{"paper_id":"00eb3c70-9d26-4aa5-ba5d-e00746266dfa","body_text":"Meta-analysis of 633,317 individuals shows associations between healthy diets and mental health in 23 low- and middle-income countries | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Meta-analysis of 633,317 individuals shows associations between healthy diets and mental health in 23 low- and middle-income countries Thalia Sparling, Cesar Cornejo, Bryan Cheng, Lisa Troy, Suneetha Kadiyala This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6530671/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract There is growing evidence of the association between poor diet quality and common mental disorders, which together contribute to global health syndemics. However, there is no synthesis quantifying associations, assessing robustness of evidence or identifying gaps in Low- and Middle-Income Countries (LMIC) where these concomitant health burdens are most prevalent. We used an Evidence and Gap Map of over 3000 systematically selected, peer-reviewed studies linking food security, diets, and nutrition to anxiety, depression, stress and mental wellbeing (2000-2024). From this, we selected studies investigating associations between healthy diet patterns and mental health symptoms measured by validated tools. Eighty-three studies from 23 countries met inclusion criteria (depression n=69; anxiety n=43; stress n=26), reporting statistical measures for 633,317 unique individuals pooled from 65 LMIC sample populations. Healthy diets were associated with less depression, anxiety, and stress. The Standardized Mean Differences (SMD), expressing effect size in standard deviation units, were -0.29 for depression (95% CI -0.35 to -0.23), -0.25 for anxiety (95% CI -0.35 to -0.16), and -0.24 for stress (95% CI -0.33 to -0.14). Results remained robust when restricted to low Risk of Bias studies: depression (SMD = -0.23, CI: -0.31 to -0.16; n=266,831), anxiety (SMD = -0.19, CI: -0.30 to -0.09; n=116,248), and stress (SMD = -0.22, CI: -0.33 to -0.11; n=12,338). Findings were consistent in direction and magnitude across study designs, dietary measurements, diagnostic tools, and country income levels. We found that mental health is better in individuals with healthy diets in LMIC. Methodological limitations (e.g., cross-sectional design) and few studies from low-income countries created evidence gaps. Low-income settings experience disproportionate health vulnerabilities; thus, building on the relationship between diet and mental health can inform actions to improve both. Health sciences/Diseases/Psychiatric disorders/Depression Health sciences/Diseases/Nutrition disorders/Malnutrition depression anxiety stress common mental disorders food intake dietary patterns systematic review multi-level meta-analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction We are living in an era of global health syndemics, where multiple diseases interact synergistically within populations under conditions of social and structural inequity 1 . The most vulnerable people bear concomitant health burdens, compounding their health risk overall. Common mental health disorders (CMDs) such as depression, anxiety and stress, contributed 17.2% of life years lost to disability in 2021 2 , with higher prevalence of CMDs reported in LMICs 3 . CMDs are drivers and consequences of other marginalisations, such as food insecurity and unhealthy diet. In turn, 35% of the global population could not afford a healthy diet in 2022. Of those, 1.68 billion (nearly 60%) live in LMIC 4 , where on average only 2.1% of government health expenditure goes to mental health 5 and public spending on agriculture is both extremely low (less than $10 per rural inhabitant in LIC) and stagnant 4 especially problematic since the burdens of food insecurity and malnutrition are greatest in these settings. In LMICs, CMDs are almost always measured through non-specialist administered screening tools 6 . The ‘gold standard’ of CMD measurement has classically been clinical interviews to determine alignment with diagnostic criteria, for instance the Diagnostic and Statistical Manual of Mental Disorders (DSM) 7 . However, in global mental health research, there is a shift away from diagnoses towards implications of poor mental health, such as functional loss or wider impacts on other health indices 8,9 . One reason is that clinical resources and expertise rare and rarely accessible in low-resource settings. Furthermore, for diverse populations with greater health vulnerability, strategies are emerging to ensure cultural and discriminant validity, address cross-cultural equivalence, and prioritise measurement of functional impairment, which can be measured through symptoms of CMDs 10,11 . Thus, instruments designed as screening tools (and validated as threshold scores’ correspondence to probability of clinical diagnosis) have in fact become stand-alone assessments of both the range and extent of common mental health symptoms in a population. Thus, in LMICs, mental health measures are usually framed through ‘symptoms’ of common mental health problems. A healthy diet consists of balanced proportions (enough to meet caloric adequacy but not exceed it) of primarily whole grains, vegetables, fruit, nuts, seeds, and where available fish, while moderating meat and dairy, and limiting consumption of sugars, saturated fats, and ultra-processed foods (UPFs) 12 . In simpler terms, a healthy diet is health-promoting and disease-preventing 13 . Healthy diets are measured through many different instruments, including adherence to (national) dietary recommendations, adherence to certain diets (e.g. the mediterranean diet (MD) or the dietary inflammatory index (DII), scoring or counts of food items grouped to indicate dietary diversity and/or certain index foods that are considered beneficial to health (e.g. dietary diversity of women, children or households). These indices and measures are validated similarly in that they correspond to increased probability of nutrient adequacy (at least in aggregate at the population level). We can only hypothesise about the proportion of the world’s population who simultaneously experience poor mental health and poor diet quality. However, there is research to show that these burdens are interlinked, and can exacerbate one another 14 . Much of this evidence is descriptive rather than causal or mechanistic 14,15 . Some robust studies show that healthy diets improve mental health 16–23 , while longitudinal studies examining bidirectional relationships show that better mental health can also contribute to eating healthier diets 24 , including in children 25,26 . Healthy diets resulting in better mental health is plausible from both physiological and social arguments. For instance, healthy diets are usually considered more diverse, containing higher quantities of foods that contain antioxidants (e.g., flavones, polyphenols, anthocyanins), long-chain fatty acids (e.g., Eicosapentaenoic acid and docosahexaenoic acid), and other essential nutrients that lower risk of nutrition-related chronic disease (e.g., diabetes, cancer) 12 . These dietary components reduce inflammation 27–30 . Diet can thus improve neurotransmission and neurocognitive function through enhanced brain plasticity via brain-derived neurotrophic factor (BDNF) 31 , which collectively can improve mental health. Potential biological pathways related to mental disorders include inflammation, oxidative stress, the gut microbiome, brain plasticity, and mitochondrial dysfunction, with evidence showing that healthy dietary patterns can beneficially modulate each of these interconnected systems 28,32,33 . Higher socioeconomic status and better lifestyle factors may also underpin both healthier eating and better mental health 34 . In the reverse direction, poor mental health can impact diet quality through multiple pathways. Depression and anxiety are associated with diminished motivation, cognitive impairments, and reduced executive function 35,36 , all if which can hinder meal planning, food shopping and preparation 37 . People experiencing poor mental health may also have altered appetite regulation and food preferences, leading to consumption of energy dense and UPFs 38 , although evidence now more strongly points to UPF consumption as a risk factor for depression 39 . The impact of mental health on diet can extend beyond the individual to the household level, as studies have shown that children of parents, particularly mothers, with CMDs are more likely to experience worse diet quality and child feeding outcomes 40–42 . A clear association between healthy diets and mental health has been shown with nationally representative 43 and longitudinal 44 data in numerous high-income countries. Less is known about the interplay between diet and mental health in LMICs, a critical gap given that they host higher burdens of both poor diet quality and symptoms of CMDs. Here, we quantify the associations reported from studies in LMICs that compare healthy diets (i.e., measured through adequacy, adherence to dietary guidelines, or dietary patterns), with depression, anxiety and stress measures, assess their robustness, and identify the evidence gaps. We report according to the MOOSE guidelines for reporting on meta-analyses, the checklist for which is attached in Supplementary Methods 1. Results Study selection The initial repository of studies came from an Evidence and Gap Map (EGM), which was systematically created from searching three databases (Medline, CAB Global Health and PsychInfo) for English language studies linking food security, and nutrition measures to measures of depression, anxiety, and stress in the general population, published from January 1, 2000, to January 31, 2024. We screened 30,896 records published from 2000 through June 2020 for the original EGM, resulting in a map and analysis of 1945 studies45. To update the EGM, we screened 13,490 additional records published from 2020 to January 2024 and included 1,107 additional records46. The updated EGM includes over 3,000 studies on this topic. From this, we systematically included a subset of studies for the meta-analysis. We selected studies on ‘healthy diet’ using standard indices or factor/cluster analysis methods, and validated measures of depression, anxiety, and stress. A total of 96 studies met inclusion criteria from the EGM, from which we excluded 13 studies: 7 had a mental health outcome different from depression, anxiety or stress (e.g., common mental disorders and mental wellbeing) 1 reported the same data and results in another included study, 4 lacked comparability (did not define healthy vs unhealthy diet patterns), 1 did not have relevant data reported (only percentiles), from which it is not possible to calculate comparable statistics. Sample populations Eighty-three studies were included (depression=69; anxiety=43; stress=26), from which we extracted 139 effect estimates (depression=70; anxiety=43; stress=26) from 65 unique sample populations (depression=58; anxiety=31; stress=20). These estimates were pooled from 633,317 unique individuals (depression=473,236; anxiety=146,217; stress=24,690). The most studied populations were adults (n=42), followed distantly by females (n=16), mid- to later-life populations (n=12), pregnant women or mothers (n=10), and adolescents (n=7). The list of included studies is presented in Supplementary Results 1. Sample settings The 83 studies covered 23 countries: 4 low-income, 11 lower-middle income and 69 upper-middle income (Fig. 2), where 1 study was across six countries, 3 of which were low-income and 3 were low-middle income. Most studies were based in Asia (n=70), of which 39 were from Iran and 17 were from China (both upper-middle income countries). The rest of Asian countries studied were Turkey, Jordan, Nepal, Bangladesh, and India. Outside of Asia, 8 studies were from South America (all from Brazil) and 6 from Africa (3 of which were from Ethiopia). The low-income country studies were from Burkina Faso, Ethiopia, Lebanon, Syria and Uganda. Although we only include results from LMICs in this meta-analysis, we identified 247 additional studies from high-income countries from the EGM, which are included for comparative purposes in Fig. 2. There were 171 countries world-wide where there was not a single study on this topic. Study characteristics There were no published studies on this topic between 2000 and 2012 and then there was a steady annual increase from 1 in 2013 to 15 in 2023. Figure 3 shows the distribution of studies per country-level income (multi-country studies were counted once per each country studied). The primary study design was cross-sectional (n=71). The rest were longitudinal (n=9), and case-control (n=3). We classified the proposed direction of association in studies, regardless of design (i.e., diet as an exposure for mental health outcomes and vice versa). Most (n=73) framed their research as studying the effect of diets on mental health symptoms, and 10 declared studying the effect of mental health symptoms on diets. All but 7 estimates were based on mental health assessments using validated instruments designed as screening tools. They generally include range of symptoms, and a ranking of how often or to what degree those symptoms are experienced. Dietary measures were based either on a priori methods, i.e., a prescribed conception of a healthy diet pattern, or a posteriori, i.e., data-driven methods. Measures were grouped into into 5 groups, described in Table 1. One study used dietary measures from two groups. Groups 1-4 used a priori measures, while Group 5 used a posteriori methods. The complete distribution of dietary measures and mental health diagnostics tools is also in Table 1. Summary of effects From 139 effect estimates, we pooled the effect for the association between healthy diets and depression, anxiety and stress (Fig. 4). We found that individuals with healthy diets showed less depressive symptoms compared to those with less healthy diets (SMD = -0.29, 95% CI -0.35 to -0.23), with a mean difference of 0.29 standard deviations in depression between groups. Similarly, healthy diets were associated with less anxiety (SMD = -0.25, 95% CI -0.35 to -0.16) and less stress (SMD = -0.24, 95% CI -0.33 to -0.14). The effect sizes from individual studies ranged from -1.50 to 0.09 for depression, -0.89 to 0.01 for anxiety, and -0.89 to 0.00 for stress. The prediction intervals (versus the confidence intervals) were more conservative in their estimates of the true population-level effect (Supplementary Results 2). Evidence strength We adapted the Quality in Prognostic Studies (QuiPS) tool, which is comprised of six domains of potential bias in observational studies. The QuiPS tool is included in Supplementary Methods 4. We report on 5 bias domains evaluated: study participation (i.e., selection bias), exposure measurement, outcome measurement, confounding, and statistical analysis and reporting. Since almost all studies were cross-sectional, we omitted ‘study attrition’ for these, and the rankings on this domain for longitudinal studies is included in Supplementary Results 3. Studies that had no ‘high’ risk of bias in any domain were grouped into an overall low risk of bias group, and those with at least one high risk of bias rating in any domain were grouped into overall ‘high’. There was strong evidence of slightly weaker pooled effect estimates from studies with overall low risk of bias in each group (depression=35; anxiety=18; stress=8): depression (SMD = -0.23, CI: -0.31 to -0.16; n=266,831), anxiety (SMD = -0.19, CI: -0.30 to -0.09; n=116,248), and stress (SMD = -0.22, CI: -0.33 to -0.11; n=12,338). Given that most studies were cross-sectional, the most likely sources of bias were study participation (domain 1) and confounding (domain 4). The ranking of potential bias from study participation was based on assessing the source of target population, method used to identify population, recruitment period, place of recruitment, inclusion and exclusion criteria, adequate study participation, and baseline characteristics. As all but 3 studies treated diet as the exposure, the risk of study participation bias stemmed from the likelihood that the relationship between diet quality and mental health is different for participants in healthy eating and non-healthy eating groups. When restricting the analysis to studies with low risk of study participation bias, the estimates remained robust: depression (SMD = -0.18, CI: -0.26 to -0.10), anxiety (SMD = -0.14, CI: -0.26 to -0.02), and stress (SMD = -0.11, CI: -0.23 to -0.01). The risk of bias from confounding was assessed based on whether important confounders (factors related both to healthy eating and mental health) were measured, defined, valid and reliably measured, had uniform method and setting across all participants, whether they followed a valid approach to deal with missing data, and if important potential confounders were accounted for in the analysis. When restricting the analysis to studies with low risk of confounding bias, we found similar results, except for stress, as there were only 2 studies in this subgroup, and with weaker evidence of these effects given the smaller sample sizes (SMD = -0.26, CI: -0.55 to 0.02), anxiety (SMD = -0.06, CI: -1.05 to 0.92), and stress (SMD = -0.25, CI: -0.73 to 0.22). Pooled estimates by overall risk of bias, as well as study participation and confounding risk of bias (all low vs. high) are in Supplementary Results 4. Outliers and influential studies We identified 2 outlier (studies where the size or direction of the effects deviate substantially from the majority) and 6 influential studies (studies that significantly impact the overall results) based on studentized residuals. However, when restricting the analysis to studies with low risk of bias, no outliers remained for any mental health group, and only 1 study remained an influential study for depression. All outliers and influential studies are further discussed in Supplementary results 5. Sub-group and sensitivity analyses Findings were consistent in direction and magnitude across country income levels, study design, and dietary measurements. By country-level income classification, low-income countries had only 3 estimates for depression, 1 for anxiety and 0 for stress, while lower-middle income countries had only 4 estimates for anxiety and 2 for stress. For any country-level income group with more than 4 studies, the results were as follows: depression in lower middle-income countries = SMD -0.39 (CI: -0.71 to -0.07; n=10); depression in upper middle-income countries = SMD -0.28 (CI: -0.34 to -0.22; n=57); anxiety in upper-middle income countries = SMD -0.20 (CI: -0.27 to -0.13); n = 38), stress in upper-middle income countries = SMD -0.26 (CI: -0.36 to -0.16; n=24 studies). When we restricted analyses by study design to the nine longitudinal studies, results were almost identical. By diet measure, most studies used a priori tools to measure diets associated with a reduction in nutrition-related chronic diseases, which alone produced stronger pooled effect sizes: depression=SMD -0.36 (CI: -0.51 to -0.21; n=18); anxiety=SMD -0.25 (CI: -0.41 to -0.09; n=14); and stress=SMD -0.32 (CI: -0.56 to -0.09; n=12) (Supplementary Results 6). Results were not sensitive to the choice of different effect size indices, and we report Cohen’s d and Fisher’s z transformation estimates in Supplementary Results 7. Discussion We found consistent associations between healthy diets and better mental health when pooling eligible studies from LMICs. The magnitude of pooled effects was similar for depression, anxiety and stress (-0.29, -0.25 and -0.24, respectively), and we are confident that these results are not due to chance and are reasonably precise. This analysis presents the most robust estimates of these relationships in LMICs to date, supported by our rigorous methodology. Our results align with other similar meta-analyses. An umbrella review of meta-analyses included very similar measures of ‘healthy diets’ but looked only at depression, in any setting or population, but did not pool effect estimates from the included meta-analyses 15 . Overall, they (Gianfredi et al.) found the methodological quality low to critically low, but did conclude that there was suggestive evidence linking healthy diets defined a posteriori with depressive symptoms or diagnosis, which was supported by a meta-analysis focusing only on a posteriori diet measures, and results from pooling 8 prospective studies on dietary patterns and incident depression 47,48 ). Gianfredi et al. found stronger evidence and effects for the links between higher adherence to the Mediterranean diet and lower scores on the DII and lower risk of depression. These conclusions are similar to quantitative estimates from Lassale et al. (2019), who found that the Mediterranean diet conferred a relative risk (RR) of depression of 0.67 from four longitudinal studies (95% CI 0.55 to 0.82), and a lower DII similarly protective (RR 0.76; 95% CI: 0.63 to 0.92) from four longitudinal studies 49 . In older populations, higher DII score was also associated with incidence of depression (OR 1.33; 95% CI 1.04 to 1.70) from prospective studies, although they did not find associations with the Mediterranean diet or ‘healthy diet’ 50 . The Lassale et al. meta-analysis of 8 prospective cohorts and 9 cross-sectional studies on DII alone estimated that diets with higher inflammatory potential increased odds of depression by 45% (95% CI 1.30 to 1.62) and anxiety by 66% (95% CI 1.41 to 1.96). They also found evidence for the protective effects of a higher HEI/AHEI score (RR 0.65; 95% CI 0.50 to 0.84) from mixed longitudinal and cross-sectional studies 49 . The Gianfredi umbrella review found no evidence for vegetarian diets, also supported by another meta-analysis on this topic 51 . These synthesis studies share certain similarities with our work: they generally find that healthier diets are associated with better mental health, and even the direction and magnitude of effects are similar. Almost all note methodological limitations and heterogeneity, as we do. These studies differ in important ways from our analysis: almost all focus only on depression whereas we include anxiety and stress as well. Some mix prevention and treatment of mental health problems; we excluded treatment research since selecting participants into a study based on poor health status fundamentally confounds the relationship we were interested in testing. Some also found differences based on diet measurement, whereas our results were similar for all included measures, even if some of the evidence of effect was weak. None of the previous analyses focus on LMIC settings. We find that most of the LMIC evidence comes from two middle-income countries: Iran and China (together 68% of included studies), even though our study includes findings from 23 LMICs. There were several studies from other places such as Brazil, Bangladesh, Ethiopia or Turkey, but these were not numerous or methodologically strong enough to make conclusions about these different contexts. Thus, despite pooling many studies, there are still important evidence gaps to fill from low-income settings, and in LMICs other than Iran and China. Nonetheless, we show robust relationships between healthy diets and mental health symptoms. We gain both confidence and novel insights from sub-analyses, which showed similar direction and magnitude of effects despite some loss of power from smaller sub-sample sizes. For instance, when including only low risk of bias studies, the effect estimates weakened only somewhat (for depression -0.29 SMD to -0.23; for anxiety -0.25 to -0.19, and almost no change for stress). Changes could be explained by stronger study designs (which were rated lower risk of bias) that account better for the many factors at play influencing these relationships, such as controlling for a history of mental health issues or using more precise or standardised measures of diets (e.g., HEI versus a factor-derived, sample-dependent dietary pattern). We found consistent effect estimates across dietary measurements, mental health screening tools and study designs. For instance, regarding dietary measurements, measuring quantitative dietary intake and calculating dietary adequacy (group 1 and group 4) could capture a different aspect of a ‘healthy diet’, than a qualitative count of different food groups (group 2 and group 3) or a factor analysis of dietary data in a population (group 5). Reassuringly, we find similar effect estimates across dietary measurements, even when similar studies did not. Additionally, we found that certain measures were more common in LMICs than others. For instance, dietary diversity, a validated measure of diet quality of women and children, is now a standard indicator in LMICs, but is less commonly used in HIC 52,53 . In contrast, very few studies used dietary adequacy or adherence to national guidelines because national diet guidelines are both less defined and less measured in LMICs. Although we are confident that our results are robust and precise, we also note the prediction intervals of our estimates, which are a better measure of population variance versus the sample variance. The prediction intervals from our results indicate that the true average effect of consuming a healthy diet, based on heterogeneity of the studies, could be protective or have no effect on mental health. The interpretation of the prediction intervals aligns with the heterogeneity of the studies included and may also reflect small but possibly real differences in sample populations and settings. Strengths Our study benefits from a rigorous methodology, including a broad, thorough search of the topic to create the EGM, and state of the art screening and coding processes to identify studies. We also cast a wide net for relevant studies, including multiple measures of common mental health issues, and a broad definition of ‘healthy diets’, which allowed us to compare associations across several different facets of these topics. We were able to include many studies, and carefully considered data dependencies, arriving at a three-level model accounting for overlapping study populations. This means that we neutralised the bias coming from multiple reporting of estimates produced from the same populations. We explored consistency and robustness of the evidence, finding clear associations. Limitations We searched the literature through three databases and until January 2024, so studies indexed in non-English repositories, not in English or published after this date are not included. However, we do not believe that the results would meaningfully change based on the number of studies included from the geographies represented and countries that often publish in their dominant language are relatively well-represented in our analysis (e.g., China, Brazil). We classified country-level income status at the time of this analysis, and thus some classification may have changed (almost always to higher income status) since the time of publication or data collection. We included studies on ‘healthy’ diets, and excluded any focused solely on ‘unhealthy’ diets, which means that we may have excluded some dietary measures that are in fact related to worse mental health. We chose this because unhealthy diets is harder to define. ‘Unhealthy’ can mean too much consumption in general, too much of the wrong foods (e.g. foods linked with health issues such as diabetes like sugars and UPFs), too little calories (e.g. wasting and thinness), or too few micronutrients (e.g. growth faltering, reduced immunity and micronutrient deficiency). There are also increasing layers of assessment for diets that correspond to planetary rather than human health, for instance the water footprints or fertiliser demands for certain crops in certain settings, or carbon emissions for specific foods or transport processes. These relationships will also have bearing on mental health status, but this was beyond the scope of this analysis. Practical implications and ways forward Our findings, together with other meta-analyses on related topics, strongly indicate that healthy diets are linked to better mental health. Findings also highlight the striking need for evidence that unpacks causal mechanisms and pathways to impact, including from studies that test what changes, modifies or mediates the relationship between diets and mental health (e.g., disentangling physiological from socioeconomic effects). We will only be able to answer these questions by design: longitudinal, prospective analyses, accounting robustly for a variety of potentially influential factors, reducing likely sources of bias, across diverse settings. Advances in intervention research on this topic are urgently needed. There are promising findings that mental health can improve through dietary interventions 54–57 , that counselling and integrated mental health interventions can improve nutrition outcomes 58,59 , that nutrition-sensitive interventions can improve mental health even when it is not a primary (or even secondary) outcome 60 , or that mental health improves through intervention components not specifically designed to improve mental health 61 . Corroborating these findings and expanding this evidence base will enable us to act on the relationships that are clear in this analysis. The potential synergies between healthy diets and better mental health could prove an important lever to address widespread burdens of poor mental health and poor diets, especially where concentrated among those most vulnerable to poverty and poor health, as well as those most at risk from increasing environmental and political instability. From a policy perspective, building upon the interdependencies between diet and mental health is nascent, but is essential for policies and programmes focused on basic wellbeing or preventative care. For example, In Ethiopia, the Ministry of Health has demonstrated a strong commitment to integrating mental health services into national health programmes, particularly into primary health care, but also into its health extension and social protection activities for food security e.g., the National Social Protection (NSP) Policy, despite implementation challenges 62,63 . For instance, mental health integration has been trialled within specific branches of the Productive Safety Net Program (PSNP) by including components such as care groups and Interpersonal Psychotherapy Group (IPT-G) for Depression 64 . Mental health screening for caretakers has also been proposed as an important part of Community Management of Acute Malnutrition programmes 65,66 , and long-understood as an important step in the primary care pathway 67 . We have learned a lot about multisectoral programming for nutrition (e.g. on nutrition-sensitive agriculture, social protection, and school meals) 68 . Our findings could be integrated with these lessons, especially policies and programmes aiming to address multiple indicators of the Sustainable Development Goals, as well as wellbeing overall, which likely to centre in the post-2030 agenda 69,70 . We found that overall, healthy diets are associated with less depression, anxiety and stress in LMICs, with more than half of the evidence emerging from cross-sectional studies in Iran and China. Despite heterogeneity and methodological weaknesses in many studies, these results lend confidence to the robustness of associations. Our findings are foundational for further inquiry: they should spur studies from more LMIC countries, and settings outside of Iran and China; they showcase the need for studies that advance our understanding of causal mechanisms and intervention research that can tell us how to activate these mechanisms through policy and programming in diverse settings. Understanding the mechanistic and contextual factors that change the relationship between diets, and more broadly food security and nutrition, with mental health would provide a lever for integrated or co-located interventions that reduce health risks among populations that experience an unequal share of concurrent burdens and marginalisations. If we are intent on addressing global health inequity, then these relationships will be key to improving wellbeing overall. Online Methods The selection of studies for this analysis consisted of 2 steps. We first relied on the inclusion and exclusion criteria of a large systematic Evidence and Gap Map (EGM), the detailed methods of which can be found in the published analysis of the EGM 45 . We then applied additional criteria to select studies from the EGM for the meta-analysis. Briefly, the EGM that was created and published first in 2022 45 including 1945 studies, then updated in 2024 to 3031 studies. We searched Medline, CAB Global Health and PsychInfo databases systematically (using the same search both times with a one-year overlap to account for indexing lags; Supplementary Methods 2) to create the EGM. We included peer-reviewed, English language studies linking food security and nutrition measures to common mental health problems (depression, anxiety, stress and mental wellbeing) in the general population, published from January 1, 2000 to January 31, 2024. The updated EGM is available here 46 . Drawing on the EGM repository, we selected a subset of studies for the meta-analysis. For healthy diets, we selected all studies fitting the eligibility criteria of the EGM, which were double-screened, with additional eligibility review by a senior researcher. All studies in the EGM were coded iteratively based on specific measures and indicators used, first by a single researcher, through iterative full-record and domain-specific checks by a senior researcher. Then, aligning with the broad definition of healthy diets proposed by Cena and Calder 12 , we used relevant measures from the ‘diets’ domain of the EGM that would capture ‘healthy’ diets, such as the Healthy Eating Index 70 (HEI) and its iterations 71 , dietary diversity 72 , the Dietary Inflammatory Index 73 , and dietary patterns identified through factor analysis or clustering of foods measured through intake questionnaires (e.g., FFQ) 74 . A full list of all diet measures is provided in Table 1. Among the MH domain in the EGM, we included all the studies that measured depression, anxiety or stress using any validated tool, such as mental health screening instruments like the Centre for Epidemiological Studies – Depression Scale (CES-D), the Depression, Anxiety and Stress Scale (DASS), the Generalised Anxiety and Depression Scale (GAD) or the State Trait Anxiety Index (STAI). In this second step, our exclusion criteria for the meta-analysis consisted of 1) Mental health different from depression, anxiety or stress, 2) Lack of comparability of diet measurements (not valid or comparable within the meta-analysis), 3) Duplicated data: same sample and same results of an already included study, and 4) Relevant estimate not reported. We excluded any study with a measure of healthy diet without an unhealthy diet comparator. The detailed list of inclusion and exclusion criteria for the EGM and the meta-analysis, including all the diet and MH measurements out of scope are detailed in the Supplementary Methods 3. We grouped eligible studies according to the dietary measurements used (see Table 1 in Results). Group 1 included adherence and adequacy includes adherence to diet recommendations and nutrient adequacy. Group 2 included dietary patterns shown to reduce nutrition-related chronic diseases, such as Dietary Approaches to Stop Hypertension (DASH), Dietary Inflammatory Index (DII), and Mediterranean diet. Group 3 was made up of diet diversity indices like Minimum Dietary Diversity of Women’s (MDD-W) and individual dietary diversity (IDDS) as well as any other Dietary Variety Scores. Group 4 Diet quality indices included all Diet Quality Indices, Global Diet Quality Score (GDQS), Global Dietary Index (GDI), and all Healthy Eating Indices (HEI). Groups (1-4) are all measured via predefined patterns and/or specific food intake ( a priori methods), thus the observed dietary pattern in the population is compared to a preexisting index. Group 5 included any measure derived from factor analysis, Principle Component Analysis (PCA) or other data-driven ( a posteriori ) approaches, which comprise comparisons between clustered patterns of food intake in a study population (usually between a ‘healthy pattern’ and an ‘unhealthy’ one, e.g., processed, western, modern, or unhealthy). For included studies, two researchers extracted data on effect sizes (since most of the descriptive characteristics of studies were already coded in the EGM), along with claim type (associational or causal), exposure scale (healthy or unhealthy), sample size, number of women, percentage of women in the sample, age group, mean age, age standard deviation, and reported statistical measures. Specifically, we extracted the following statistical effect sizes: Beta Coefficients (β), Odds Ratio (OR), Adjusted Odds Ratio (AOR), Risk Ratio (RR), Hazard Ratio (HR), Pearson’s Correlation Coefficient (r), Mean Differences. We made sure to capture the direction of the ‘healthy’ effect on mental health symptoms. Effect sizes were converted to standardized metrics to ensure comparability across studies. For odds ratios (OR), risk ratios (RR), and prevalence ratios (PR), the Chinn transformation 75 was applied to obtain Cohen’s d. β coefficients were standardized using the pooled standard deviation, while mean differences were converted by dividing the difference between group means by the pooled standard deviation. Hazard ratios were adjusted based on their distribution properties. To correct for small sample bias, Hedges’ g was computed 76 . Pearson’s r was derived from standardized mean differences, and Fisher’s z transformation was applied to normalize correlations. As the primary effect size index, we selected the Standardized Mean Difference (SMD) estimated by Hedges’ g due to its robustness in meta-analyses, allowing for comparisons of means and regression coefficients across diverse study designs 77 , and easier interpretability 76 . When studies used data from the same sample population, but reported a different sample size, we extracted the highest number to estimate sample populations. To estimate pooled effect sizes, we conducted a three-level meta-analysis using a robust variance estimation (RVE) framework. The three-level meta-analysis dealt with the dependency coming from studies drawing estimates from the same populations 78 , and we modelled our data to account for differences at the individual effect sizes, within sample population and between sample populations. The RVE allows adjustment for the standard errors and improves the statistical inferences when we face a data dependency issue 79 . The model was estimated using the restricted maximum-likelihood (REML) estimator 80 , incorporating nested random effects at the study and effect-size levels. Variance components were estimated using a structured variance-covariance matrix with a predefined correlation coefficient (ρ=0.5) to model dependence among estimates derived from the same dataset. Observations were weighted by their inverse variance, and heterogeneity was assessed through variance decomposition (I²) across levels. We considered both confidence intervals and prediction intervals, the latter of which is a more transparent indicator of heterogeneity at a population level. To evaluate the robustness of the findings, sensitivity analyses included alternative specifications of the correlation parameter, exclusion of influential studies, and comparisons across effect size metrics (Hedges’ g, Cohen’s d, Fisher’s z). Outlier and influential study detection were performed using Studentized residuals and Cook’s distances. Studies were flagged as potential outliers if their Studentized residual exceeded the 100 × (1 - 0.05/ (2 × k))th percentile of a standard normal distribution, applying a Bonferroni correction for multiple comparisons with a two-sided α=0.05 across k studies. Influential studies were identified using Cook’s distance, where values exceeding the median plus six times the interquartile range (IQR) were considered indicative of undue influence on model estimates. We tested alternative approaches to deal with the dependency of the data 81 . Model comparison was conducted by testing a reduced two-level model (removing the second random effect) against the full three-level specification to assess whether modelling within-study dependence significantly improved model fit 82 . The full three-level model yielded lower Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and corrected AIC (AICc) values, alongside a higher log-likelihood (logLik), indicating a superior fit 83 . The likelihood ratio test (LRT) comparing the two models returned a p-value of 0.0831, suggesting that while the improvement in fit was not statistically significant at the conventional 5% level, it was marginally close. Given the hierarchical nature of the data, where multiple effect sizes stem from the same study population, the three-level model was retained as the preferred approach to appropriately account for dependency and reduce bias in pooled estimates 82 . A senior researcher assessed the risk of bias (RoB) for each study using the “Quality In Prognosis Studies” (QUIPS) tool 84 , with any uncertainty checked by a second researcher. Although the QUIPS tool is designed for prognostic studies, prognoses are similar to risks in epidemiology. Furthermore, most of the included studies are cross-sectional, and there is no existing tool for these that covers all important domains of possible bias 85 . The QUIPS covers many of the important domains identified, including study participation, study attrition (omitted for cross-sectional studies), prognostic factor (exchanged for exposure) measurement, outcome measurement, study confounding, and statistical analysis and reporting. For each of these domains, studies were ranked as having low, moderate or high risk of bias. Then, we grouped all studies in 2 groups: high RoB (has at least 1 domain with high RoB) and low RoB (no domain has high RoB). We examined the publication trends over time; used the country-level income classification as defined by the World Bank at the time of this analysis (2024) 86 to examine effects by low-, lower-middle, and upper-middle income country status; and analysed the geographic spread across regions and countries. We included literature where healthy diets and mental health were associated, including when the hypothesised exposure was healthy diets or mental health, and vice versa for the outcome or dependent variable. We used the ‘hypothesis direction’ classification in the EGM to examine these different groups, as well as analysing potentially differential effects for different population groups (e.g. adults, elderly, adolescents, or parents paired with their children). We examined pooled estimates by dietary measure group (Table 1). We conducted several sensitivity analyses by different study characteristics: study design, dietary measurements and income level. All our sensitivity analyses were conducted first on the entire set of studies, then restricting to low RoB studies, and restricting to the high RoB studies. Declarations Funding: This work is funded through the Innovative Methods and Metrics for Agriculture and Nutrition Action (IMMANA) programme, led by the London School of Hygiene & Tropical Medicine (LSHTM), in partnership with Tufts University and the University of Sheffield. IMMANA is co-funded with UK International Development from the UK government and by the Gates Foundation INV-002962 / OPP1211308. The conclusions and opinions expressed in this work are those of the author(s) alone and shall not be attributed to the Foundation. Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission. Please note works submitted as a preprint have not undergone a peer review process Author contributions: T.M.S.: conceptualization, methodology, investigation, data curation, writing (original draft), writing (review and editing), project administration. C.C.: methodology, data curation, statistical- formal analysis and visualisation, software and programming, writing (original draft), writing (review and editing). B.C.: methodology, investigation, data curation. L.M.T.: writing (review and editing). S.K.: funding acquisition, supervision, writing (review and editing). Acknowledgements We thank the IMMANA team for their ideas, logistical and dissemination support. We thank Claudia Offner and Megan Deeney for their extensive work managing and contributing to the initial iteration of the Evidence and Gap Map, as well as Xuerui Han, Zhuozhi Lin and Chiara Lier for working to screen and code records from this iteration We thank Leisha Beardmore for their contribution to hiring and managing the team who updated the EGM, and to Corina Zhao, Tala Chehaitly, Yifei Li, Chantel Yenyu Ku, and Brena Bessa for their substantial efforts in screening and coding records for the EGM update. References Singer, M., Bulled, N., Ostrach, B. & Mendenhall, E. Syndemics and the biosocial conception of health. The Lancet 389 , 941–950 (2017). Institute for Health Metrics and Evaluation (IHME). Global Burden of Disease 2021: Findings from the GBD 2021 Study . https://www.healthdata.org/research-analysis/library/global-burden-disease-2021-findings-gbd-2021-study (2024). Murray, C. J. L. Findings from the Global Burden of Disease Study 2021. The Lancet 403 , 2259–2262 (2024). FAO, IFAD, UNICEF, WFP & WHO. The State of Food Security and Nutrition in the World 2024 . (FAO; IFAD; UNICEF; WFP; WHO;, 2024). doi:10.4060/cd1254en. World Health Organization. Government expenditures on mental health as a percentage of total government expenditures on health (%). WHO Global Health Observatory data repository. Ali, a-C., Ryan, G. & De Silva, M. J. Validated Screening Tools for Common Mental Disorders in Low and Middle Income Countries: A Systematic Review. PLOS ONE 11 , e0156939 (2016). American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders . (American Psychiatric Association, 2013). doi:10.1176/appi.books.9780890425596. Kirk, S. A., Cohen, D. & Gomory, T. DSM-5: The Delayed Demise of Descriptive Diagnosis. in The DSM-5 in Perspective (eds. Demazeux, S. & Singy, P.) vol. 10 63–81 (Springer Netherlands, Dordrecht, 2015). Lacasse, J. R. After DSM-5: A Critical Mental Health Research Agenda for the 21st Century. Res. Soc. Work Pract. 24 , 5–10 (2014). Tennyson, R. L., Kemp, C. G. & Rao, D. Challenges and strategies for implementing mental health measurement for research in low-resource settings. Int. Health 8 , 374–380 (2016). Watson, L. K., Kaiser, B. N., Giusto, A. M., Ayuku, D. & Puffer, E. S. Validating mental health assessment in Kenya using an innovative gold standard. Int. J. Psychol. J. Int. Psychol. 55 , 425–434 (2020). Cena, H. & Calder, P. C. Defining a Healthy Diet: Evidence for the Role of Contemporary Dietary Patterns in Health and Disease. Nutrients 12 , 334 (2020). Neufeld, L. M., Hendriks, S. & Hugas, M. Healthy Diet: A Definition for the United Nations Food Systems Summit 2021. in Science and Innovations for Food Systems Transformation (eds. Von Braun, J., Afsana, K., Fresco, L. O. & Hassan, M. H. A.) 21–30 (Springer International Publishing, Cham, 2023). doi:10.1007/978-3-031-15703-5_3. Sparling, T. M. et al. Global Mental Health and Nutrition: Moving Toward a Convergent Research Agenda. Front. Public Health 9 , 722290 (2021). Gianfredi, V. et al. Association between dietary patterns and depression: an umbrella review of meta-analyses of observational studies and intervention trials. Nutr. Rev. 81 , 346–359 (2023). Adjibade, M. et al. Prospective association between adherence to dietary recommendations and incident depressive symptoms in the French NutriNet-Santé cohort. Br. J. Nutr. 120 , 290–300 (2018). Bizzozero-Peroni, B. et al. Proinflammatory dietary pattern and depression risk in older adults: Prospective analyses from the Seniors-ENRICA studies. Clin. Nutr. 41 , 2614–2620 (2022). Cherian, L. et al. DASH and Mediterranean-Dash Intervention for Neurodegenerative Delay (MIND) Diets Are Associated With Fewer Depressive Symptoms Over Time. J. Gerontol. A. Biol. Sci. Med. Sci. 76 , 151–156 (2021). Collin, C. et al. Adherence to dietary guidelines as a protective factor against chronic or recurrent depressive symptoms in the French SU.VI.MAX cohort. Prev. Med. 91 , 335–343 (2016). Jacka, F. N. et al. A Prospective Study of Diet Quality and Mental Health in Adolescents. PLoS ONE 6 , e24805 (2011). Jiang, W. et al. The relationship of dietary diversity score with depression and anxiety among prenatal and post‐partum women. J. Obstet. Gynaecol. Res. 44 , 1929–1936 (2018). Recchia, D. et al. Associations between long-term adherence to healthy diet and recurrent depressive symptoms in Whitehall II Study. Eur. J. Nutr. 59 , 1031–1041 (2020). Wang, J. et al. Dietary inflammatory index and depression: a meta-analysis. Public Health Nutr. 22 , 654–660 (2019). Oftedal, S., Glozier, N., Holliday, E. G. & Duncan, M. J. Diet quality and depressive symptoms. Assessing the direction of the association in a population-based cohort study. J. Affect. Disord. 274 , 347–353 (2020). Arvidsson, L. et al. Bidirectional associations between psychosocial well-being and adherence to healthy dietary guidelines in European children: prospective findings from the IDEFICS study. BMC Public Health 17 , 926 (2017). Jacka, F. N., Cherbuin, N., Anstey, K. J. & Butterworth, P. Does reverse causality explain the relationship between diet and depression? J. Affect. Disord. 175 , 248–250 (2015). Marx, W., Moseley, G., Berk, M. & Jacka, F. Nutritional psychiatry: the present state of the evidence. Proc. Nutr. Soc. 76 , 427–436 (2017). Berk, M. et al. So depression is an inflammatory disease, but where does the inflammation come from? BMC Med. 11 , 200 (2013). Schwingshackl, L. & Hoffmann, G. Mediterranean dietary pattern, inflammation and endothelial function: A systematic review and meta-analysis of intervention trials. Nutr. Metab. Cardiovasc. Dis. 24 , 929–939 (2014). Estruch, R. Anti-inflammatory effects of the Mediterranean diet: the experience of the PREDIMED study. Proc. Nutr. Soc. 69 , 333–340 (2010). Guimarães, L. R. et al. Serum levels of brain-derived neurotrophic factor in schizophrenia on a hypocaloric diet. Prog. Neuropsychopharmacol. Biol. Psychiatry 32 , 1595–1598 (2008). Moylan, S. et al. Oxidative & nitrosative stress in depression: Why so much stress? Neurosci. Biobehav. Rev. 45 , 46–62 (2014). Zainuddin, M. S. A. & Thuret, S. Retracted: Nutrition, adult hippocampal neurogenesis and mental health. Br. Med. Bull. 103 , 89–114 (2012). Carroll, S. J., Dale, M. J., Niyonsenga, T., Taylor, A. W. & Daniel, M. Associations between area socioeconomic status, individual mental health, physical activity, diet and change in cardiometabolic risk amongst a cohort of Australian adults: A longitudinal path analysis. PLOS ONE 15 , e0233793 (2020). Snyder, H. R. Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: A meta-analysis and review. Psychol. Bull. 139 , 81–132 (2013). Arnsten, A. F. T. & Rubia, K. Neurobiological Circuits Regulating Attention, Cognitive Control, Motivation, and Emotion: Disruptions in Neurodevelopmental Psychiatric Disorders. J. Am. Acad. Child Adolesc. Psychiatry 51 , 356–367 (2012). Jacka, F. N., Cherbuin, N., Anstey, K. J., Sachdev, P. & Butterworth, P. Western diet is associated with a smaller hippocampus: a longitudinal investigation. BMC Med. 13 , 215 (2015). Christensen, L. & Brooks, A. Changing Food Preference as a Function of Mood. J. Psychol. 140 , 293–306 (2006). Samuthpongtorn, C. et al. Consumption of Ultraprocessed Food and Risk of Depression. JAMA Netw. Open 6 , e2334770 (2023). Miller, L. C. et al. Maternal depression is associated with less dietary diversity among rural Nepali children. Matern. Child. Nutr. 17 , e13221 (2021). Anato, A., Baye, K., Tafese, Z. & Stoecker, B. J. Maternal depression is associated with child undernutrition: A cross‐sectional study in Ethiopia. Matern. Child. Nutr. 16 , e12934 (2020). Chee Din, M. A., Mohd Fahmi Teng, N. I. & Abdul Manaf, Z. Maternal depression and child feeding practices: Determinants to malnutrition among young children in Malaysian rural area. Womens Health 19 , 17455057221147800 (2023). Fismen, A.-S. et al. Associations between eating habits and mental health among adolescents in five nordic countries: a cross-sectional survey. BMC Public Health 24 , 2640 (2024). Dabravolskaj, J. et al. Association Between Diet and Mental Health Outcomes in a Sample of 13,887 Adolescents in Canada. Prev. Chronic. Dis. 21 , 240187 (2024). Sparling, T. M. et al. Systematic evidence and gap map of research linking food security and nutrition to mental health. Nat. Commun. 13 , 4608 (2022). Sparling, T. M. & Cornejo, C. Systematic evidence and gap map of research linking food security and nutrition to mental health. ANH Academy https://www.anh-academy.org/mentalhealth-egm-population.htm (2025). Xu, Y. et al. Role of dietary factors in the prevention and treatment for depression: an umbrella review of meta-analyses of prospective studies. Transl. Psychiatry 11 , 478 (2021). Li, Y. et al. Dietary patterns and depression risk: A meta-analysis. Psychiatry Res. 253 , 373–382 (2017). Lassale, C. et al. Healthy dietary indices and risk of depressive outcomes: a systematic review and meta-analysis of observational studies. Mol. Psychiatry 24 , 965–986 (2019). Matison, A. P., Mather, K. A., Flood, V. M. & Reppermund, S. Associations between nutrition and the incidence of depression in middle-aged and older adults: A systematic review and meta-analysis of prospective observational population-based studies. Ageing Res. Rev. 70 , 101403 (2021). Askari, M. et al. Vegetarian diet and the risk of depression, anxiety, and stress symptoms: a systematic review and meta-analysis of observational studies. Crit. Rev. Food Sci. Nutr. 62 , 261–271 (2022). Herforth, A. The Sustainable Development Goals have a new diet indicator! Both the metric and the methods are key | ANH Academy. https://www.anh-academy.org/community/blogs/the-sustainable-development-goals-have-a-new-diet-indicator-both-the-metric-and-the. Heidari-Beni, M., Hemati, Z. & Qorbani, M. The Dietary Diversity Score: Methods, Indicators, and Applications to General Population. in Biomarkers in Nutrition (eds. Patel, V. B. & Preedy, V. R.) 293–306 (Springer International Publishing, Cham, 2022). doi:10.1007/978-3-031-07389-2_17. Cetrone, H. M. et al. Food security mediates the decrease in women’s depressive symptoms in a participatory nutrition-sensitive agroecology intervention in rural Tanzania. Public Health Nutr. 24 , 4682–4692 (2021). Casas, I. et al. Effects of a Mediterranean Diet Intervention on Maternal Stress, Well-Being, and Sleep Quality throughout Gestation—The IMPACT-BCN Trial. Nutrients 15 , 2362 (2023). Parletta, N. et al. A Mediterranean-style dietary intervention supplemented with fish oil improves diet quality and mental health in people with depression: A randomized controlled trial (HELFIMED). Nutr. Neurosci. 22 , 474–487 (2019). Jacka, F. N. et al. A randomised controlled trial of dietary improvement for adults with major depression (the ‘SMILES’ trial). BMC Med. 15 , 23 (2017). Watanabe, J. et al. Effects of ‘SPRAT’ programme for dietary and lifestyle education to improve psychosomatic symptoms and dietary habits among adolescents: a cluster randomised controlled trial. BMC Public Health 22 , 461 (2022). Haque, N. B., Mihrshahi, S. & Haider, R. Peer counselling as an approach to improve complementary feeding practices: a narrative review. J. Health Popul. Nutr. 42 , 60 (2023). Cetrone, H. et al. A Participatory Agroecological Intervention Reduces Women’s Risk of Probable Depression Through Improvements in Food Security in Singida, Tanzania. Curr. Dev. Nutr. 4 , nzaa053_024 (2020). Verdeli, H. et al. Alleviating maternal depression to improve IYCF and WASH behaviors. in (CORE Group, Online, 2021). Engdawork, K. et al. “Space to see the future”? A political economy analysis of child and adolescent mental health and well-being in Ethiopia including routes for change. Front. Sociol. 9 , 1488619 (2025). Yitbarek, K. et al. Barriers and Facilitators for Implementing Mental Health Services into the Ethiopian Health Extension Program: A Qualitative Study. Risk Manag. Healthc. Policy Volume 14 , 1199–1210 (2021). Gilligan, D. O., Hirvonen, K., Leight, J. & Tesfaye, H. Implementer-Led Evaluation and Learning (IMPEL) Evaluation of SPIR II RFSA - Midline Survey Report . https://hdl.handle.net/10568/144025 (2024). Abdullahi, A. T., Farouk, Z. L. & Imam, A. Common mental disorders in mothers of children attending out-patient malnutrition clinics in rural North-western Nigeria: a cross-sectional study. BMC Public Health 21 , 185 (2021). Le Roch, K., Tonon, B. & Acharya, O. What influence mothers’ mental health and health care seeking behaviors for their malnourished children in Nepal: building evidence for a broader perspective. Eur. Psychiatry 67 , S55–S55 (2024). World Health Organization & World Organization of Family Doctors. Integrating mental health into primary care : a global perspective. 206 (2008). Storhaug, I. G. et al. Making the most of existing research: an evidence gap map of the effects of food systems interventions in low-income and middle-income countries. BMJ Open 12 , e055062 (2022). High-Level Committee on Programmes (HLCP). Valuing What Counts - UN System-wide Contribution on Beyond GDP (advance unedited). (2022). Jansen, A., Wang, R., Behrens, P. & Hoekstra, R. Beyond GDP: a review and conceptual framework for measuring sustainable and inclusive wellbeing. Lancet Planet. Health 8 , e695–e705 (2024). Online Methods References Sparling, T. M. et al. Systematic evidence and gap map of research linking food security and nutrition to mental health. Nat Commun 13 , 4608 (2022). Sparling, T. M. & Cornejo, C. Systematic evidence and gap map of research linking food security and nutrition to mental health. ANH Academy https://www.anh-academy.org/mentalhealth-egm-population.htm (2025). Cena, H. & Calder, P. C. Defining a Healthy Diet: Evidence for the Role of Contemporary Dietary Patterns in Health and Disease. Nutrients 12 , 334 (2020). Guenther, P. M., Reedy, J. & Krebs-Smith, S. M. Development of the Healthy Eating Index-2005. Journal of the American Dietetic Association 108 , 1896–1901 (2008). Shams-White, M. M. et al. Healthy Eating Index-2020: Review and Update Process to Reflect the Dietary Guidelines for Americans, 2020-2025. Journal of the Academy of Nutrition and Dietetics 123 , 1280–1288 (2023). Hoddinott, J. & Yohannes, Y. Dietary Diversity as a Household Food Security Indicator. Food and Nutrition Technical Assistance Project (1825). Cavicchia, P. P. et al. A New Dietary Inflammatory Index Predicts Interval Changes in Serum High-Sensitivity C-Reactive Protein1–3. The Journal of Nutrition 139 , 2365–2372 (2009). Kant, A. K. Dietary patterns and health outcomes. Journal of the American Dietetic Association 104 , 615–635 (2004). Chinn, S. A simple method for converting an odds ratio to effect size for use in meta-analysis. Statist. Med. 19 , 3127–3131 (2000). Hedges, L. V. Distribution Theory for Glass’s Estimator of Effect size and Related Estimators. Journal of Educational Statistics 6 , 107–128 (1981). Borenstein, M., Hedges, L. V., Higgins, J. & Rothstein, H. R. Introduction to Meta-Analysis . (Wiley, Hoboken, NJ Chichester, 2021). Assink, M. & Wibbelink, C. J. M. Fitting three-level meta-analytic models in R: A step-by-step tutorial. TQMP 12 , 154–174 (2016). Hedges, L. V., Tipton, E. & Johnson, M. C. Robust variance estimation in meta‐regression with dependent effect size estimates. Research Synthesis Methods 1 , 39–65 (2010). Viechtbauer, W. Bias and Efficiency of Meta-Analytic Variance Estimators in the Random-Effects Model. Journal of Educational and Behavioral Statistics 30 , 261–293 (2005). Cinar, O., Umbanhowar, J., Hoeksema, J. D. & Viechtbauer, W. Using information‐theoretic approaches for model selection in meta‐analysis. Research Synthesis Methods 12 , 537–556 (2021). Harrer, M., Cuijpers, P., A, F. T. & Ebert, D. D. Doing Meta-Analysis With R: A Hands-On Guide . (Chapman & Hall/CRC Press, Boca Raton, FL and London, 2021). Neyman, J. & Pearson, E. S. On the Problem of the Most Efficient Tests of Statistical Hypotheses. in Breakthroughs in Statistics (eds. Kotz, S. & Johnson, N. L.) 73–108 (Springer New York, New York, NY, 1992). doi:10.1007/978-1-4612-0919-5_6. Hayden, J. A., Van Der Windt, D. A., Cartwright, J. L., Côté, P. & Bombardier, C. Assessing Bias in Studies of Prognostic Factors. Ann Intern Med 158 , 280–286 (2013). Kelly, S. E. et al. A scoping review shows that no single existing risk of bias assessment tool considers all sources of bias for cross-sectional studies. Journal of Clinical Epidemiology 172 , 111408 (2024). World Bank. World Bank Country and Lending Groups – World Bank Data Help Desk. World Bank https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups (2024). Tables Table 1 | Distribution of healthy diets measures and mental health diagnostic tools. Dietary measures were grouped into 5 groups. Mental health measures are presented per outcome, but each tool can measure more than one outcome. Groups Dietary measurements Number of studies 1 Group 1: Adherence and adequacy (G1) 2 - Adherence to diet recommendations - Nutrient adequacy 2 Group 2: Dietary patterns reducing nutrition-related chronic diseases (G2) 2 - Dietary Approaches to Stop Hypertension (DASH) - Dietary Inflammatory Index - Mediterranean diet 22 Groups 3: Diet diversity indices (G3) 2 - Dietary diversity (MDD, IDDS) - Any other Dietary Variety Scores 15 Group 4: Diet quality indices (G4) 2 - Diet Quality Indexes (all) - Global Diet Quality Score - Global Dietary Index - Healthy Eating Indexes (all) 19 Group 5: Factor analysis and others (G5) 3 All comparisons between a healthy and an unhealthy diet. Unhealthy is a group with a processed, western, modern, traditional, or unhealthy diet determined for each study. 26 Mental Health Outcome (number of studies 4 ) Validated diagnostic tool Number of studies 4 Depression (69) Beck Depression Inventory (BDI) 7 Center for Epidemiological Studies - Depression scale (CES-D) 7 Clinical/diagnostic interview (CIDI - SF) 2 Depression, Anxiety and Stress Scale (DASS) 14 Edinburgh Postpartum Depression Scale (EPDS) 7 Geriatric Depression Scale (GDS) 2 Depression subscale of the Hospital Anxiety and Depression Scale (HADS-D) 1 6-item Kutcher Adolescent Depression Scale (KADS-6) 1 Mini International Neuropsychiatric Interview (MINI) 2 Multidimensional Sub-health Questionnaire of Adolescents (MSQA) 1 Patient Health Questionnaire (PHQ-9) 11 Primary Care Evaluation of Mental Disorders (PRIME-MD) 1 Self-Rating Depression Scale (SDS) 1 Zung self-rating scale 2 Anxiety (43) Coronavirus Anxiety Scale (CAS) 1 Depression, Anxiety and Stress Scale (DASS) 15 General Anxiety Disorder Scale (GAD) 10 Hospital Anxiety and Depression Scale (HADS-A) 11 Mini International Neuropsychiatric Interview (MINI) 1 Multidimensional Sub-health Questionnaire of Adolescents (MSQA) 1 Primary Care Evaluation of Mental Disorders (PRIME-MD) 1 Zung Self-reported Anxiety Scale (Zung SAS) 1 State-Trait Anxiety Inventory (STAI) 1 Stress (26) Depression, Anxiety and Stress Scale (DASS) 14 General Health Questionnaire (GHQ) 1 Hospital Anxiety and Depression Scale (HADS) 8 Perceived Stress Scale (PSS) 3 1 One study measured diets with both a Group 2 and a Group 4 measurement, so it is double-counted. 2 A priori measure. 3 A posterior measure. 4 Several studies reported 2 or 3 mental health outcomes, so the count exceeds the total number of studies (n = 83). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementarymaterialsHDMHMetaanalysis2025FINAL.docx Supplementary material_Meta-analysis on healthy diets and mental health in LMICs Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-6530671\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":false,\"archivedVersions\":[],\"articleType\":\"Article\",\"associatedPublications\":[],\"authors\":[{\"id\":449220621,\"identity\":\"2a9ebb3c-e99f-4a2e-90cf-e68096d50871\",\"order_by\":0,\"name\":\"Thalia Sparling\",\"email\":\"data:image/png;base64,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\",\"orcid\":\"https://orcid.org/0000-0002-8071-3232\",\"institution\":\"London School of Hygiene and Tropical Medicine\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Thalia\",\"middleName\":\"\",\"lastName\":\"Sparling\",\"suffix\":\"\"},{\"id\":449220622,\"identity\":\"ef94e83c-2022-4b19-b336-84ce7ae541c1\",\"order_by\":1,\"name\":\"Cesar Cornejo\",\"email\":\"\",\"orcid\":\"https://orcid.org/0009-0006-1042-8838\",\"institution\":\"London School of Hygiene and Tropical Medicine\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Cesar\",\"middleName\":\"\",\"lastName\":\"Cornejo\",\"suffix\":\"\"},{\"id\":449220623,\"identity\":\"d0d165ef-cd6e-42aa-afc9-98d6324f4894\",\"order_by\":2,\"name\":\"Bryan Cheng\",\"email\":\"\",\"orcid\":\"https://orcid.org/0000-0002-4052-7515\",\"institution\":\"Teachers College, Columbia University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Bryan\",\"middleName\":\"\",\"lastName\":\"Cheng\",\"suffix\":\"\"},{\"id\":449220624,\"identity\":\"5041fdf6-32e6-4f1e-ad1c-1e17f72ccc83\",\"order_by\":3,\"name\":\"Lisa Troy\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"School of Public Health and Health Sciences, University of Massachusetts\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Lisa\",\"middleName\":\"\",\"lastName\":\"Troy\",\"suffix\":\"\"},{\"id\":449220625,\"identity\":\"18484621-ed99-4dc4-8ff9-a2826ba40067\",\"order_by\":4,\"name\":\"Suneetha Kadiyala\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"London School of Hygiene and Tropical Medicine\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Suneetha\",\"middleName\":\"\",\"lastName\":\"Kadiyala\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2025-04-25 16:50:10\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-6530671/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-6530671/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":81672904,\"identity\":\"38d545c0-dfc6-4ad7-87b3-3eadacb8895e\",\"added_by\":\"auto\",\"created_at\":\"2025-04-30 06:36:55\",\"extension\":\"png\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":280573,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003ePRISMA diagram. \\u003c/strong\\u003e\\u003cem\\u003eAdapted Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart representing the 2-step study selection process. Step 1 reports Evidence Gap Map studies at the identification, screening and final inclusion stages. Step 2 reports the meta-analysis studies at each stage, which started by identifying eligible studies \\u003c/em\\u003e\\u003cem\\u003e\\u003cstrong\\u003eincluded in\\u003c/strong\\u003e\\u003c/em\\u003e\\u003cem\\u003e the Evidence Gap Map.\\u003c/em\\u003e\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage2.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/ab462d1b7476364772a84b7b.png\"},{\"id\":81672905,\"identity\":\"2f7928df-b519-48b1-9d6d-6fc6323d61db\",\"added_by\":\"auto\",\"created_at\":\"2025-04-30 06:36:55\",\"extension\":\"png\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":819312,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eGeographic, temporal and income-level distribution \\u003c/strong\\u003eof association between healthy diets and mental health symptoms. For comparison, the graph includes high-income country studies included in the EGM, but not included in the meta-analysis.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage3.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/0dab7cb53d9448f187772bf8.png\"},{\"id\":81673736,\"identity\":\"97b173bb-1d91-41a2-bfa0-c8f9601ac728\",\"added_by\":\"auto\",\"created_at\":\"2025-04-30 06:52:55\",\"extension\":\"png\",\"order_by\":3,\"title\":\"Figure 3\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":89680,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003ePublication trend by income classification. \\u003c/strong\\u003eDistribution of studies of association between healthy diets and mental health symptoms published by year between 2000 and 2024. n indicates the total number of studies, not including 2024. For comparison, the graph includes the high-income country studies included in the EGM for comparison, which are not included in the meta-analysis.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage4.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/702a28b868123116f0066b8b.png\"},{\"id\":81672909,\"identity\":\"2a7ce830-d6fa-4095-8262-5db874a00fd9\",\"added_by\":\"auto\",\"created_at\":\"2025-04-30 06:36:56\",\"extension\":\"jpeg\",\"order_by\":4,\"title\":\"Figure 4\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":1804056,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eSummary effect and risk of bias (RoB). \\u003c/strong\\u003eForest plot and risk of bias per domain of all studies by mental health by mental health outcome: n=number of observations. D1 = Bias due to study participation, D2 = Bias due to exposure measurement, D3 = Bias due to outcome measurement, D4= Bias due to confounding, D5 = Bias in statistical analysis and reporting, Overall = No high risk of bias in any domain. D1 to D5 RoB: Green = low RoB; Yellow = unclear RoB; Red = high RoB. Overall RoB: Orange = High RoB; Light Green = Low RoB.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage5.jpeg\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/8a380b0c1c0ea980096c841c.jpeg\"},{\"id\":81672908,\"identity\":\"a99a1ba8-f881-438a-be95-9dbf0cfbd8e4\",\"added_by\":\"auto\",\"created_at\":\"2025-04-30 06:36:55\",\"extension\":\"png\",\"order_by\":5,\"title\":\"Figure 5\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":1140527,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eSummary effect of low risk of bias studies.\\u003c/strong\\u003eForest plot and risk of bias (RoB) per domain of all studies with low risk of bias by mental health outcome. D1 = Bias due to study participation, D2 = Bias due to exposure measurement, D3 = Bias due to outcome measurement, D4= Bias due to confounding, D5 = Bias in statistical analysis and reporting, Overall = No high risk of bias in any domain. CI: Confidence Interval; SMD: Standardised mean difference. Green = low RoB; Yellow = unclear RoB; Light Green = Low Overall RoB.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"floatimage6.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/4e2da2999b7c6892faa46f60.png\"},{\"id\":81842255,\"identity\":\"84824600-b86c-4656-8076-bc97896743df\",\"added_by\":\"auto\",\"created_at\":\"2025-05-02 16:28:27\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":4436398,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/2e233e09-65dd-4f0d-8b5d-9786123053c5.pdf\"},{\"id\":81672917,\"identity\":\"9fdd1a5b-e4c0-46ea-87ce-22275ab19edb\",\"added_by\":\"auto\",\"created_at\":\"2025-04-30 06:36:56\",\"extension\":\"docx\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":2791580,\"visible\":true,\"origin\":\"\",\"legend\":\"Supplementary material_Meta-analysis on healthy diets and mental health in LMICs\",\"description\":\"\",\"filename\":\"SupplementarymaterialsHDMHMetaanalysis2025FINAL.docx\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-6530671/v1/b0e00c1ac8888fb57308568f.docx\"}],\"financialInterests\":\"There is \\u003cb\\u003eNO\\u003c/b\\u003e Competing Interest.\",\"formattedTitle\":\"Meta-analysis of 633,317 individuals shows associations between healthy diets and mental health in 23 low- and middle-income countries\",\"fulltext\":[{\"header\":\"Introduction\",\"content\":\"\\u003cp\\u003eWe are living in an era of global health syndemics, where multiple diseases interact synergistically within populations under conditions of social and structural inequity\\u003csup\\u003e1\\u003c/sup\\u003e. \\u0026nbsp;The most vulnerable people bear concomitant health burdens, compounding their health risk overall. Common mental health disorders (CMDs) such as depression, anxiety and stress, contributed 17.2% of life years lost to disability in 2021\\u003csup\\u003e2\\u003c/sup\\u003e, with higher prevalence of CMDs reported in LMICs\\u003csup\\u003e3\\u003c/sup\\u003e. CMDs are drivers and consequences of other marginalisations, such as food insecurity and unhealthy diet. In turn, 35% of the global population could not afford a healthy diet in 2022. \\u0026nbsp;Of those, 1.68 billion (nearly 60%) live in LMIC\\u003csup\\u003e4\\u003c/sup\\u003e, where on average only 2.1% of government health expenditure goes to mental health\\u003csup\\u003e5\\u003c/sup\\u003e and public spending on agriculture is both extremely low (less than $10 per rural inhabitant in LIC) and stagnant\\u003csup\\u003e4\\u003c/sup\\u003e especially problematic since the burdens of food insecurity and malnutrition are greatest in these settings.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eIn LMICs, CMDs are almost always measured through non-specialist administered screening tools\\u003csup\\u003e6\\u003c/sup\\u003e. The \\u0026lsquo;gold standard\\u0026rsquo; of CMD measurement has classically been clinical interviews to determine alignment with diagnostic criteria, for instance the Diagnostic and Statistical Manual of Mental Disorders (DSM)\\u003csup\\u003e7\\u003c/sup\\u003e. However, in global mental health research, there is a shift away from diagnoses towards implications of poor mental health, such as functional loss or wider impacts on other health indices\\u003csup\\u003e8,9\\u003c/sup\\u003e. One reason is that clinical resources and expertise rare and rarely accessible in low-resource settings. Furthermore, for diverse populations with greater health vulnerability, strategies are emerging to ensure cultural and discriminant validity, address cross-cultural equivalence, and prioritise measurement of functional impairment, which can be measured through symptoms of CMDs\\u003csup\\u003e10,11\\u003c/sup\\u003e. Thus, instruments designed as screening tools (and validated as threshold scores\\u0026rsquo; correspondence to probability of clinical diagnosis) have in fact become stand-alone assessments of both the range and extent of common mental health symptoms in a population. Thus, in LMICs, mental health measures are usually framed through \\u0026lsquo;symptoms\\u0026rsquo; of common mental health problems. \\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eA healthy diet consists of balanced proportions (enough to meet caloric adequacy but not exceed it) of primarily whole grains, vegetables, fruit, nuts, seeds, and where available fish, while moderating meat and dairy, and limiting consumption of sugars, saturated fats, and ultra-processed foods (UPFs)\\u003csup\\u003e12\\u003c/sup\\u003e. \\u0026nbsp;In simpler terms, a healthy diet is health-promoting and disease-preventing\\u003csup\\u003e13\\u003c/sup\\u003e. Healthy diets are measured through many different instruments, including adherence to (national) dietary recommendations, adherence to certain diets (e.g. the mediterranean diet (MD) or the dietary inflammatory index (DII), scoring or counts of food items grouped to indicate dietary diversity and/or certain index foods that are considered beneficial to health (e.g. dietary diversity of women, children or households). \\u0026nbsp; These indices and measures are validated similarly in that they correspond to increased probability of nutrient adequacy (at least in aggregate at the population level).\\u003c/p\\u003e\\n\\u003cp\\u003eWe can only hypothesise about the proportion of the world\\u0026rsquo;s population who simultaneously experience poor mental health and poor diet quality. However, there is research to show that these burdens are interlinked, and can exacerbate one another\\u003csup\\u003e14\\u003c/sup\\u003e. Much of this evidence is descriptive rather than causal or mechanistic\\u003csup\\u003e14,15\\u003c/sup\\u003e. Some robust studies show that healthy diets improve mental health\\u003csup\\u003e16\\u0026ndash;23\\u003c/sup\\u003e, while longitudinal studies examining bidirectional relationships show that better mental health can also contribute to eating healthier diets\\u003csup\\u003e24\\u003c/sup\\u003e, including in children\\u003csup\\u003e25,26\\u003c/sup\\u003e. Healthy diets resulting in better mental health is plausible from both physiological and social arguments. For instance, healthy diets are usually considered more diverse, containing higher quantities of foods that contain antioxidants (e.g., flavones, polyphenols, anthocyanins), long-chain fatty acids (e.g., Eicosapentaenoic acid and docosahexaenoic acid), and other essential nutrients that lower risk of nutrition-related chronic disease (e.g., diabetes, cancer)\\u003csup\\u003e12\\u003c/sup\\u003e. These dietary components reduce inflammation\\u003csup\\u003e27\\u0026ndash;30\\u003c/sup\\u003e. \\u0026nbsp;Diet can thus improve neurotransmission and neurocognitive function through enhanced brain plasticity via brain-derived neurotrophic factor (BDNF)\\u003csup\\u003e31\\u003c/sup\\u003e, which collectively can improve mental health. Potential biological pathways related to mental disorders include inflammation, oxidative stress, the gut microbiome, brain plasticity, and mitochondrial dysfunction, with evidence showing that healthy dietary patterns can beneficially modulate each of these interconnected systems\\u003csup\\u003e28,32,33\\u003c/sup\\u003e. Higher socioeconomic status and better lifestyle factors may also underpin both healthier eating and better mental health\\u003csup\\u003e34\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eIn the reverse direction, poor mental health can impact diet quality through multiple pathways. Depression and anxiety are associated with diminished motivation, cognitive impairments, and reduced executive function\\u003csup\\u003e35,36\\u003c/sup\\u003e, all if which can hinder meal planning, food shopping and preparation\\u003csup\\u003e37\\u003c/sup\\u003e. People experiencing poor mental health may also have altered appetite regulation and food preferences, leading to consumption of energy dense and UPFs\\u003csup\\u003e38\\u003c/sup\\u003e, although evidence now more strongly points to UPF consumption as a risk factor for depression\\u003csup\\u003e39\\u003c/sup\\u003e. The impact of mental health on diet can extend beyond the individual to the household level, as studies have shown that children of parents, particularly mothers, with CMDs are more likely to experience worse diet quality and child feeding outcomes\\u003csup\\u003e40\\u0026ndash;42\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eA clear association between healthy diets and mental health has been shown with nationally representative\\u003csup\\u003e43\\u003c/sup\\u003e and longitudinal\\u003csup\\u003e44\\u003c/sup\\u003e data in numerous high-income countries. Less is known about the interplay between diet and mental health in LMICs, a critical gap given that they host higher burdens of both poor diet quality and symptoms of CMDs. Here, we quantify the associations reported from studies in LMICs that compare healthy diets (i.e., measured through adequacy, adherence to dietary guidelines, or dietary patterns), with depression, anxiety and stress measures, assess their robustness, and identify the evidence gaps. We report according to the MOOSE guidelines for reporting on meta-analyses, the checklist for which is attached in Supplementary Methods 1.\\u003c/p\\u003e\"},{\"header\":\"Results\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eStudy selection\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe initial repository of studies came from an Evidence and Gap Map (EGM), which was systematically created from searching three databases (Medline, CAB Global Health and PsychInfo) for English language studies linking food security, and nutrition measures to measures of depression, anxiety, and stress in the general population, published from January 1, 2000, to January 31, 2024. \\u0026nbsp;We screened 30,896 records published from 2000 through June 2020 for the original EGM, resulting in a map and analysis of 1945 studies45. To update the EGM, we screened 13,490 additional records published from 2020 to January 2024 and included 1,107 additional records46. The updated EGM includes over 3,000 studies on this topic. From this, we systematically included a subset of studies for the meta-analysis. We selected studies on \\u0026lsquo;healthy diet\\u0026rsquo; using standard indices or factor/cluster analysis methods, and validated measures of depression, anxiety, and stress. A total of 96 studies met inclusion criteria from the EGM, from which we excluded 13 studies: 7 had a mental health outcome different from depression, anxiety or stress (e.g., common mental disorders and mental wellbeing) 1 reported the same data and results in another included study, 4 lacked comparability (did not define healthy vs unhealthy diet patterns), 1 did not have relevant data reported (only percentiles), from which it is not possible to calculate comparable statistics.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eSample populations \\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eEighty-three studies were included (depression=69; anxiety=43; stress=26), from which we extracted 139 effect estimates (depression=70; anxiety=43; stress=26) from 65 unique sample populations (depression=58; anxiety=31; stress=20). These estimates were pooled from 633,317 unique individuals (depression=473,236; anxiety=146,217; stress=24,690). The most studied populations were adults (n=42), followed distantly by females (n=16), mid- to later-life populations (n=12), pregnant women or mothers (n=10), and adolescents (n=7). The list of included studies is presented in Supplementary Results 1.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eSample settings\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe 83 studies covered 23 countries: 4 low-income, 11 lower-middle income and 69 upper-middle income (Fig. 2), where 1 study was across six countries, 3 of which were low-income and 3 were low-middle income. Most studies were based in Asia (n=70), of which 39 were from Iran and 17 were from China (both upper-middle income countries). The rest of Asian countries studied were Turkey, Jordan, Nepal, Bangladesh, and India. Outside of Asia, 8 studies were from South America (all from Brazil) and 6 from Africa (3 of which were from Ethiopia). The low-income country studies were from Burkina Faso, Ethiopia, Lebanon, Syria and Uganda. Although we only include results from LMICs in this meta-analysis, we identified 247 additional studies from high-income countries from the EGM, which are included for comparative purposes in Fig. 2. There were 171 countries world-wide where there was not a single study on this topic.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eStudy characteristics\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThere were no published studies on this topic between 2000 and 2012 and then there was a steady annual increase from 1 in 2013 to 15 in 2023. Figure 3 shows the distribution of studies per country-level income (multi-country studies were counted once per each country studied). The primary study design was cross-sectional (n=71). The rest were longitudinal (n=9), and case-control (n=3). We classified the proposed direction of association in studies, regardless of design (i.e., diet as an exposure for mental health outcomes and vice versa). Most (n=73) framed their research as studying the effect of diets on mental health symptoms, and 10 declared studying the effect of mental health symptoms on diets.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eAll but 7 estimates were based on mental health assessments using validated instruments designed as screening tools. They generally include range of symptoms, and a ranking of how often or to what degree those symptoms are experienced. Dietary measures were based either on a priori methods, i.e., a prescribed conception of a healthy diet pattern, or a posteriori, i.e., data-driven methods. \\u0026nbsp;Measures were grouped into into 5 groups, described in Table 1. One study used dietary measures from two groups. \\u0026nbsp;Groups 1-4 used a priori measures, while Group 5 used a posteriori methods. The complete distribution of dietary measures and mental health diagnostics tools is also in Table 1.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eSummary of effects\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eFrom 139 effect estimates, we pooled the effect for the association between healthy diets and depression, anxiety and stress (Fig. 4). We found that individuals with healthy diets showed less depressive symptoms compared to those with less healthy diets (SMD = -0.29, 95% CI -0.35 to -0.23), with a mean difference of 0.29 standard deviations in depression between groups. Similarly, healthy diets were associated with less anxiety (SMD = -0.25, 95% CI -0.35 to -0.16) and less stress (SMD = -0.24, 95% CI -0.33 to -0.14). The effect sizes from individual studies ranged from -1.50 to 0.09 for depression, -0.89 to 0.01 for anxiety, and -0.89 to 0.00 for stress. The prediction intervals (versus the confidence intervals) were more conservative in their estimates of the true population-level effect (Supplementary Results 2).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eEvidence strength\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eWe adapted the Quality in Prognostic Studies (QuiPS) tool, which is comprised of six domains of potential bias in observational studies. The QuiPS tool is included in Supplementary Methods 4. We report on 5 bias domains evaluated: study participation (i.e., selection bias), exposure measurement, outcome measurement, confounding, and statistical analysis and reporting. Since almost all studies were cross-sectional, we omitted \\u0026lsquo;study attrition\\u0026rsquo; for these, and the rankings on this domain for longitudinal studies is included in Supplementary Results 3. \\u0026nbsp;Studies that had no \\u0026lsquo;high\\u0026rsquo; risk of bias in any domain were grouped into an overall low risk of bias group, and those with at least one high risk of bias rating in any domain were grouped into overall \\u0026lsquo;high\\u0026rsquo;. There was strong evidence of slightly weaker pooled effect estimates from studies with overall low risk of bias in each group (depression=35; anxiety=18; stress=8): depression (SMD = -0.23, CI: -0.31 to -0.16; n=266,831), anxiety (SMD = -0.19, CI: -0.30 to -0.09; n=116,248), and stress (SMD = -0.22, CI: -0.33 to -0.11; n=12,338).\\u003c/p\\u003e\\n\\u003cp\\u003eGiven that most studies were cross-sectional, the most likely sources of bias were study participation (domain 1) and confounding (domain 4). The ranking of potential bias from study participation was based on assessing the source of target population, method used to identify population, recruitment period, place of recruitment, inclusion and exclusion criteria, adequate study participation, and baseline characteristics. As all but 3 studies treated diet as the exposure, the risk of study participation bias stemmed from the likelihood that the relationship between diet quality and mental health is different for participants in healthy eating and non-healthy eating groups. When restricting the analysis to studies with low risk of study participation bias, the estimates remained robust: depression (SMD = -0.18, CI: -0.26 to -0.10), anxiety (SMD = -0.14, CI: -0.26 to -0.02), and stress (SMD = -0.11, CI: -0.23 to -0.01). The risk of bias from confounding was assessed based on whether important confounders (factors related both to healthy eating and mental health) were measured, defined, valid and reliably measured, had uniform method and setting across all participants, whether they followed a valid approach to deal with missing data, and if important potential confounders were accounted for in the analysis. When restricting the analysis to studies with low risk of confounding bias, we found similar results, except for stress, as there were only 2 studies in this subgroup, and with weaker evidence of these effects given the smaller sample sizes (SMD = -0.26, CI: -0.55 to 0.02), anxiety (SMD = -0.06, CI: -1.05 to 0.92), and stress (SMD = -0.25, CI: -0.73 to 0.22). Pooled estimates by overall risk of bias, as well as study participation and confounding risk of bias (all low vs. high) are in Supplementary Results 4.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eOutliers and influential studies\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eWe identified 2 outlier (studies where the size or direction of the effects deviate substantially from the majority) and 6 influential studies (studies that significantly impact the overall results) based on studentized residuals. However, when restricting the analysis to studies with low risk of bias, no outliers remained for any mental health group, and only 1 study remained an influential study for depression. All outliers and influential studies are further discussed in Supplementary results 5.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eSub-group and sensitivity analyses\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eFindings were consistent in direction and magnitude across country income levels, study design, and dietary measurements. By country-level income classification, low-income countries had only 3 estimates for depression, 1 for anxiety and 0 for stress, while lower-middle income countries had only 4 estimates for anxiety and 2 for stress. For any country-level income group with more than 4 studies, the results were as follows: depression in lower middle-income countries = SMD -0.39 (CI: -0.71 to -0.07; n=10); depression in upper middle-income countries = SMD -0.28 (CI: -0.34 to -0.22; n=57); anxiety in upper-middle income countries = SMD -0.20 (CI: -0.27 to -0.13); n = 38), stress in upper-middle income countries = SMD -0.26 (CI: -0.36 to -0.16; n=24 studies). When we restricted analyses by study design to the nine longitudinal studies, results were almost identical. By diet measure, most studies used a priori tools to measure diets associated with a reduction in nutrition-related chronic diseases, which alone produced stronger pooled effect sizes: depression=SMD -0.36 (CI: -0.51 to -0.21; n=18); anxiety=SMD -0.25 (CI: -0.41 to -0.09; n=14); and stress=SMD -0.32 (CI: -0.56 to -0.09; n=12) (Supplementary Results 6). Results were not sensitive to the choice of different effect size indices, and we report Cohen\\u0026rsquo;s d and Fisher\\u0026rsquo;s z transformation estimates in Supplementary Results 7.\\u003c/p\\u003e\"},{\"header\":\"Discussion\",\"content\":\"\\u003cp\\u003eWe found consistent associations between healthy diets and better mental health when pooling eligible studies from LMICs. The magnitude of pooled effects was similar for depression, anxiety and stress (-0.29, -0.25 and -0.24, respectively), and we are confident that these results are not due to chance and are reasonably precise. This analysis presents the most robust estimates of these relationships in LMICs to date, supported by our rigorous methodology.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eOur results align with other similar meta-analyses. An umbrella review of meta-analyses included very similar measures of \\u0026lsquo;healthy diets\\u0026rsquo; but looked only at depression, in any setting or population, but did not pool effect estimates from the included meta-analyses\\u003csup\\u003e15\\u003c/sup\\u003e. \\u0026nbsp;Overall, they (Gianfredi et al.) found the methodological quality low to critically low, but did conclude that there was suggestive evidence linking healthy diets defined \\u003cem\\u003ea posteriori\\u003c/em\\u003e with depressive symptoms or diagnosis, which was supported by a meta-analysis focusing only on \\u003cem\\u003ea posteriori\\u003c/em\\u003e diet measures, and results from pooling 8 prospective studies on dietary patterns and incident depression\\u003csup\\u003e47,48\\u003c/sup\\u003e). Gianfredi et al. found stronger evidence and effects for the links between higher adherence to the Mediterranean diet and lower scores on the DII and lower risk of depression. \\u0026nbsp;These conclusions are similar to quantitative estimates from Lassale et al. (2019), who found that the Mediterranean diet conferred a relative risk (RR) of depression of 0.67 from four longitudinal studies (95% CI 0.55 to 0.82), and a lower DII similarly protective (RR 0.76; 95% CI: 0.63 to 0.92) from four longitudinal studies\\u003csup\\u003e49\\u003c/sup\\u003e. In older populations, higher DII score was also associated with incidence of depression (OR 1.33; 95% CI 1.04 to 1.70) from prospective studies, although they did not find associations with the Mediterranean diet or \\u0026lsquo;healthy diet\\u0026rsquo;\\u003csup\\u003e50\\u003c/sup\\u003e. The Lassale et al. meta-analysis of 8 prospective cohorts and 9 cross-sectional studies on DII alone estimated that diets with higher inflammatory potential increased odds of depression by 45% (95% CI 1.30\\u0026thinsp;to 1.62) and anxiety by 66% (95% CI 1.41\\u0026thinsp;to 1.96). They\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003ealso found evidence for the protective effects of a higher HEI/AHEI score (RR 0.65; 95% CI 0.50 to 0.84) from mixed longitudinal and cross-sectional studies\\u003csup\\u003e49\\u003c/sup\\u003e. The Gianfredi umbrella review found no evidence for vegetarian diets, also supported by another meta-analysis on this topic\\u003csup\\u003e51\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eThese synthesis studies share certain similarities with our work: they generally find that healthier diets are associated with better mental health, and even the direction and magnitude of effects are similar. Almost all note methodological limitations and heterogeneity, as we do. \\u0026nbsp; These studies differ in important ways from our analysis: almost all focus only on depression whereas we include anxiety and stress as well. Some mix prevention and treatment of mental health problems; we excluded treatment research since selecting participants into a study based on poor health status fundamentally confounds the relationship we were interested in testing. Some also found differences based on diet measurement, whereas our results were similar for all included measures, even if some of the evidence of effect was weak. None of the previous analyses focus on LMIC settings.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eWe find that most of the LMIC evidence comes from two middle-income countries: Iran and China (together 68% of included studies), even though our study includes findings from 23 LMICs. \\u0026nbsp;There were several studies from other places such as Brazil, Bangladesh, Ethiopia or Turkey, but these were not numerous or methodologically strong enough to make conclusions about these different contexts. Thus, despite pooling many studies, there are still important evidence gaps to fill from low-income settings, and in LMICs other than Iran and China. Nonetheless, we show robust relationships between healthy diets and mental health symptoms.\\u003c/p\\u003e\\n\\u003cp\\u003eWe gain both confidence and novel insights from sub-analyses, which showed similar direction and magnitude of effects despite some loss of power from smaller sub-sample sizes. For instance, when including only low risk of bias studies, the effect estimates weakened only somewhat (for depression -0.29 SMD to -0.23; for anxiety -0.25 to -0.19, and almost no change for stress). Changes could be explained by stronger study designs (which were rated lower risk of bias) that account better for the many factors at play influencing these relationships, such as controlling for a history of mental health issues or using more precise or standardised measures of diets (e.g., HEI versus a factor-derived, sample-dependent dietary pattern).\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eWe found consistent effect estimates across dietary measurements, mental health screening tools and study designs. For instance, regarding dietary measurements, measuring quantitative dietary intake and calculating dietary adequacy (group 1 and group 4) could capture a different aspect of a \\u0026lsquo;healthy diet\\u0026rsquo;, than a qualitative count of different food groups (group 2 and group 3) or a factor analysis of dietary data in a population (group 5). Reassuringly, we find similar effect estimates across dietary measurements, even when similar studies did not. Additionally, we found that certain measures were more common in LMICs than others. For instance, dietary diversity, a validated measure of diet quality of women and children, is now a standard indicator in LMICs, but is less commonly used in HIC\\u003csup\\u003e52,53\\u003c/sup\\u003e. In contrast, very few studies used dietary adequacy or adherence to national guidelines because national diet guidelines are both less defined and less measured in LMICs.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eAlthough we are confident that our results are robust and precise, we also note the prediction intervals of our estimates, which are a better measure of population variance versus the sample variance. The prediction intervals from our results indicate that the true average effect of consuming a healthy diet, based on heterogeneity of the studies, could be protective or have no effect on mental health. The interpretation of the prediction intervals aligns with the heterogeneity of the studies included and may also reflect small but possibly real differences in sample populations and settings. \\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eStrengths\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eOur study benefits from a rigorous methodology, including a broad, thorough search of the topic to create the EGM, and state of the art screening and coding processes to identify studies. \\u0026nbsp;We also cast a wide net for relevant studies, including multiple measures of common mental health issues, and a broad definition of \\u0026lsquo;healthy diets\\u0026rsquo;, which allowed us to compare associations across several different facets of these topics. \\u0026nbsp;We were able to include many studies, and carefully considered data dependencies, arriving at a three-level model accounting for overlapping study populations. This means that we neutralised the bias coming from multiple reporting of estimates produced from the same populations. We explored consistency and robustness of the evidence, finding clear associations.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eLimitations\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eWe searched the literature through three databases and until January 2024, so studies indexed in non-English repositories, not in English or published after this date are not included. However, we do not believe that the results would meaningfully change based on the number of studies included from the geographies represented and countries that often publish in their dominant language are relatively well-represented in our analysis (e.g., China, Brazil). We classified country-level income status at the time of this analysis, and thus some classification may have changed (almost always to higher income status) since the time of publication or data collection.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eWe included studies on \\u0026lsquo;healthy\\u0026rsquo; diets, and excluded any focused solely on \\u0026lsquo;unhealthy\\u0026rsquo; diets, which means that we may have excluded some dietary measures that are in fact related to worse mental health. \\u0026nbsp;We chose this because unhealthy diets is harder to define. \\u0026nbsp;\\u0026lsquo;Unhealthy\\u0026rsquo; can mean too much consumption in general, too much of the wrong foods (e.g. foods linked with health issues such as diabetes like sugars and UPFs), too little calories (e.g. wasting and thinness), or too few micronutrients (e.g. growth faltering, reduced immunity and micronutrient deficiency). There are also increasing layers of assessment for diets that correspond to planetary rather than human health, for instance the water footprints or fertiliser demands for certain crops in certain settings, or carbon emissions for specific foods or transport processes. \\u0026nbsp;These relationships will also have bearing on mental health status, but this was beyond the scope of this analysis.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003ePractical implications and ways forward\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eOur findings, together with other meta-analyses on related topics, strongly indicate that healthy diets are linked to better mental health. Findings also highlight the striking need for evidence that unpacks causal mechanisms and pathways to impact, including from studies that test what changes, modifies or mediates the relationship between diets and mental health (e.g., disentangling physiological from socioeconomic effects). We will only be able to answer these questions by design: longitudinal, prospective analyses, accounting robustly for a variety of potentially influential factors, reducing likely sources of bias, across diverse settings.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eAdvances in intervention research on this topic are urgently needed. There are promising findings that mental health can improve through dietary interventions\\u003csup\\u003e54\\u0026ndash;57\\u003c/sup\\u003e, that counselling and integrated mental health interventions can improve nutrition outcomes\\u003csup\\u003e58,59\\u003c/sup\\u003e, that nutrition-sensitive interventions can improve mental health even when it is not a primary (or even secondary) outcome\\u003csup\\u003e60\\u003c/sup\\u003e, or that mental health improves through intervention components not specifically designed to improve mental health\\u003csup\\u003e61\\u003c/sup\\u003e. Corroborating these findings and expanding this evidence base will enable us to act on the relationships that are clear in this analysis. The potential synergies between healthy diets and better mental health could prove an important lever to address widespread burdens of poor mental health and poor diets, especially where concentrated among those most vulnerable to poverty and poor health, as well as those most at risk from increasing environmental and political instability.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eFrom a policy perspective, building upon the interdependencies between diet and mental health is nascent, but is essential for policies and programmes focused on basic wellbeing or preventative care. For example, In Ethiopia, the Ministry of Health has demonstrated a strong commitment to integrating mental health services into national health programmes, particularly into primary health care, but also into its health extension and social protection activities for food security e.g., the National Social Protection (NSP) Policy, despite implementation challenges\\u003csup\\u003e62,63\\u003c/sup\\u003e. For instance, mental health integration has been trialled within specific branches of the Productive Safety Net Program (PSNP) by including components such as care groups and Interpersonal Psychotherapy Group (IPT-G) for Depression\\u003csup\\u003e64\\u003c/sup\\u003e. Mental health screening for caretakers has also been proposed as an important part of Community Management of Acute Malnutrition programmes\\u003csup\\u003e65,66\\u003c/sup\\u003e, and long-understood as an important step in the primary care pathway\\u003csup\\u003e67\\u003c/sup\\u003e. We have learned a lot about multisectoral programming for nutrition (e.g. on nutrition-sensitive agriculture, social protection, and school meals)\\u003csup\\u003e68\\u003c/sup\\u003e. \\u0026nbsp;Our findings could be integrated with these lessons, especially policies and programmes aiming to address multiple indicators of the Sustainable Development Goals, as well as wellbeing overall, which likely to centre in the post-2030 agenda\\u003csup\\u003e69,70\\u003c/sup\\u003e. \\u0026nbsp;\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eWe found that overall, healthy diets are associated with less depression, anxiety and stress in LMICs, with more than half of the evidence emerging from cross-sectional studies in Iran and China. \\u0026nbsp;Despite heterogeneity and methodological weaknesses in many studies, these results lend confidence to the robustness of associations. \\u0026nbsp; Our findings are foundational for further inquiry: they should spur studies from more LMIC countries, and settings outside of Iran and China; they showcase the need for studies that advance our understanding of causal mechanisms and intervention research that can tell us how to activate these mechanisms through policy and programming in diverse settings.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eUnderstanding the mechanistic and contextual factors that change the relationship between diets, and more broadly food security and nutrition, with mental health would provide a lever for integrated or co-located interventions that reduce health risks among populations that experience an unequal share of concurrent burdens and marginalisations. If we are intent on addressing global health inequity, then these relationships will be key to improving wellbeing overall. \\u0026nbsp;\\u003c/p\\u003e\"},{\"header\":\"Online Methods\",\"content\":\"\\u003cp\\u003eThe selection of studies for this analysis consisted of 2 steps. We first relied on the inclusion and exclusion criteria of a large systematic Evidence and Gap Map (EGM), the detailed methods of which can be found in the published analysis of the EGM\\u003csup\\u003e45\\u003c/sup\\u003e. We then applied additional criteria to select studies from the EGM for the meta-analysis. \\u0026nbsp;Briefly, the EGM that was created and published first in 2022\\u003csup\\u003e45\\u003c/sup\\u003e including 1945 studies, then updated in 2024 to 3031 studies. We searched Medline, CAB Global Health and PsychInfo databases systematically (using the same search both times with a one-year overlap to account for indexing lags; Supplementary Methods 2) to create the EGM. \\u0026nbsp; We included peer-reviewed, English language studies linking food security and nutrition measures to common mental health problems (depression, anxiety, stress and mental wellbeing) in the general population, published from January 1, 2000 to January 31, 2024. The updated EGM is available here\\u003csup\\u003e46\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eDrawing on the EGM repository, we selected a subset of studies for the meta-analysis. \\u0026nbsp;For healthy diets, we selected all studies fitting the eligibility criteria of the EGM, which were double-screened, with additional eligibility review by a senior researcher. \\u0026nbsp;All studies in the EGM were coded iteratively based on specific measures and indicators used, first by a single researcher, through iterative full-record and domain-specific checks by a senior researcher. Then, aligning with the broad definition of healthy diets proposed by Cena and Calder\\u003csup\\u003e12\\u003c/sup\\u003e, we used relevant measures from the \\u0026lsquo;diets\\u0026rsquo; domain of the EGM that would capture \\u0026lsquo;healthy\\u0026rsquo; diets, such as the Healthy Eating Index\\u003csup\\u003e70\\u003c/sup\\u003e (HEI) and its iterations\\u003csup\\u003e71\\u003c/sup\\u003e, dietary diversity\\u003csup\\u003e72\\u003c/sup\\u003e, the Dietary Inflammatory Index\\u003csup\\u003e73\\u003c/sup\\u003e, and dietary patterns identified through factor analysis or clustering of foods measured through intake questionnaires (e.g., FFQ)\\u003csup\\u003e74\\u003c/sup\\u003e. \\u0026nbsp;A full list of all diet measures is provided in Table 1.\\u003c/p\\u003e\\n\\u003cp\\u003eAmong the MH domain in the EGM, we included all the studies that measured depression, anxiety or stress using any validated tool, such as mental health screening instruments like the Centre for Epidemiological Studies \\u0026ndash; Depression Scale (CES-D), the Depression, Anxiety and Stress Scale (DASS), the Generalised Anxiety and Depression Scale (GAD) or the State Trait Anxiety Index (STAI). In this second step, our exclusion criteria for the meta-analysis consisted of 1) Mental health different from depression, anxiety or stress, 2) Lack of comparability of diet measurements (not valid or comparable within the meta-analysis), 3) Duplicated data: same sample and same results of an already included study, and 4) Relevant estimate not reported. We excluded any study with a measure of healthy diet without an unhealthy diet comparator. The detailed list of inclusion and exclusion criteria for the EGM and the meta-analysis, including all the diet and MH measurements out of scope are detailed in the Supplementary Methods 3.\\u003c/p\\u003e\\n\\u003cp\\u003eWe grouped eligible studies according to the dietary measurements used (see Table 1 in Results). Group 1 included adherence and adequacy includes adherence to diet recommendations and nutrient adequacy. Group 2 included dietary patterns shown to reduce nutrition-related chronic diseases, such as Dietary Approaches to Stop Hypertension (DASH), Dietary Inflammatory Index (DII), and Mediterranean diet. Group 3 was made up of diet diversity indices like Minimum Dietary Diversity of Women\\u0026rsquo;s (MDD-W) and individual dietary diversity (IDDS) as well as any other Dietary Variety Scores. Group 4 Diet quality indices included all Diet Quality Indices, Global Diet Quality Score (GDQS), Global Dietary Index (GDI), and all Healthy Eating Indices (HEI). Groups (1-4) are all measured via predefined patterns and/or specific food intake (\\u003cem\\u003ea priori\\u003c/em\\u003e methods), thus the observed dietary pattern in the population is compared to a preexisting index. Group 5 included any measure derived from factor analysis, Principle Component Analysis (PCA) or other data-driven (\\u003cem\\u003ea posteriori\\u003c/em\\u003e) approaches, which comprise comparisons between clustered patterns of food intake in a study population (usually between a \\u0026lsquo;healthy pattern\\u0026rsquo; and an \\u0026lsquo;unhealthy\\u0026rsquo; one, e.g., processed, western, modern, or unhealthy).\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eFor included studies, two researchers extracted data on effect sizes (since most of the descriptive characteristics of studies were already coded in the EGM), along with claim type (associational or causal), exposure scale (healthy or unhealthy), sample size, number of women, percentage of women in the sample, age group, mean age, age standard deviation, and reported statistical measures. Specifically, we extracted the following statistical effect sizes: Beta Coefficients (\\u0026beta;), Odds Ratio (OR), Adjusted Odds Ratio (AOR), Risk Ratio (RR), Hazard Ratio (HR), Pearson\\u0026rsquo;s Correlation Coefficient (r), Mean Differences. We made sure to capture the direction of the \\u0026lsquo;healthy\\u0026rsquo; effect on mental health symptoms.\\u003c/p\\u003e\\n\\u003cp\\u003eEffect sizes were converted to standardized metrics to ensure comparability across studies. For odds ratios (OR), risk ratios (RR), and prevalence ratios (PR), the Chinn transformation\\u003csup\\u003e75\\u003c/sup\\u003e was applied to obtain Cohen\\u0026rsquo;s d. \\u0026beta; coefficients were standardized using the pooled standard deviation, while mean differences were converted by dividing the difference between group means by the pooled standard deviation. Hazard ratios were adjusted based on their distribution properties. To correct for small sample bias, Hedges\\u0026rsquo; g was computed\\u003csup\\u003e76\\u003c/sup\\u003e. Pearson\\u0026rsquo;s r was derived from standardized mean differences, and Fisher\\u0026rsquo;s z transformation was applied to normalize correlations. As the primary effect size index, we selected the Standardized Mean Difference (SMD) estimated by Hedges\\u0026rsquo; g due to its robustness in meta-analyses, allowing for comparisons of means and regression coefficients across diverse study designs\\u003csup\\u003e77\\u003c/sup\\u003e, and easier interpretability\\u003csup\\u003e76\\u003c/sup\\u003e.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eWhen studies used data from the same sample population, but reported a different sample size, we extracted the highest number to estimate sample populations. To estimate pooled effect sizes, we conducted a three-level meta-analysis using a robust variance estimation (RVE) framework. The three-level meta-analysis dealt with the dependency coming from studies drawing estimates from the same populations\\u003csup\\u003e78\\u003c/sup\\u003e, and we modelled our data to account for differences at the individual effect sizes, within sample population and between sample populations. The RVE allows adjustment for the standard errors and improves the statistical inferences when we face a data dependency issue\\u003csup\\u003e79\\u003c/sup\\u003e. The model was estimated using the restricted maximum-likelihood (REML) estimator\\u003csup\\u003e80\\u003c/sup\\u003e, incorporating nested random effects at the study and effect-size levels. Variance components were estimated using a structured variance-covariance matrix with a predefined correlation coefficient (\\u0026rho;=0.5) to model dependence among estimates derived from the same dataset. Observations were weighted by their inverse variance, and heterogeneity was assessed through variance decomposition (I\\u0026sup2;) across levels. We considered both confidence intervals and prediction intervals, the latter of which is a more transparent indicator of heterogeneity at a population level. To evaluate the robustness of the findings, sensitivity analyses included alternative specifications of the correlation parameter, exclusion of influential studies, and comparisons across effect size metrics (Hedges\\u0026rsquo; g, Cohen\\u0026rsquo;s d, Fisher\\u0026rsquo;s z).\\u003c/p\\u003e\\n\\u003cp\\u003eOutlier and influential study detection were performed using Studentized residuals and Cook\\u0026rsquo;s distances. Studies were flagged as potential outliers if their Studentized residual exceeded the 100 \\u0026times; (1 - 0.05/ (2 \\u0026times; k))th percentile of a standard normal distribution, applying a Bonferroni correction for multiple comparisons with a two-sided \\u0026alpha;=0.05 across k studies. Influential studies were identified using Cook\\u0026rsquo;s distance, where values exceeding the median plus six times the interquartile range (IQR) were considered indicative of undue influence on model estimates.\\u003c/p\\u003e\\n\\u003cp\\u003eWe tested alternative approaches to deal with the dependency of the data\\u003csup\\u003e81\\u003c/sup\\u003e. Model comparison was conducted by testing a reduced two-level model (removing the second random effect) against the full three-level specification to assess whether modelling within-study dependence significantly improved model fit\\u003csup\\u003e82\\u003c/sup\\u003e. The full three-level model yielded lower Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and corrected AIC (AICc) values, alongside a higher log-likelihood (logLik), indicating a superior fit\\u003csup\\u003e83\\u003c/sup\\u003e. The likelihood ratio test (LRT) comparing the two models returned a p-value of 0.0831, suggesting that while the improvement in fit was not statistically significant at the conventional 5% level, it was marginally close. Given the hierarchical nature of the data, where multiple effect sizes stem from the same study population, the three-level model was retained as the preferred approach to appropriately account for dependency and reduce bias in pooled estimates\\u003csup\\u003e82\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eA senior researcher assessed the risk of bias (RoB) for each study using the \\u0026ldquo;Quality In Prognosis Studies\\u0026rdquo; (QUIPS) tool\\u003csup\\u003e84\\u003c/sup\\u003e, with any uncertainty checked by a second researcher. Although the QUIPS tool is designed for prognostic studies, prognoses are similar to risks in epidemiology. \\u0026nbsp;Furthermore, most of the included studies are cross-sectional, and there is no existing tool for these that covers all important domains of possible bias\\u003csup\\u003e85\\u003c/sup\\u003e. \\u0026nbsp;The QUIPS covers many of the important domains identified, including study participation, study attrition (omitted for cross-sectional studies), prognostic factor (exchanged for exposure) measurement, outcome measurement, study confounding, and statistical analysis and reporting. For each of these domains, studies were ranked as having low, moderate or high risk of bias. Then, we grouped all studies in 2 groups: high RoB (has at least 1 domain with high RoB) and low RoB (no domain has high RoB).\\u003c/p\\u003e\\n\\u003cp\\u003eWe examined the publication trends over time; used the country-level income classification as defined by the World Bank at the time of this analysis (2024)\\u003csup\\u003e86\\u003c/sup\\u003e to examine effects by low-, lower-middle, and upper-middle income country status; and analysed the geographic spread across regions and countries. We included literature where healthy diets and mental health were associated, including when the hypothesised exposure was healthy diets or mental health, and vice versa for the outcome or dependent variable. \\u0026nbsp; We used the \\u0026lsquo;hypothesis direction\\u0026rsquo; classification in the EGM to examine these different groups, as well as analysing potentially differential effects for different population groups (e.g. adults, elderly, adolescents, or parents paired with their children). We examined pooled estimates by dietary measure group (Table 1).\\u003c/p\\u003e\\n\\u003cp\\u003eWe conducted several sensitivity analyses by different study characteristics: study design, dietary measurements and income level. All our sensitivity analyses were conducted first on the entire set of studies, then restricting to low RoB studies, and restricting to the high RoB studies.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eFunding:\\u0026nbsp;\\u003c/strong\\u003eThis work is funded through the Innovative Methods and Metrics for Agriculture and Nutrition Action (IMMANA) programme, led by the London School of Hygiene \\u0026amp; Tropical Medicine (LSHTM), in partnership with Tufts University and the University of Sheffield. IMMANA is co-funded with UK International Development from the UK government and by the Gates Foundation INV-002962 / OPP1211308. The conclusions and opinions expressed in this work are those of the author(s) alone and shall not be attributed to the Foundation. Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission. Please note works submitted as a preprint have not undergone a peer review process\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAuthor contributions:\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eT.M.S.: conceptualization, methodology, investigation, data curation, writing (original draft), writing (review and editing), project administration. C.C.: methodology, data curation, statistical- formal analysis and visualisation, software and programming, writing (original draft), writing (review and editing). B.C.: methodology, investigation, data curation. L.M.T.: writing (review and editing). \\u0026nbsp;S.K.: funding acquisition, supervision, writing (review and editing).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAcknowledgements\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eWe thank the IMMANA team for their ideas, logistical and dissemination support. We thank Claudia Offner and Megan Deeney for their extensive work managing and contributing to the initial iteration of the Evidence and Gap Map, as well as Xuerui Han, Zhuozhi Lin and Chiara Lier for working to screen and code records from this iteration We thank Leisha Beardmore for their contribution to hiring and managing the team who updated the EGM, and to Corina Zhao, Tala Chehaitly, Yifei Li, Chantel Yenyu Ku, and Brena Bessa for their substantial efforts in screening and coding records for the EGM update.\\u003c/p\\u003e\"},{\"header\":\"References\",\"content\":\"\\u003col\\u003e\\n \\u003cli\\u003eSinger, M., Bulled, N., Ostrach, B. \\u0026amp; Mendenhall, E. Syndemics and the biosocial conception of health. \\u003cem\\u003eThe Lancet\\u003c/em\\u003e \\u003cstrong\\u003e389\\u003c/strong\\u003e, 941\\u0026ndash;950 (2017).\\u003c/li\\u003e\\n \\u003cli\\u003eInstitute for Health Metrics and Evaluation (IHME). \\u003cem\\u003eGlobal Burden of Disease 2021: Findings from the GBD 2021 Study\\u003c/em\\u003e. https://www.healthdata.org/research-analysis/library/global-burden-disease-2021-findings-gbd-2021-study (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eMurray, C. J. L. Findings from the Global Burden of Disease Study 2021. \\u003cem\\u003eThe Lancet\\u003c/em\\u003e \\u003cstrong\\u003e403\\u003c/strong\\u003e, 2259\\u0026ndash;2262 (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eFAO, IFAD, UNICEF, WFP \\u0026amp; WHO. \\u003cem\\u003eThe State of Food Security and Nutrition in the World 2024\\u003c/em\\u003e. (FAO; IFAD; UNICEF; WFP; WHO;, 2024). doi:10.4060/cd1254en.\\u003c/li\\u003e\\n \\u003cli\\u003eWorld Health Organization. Government expenditures on mental health as a percentage of total government expenditures on health (%). WHO Global Health Observatory data repository.\\u003c/li\\u003e\\n \\u003cli\\u003eAli, \\u0026nbsp;a-C., Ryan, G. \\u0026amp; De Silva, M. J. Validated Screening Tools for Common Mental Disorders in Low and Middle Income Countries: A Systematic Review. \\u003cem\\u003ePLOS ONE\\u003c/em\\u003e \\u003cstrong\\u003e11\\u003c/strong\\u003e, e0156939 (2016).\\u003c/li\\u003e\\n \\u003cli\\u003eAmerican Psychiatric Association. \\u003cem\\u003eDiagnostic and Statistical Manual of Mental Disorders\\u003c/em\\u003e. (American Psychiatric Association, 2013). doi:10.1176/appi.books.9780890425596.\\u003c/li\\u003e\\n \\u003cli\\u003eKirk, S. A., Cohen, D. \\u0026amp; Gomory, T. DSM-5: The Delayed Demise of Descriptive Diagnosis. in \\u003cem\\u003eThe DSM-5 in Perspective\\u003c/em\\u003e (eds. Demazeux, S. \\u0026amp; Singy, P.) vol. 10 63\\u0026ndash;81 (Springer Netherlands, Dordrecht, 2015).\\u003c/li\\u003e\\n \\u003cli\\u003eLacasse, J. R. After DSM-5: A Critical Mental Health Research Agenda for the 21st Century. \\u003cem\\u003eRes. Soc. Work Pract.\\u003c/em\\u003e \\u003cstrong\\u003e24\\u003c/strong\\u003e, 5\\u0026ndash;10 (2014).\\u003c/li\\u003e\\n \\u003cli\\u003eTennyson, R. L., Kemp, C. G. \\u0026amp; Rao, D. Challenges and strategies for implementing mental health measurement for research in low-resource settings. \\u003cem\\u003eInt. Health\\u003c/em\\u003e \\u003cstrong\\u003e8\\u003c/strong\\u003e, 374\\u0026ndash;380 (2016).\\u003c/li\\u003e\\n \\u003cli\\u003eWatson, L. K., Kaiser, B. N., Giusto, A. M., Ayuku, D. \\u0026amp; Puffer, E. S. Validating mental health assessment in Kenya using an innovative gold standard. \\u003cem\\u003eInt. J. Psychol. J. Int. Psychol.\\u003c/em\\u003e \\u003cstrong\\u003e55\\u003c/strong\\u003e, 425\\u0026ndash;434 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eCena, H. \\u0026amp; Calder, P. C. Defining a Healthy Diet: Evidence for the Role of Contemporary Dietary Patterns in Health and Disease. \\u003cem\\u003eNutrients\\u003c/em\\u003e \\u003cstrong\\u003e12\\u003c/strong\\u003e, 334 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eNeufeld, L. M., Hendriks, S. \\u0026amp; Hugas, M. Healthy Diet: A Definition for the United Nations Food Systems Summit 2021. in \\u003cem\\u003eScience and Innovations for Food Systems Transformation\\u003c/em\\u003e (eds. Von Braun, J., Afsana, K., Fresco, L. O. \\u0026amp; Hassan, M. H. A.) 21\\u0026ndash;30 (Springer International Publishing, Cham, 2023). doi:10.1007/978-3-031-15703-5_3.\\u003c/li\\u003e\\n \\u003cli\\u003eSparling, T. M. \\u003cem\\u003eet al.\\u003c/em\\u003e Global Mental Health and Nutrition: Moving Toward a Convergent Research Agenda. \\u003cem\\u003eFront. Public Health\\u003c/em\\u003e \\u003cstrong\\u003e9\\u003c/strong\\u003e, 722290 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eGianfredi, V. \\u003cem\\u003eet al.\\u003c/em\\u003e Association between dietary patterns and depression: an umbrella review of meta-analyses of observational studies and intervention trials. \\u003cem\\u003eNutr. Rev.\\u003c/em\\u003e \\u003cstrong\\u003e81\\u003c/strong\\u003e, 346\\u0026ndash;359 (2023).\\u003c/li\\u003e\\n \\u003cli\\u003eAdjibade, M. \\u003cem\\u003eet al.\\u003c/em\\u003e Prospective association between adherence to dietary recommendations and incident depressive symptoms in the French NutriNet-Sant\\u0026eacute; cohort. \\u003cem\\u003eBr. J. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e120\\u003c/strong\\u003e, 290\\u0026ndash;300 (2018).\\u003c/li\\u003e\\n \\u003cli\\u003eBizzozero-Peroni, B. \\u003cem\\u003eet al.\\u003c/em\\u003e Proinflammatory dietary pattern and depression risk in older adults: Prospective analyses from the Seniors-ENRICA studies. \\u003cem\\u003eClin. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e41\\u003c/strong\\u003e, 2614\\u0026ndash;2620 (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eCherian, L. \\u003cem\\u003eet al.\\u003c/em\\u003e DASH and Mediterranean-Dash Intervention for Neurodegenerative Delay (MIND) Diets Are Associated With Fewer Depressive Symptoms Over Time. \\u003cem\\u003eJ. Gerontol. A. Biol. Sci. Med. Sci.\\u003c/em\\u003e \\u003cstrong\\u003e76\\u003c/strong\\u003e, 151\\u0026ndash;156 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eCollin, C. \\u003cem\\u003eet al.\\u003c/em\\u003e Adherence to dietary guidelines as a protective factor against chronic or recurrent depressive symptoms in the French SU.VI.MAX cohort. \\u003cem\\u003ePrev. Med.\\u003c/em\\u003e \\u003cstrong\\u003e91\\u003c/strong\\u003e, 335\\u0026ndash;343 (2016).\\u003c/li\\u003e\\n \\u003cli\\u003eJacka, F. N. \\u003cem\\u003eet al.\\u003c/em\\u003e A Prospective Study of Diet Quality and Mental Health in Adolescents. \\u003cem\\u003ePLoS ONE\\u003c/em\\u003e \\u003cstrong\\u003e6\\u003c/strong\\u003e, e24805 (2011).\\u003c/li\\u003e\\n \\u003cli\\u003eJiang, W. \\u003cem\\u003eet al.\\u003c/em\\u003e The relationship of dietary diversity score with depression and anxiety among prenatal and post‐partum women. \\u003cem\\u003eJ. Obstet. Gynaecol. Res.\\u003c/em\\u003e \\u003cstrong\\u003e44\\u003c/strong\\u003e, 1929\\u0026ndash;1936 (2018).\\u003c/li\\u003e\\n \\u003cli\\u003eRecchia, D. \\u003cem\\u003eet al.\\u003c/em\\u003e Associations between long-term adherence to healthy diet and recurrent depressive symptoms in Whitehall II Study. \\u003cem\\u003eEur. J. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e59\\u003c/strong\\u003e, 1031\\u0026ndash;1041 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eWang, J. \\u003cem\\u003eet al.\\u003c/em\\u003e Dietary inflammatory index and depression: a meta-analysis. \\u003cem\\u003ePublic Health Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e22\\u003c/strong\\u003e, 654\\u0026ndash;660 (2019).\\u003c/li\\u003e\\n \\u003cli\\u003eOftedal, S., Glozier, N., Holliday, E. G. \\u0026amp; Duncan, M. J. Diet quality and depressive symptoms. Assessing the direction of the association in a population-based cohort study. \\u003cem\\u003eJ. Affect. Disord.\\u003c/em\\u003e \\u003cstrong\\u003e274\\u003c/strong\\u003e, 347\\u0026ndash;353 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eArvidsson, L. \\u003cem\\u003eet al.\\u003c/em\\u003e Bidirectional associations between psychosocial well-being and adherence to healthy dietary guidelines in European children: prospective findings from the IDEFICS study. \\u003cem\\u003eBMC Public Health\\u003c/em\\u003e \\u003cstrong\\u003e17\\u003c/strong\\u003e, 926 (2017).\\u003c/li\\u003e\\n \\u003cli\\u003eJacka, F. N., Cherbuin, N., Anstey, K. J. \\u0026amp; Butterworth, P. Does reverse causality explain the relationship between diet and depression? \\u003cem\\u003eJ. Affect. Disord.\\u003c/em\\u003e \\u003cstrong\\u003e175\\u003c/strong\\u003e, 248\\u0026ndash;250 (2015).\\u003c/li\\u003e\\n \\u003cli\\u003eMarx, W., Moseley, G., Berk, M. \\u0026amp; Jacka, F. Nutritional psychiatry: the present state of the evidence. \\u003cem\\u003eProc. Nutr. Soc.\\u003c/em\\u003e \\u003cstrong\\u003e76\\u003c/strong\\u003e, 427\\u0026ndash;436 (2017).\\u003c/li\\u003e\\n \\u003cli\\u003eBerk, M. \\u003cem\\u003eet al.\\u003c/em\\u003e So depression is an inflammatory disease, but where does the inflammation come from? \\u003cem\\u003eBMC Med.\\u003c/em\\u003e \\u003cstrong\\u003e11\\u003c/strong\\u003e, 200 (2013).\\u003c/li\\u003e\\n \\u003cli\\u003eSchwingshackl, L. \\u0026amp; Hoffmann, G. Mediterranean dietary pattern, inflammation and endothelial function: A systematic review and meta-analysis of intervention trials. \\u003cem\\u003eNutr. Metab. Cardiovasc. Dis.\\u003c/em\\u003e \\u003cstrong\\u003e24\\u003c/strong\\u003e, 929\\u0026ndash;939 (2014).\\u003c/li\\u003e\\n \\u003cli\\u003eEstruch, R. Anti-inflammatory effects of the Mediterranean diet: the experience of the PREDIMED study. \\u003cem\\u003eProc. Nutr. Soc.\\u003c/em\\u003e \\u003cstrong\\u003e69\\u003c/strong\\u003e, 333\\u0026ndash;340 (2010).\\u003c/li\\u003e\\n \\u003cli\\u003eGuimar\\u0026atilde;es, L. R. \\u003cem\\u003eet al.\\u003c/em\\u003e Serum levels of brain-derived neurotrophic factor in schizophrenia on a hypocaloric diet. \\u003cem\\u003eProg. Neuropsychopharmacol. Biol. Psychiatry\\u003c/em\\u003e \\u003cstrong\\u003e32\\u003c/strong\\u003e, 1595\\u0026ndash;1598 (2008).\\u003c/li\\u003e\\n \\u003cli\\u003eMoylan, S. \\u003cem\\u003eet al.\\u003c/em\\u003e Oxidative \\u0026amp; nitrosative stress in depression: Why so much stress? \\u003cem\\u003eNeurosci. Biobehav. Rev.\\u003c/em\\u003e \\u003cstrong\\u003e45\\u003c/strong\\u003e, 46\\u0026ndash;62 (2014).\\u003c/li\\u003e\\n \\u003cli\\u003eZainuddin, M. S. A. \\u0026amp; Thuret, S. Retracted: Nutrition, adult hippocampal neurogenesis and mental health. \\u003cem\\u003eBr. Med. Bull.\\u003c/em\\u003e \\u003cstrong\\u003e103\\u003c/strong\\u003e, 89\\u0026ndash;114 (2012).\\u003c/li\\u003e\\n \\u003cli\\u003eCarroll, S. J., Dale, M. J., Niyonsenga, T., Taylor, A. W. \\u0026amp; Daniel, M. Associations between area socioeconomic status, individual mental health, physical activity, diet and change in cardiometabolic risk amongst a cohort of Australian adults: A longitudinal path analysis. \\u003cem\\u003ePLOS ONE\\u003c/em\\u003e \\u003cstrong\\u003e15\\u003c/strong\\u003e, e0233793 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eSnyder, H. R. Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: A meta-analysis and review. \\u003cem\\u003ePsychol. Bull.\\u003c/em\\u003e \\u003cstrong\\u003e139\\u003c/strong\\u003e, 81\\u0026ndash;132 (2013).\\u003c/li\\u003e\\n \\u003cli\\u003eArnsten, A. F. T. \\u0026amp; Rubia, K. Neurobiological Circuits Regulating Attention, Cognitive Control, Motivation, and Emotion: Disruptions in Neurodevelopmental Psychiatric Disorders. \\u003cem\\u003eJ. Am. Acad. Child Adolesc. Psychiatry\\u003c/em\\u003e \\u003cstrong\\u003e51\\u003c/strong\\u003e, 356\\u0026ndash;367 (2012).\\u003c/li\\u003e\\n \\u003cli\\u003eJacka, F. N., Cherbuin, N., Anstey, K. J., Sachdev, P. \\u0026amp; Butterworth, P. Western diet is associated with a smaller hippocampus: a longitudinal investigation. \\u003cem\\u003eBMC Med.\\u003c/em\\u003e \\u003cstrong\\u003e13\\u003c/strong\\u003e, 215 (2015).\\u003c/li\\u003e\\n \\u003cli\\u003eChristensen, L. \\u0026amp; Brooks, A. Changing Food Preference as a Function of Mood. \\u003cem\\u003eJ. Psychol.\\u003c/em\\u003e \\u003cstrong\\u003e140\\u003c/strong\\u003e, 293\\u0026ndash;306 (2006).\\u003c/li\\u003e\\n \\u003cli\\u003eSamuthpongtorn, C. \\u003cem\\u003eet al.\\u003c/em\\u003e Consumption of Ultraprocessed Food and Risk of Depression. \\u003cem\\u003eJAMA Netw. Open\\u003c/em\\u003e \\u003cstrong\\u003e6\\u003c/strong\\u003e, e2334770 (2023).\\u003c/li\\u003e\\n \\u003cli\\u003eMiller, L. C. \\u003cem\\u003eet al.\\u003c/em\\u003e Maternal depression is associated with less dietary diversity among rural Nepali children. \\u003cem\\u003eMatern. Child. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e17\\u003c/strong\\u003e, e13221 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eAnato, A., Baye, K., Tafese, Z. \\u0026amp; Stoecker, B. J. Maternal depression is associated with child undernutrition: A cross‐sectional study in Ethiopia. \\u003cem\\u003eMatern. Child. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e16\\u003c/strong\\u003e, e12934 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eChee Din, M. A., Mohd Fahmi Teng, N. I. \\u0026amp; Abdul Manaf, Z. Maternal depression and child feeding practices: Determinants to malnutrition among young children in Malaysian rural area. \\u003cem\\u003eWomens Health\\u003c/em\\u003e \\u003cstrong\\u003e19\\u003c/strong\\u003e, 17455057221147800 (2023).\\u003c/li\\u003e\\n \\u003cli\\u003eFismen, A.-S. \\u003cem\\u003eet al.\\u003c/em\\u003e Associations between eating habits and mental health among adolescents in five nordic countries: a cross-sectional survey. \\u003cem\\u003eBMC Public Health\\u003c/em\\u003e \\u003cstrong\\u003e24\\u003c/strong\\u003e, 2640 (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eDabravolskaj, J. \\u003cem\\u003eet al.\\u003c/em\\u003e Association Between Diet and Mental Health Outcomes in a Sample of 13,887 Adolescents in Canada. \\u003cem\\u003ePrev. Chronic. Dis.\\u003c/em\\u003e \\u003cstrong\\u003e21\\u003c/strong\\u003e, 240187 (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eSparling, T. M. \\u003cem\\u003eet al.\\u003c/em\\u003e Systematic evidence and gap map of research linking food security and nutrition to mental health. \\u003cem\\u003eNat. Commun.\\u003c/em\\u003e \\u003cstrong\\u003e13\\u003c/strong\\u003e, 4608 (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eSparling, T. M. \\u0026amp; Cornejo, C. Systematic evidence and gap map of research linking food security and nutrition to mental health. \\u003cem\\u003eANH Academy\\u003c/em\\u003e https://www.anh-academy.org/mentalhealth-egm-population.htm (2025).\\u003c/li\\u003e\\n \\u003cli\\u003eXu, Y. \\u003cem\\u003eet al.\\u003c/em\\u003e Role of dietary factors in the prevention and treatment for depression: an umbrella review of meta-analyses of prospective studies. \\u003cem\\u003eTransl. Psychiatry\\u003c/em\\u003e \\u003cstrong\\u003e11\\u003c/strong\\u003e, 478 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eLi, Y. \\u003cem\\u003eet al.\\u003c/em\\u003e Dietary patterns and depression risk: A meta-analysis. \\u003cem\\u003ePsychiatry Res.\\u003c/em\\u003e \\u003cstrong\\u003e253\\u003c/strong\\u003e, 373\\u0026ndash;382 (2017).\\u003c/li\\u003e\\n \\u003cli\\u003eLassale, C. \\u003cem\\u003eet al.\\u003c/em\\u003e Healthy dietary indices and risk of depressive outcomes: a systematic review and meta-analysis of observational studies. \\u003cem\\u003eMol. Psychiatry\\u003c/em\\u003e \\u003cstrong\\u003e24\\u003c/strong\\u003e, 965\\u0026ndash;986 (2019).\\u003c/li\\u003e\\n \\u003cli\\u003eMatison, A. P., Mather, K. A., Flood, V. M. \\u0026amp; Reppermund, S. Associations between nutrition and the incidence of depression in middle-aged and older adults: A systematic review and meta-analysis of prospective observational population-based studies. \\u003cem\\u003eAgeing Res. Rev.\\u003c/em\\u003e \\u003cstrong\\u003e70\\u003c/strong\\u003e, 101403 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eAskari, M. \\u003cem\\u003eet al.\\u003c/em\\u003e Vegetarian diet and the risk of depression, anxiety, and stress symptoms: a systematic review and meta-analysis of observational studies. \\u003cem\\u003eCrit. Rev. Food Sci. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e62\\u003c/strong\\u003e, 261\\u0026ndash;271 (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eHerforth, A. The Sustainable Development Goals have a new diet indicator! Both the metric and the methods are key | ANH Academy. https://www.anh-academy.org/community/blogs/the-sustainable-development-goals-have-a-new-diet-indicator-both-the-metric-and-the.\\u003c/li\\u003e\\n \\u003cli\\u003eHeidari-Beni, M., Hemati, Z. \\u0026amp; Qorbani, M. The Dietary Diversity Score: Methods, Indicators, and Applications to General Population. in \\u003cem\\u003eBiomarkers in Nutrition\\u003c/em\\u003e (eds. Patel, V. B. \\u0026amp; Preedy, V. R.) 293\\u0026ndash;306 (Springer International Publishing, Cham, 2022). doi:10.1007/978-3-031-07389-2_17.\\u003c/li\\u003e\\n \\u003cli\\u003eCetrone, H. M. \\u003cem\\u003eet al.\\u003c/em\\u003e Food security mediates the decrease in women\\u0026rsquo;s depressive symptoms in a participatory nutrition-sensitive agroecology intervention in rural Tanzania. \\u003cem\\u003ePublic Health Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e24\\u003c/strong\\u003e, 4682\\u0026ndash;4692 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eCasas, I. \\u003cem\\u003eet al.\\u003c/em\\u003e Effects of a Mediterranean Diet Intervention on Maternal Stress, Well-Being, and Sleep Quality throughout Gestation\\u0026mdash;The IMPACT-BCN Trial. \\u003cem\\u003eNutrients\\u003c/em\\u003e \\u003cstrong\\u003e15\\u003c/strong\\u003e, 2362 (2023).\\u003c/li\\u003e\\n \\u003cli\\u003eParletta, N. \\u003cem\\u003eet al.\\u003c/em\\u003e A Mediterranean-style dietary intervention supplemented with fish oil improves diet quality and mental health in people with depression: A randomized controlled trial (HELFIMED). \\u003cem\\u003eNutr. Neurosci.\\u003c/em\\u003e \\u003cstrong\\u003e22\\u003c/strong\\u003e, 474\\u0026ndash;487 (2019).\\u003c/li\\u003e\\n \\u003cli\\u003eJacka, F. N. \\u003cem\\u003eet al.\\u003c/em\\u003e A randomised controlled trial of dietary improvement for adults with major depression (the \\u0026lsquo;SMILES\\u0026rsquo; trial). \\u003cem\\u003eBMC Med.\\u003c/em\\u003e \\u003cstrong\\u003e15\\u003c/strong\\u003e, 23 (2017).\\u003c/li\\u003e\\n \\u003cli\\u003eWatanabe, J. \\u003cem\\u003eet al.\\u003c/em\\u003e Effects of \\u0026lsquo;SPRAT\\u0026rsquo; programme for dietary and lifestyle education to improve psychosomatic symptoms and dietary habits among adolescents: a cluster randomised controlled trial. \\u003cem\\u003eBMC Public Health\\u003c/em\\u003e \\u003cstrong\\u003e22\\u003c/strong\\u003e, 461 (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eHaque, N. B., Mihrshahi, S. \\u0026amp; Haider, R. Peer counselling as an approach to improve complementary feeding practices: a narrative review. \\u003cem\\u003eJ. Health Popul. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e42\\u003c/strong\\u003e, 60 (2023).\\u003c/li\\u003e\\n \\u003cli\\u003eCetrone, H. \\u003cem\\u003eet al.\\u003c/em\\u003e A Participatory Agroecological Intervention Reduces Women\\u0026rsquo;s Risk of Probable Depression Through Improvements in Food Security in Singida, Tanzania. \\u003cem\\u003eCurr. Dev. Nutr.\\u003c/em\\u003e \\u003cstrong\\u003e4\\u003c/strong\\u003e, nzaa053_024 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eVerdeli, H. \\u003cem\\u003eet al.\\u003c/em\\u003e Alleviating maternal depression to improve IYCF and WASH behaviors. in (CORE Group, Online, 2021).\\u003c/li\\u003e\\n \\u003cli\\u003eEngdawork, K. \\u003cem\\u003eet al.\\u003c/em\\u003e \\u0026ldquo;Space to see the future\\u0026rdquo;? A political economy analysis of child and adolescent mental health and well-being in Ethiopia including routes for change. \\u003cem\\u003eFront. Sociol.\\u003c/em\\u003e \\u003cstrong\\u003e9\\u003c/strong\\u003e, 1488619 (2025).\\u003c/li\\u003e\\n \\u003cli\\u003eYitbarek, K. \\u003cem\\u003eet al.\\u003c/em\\u003e Barriers and Facilitators for Implementing Mental Health Services into the Ethiopian Health Extension Program: A Qualitative Study. \\u003cem\\u003eRisk Manag. Healthc. Policy\\u003c/em\\u003e \\u003cstrong\\u003eVolume 14\\u003c/strong\\u003e, 1199\\u0026ndash;1210 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eGilligan, D. O., Hirvonen, K., Leight, J. \\u0026amp; Tesfaye, H. \\u003cem\\u003eImplementer-Led Evaluation and Learning (IMPEL) Evaluation of SPIR II RFSA - Midline Survey Report\\u003c/em\\u003e. https://hdl.handle.net/10568/144025 (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eAbdullahi, A. T., Farouk, Z. L. \\u0026amp; Imam, A. Common mental disorders in mothers of children attending out-patient malnutrition clinics in rural North-western Nigeria: a cross-sectional study. \\u003cem\\u003eBMC Public Health\\u003c/em\\u003e \\u003cstrong\\u003e21\\u003c/strong\\u003e, 185 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eLe Roch, K., Tonon, B. \\u0026amp; Acharya, O. What influence mothers\\u0026rsquo; mental health and health care seeking behaviors for their malnourished children in Nepal: building evidence for a broader perspective. \\u003cem\\u003eEur. Psychiatry\\u003c/em\\u003e \\u003cstrong\\u003e67\\u003c/strong\\u003e, S55\\u0026ndash;S55 (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eWorld Health Organization \\u0026amp; World Organization of Family Doctors. Integrating mental health into primary care : a global perspective. 206 (2008).\\u003c/li\\u003e\\n \\u003cli\\u003eStorhaug, I. G. \\u003cem\\u003eet al.\\u003c/em\\u003e Making the most of existing research: an evidence gap map of the effects of food systems interventions in low-income and middle-income countries. \\u003cem\\u003eBMJ Open\\u003c/em\\u003e \\u003cstrong\\u003e12\\u003c/strong\\u003e, e055062 (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eHigh-Level Committee on Programmes (HLCP). Valuing What Counts - UN System-wide Contribution on Beyond GDP (advance unedited). (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eJansen, A., Wang, R., Behrens, P. \\u0026amp; Hoekstra, R. Beyond GDP: a review and conceptual framework for measuring sustainable and inclusive wellbeing. \\u003cem\\u003eLancet Planet. Health\\u003c/em\\u003e \\u003cstrong\\u003e8\\u003c/strong\\u003e, e695\\u0026ndash;e705 (2024).\\u003c/li\\u003e\\u003c/ol\\u003e\\n \\u003ch2\\u003eOnline Methods References\\u003c/strong\\u003e\\u003c/h2\\u003e\\n \\u003col start=71\\u003e \\u003cli\\u003eSparling, T. M. \\u003cem\\u003eet al.\\u003c/em\\u003e Systematic evidence and gap map of research linking food security and nutrition to mental health. \\u003cem\\u003eNat Commun\\u003c/em\\u003e \\u003cstrong\\u003e13\\u003c/strong\\u003e, 4608 (2022).\\u003c/li\\u003e\\n \\u003cli\\u003eSparling, T. M. \\u0026amp; Cornejo, C. Systematic evidence and gap map of research linking food security and nutrition to mental health. \\u003cem\\u003eANH Academy\\u003c/em\\u003e https://www.anh-academy.org/mentalhealth-egm-population.htm (2025).\\u003c/li\\u003e\\n \\u003cli\\u003eCena, H. \\u0026amp; Calder, P. C. Defining a Healthy Diet: Evidence for the Role of Contemporary Dietary Patterns in Health and Disease. \\u003cem\\u003eNutrients\\u003c/em\\u003e \\u003cstrong\\u003e12\\u003c/strong\\u003e, 334 (2020).\\u003c/li\\u003e\\n \\u003cli\\u003eGuenther, P. M., Reedy, J. \\u0026amp; Krebs-Smith, S. M. Development of the Healthy Eating Index-2005. \\u003cem\\u003eJournal of the American Dietetic Association\\u003c/em\\u003e \\u003cstrong\\u003e108\\u003c/strong\\u003e, 1896\\u0026ndash;1901 (2008).\\u003c/li\\u003e\\n \\u003cli\\u003eShams-White, M. M. \\u003cem\\u003eet al.\\u003c/em\\u003e Healthy Eating Index-2020: Review and Update Process to Reflect the Dietary Guidelines for Americans, 2020-2025. \\u003cem\\u003eJournal of the Academy of Nutrition and Dietetics\\u003c/em\\u003e \\u003cstrong\\u003e123\\u003c/strong\\u003e, 1280\\u0026ndash;1288 (2023).\\u003c/li\\u003e\\n \\u003cli\\u003eHoddinott, J. \\u0026amp; Yohannes, Y. Dietary Diversity as a Household Food Security Indicator. \\u003cem\\u003eFood and Nutrition Technical Assistance Project\\u003c/em\\u003e (1825).\\u003c/li\\u003e\\n \\u003cli\\u003eCavicchia, P. P. \\u003cem\\u003eet al.\\u003c/em\\u003e A New Dietary Inflammatory Index Predicts Interval Changes in Serum High-Sensitivity C-Reactive Protein1\\u0026ndash;3. \\u003cem\\u003eThe Journal of Nutrition\\u003c/em\\u003e \\u003cstrong\\u003e139\\u003c/strong\\u003e, 2365\\u0026ndash;2372 (2009).\\u003c/li\\u003e\\n \\u003cli\\u003eKant, A. K. Dietary patterns and health outcomes. \\u003cem\\u003eJournal of the American Dietetic Association\\u003c/em\\u003e \\u003cstrong\\u003e104\\u003c/strong\\u003e, 615\\u0026ndash;635 (2004).\\u003c/li\\u003e\\n \\u003cli\\u003eChinn, S. A simple method for converting an odds ratio to effect size for use in meta-analysis. \\u003cem\\u003eStatist. Med.\\u003c/em\\u003e \\u003cstrong\\u003e19\\u003c/strong\\u003e, 3127\\u0026ndash;3131 (2000).\\u003c/li\\u003e\\n \\u003cli\\u003eHedges, L. V. Distribution Theory for Glass\\u0026rsquo;s Estimator of Effect size and Related Estimators. \\u003cem\\u003eJournal of Educational Statistics\\u003c/em\\u003e \\u003cstrong\\u003e6\\u003c/strong\\u003e, 107\\u0026ndash;128 (1981).\\u003c/li\\u003e\\n \\u003cli\\u003eBorenstein, M., Hedges, L. V., Higgins, J. \\u0026amp; Rothstein, H. R. \\u003cem\\u003eIntroduction to Meta-Analysis\\u003c/em\\u003e. (Wiley, Hoboken, NJ Chichester, 2021).\\u003c/li\\u003e\\n \\u003cli\\u003eAssink, M. \\u0026amp; Wibbelink, C. J. M. Fitting three-level meta-analytic models in R: A step-by-step tutorial. \\u003cem\\u003eTQMP\\u003c/em\\u003e \\u003cstrong\\u003e12\\u003c/strong\\u003e, 154\\u0026ndash;174 (2016).\\u003c/li\\u003e\\n \\u003cli\\u003eHedges, L. V., Tipton, E. \\u0026amp; Johnson, M. C. Robust variance estimation in meta‐regression with dependent effect size estimates. \\u003cem\\u003eResearch Synthesis Methods\\u003c/em\\u003e \\u003cstrong\\u003e1\\u003c/strong\\u003e, 39\\u0026ndash;65 (2010).\\u003c/li\\u003e\\n \\u003cli\\u003eViechtbauer, W. Bias and Efficiency of Meta-Analytic Variance Estimators in the Random-Effects Model. \\u003cem\\u003eJournal of Educational and Behavioral Statistics\\u003c/em\\u003e \\u003cstrong\\u003e30\\u003c/strong\\u003e, 261\\u0026ndash;293 (2005).\\u003c/li\\u003e\\n \\u003cli\\u003eCinar, O., Umbanhowar, J., Hoeksema, J. D. \\u0026amp; Viechtbauer, W. Using information‐theoretic approaches for model selection in meta‐analysis. \\u003cem\\u003eResearch Synthesis Methods\\u003c/em\\u003e \\u003cstrong\\u003e12\\u003c/strong\\u003e, 537\\u0026ndash;556 (2021).\\u003c/li\\u003e\\n \\u003cli\\u003eHarrer, M., Cuijpers, P., A, F. T. \\u0026amp; Ebert, D. D. \\u003cem\\u003eDoing Meta-Analysis With R: A Hands-On Guide\\u003c/em\\u003e. (Chapman \\u0026amp; Hall/CRC Press, Boca Raton, FL and London, 2021).\\u003c/li\\u003e\\n \\u003cli\\u003eNeyman, J. \\u0026amp; Pearson, E. S. On the Problem of the Most Efficient Tests of Statistical Hypotheses. in \\u003cem\\u003eBreakthroughs in Statistics\\u003c/em\\u003e (eds. Kotz, S. \\u0026amp; Johnson, N. L.) 73\\u0026ndash;108 (Springer New York, New York, NY, 1992). doi:10.1007/978-1-4612-0919-5_6.\\u003c/li\\u003e\\n \\u003cli\\u003eHayden, J. A., Van Der Windt, D. A., Cartwright, J. L., C\\u0026ocirc;t\\u0026eacute;, P. \\u0026amp; Bombardier, C. Assessing Bias in Studies of Prognostic Factors. \\u003cem\\u003eAnn Intern Med\\u003c/em\\u003e \\u003cstrong\\u003e158\\u003c/strong\\u003e, 280\\u0026ndash;286 (2013).\\u003c/li\\u003e\\n \\u003cli\\u003eKelly, S. E. \\u003cem\\u003eet al.\\u003c/em\\u003e A scoping review shows that no single existing risk of bias assessment tool considers all sources of bias for cross-sectional studies. \\u003cem\\u003eJournal of Clinical Epidemiology\\u003c/em\\u003e \\u003cstrong\\u003e172\\u003c/strong\\u003e, 111408 (2024).\\u003c/li\\u003e\\n \\u003cli\\u003eWorld Bank. World Bank Country and Lending Groups \\u0026ndash; World Bank Data Help Desk. \\u003cem\\u003eWorld Bank\\u003c/em\\u003e https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups (2024).\\u003c/li\\u003e\\n\\u003c/ol\\u003e\"},{\"header\":\"Tables\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eTable 1 | Distribution of healthy diets measures and mental health diagnostic tools.\\u003c/strong\\u003e Dietary measures were grouped into 5 groups. Mental health measures are presented per outcome, but each tool can measure more than one outcome.\\u003c/p\\u003e\\n\\u003ctable border=\\\"1\\\" cellspacing=\\\"0\\\" cellpadding=\\\"0\\\" width=\\\"587\\\"\\u003e\\n \\u003ctbody\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eGroups\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eDietary measurements\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eNumber of studies\\u003csup\\u003e1\\u003c/sup\\u003e\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eGroup 1: Adherence and adequacy (G1)\\u003csup\\u003e2\\u003c/sup\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e- Adherence to diet recommendations\\u003c/p\\u003e\\n \\u003cp\\u003e- Nutrient adequacy\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eGroup 2: Dietary patterns reducing nutrition-related chronic diseases (G2)\\u003csup\\u003e\\u0026nbsp;2\\u003c/sup\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e- Dietary Approaches to Stop Hypertension (DASH)\\u003c/p\\u003e\\n \\u003cp\\u003e- Dietary Inflammatory Index\\u003c/p\\u003e\\n \\u003cp\\u003e- Mediterranean diet\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e22\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eGroups 3: Diet diversity indices (G3)\\u003csup\\u003e\\u0026nbsp;2\\u003c/sup\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e- Dietary diversity (MDD, IDDS)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003cp\\u003e- Any other Dietary Variety Scores\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e15\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eGroup 4: Diet quality indices (G4)\\u003csup\\u003e\\u0026nbsp;2\\u003c/sup\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e- Diet Quality Indexes (all)\\u003c/p\\u003e\\n \\u003cp\\u003e- Global Diet Quality Score\\u003c/p\\u003e\\n \\u003cp\\u003e- Global Dietary Index\\u003c/p\\u003e\\n \\u003cp\\u003e- Healthy Eating Indexes (all)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e19\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eGroup 5: Factor analysis and others (G5)\\u003csup\\u003e\\u0026nbsp;3\\u003c/sup\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eAll comparisons between a healthy and an unhealthy diet. Unhealthy is a group with a processed, western, modern, traditional, or unhealthy diet determined for each study.\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e26\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eMental Health Outcome\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003e(number of studies\\u003csup\\u003e4\\u003c/sup\\u003e)\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eValidated diagnostic tool\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e\\u003cstrong\\u003eNumber of studies\\u003csup\\u003e4\\u003c/sup\\u003e\\u003c/strong\\u003e\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"14\\\" valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eDepression (69)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eBeck Depression Inventory (BDI)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e7\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eCenter for Epidemiological Studies - Depression scale (CES-D)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e7\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eClinical/diagnostic interview (CIDI - SF)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eDepression, Anxiety and Stress Scale (DASS)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e14\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eEdinburgh Postpartum Depression Scale (EPDS)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e7\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eGeriatric Depression Scale (GDS)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eDepression subscale of the Hospital Anxiety and Depression Scale (HADS-D)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003e6-item Kutcher Adolescent Depression Scale (KADS-6)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eMini International Neuropsychiatric Interview (MINI)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eMultidimensional Sub-health Questionnaire of Adolescents (MSQA)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003ePatient Health Questionnaire (PHQ-9)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e11\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003ePrimary Care Evaluation of Mental Disorders (PRIME-MD)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eSelf-Rating Depression Scale (SDS)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eZung self-rating scale\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e2\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"9\\\" valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eAnxiety (43)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eCoronavirus Anxiety Scale (CAS)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eDepression, Anxiety and Stress Scale (DASS)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e15\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eGeneral Anxiety Disorder Scale (GAD)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e10\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eHospital Anxiety and Depression Scale (HADS-A)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e11\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eMini International Neuropsychiatric Interview (MINI)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eMultidimensional Sub-health Questionnaire of Adolescents (MSQA)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003ePrimary Care Evaluation of Mental Disorders (PRIME-MD)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eZung Self-reported Anxiety Scale (Zung SAS) \\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eState-Trait Anxiety Inventory (STAI)\\u0026nbsp;\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd rowspan=\\\"4\\\" valign=\\\"top\\\" style=\\\"width: 162px;\\\"\\u003e\\n \\u003cp\\u003eStress (26)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eDepression, Anxiety and Stress Scale (DASS)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e14\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eGeneral Health Questionnaire (GHQ)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e1\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003eHospital Anxiety and Depression Scale (HADS)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e8\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 406px;\\\"\\u003e\\n \\u003cp\\u003ePerceived Stress Scale (PSS)\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003ctd valign=\\\"top\\\" style=\\\"width: 19px;\\\"\\u003e\\n \\u003cp\\u003e3\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003ctr\\u003e\\n \\u003ctd colspan=\\\"3\\\" valign=\\\"top\\\" style=\\\"width: 587px;\\\"\\u003e\\n \\u003cp\\u003e1 One study measured diets with both a Group 2 and a Group 4 measurement, so it is double-counted. 2 A priori measure. 3 A posterior measure. 4 Several studies reported 2 or 3 mental health outcomes, so the count exceeds the total number of studies (n = 83).\\u003c/p\\u003e\\n \\u003c/td\\u003e\\n \\u003c/tr\\u003e\\n \\u003c/tbody\\u003e\\n\\u003c/table\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":true,\"hideJournal\":false,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":false,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"nature-portfolio\",\"isNatureJournal\":true,\"hasQc\":false,\"allowDirectSubmit\":false,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"\",\"title\":\"Nature Portfolio\",\"twitterHandle\":\"\",\"acdcEnabled\":false,\"dfaEnabled\":false,\"editorialSystem\":\"ejp\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":false},\"keywords\":\"depression, anxiety, stress, common mental disorders, food intake, dietary patterns, systematic review, multi-level meta-analysis\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-6530671/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-6530671/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"There is growing evidence of the association between poor diet quality and common mental disorders, which together contribute to global health syndemics. However, there is no synthesis quantifying associations, assessing robustness of evidence or identifying gaps in Low- and Middle-Income Countries (LMIC) where these concomitant health burdens are most prevalent. \\r\\n\\r\\nWe used an Evidence and Gap Map of over 3000 systematically selected, peer-reviewed studies linking food security, diets, and nutrition to anxiety, depression, stress and mental wellbeing (2000-2024). From this, we selected studies investigating associations between healthy diet patterns and mental health symptoms measured by validated tools.\\r\\n\\r\\nEighty-three studies from 23 countries met inclusion criteria (depression n=69; anxiety n=43; stress n=26), reporting statistical measures for 633,317 unique individuals pooled from 65 LMIC sample populations. Healthy diets were associated with less depression, anxiety, and stress. The Standardized Mean Differences (SMD), expressing effect size in standard deviation units, were -0.29 for depression (95% CI -0.35 to -0.23), -0.25 for anxiety (95% CI -0.35 to -0.16), and -0.24 for stress (95% CI -0.33 to -0.14). Results remained robust when restricted to low Risk of Bias studies: depression (SMD = -0.23, CI: -0.31 to -0.16; n=266,831), anxiety (SMD = -0.19, CI: -0.30 to -0.09; n=116,248), and stress (SMD = -0.22, CI: -0.33 to -0.11; n=12,338). Findings were consistent in direction and magnitude across study designs, dietary measurements, diagnostic tools, and country income levels. We found that mental health is better in individuals with healthy diets in LMIC. Methodological limitations (e.g., cross-sectional design) and few studies from low-income countries created evidence gaps. Low-income settings experience disproportionate health vulnerabilities; thus, building on the relationship between diet and mental health can inform actions to improve both.\",\"manuscriptTitle\":\"Meta-analysis of 633,317 individuals shows associations between healthy diets and mental health in 23 low- and middle-income countries\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2025-04-30 06:36:51\",\"doi\":\"10.21203/rs.3.rs-6530671/v1\",\"editorialEvents\":[],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"nature-health\",\"isNatureJournal\":true,\"hasQc\":false,\"allowDirectSubmit\":false,\"externalIdentity\":\"nathealth\",\"sideBox\":\"Learn more about [Nature Health](https://www.nature.com/naturehealth/)\",\"snPcode\":\"44360\",\"submissionUrl\":\"https://mts-nathealth.nature.com\",\"title\":\"Nature Health\",\"twitterHandle\":\"\",\"acdcEnabled\":true,\"dfaEnabled\":true,\"editorialSystem\":\"ejp\",\"reportingPortfolio\":\"Nature\",\"inReviewEnabled\":true,\"inReviewRevisionsEnabled\":false}}],\"origin\":\"\",\"ownerIdentity\":\"55e1ee84-fa7b-4d23-b7e8-b11464c76178\",\"owner\":[],\"postedDate\":\"April 30th, 2025\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"under-review\",\"subjectAreas\":[{\"id\":47802941,\"name\":\"Health sciences/Diseases/Psychiatric disorders/Depression\"},{\"id\":47802942,\"name\":\"Health sciences/Diseases/Nutrition disorders/Malnutrition\"}],\"tags\":[],\"updatedAt\":\"2025-07-29T16:26:20+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2025-04-30 06:36:51\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-6530671\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-6530671\",\"identity\":\"rs-6530671\",\"version\":[\"v1\"]},\"buildId\":\"8U1c8b4HqxoKbykW_rLl7\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}