study findings in the context of the diathesis-stress model [28]. Stress-diathesis models are models that can facilitate our understanding of how predispositional factors from various domains can cause susceptibility to psychopathology and eventually lead to conditions that are sufficient for the development of a mental health disorder [29]. These models may encompass multiple factors contributing to psychopathology, including biological vulnerabilities, psychological susceptibilities, social variables, environmental variables and developmental experiences [29].
The review aimed to interpret findings while taking into consideration the various risk factors that can affect students with a biological, psychological or social vulnerability to mental illness or to bad diet quality. This was in line with both the stress-diathesis model and the biopsychosocial model of health and illness [30,31]. Biopsychosocial factors linked to mental health include stress, stressful life events, body image, physical activity, sleep, social support, use of alcohol or illicit drugs. There are also biopsychosocial factors linked to diet, such as availability and access of pre-prepared meals/fast foods on campus, lack of cooking skills, lack of culinary and basic nutritional knowledge, no previous hands-on involvement in food preparation in the family environment, limited resources including money for shopping, no easy access to healthy food and lack of companionship during meal times [32]. Hence, we aimed to look for moderators or mediators of the association between diet and mental health, in recognition that mental health and diet quality difficulties are multifaceted and underpinned by complex biopsychosocial processes.
Scoping searches did not identify any previous systematic literature reviews appraising both directions of the influence of diet quality on the mental health of university students, and vice versa. However, one previous systematic review appraising the association of mental health with the diet quality of students [20] was identified, which was published in 2018; further scoping searches showed that at least eight relevant studies were published since 2018. In 2021, another systematic review appraised the opposite direction of the association, i.e. the influence of diet on the mental health of students [33]; further scoping searches showed that at least six relevant studies were published since the data search of this review was done. None of these reviews assessed both directions of the association between diet quality and mental health. Even though studies have been treating associations with diet and mental health, and associations with mental health and diet as separate, in reality, most research cannot establish whether one is predicting the other, as they are associations. In view of this, we feel that a full picture can only be obtained by including studies in the review that have assessed either direction of the association.
Hence, the current review aims to provide knowledge by appraising studies investigating the influence of diet quality on the mental health of university students, and vice versa. This is important as the findings may have implications in health education practice. The review aims to appraise studies that have been assessed by the previous reviews [20,33], as well as studies that have never been appraised before.
Given the fact that scoping searches indicated the majority of studies to be cross-sectional, the review The association of diet quality with the mental health of university students did not aim to answer the question of causation in regard to the relationship between diet and mental health.

Methods
This systematic review followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement [34] and was registered in the PROSPERO International Prospective Register of Systematic Reviews (number CRD42020196336 at www.crd.york.ac.uk/PROSPERO). There were no discrepancies between the initial protocol and the processes that were followed.

Search strategy
A search of the literature was performed on 1 July 2020 (date range for searches was from inception to 1 July 2020). The searches were re-run on 15 July 2022 in order to update the review with recent studies. The databases PubMed, CINAHL, EMBASE, PsycINFO, The Cochrane Library and Web of Science were searched by using the following search terms: Student* AND (Diet* OR Nutrition OR Eat* OR Food OR Weight gain) AND (Mood OR Depress* OR Anxiety OR Stress OR Mental health).
'Weight loss' was not included in search terms as previous studies have reported a weight gain (rather than weight loss) in students transitioning to university. Moreover, weight loss due to depression would be more relevant to lack of appetite, which is not the focus of this review.
Both medical subject headings and free-text terms were incorporated, which were adapted according to the database searched. Google Scholar, OpenGrey and ResearchGate were also searched in order to identify any relevant grey literature. This strategy was followed in order to ensure a broad coverage of studies. The reference lists of the included studies and reviews were hand searched in order to identify any additional papers of relevance. Where further information was required, authors of retrieved studies were contacted.

Inclusion and exclusion criteria
The review included studies published in peerreviewed journals or grey literature, including cross-sectional and longitudinal studies, as well as review studies of observational evidence. Scoping searches did not identify any relevant intervention studies, randomized controlled trials or qualitative studies. Intervention studies examining only individual foods/nutrients or focussing only on a single food component were not considered eligible, as the focus is on whole of diet associations and effects. Hence, the focus was on observational studies, as the review aimed to gain insight into the associations of diet quality with the mental health (and vice versa) of university students in their natural environment without any external interventions.
Publication languages included English, Greek and Spanish (as these are the authors' languages). There were no publication period restrictions.
Studies were included if they involved participants who were university students of any ethnic origin, gender and age, studying in any country, with or without a mental health diagnosis (such as depression and/or anxiety).
Studies were excluded if they involved participants who were not university students, if they studied associations of mental health with single macro/micronutrients rather than overall diet quality, or if they studied associations of mental health with nutritional supplements rather than diet quality. Studies were also excluded if they focussed on disordered eating behaviours rather than on diet quality.

Main outcomes
The main outcomes included depression or anxiety or depressive/anxiety symptoms or other mental health symptoms (assessed by relevant scales, or as experienced subjectively by participants, or as diagnosed by health professionals) and diet quality. The review included studies using diet measures (such as food frequency questionnaires) and/or diet quality scores obtained from diet quality instruments.

Screening
Titles were screened by author S.S. for inclusion, followed by screening of abstract and then content. Full texts were obtained in cases where abstract eligibility was considered uncertain or if title eligibility was considered uncertain and abstracts were not available. Studies were included at the abstract screening stage if abstracts were in English, Spanish or Greek, but studies were excluded if their full texts were in other language. Authors of studies were contacted when there was not enough information to decide whether a study met the inclusion criteria. All screening steps were discussed with author G.P.A. The studies meeting the inclusion criteria were selected for the review as summarized in Fig. 1.

Data extraction
Data were extracted from observational studies by using the relevant sections of the Cochrane good practice data extraction form. Data were extracted from reviews by using a modified version of the National Institute for health and Care Excellence extraction form [35]. The data were extracted in an electronic format in order to achieve effective time management and reduce any errors during data entry.
The extracted data included the following: authors, year of publication, setting, study design, sample size, geographical location, follow-up time (if applicable), demographic and clinical characteristics of participants, measures used (where applicable) and main findings (dietary assessment tool used and score used, assessment of depression and/or anxiety, depressive and anxiety symptoms scale and threshold used), confounders used and relevant statistics. In cases where various analyses were completed, the analysis that had taken the largest number of confounders into consideration was used.

Risk of bias/quality assessment
The quality of studies was scored into high, medium and low quality by using the Newcastle-Ottawa Quality Assessment Scale (adapted for cross-sectional studies) [36]. This instrument has a highest score of 10, with 5 points being allocated to selection (representativeness of the sample, sample size, non-respondents and ascertainment of the exposure), 2 points being allocated to comparability and 3 points being allocated to outcome (including assessment of outcome and statistical tests). The guidance of the Centre for Reviews and Dissemination [37] was used for appraising the quality of review papers. Where appropriate, discussion between the authors was used to resolve any uncertainties.

Strategy for data synthesis
A narrative synthesis review [38] of observational studies (and of reviews of observational studies) reporting associations of diet quality with mood and mental health of university students (with or without an established mental health diagnosis), and vice versa, was performed. This was considered to be the best approach to analyse the observational data available.

Results
Following title screening of 13 290 articles, 167 abstracts were read in full and assessed against the inclusion and exclusion criteria. Eighty-six full-text articles that met the inclusion criteria were then retrieved, and the full texts were subsequently screened against the criteria. Authors of three papers were contacted to obtain further information [39][40][41]. The final number of papers that were included in the review was 68 (Fig. 1).
Of the included studies, 44 primary studies investigated the influence of diet on mental health (of which 43 were cross-sectional and one was longitudinal). There was also one review identified that investigated this direction of the association. In terms of study quality (as measured by the Newcastle-Ottawa Quality Assessment Scale and the guidance of the Centre for Reviews and Dissemination), 15 studies were evaluated as of The association of diet quality with the mental health of university students high quality, 29 were of medium quality and one was of low quality (Online Supplemental Materials 1-3). There were no relevant randomized controlled trials, intervention studies or qualitative studies identified.
In terms of studies investigating the influence of mental health on diet, one systematic review and 22 primary studies were evaluated (of which 18 were cross-sectional and four were longitudinal). Three studies were considered of high quality, 19 studies were of medium quality and one study was of low quality (Online Supplemental Materials 1-3). The inclusion criteria were not met by any relevant randomized controlled trials, intervention studies or qualitative studies.
Where diet quality instruments were used, the most common measure was the Healthy Eating Index (HEI). HEI is a measure of diet quality that assesses how well food intake aligns with key recommendations of the Dietary Guidelines for Americans [42]. In terms of mental health instruments, the most frequently used instrument was the Depression, Anxiety and Stress Scale (DASS-21), which is a set of three self-report scales designed to measure the emotional states of depression, anxiety and stress [43]. The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students The association of diet quality with the mental health of university students  Not applicable (N/A). Hypothesis: Good diet quality will have a beneficial effect on mental health parameters, and/or bad diet quality will have a detrimental effect on mental health parameters. Hypothesis outcomes: (i) Hypothesis accepted.
(ii) Hypothesis rejected-good diet quality had an adverse effect on mental health.
(iii) Hypothesis rejected-no association between diet quality and mental health. a If applicable.

Diet quality associations with mental health
In regard to geographical settings of studies investigating the influence of diet quality on mental health, 38% of the studies took place in Europe, 31% in the United States and Canada, 12% in Asia, 12% in the Middle East, 5% in Latin America and 2% in Africa. The studies also varied in terms of the number of study participants, from 36 to 68 559 (Table I).
In order to describe the results, we organized the studies into those that used a diet quality tool as a predictor of diet quality and those that used other kinds of tools, such as food frequency questionnaires. We also organized the results in terms of outcomes, including depression, anxiety and stress (Table II).
Nine studies examined the influence of diet quality on anxiety, of which eight found significant associations. Of these studies, four used a diet quality measure and all found a negative association of diet quality with anxiety [46,49,64,65]. Studies using questionnaires also showed that unhealthy diet was positively associated with anxiety [57, 58,60,63].
Thirteen studies looked into the influence of diet quality on stress, of which 11 found significant associations. Six studies using a diet quality measure found a negative association of diet quality with stress [46,[64][65][66][67][68]. Additionally, studies using questionnaires found that unhealthy diets were positively correlated with stress [51,55,59,62,63], as well as that healthy diets were negatively correlated with stress [51,55,59].
There were 13 identified studies investigating the influence of diet quality on general mental wellbeing, of which 12 found significant associations. Out of the six studies that used a diet quality measure, four concluded that poor diet quality was negatively associated with mental well-being [39,46,69,70], and one concluded that good diet quality was positively associated with mental well-being [71]. One study found a positive association of diet quality with positive emotional state, but no association with negative emotional state [72]. Studies using questionnaires reported unhealthy diets to be associated with bad mental well-being [52,73,74], as well as healthy diets to be associated with good mental well-being [52,[74][75][76][77].
Other mental health parameters that were examined by studies to determine whether they are influenced by diet included post-traumatic stress disorder, academic stress, positive self-concept and psychological resilience. All of these studies reported results towards the expected direction. Specifically, two studies reported that a healthier diet was associated with fewer post-traumatic stress symptoms in university students [52,54], one study reported that high adherence to the Mediterranean diet decreased academic stress in regard to students communicating their own ideas [78], two studies reported that the Mediterranean diet was associated with more positive self-concept [79,80] and one study suggested that better diet quality was associated with better psychological resilience [81].
In terms of effect sizes for studies investigating the association of diet quality with mental health, it was possible to retrieve information for 31 out of the 35 studies that found significant associations (Table II). It was observed that effect sizes were small for 22 studies, moderate for five studies and large for four studies (Table II).

Mental health associations with diet quality
Twenty-two primary studies and one systematic literature review investigated the associations of mental health with diet quality. The identified studies took place in various locations (Table III). Specifically, 32% of the studies were conducted in the United States and Canada, 29% in the Middle East, The association of diet quality with the mental health of university students 14% in Europe, 10% in Latin America, 10% in Asia and 5% in Australia. The number of study participants varied from 88 to 2810 (excluding the systematic literature review). The main findings of these studies are summarized in Table IV. Out of the four studies investigating the influence of depression on diet quality, two found significant associations. In particular, one study found evidence that depressed women were more likely to follow unhealthy diets; however, no associations were found for men [82]. Moreover, one study found that increased sugar intake was associated with more symptoms of depression, but there were no associations of depression scores with the overall diet quality scores [83,84].
Thirteen studies investigated the influence of stress on diet quality, of which 12 found significant associations. Of the studies that showed significant associations, two used a diet quality tool and found that high perceived stress was associated with low diet quality scores [85], as well as that low perceived stress was associated with high diet quality scores [86]. The remaining 10 studies used questionnaires and similarly showed that high perceived stress was associated with unhealthy diets [87][88][89][90][91][92][93][94], as well as that low perceived stress was associated with healthy diets [89,95]. Moreover, evidence from two longitudinal studies suggested that increasing stress over a period of time can be detrimental on diet [91,96].
Other mental health conditions that have been investigated to determine whether they can influence diet quality included anxiety, test anxiety, academic stress and menstrual distress. In particular, three cross-sectional studies investigated the influence of anxiety on diet quality by using a diet quality tool. One found that anxiety was associated with a greater risk of low macronutrient quality [84]; the other two studies found no associations of stress with the overall diet quality score [83,97].
In terms of the influence of test anxiety and academic stress on diet quality, all four studies showed significant associations with diet quality [98][99][100][101]. Specifically, it was found that test anxiety and academic stress negatively predicted adherence to the Mediterranean diet [98]; there was also evidence from two longitudinal studies to suggest that as test anxiety/academic stress increased, the intake of unhealthy food also increased [99,100]. The above findings were also supported by the results of a cross-sectional study [101].
Finally, one study investigated the influence of menstrual distress on diet [102] and found that negative mood during the menstrual/premenstrual phases was associated with diet changes. For example, negative mood was positively associated with ingestion of tea, coffee and carbonated drinks, as female students may have been trying to stimulate their nervous system to alleviate their negative mood through these diet changes.
In terms of effect sizes, it was possible to retrieve information about effect sizes for 10 studies (Table IV). The effect sizes observed were small for eight studies and moderate for two studies.

Review papers
No previous systematic reviews appraising both directions of the association of diet quality with mental health were identified.
One previous review [33] aimed to examine the influence of diet on depression and anxiety among college students. In contrast to our review, this review filtered out articles published before 2000, and only included students enrolled in at least 2year programmes. Moreover, this review focussed only on depression and anxiety, but no other aspects of mental health. This review assessed 16 crosssectional studies, of which 14 fulfilled our inclusion criteria and have also been included in our review [44, 47, 48, 50, 53-55, 58, 59, 73, 82-84]. The results of this review were in line with our results, as the authors concluded that most of the cross-sectional studies found a positive influence of healthy diets on depression and anxiety, with a few studies finding inconsistent results.
One previous systematic review appraising the influence of stress on the diet of students [20] was identified. In contrast to our review, the authors reviewed the influence of stress but no other mental health disorders. The authors also included studies involving disordered eating and maladaptive 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 The association of diet quality with the mental health of university students 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  The association of diet quality with the mental health of university students   The association of diet quality with the mental health of university students  The association of diet quality with the mental health of university students  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. a If applicable.
weight-related behaviours, which was not the scope of our review. The stress and dietary intake section of the review identified 12 studies, all of which were considered to be of relevance and have been included in the current review [51,55,59,66,85,87,90,[92][93][94][95]99]. The conclusions of this review were in line with our finding that there is a positive association of stress with unhealthy diet, as well as a negative association of stress with healthy diet in university students.

Main findings
Sixty-six primary studies and two systematic literature review were reviewed in total. Majority of primary studies (n = 53) and the two reviews showed results in the expected direction, where good diet quality was associated with good mental health, and good mental health was related with good diet quality. In terms of dietary parameters, students consuming high-quality foods (including fruits, vegetables, nuts and fish) reported fewer mental health symptoms, compared with students who had a high intake of pro-inflammatory foods (such as processed meat, refined carbohydrates, desserts and sweetened beverages), who reported more mental health issues. 36 out of 45 studies supported this association. This was in line with the stress-diathesis model, with the unhealthy foods acting as the stressor and being associated with the development of mental health difficulties. From a mechanistic point of view, there are various processes that could be mediating the relationship between diet quality and mental health. These include inflammation which is a core aetiological feature of depression [103]; various components of a healthy diet may be reducing inflammation [104]. Other candidate biological mechanisms include oxidative stress [105], maintenance of beneficial gut microbes [106], adult hippocampal neurogenesis [107], modulation of the tryptophan-kynurenine metabolism [108], maintenance of mitochondrial biogenesis [109] and regulation of epigenetic processes [110,111].
No studies found an adverse influence of good diet quality nor a beneficial influence of bad diet quality on mental health. These results confirm that unhealthy diets such as western diets have no benefits on the mental health of students, which is in line with the findings of previous studies [107].
Majority of studies (19 out of 23) also showed associations of mental health with diet quality, which were in the expected direction. The most compelling evidence was in favour of high levels of stress having an unhealthy influence on diet. Despite a lack of longitudinal studies, the four identified longitudinal studies that either investigated the influence of stress or academic stress on diet quality showed that stress is associated with unhealthy diets over time. These results are in line with previous research in adults, which showed that stress was associated with unhealthy eating [112], including intake of foods with high sugar content.
A potential mechanism mediating this association may be the hypothalamus-pituitary-adrenal axis being hyperactive in depression and anxiety, leading to an increase in serum cortisol. As a consequence, appetite may be increased with a preference for energy-dense foods at the expense of healthy foods [113,114]. Stimulation of the appetite-stimulating hormone ghrelin during stress may also be of relevance [115].
The most common country of study origin was the United States, where the characteristics of university life may be different than other developed or third world countries. However, studies in countries of lower socio-economic status were included in the review achieving good global coverage.
In terms of interpreting findings in the context of the stress-diathesis model, some of the studies recognized the complex relationships of biopsychosocial factors on both mental health and diet. Examples of such factors affecting mental health that we identified included stressful life events, body image, physical activity, sleep, social support, use of alcohol or illicit drugs. We also identified factors linked to diet, such as availability and access of pre-prepared meals/fast foods on campus, limited resources including money for shopping, no easy access to healthy food and lack of companionship during meal times.
The review identified that more studies used food frequency questionnaires rather than diet quality measures. This means when only food frequency questionnaires were used, it was not always possible to obtain a clear picture in regard to the extent that diets of participants conformed to dietary recommendations. In contrast, scores of validated diet quality instruments based on dietary recommendations were easier to interpret and contributed to a better understanding of the associations between diet quality and mental health.

Limitations
Strength of this review is the fact that an exhaustive review of the literature was performed, including grey literature. However, it is not possible to exclude the possibility of having missed studies due to publication bias, in cases where non-significant results had not been published.
Some studies included participants who were 17 years old, which would classify as adolescents. Even though these studies were limited, the possibility of an age bias from these studies cannot be excluded.
This review largely relied on cross-sectional studies, which assessed diet and mental health at a single point of time. This means that it has not always been possible to make definite conclusions about the direction of associations or about the changes in diet quality or mental health over time.

Implications for health education practice
Even though the effect sizes of the included studies were mostly small or moderate, the observed results still have implications in health education practice. Given the evidence that unhealthy dietary practices are associated with worsening mental health of students, mental health education at university should aim to raise awareness of this association. This could be done along with input from nutritionists and could involve the introduction of relevant modules in health courses, online courses, as well as the use of leaflets and posters in campuses [20]. This review has not identified any relevant whole-diet intervention or randomized controlled studies to improve mental health of students through diet, or vice versa. Results for an ongoing randomized controlled trial [116] are awaited, which involves the use of a web-based wellness platform to support healthy living of students by focussing on nutrition and physical activity.
An approach involving web interventions or lectures could also be followed to educate students about the relation of stress with unhealthy dietary habits. It has been suggested that the introduction of mindfulness-based stress reduction techniques and mindful eating may be effective techniques to address this issue [20,117].
Universities may need to review how they operate, what economic and planning decisions they make in terms of which franchises they allow onto campuses, influence what is sold in shops and raise awareness of the importance of diet quality. By helping students improve their diet quality, they may experience fewer mental health issues, as well as fewer and less severe depressive episodes. Similarly, by helping students deal with stress, they may be able to experience healthier dietary habits during their university education.

Future recommendations
Most of the identified studies were cross-sectional, as they assessed diet and mental health at a single point of time. We recommend that future studies use a longitudinal design when possible, enabling researchers to determine the direction of any detected associations. However, we recognize that longitudinal studies can pose challenges in terms of time and costs required.
Apart from the available cross-sectional studies establishing some associations, we also need wellpowered clinical trials to further assess the associations of diet quality with the prevention, severity and relapse of depression, stress and other mental health issues of students. The findings would inform the design of further studies including randomized controlled trials and intervention studies. Such studies may provide more insight about the relationship between the three factors of stress, diet and emotion. However, it is recognized that designing and executing such studies may be challenging and that difficulties including randomization may be anticipated.
We used the stress-diathesis model to interrogate the data, where the diatheses could be of biological, psychological or social nature. We noticed that studies did not always include data relevant to this model. For example, biological factors such as genetic predispositions or gut microbiota of students were not reported. Psychological factors such as perfectionism traits were also not reported. Moreover, social factors were not always reported, such as lack of ability to form group memberships, lack of culinary and basic nutritional knowledge and lack of resources and access to healthy food. Hence, the included studies only partially covered the stress-diathesis model, meaning there is scope for future studies to use the stress-diathesis model as a reference.
In terms of dietary instruments, the HEI was the most common instrument used; however, it is a long instrument and students may not always engage. A previous study assessing the relationship of diet quality and mental health in adolescence has highlighted the need of a brief, validated measure of diet quality to be used in studies involving adolescents or young adults [118]. An example of such a measure might be the short-form food frequency questionnaire [17], which may maximize the student engagement for future studies. In terms of mental health measures, the DASS-21 was the most commonly used measure and is a good option as it provides information about the three mental health aspects of depression, anxiety and stress. Results of studies with consistent tools for diet quality and mental health might enable the execution of a meta-analysis in the future.

Conclusions
The review results show observational evidence that a healthy diet of university students is associated with better mental health, as students who scored favourably in the diet quality instruments also scored favourably in depression, anxiety and stress scales. The opposite applies for university students following unhealthy diets, as unhealthy diet in this group of students is associated with depression, stress, anxiety and other mental health issues. There is also observational evidence to suggest that stress experienced by university students is associated with deterioration in their diet quality, including a reduction in the intake of fruits and vegetables and an increase in the consumption of sweets and fast food.
In order to establish the effectiveness of potential interventions for maintaining a healthy diet and good mental health of students, further observational studies, as well as randomized control trials, would be required. This would allow the determination of whether interventions to improve diet quality at the university level could reduce mental health issues and whether providing support to students under stress may lead to healthier dietary habits when living on campuses.

Supplementary data
Supplementary data are available at HEAL online.

Funding
This research received no specific grant from any funding agency in the public, commercial, or notfor-profit sectors.