Diet Quality According to Mental Status and Associated Factors during Adulthood in Spain

Common mental disorders (CMD) are characterized by non-psychotic depressive symptoms, anxiety and somatic complaints, which affect the performance of daily activities. This study aimed to analyze prevalence of diet quality among adults with and without CMD from 2006 to 2017, to study the frequency of food consumption and diet quality according to mental status and age, and to determine which sociodemographic, lifestyle and health-related factors are associated with poor/moderate diet quality, according to mental status. A nationwide cross-sectional study was performed in adults with (n = 12,545) and without CMD (n = 48,079). The data were obtained from three Spanish National Health Surveys (2006, 2011/2012 and 2017). Two logistic regression analyses were used to identify factors associated with diet quality in people with and without CMD. Among those with CMD, the probability of having poor/moderate diet quality was significantly lower for overweight or obese people and those who took part in leisure-time physical activity. Among those without CMD, university graduates were less likely to have a poor/moderate diet quality. Good diet quality was observed more in older adults (≥65 years old) than in emerging (18–24 years old) or young adults (25–44 years old), regardless of mental status.


Introduction
Mental disorders constitute a rising public health concern throughout the world and are responsible for major social and economic issues affecting all age groups [1,2]. In the wide spectrum of mental illnesses, the most prevalent conditions are depressive and anxiety disorders, defined by the World Health Organization as common mental disorders (CMD) [3]. Although CMD are situated among the top 25 causes of the global burden of disease [4], they are in most cases, preventable and treatable [5][6][7]. Nevertheless, reducing the prevalence of CMD continues to pose a major challenge for health systems worldwide [8].
Many studies have examined the association between diet and depression, with only a limited exploration of anxiety and more severe mental illnesses. However, it is now clear that observational research nutritional psychiatry needs to be extended into this area [9][10][11]. Moreover, dietary intervention studies could not only provide far-reaching guidelines for the prevention and treatment of CMD in the future, but they could also further our current understanding of the associations between diet quality and CMD from an epidemiological viewpoint [12].
The influence of diet and nutrition on mental health and wellbeing is an emerging research avenue [13], and a poor diet quality has recently been revealed as a risk factor for CMD [14,15]. Over recent years, many studies have investigated the role of diet in the development of depression due to its influence on inflammatory pathways [16,17].
as the first-stage units, family dwellings as the second-stage units and surveyable people present at home the third. SNHS collected data via in-home interviews which included specific questions on sociodemographic characteristics, mental health status and diet quality. Interviews were conducted by a suitably approved interviewer and supplemented in some cases by a follow-up telephone interview. Further details on the SNHS methodology can be obtained from the National Institute of Statistics [49][50][51].
From the SNHS database, we selected people aged ≥ 18 years old. From the initial 72

Diet Quality
The SHEI tool was used to study diet quality as a dependent variable [44]. The SHEI is an instrument designed to measure how well diets meet the recommendations of the SSCN dietary guidelines [45].
The SHEI questionnaire consists of ten food representative groups from the dietary guidelines, as follows: (a) daily consumption: (1) bread/grains, (2) vegetables (i.e., leafy vegetables, salads) (3) fruit (excluding juices) and (4) dairy products (yoghurt, cheese, milk). (b) Weekly consumption: (5) meat (lamb, pork, chicken) and (6) legumes. (c) Occasional food consumption: (7) cold meats and cuts, (8) sweets (pastries, cereals with sugar, biscuits, jams) and (9) soft drinks with sugar, and the last was (10) the variety of the diet, according to the SSCN recommendations for a healthy diet. The ten food items were worded in exactly the same way in the three SNHS. These items were divided into five response options, in the following order according to the frequency of food consumption: (i) never or hardly ever, (ii) <1 a week, (iii) 1-2 a week, (iv) ≥3 times a week, but not daily, (iv) and daily.
Each of the ten food groups was scored from 0 to 10 points (Supplementary Table  S1), with the highest score in a food group denoting maximum compliance with the SSCN recommendations [45].
The overall score of the SHEI results was based exclusively on the frequency of food consumption rather than the quantity consumed and was calculated from the sum of the frequency of consumption of the ten representative food groups and ranged from 0 to 100 points: the 'poor diet quality' category corresponds with the lowest scores (SHEI score < 51 points), the 'moderate diet quality' category corresponds with the middle scores (SHEI score 51-80 points) and 'good diet quality' category corresponds with the highest scores (SHEI score > 80 points) [44].
Questions were answered in a 4-point Likert-type response format, from 0 points ('more than usual') to 3 points ('much less than usual') and were scored on a bimodal response scale (0-0-1-1), according to the original GHQ method [50]. The total score there-fore ranges from 0 to 12 points. A cut-off of ≥3 points was chosen to estimate the pro-portion of participants without CMD (<3 points) and with CMD (≥3 points) [55].

Sociodemographic Characteristics
Sociodemographic factors were analyzed as independent variables and were gathered by asking subjects the following questions: 'What is your gender?' (female, male), 'What is your marital status?' (never-married, married, widowed and separated/divorced), 'What is your educational level?' (without formal education, completed primary studies, completed secondary studies or professional training, and completed university studies), 'What is your nationality?' (Spanish, foreign), 'What is the size of your town of your residence?' (<10,000 inhabitants and ≥10,000 inhabitants), 'How old are you?' (the age was divided into four major life stages encompassing most of the human lifespan: emerging adults (18-24 years old) as proposed by Arnett [56], young adults (25-44 years old), middle-aged adults (45-64 years old) and older adults (≥65 years), as proposed by Erikson [57]) and 'What is/was the occupation, profession or trade you perform or performed in your last job?' (occupation was classified into the occupation social class categories reported by the Spanish Society of Epidemiology [58] as Classes I and II, Classes III and IV, Classes V and VI).

Health-Related Variables
Information on health-related variables was collected through the following questions: 'Can you tell me if you currently smoke?' (yes, no), 'Have you consumed any alcoholic drinks in the last twelve months?' (yes, no). In addition, self-perceived health status was assessed with the following question: 'What is your health status?', to which the participants could answer: very good, good, fair, poor or very poor.

Lifestyle Variables
The SNHS measured lifestyle variables on the basis of the following questions: 'Do you take part in any physical activity in your work or your main activity?' (yes, no) and 'Do you engage in any physical activity during your leisure time?' (yes, no).

Ethical Aspects
Data from the three SNHS (2006, 2011/2012 and 2017) are available to the public and are stored in anonymized microdata [46][47][48]; as a result, no special permits were required for their use. According to Spanish law, the approval of an Ethics Committee was not required. This data is shown in the Supplementary File.

Statistical Analysis
Statistical methods used in data analysis were descriptive statistics, inferential statistics and regression technique. Descriptive statistics were presented with a count (n/%). We performed the Kolmogorov-Smirnov normality test before any comparison of quantitative variables between the groups. Inferential statistics used were the Chi-Square test, or Fisher's exact test if the number of expected frequencies was >5 between qualitative variables; Student's t-test to compare the means between two independent groups as a parametric test, and the Mann-Whitney U test as a non-parametric test. The kind of regression technique used was logistic regression. In particular, two binary logistic regressions in subjects with and without CMD were performed to identify the factors related to diet quality in each group. For the purpose of the analysis, the diet quality was recoded as a categorical variable into 'good diet quality' (if the SHEI score was above than 80 points) and 'poor diet quality/moderate diet' (if the SHEI score was below or equal to 80 points). The significant variables (p < 0.05) obtained in each univariate analysis were modelled in binary logistic regressions. The goodness of fit was corroborated with the Hosmer-Lemeshow test. The Wald statistic was used as a contrasting statistic. In addition, the crude and adjusted odds ratios (OR) were calculated with 95% confidence intervals. The presence of confounding and interaction were examined. The hypothesis tests were two-tailed and statistical significance was fixed at an alpha error of below 5%. The weighting coefficients included in the SNHS were applied in all the analyses to ensure representativeness. For the statistical analysis, the IBM SPSS Statistics version 25 program (IBM Corp, Armonk, NY, USA), licensed to the University of Seville (Spain) was used.

Characteristics of Participants
A total of 60,624 adults (48,079 people without CMD and 12,545 people with CMD) participated in the present study. Females were much more likely to be represented among those without CMD (75.12%) than those with CMD (24.88%). Participants who were active during their leisure time were much more likely to be represented among those without CMD (83.04%) than those with CMD (16.96%). In contrast, adults who perceived 'poor' their health status were much more likely to be represented among those with CMD (56.52%) than those without CMD (43.48%). Other sociodemographic, lifestyle and health-related factors are shown in Table 1.    As regards weekly consumption, 1-2 weekly consumption of meat was lower among individuals without CMD (27.22% vs. 31.42% p < 0.001), while for legumes, it was higher among individuals without CMD (67.01% vs. 57.66% p < 0.001).

Frequency of Food Consumption and Diet Quality According to Mental Status
As for occasional food consumption, the consumption 'never or hardly ever' of cold meats and cuts, sweets, soft drinks with sugar consumption was lower among participants without CMD (13.13% vs. 19.03% p < 0.001; 17.01% vs. 20.93% p < 0.001; 48.92% vs. 55.36% p < 0.001, respectively).
The relationships between overall diet quality score, food group scores and presence/absence of CMD are shown in Table 3. People without CMD meet more food-based dietary guidelines of the SSCN in relation to bread/grains, fruit, dairy products and legumes consumptions in comparison with people with CMD. In contrast, people with CMD meet more food-base dietary guidelines of the SSCN in relation to cold meats and cuts, sweets and soft drinks with sugar consumption in comparison with people without CMD. In total, the overall diet quality was better in people with CMD.

Comparison of the Frequency of Food Consumption and Diet Quality across Age Groups within Individuals with CMD
Older adults with CMD had a higher prevalence of daily consumption of bread/grains, fruit and dairy products compared to other age groups. Nonetheless, vegetable consumption was higher among middle-aged adults than other age groups ( Figure 1A).
In relation to weekly consumption, older adults with CMD had a significantly higher 1-2 weekly consumption of meat than other age groups. Meanwhile, the consumption of legumes 1-2 a week was more frequent in middle-aged adults with CMD compared to other age groups ( Figure 1B).
As far as occasional food consumption is concerned, the prevalence of older adults with CMD who never or hardly ever consumed cold meats and cuts, sweets and soft drinks with sugar was higher than other age groups ( Figure 1C).

Comparison of the Frequency of Food Consumption and Diet Quality across Age Groups within Individuals without CMD
Among participants without CMD, older adults met the existing dietary guidelines in terms of the daily consumption of bread/grains, vegetables, fruit and dairy products compared with the other age groups (Figure 2A).
As regards weekly consumption, the prevalence of 1-2 weekly consumption units of meat was higher among older adults without CMD. Nevertheless, middle-aged adults had a significantly higher 1-2 weekly consumption of legumes ( Figure 2B). As for diet quality, moderate diet quality was higher in emerging adults without CMD (emerging adults: 78.62%, young adults: 75.07%, middle-aged adults: 61.29%, older adults: 48.42%; p < 0.001) than other age groups, as was poor diet quality (emerging adults: 10.38%, young adults: 3.67%, middle-aged adults: 0.98%, older adults: 0.30%; p < 0.001). In contrast, good diet quality was more prevalent in older adults without CMD (emerging adults: 11.00%, young adults: 21.26%, middle-aged adults: 37.73%, older adults: 51.28%; p < 0.001). In relation to occasional food consumption, the prevalence of older adults without CMD who never or hardly ever reported a daily consumption of cold meats and cuts, sweets and soft drinks with sugar was higher in comparison with other age groups ( Figure 2C).

Comparison of the Frequency of Food Consumption and Diet Quality across Age Group in Individuals with and without CMD
As can be seen in Table 4, a higher frequency of daily consumption of bread/grains was observed among emerging and young adults without CMD in comparison with those with CMD. Moreover, a higher frequency of daily consumption of bread/grains was observed among middle-aged and older adults with CMD compared to those without CMD.  As regards weekly consumption, the consumption of meat was higher among middleaged and older individuals with CMD in comparison with those without CMD. Similarly, 1-2 weekly consumption of legumes was higher among emerging and young adults without CMD compared to those with CMD.
As for occasional food consumption, the consumption 'never or hardly ever' of soft drinks with sugar consumption was higher among emerging and young adults without CMD in comparison with those with CMD. In contrast, the consumption 'never or hardly ever' of soft drinks with sugar was more prevalent in older adults with CMD compared to those without CMD.
In relation to diet quality, the prevalence of good diet quality among older adults with CMD was higher than those without CMD. On the other hand, the prevalence of poor diet quality among emerging adults without CMD was higher than those with CMD.

Association between Sociodemographic Characteristics, Lifestyle Behavior, Health-Related Variables and Diet Quality According to Mental Status
Among participants with CMD (Table 5), the adjusted logistic regression model indicated that the likelihood of having a poor or moderate diet quality was greater in individuals without formal education (OR = 1.24, 95% CI 1.09-1.40) and primary education (OR = 1.17, 95% CI 1.06-1.29). In contrast, lower rates of poor or moderate diet quality were linked to overweight (OR = 0.83, 95% CI 0.76-0.91) and obese (OR = 0.84, 95% CI 0.76-0.94) status, and those who took part in leisure-time physical activity (OR = 0.78, 95% CI 0.72-0.84).  The adjusted logistic regression model in Table 6 shows that the probability of having a poor or moderate diet quality was lower in people with university studies (OR = 0.90, 95% CI 0.84-0.95), and those belonging to Social Classes I and II (OR = 0.89, 95% CI 0.84-0.94). In contrast, higher rates of poor or moderate diet quality were linked to Social Classes V and VI (OR = 1.08, 95% CI 1.03-1.13), underweight status (OR = 1.47, 95% CI 1.23-1.77), people who had consumed alcohol (OR = 1.04, 95% CI 1.01-1.09), and those who took part in no leisure-time physical activity (OR = 1.47, 95% CI 1.40-1.53). Furthermore, the probability of having a poor or moderate diet quality was greater when the perceived health was fair (OR = 0.93, 95% CI 0.88-0.98) or very poor (OR = 0.80, 95% CI 0.64-0.99).  1 OR = odds ratio; Ora = odds ratio adjusted for all sociodemographic, lifestyle and health-related factors; CI 95% = confidence interval; n = number of people with a poor or moderate diet quality; Hosmer-Lemeshow test χ 2 = 19.53, p = 0.12; Nagelkerke's R 2 : 0.14; p-value < 0.001.

Main Findings
The present study is unique in that it shows the relationship between a large number of characteristics and diet quality in a large sample of adults with and without CMD living in Spain, from a survey conducted in three waves (2006, 2011/2012 and 2017).
In our study, the prevalence of diet quality in need of improvement among people with and without CMD was higher in 2017 than in 2011/12. This may be due to the fact that in recent years, food habits and consumption in Spain are moving away from the traditional Mediterranean diet towards an increasingly "westernized" diet [60]. The Spanish diet is becoming saltier and sweeter due to the incorporation of more highly processed foods and changes in dietary habits. Specifically, the report on food consumption in Spain 2017 [61], conducted by the Ministry of Agriculture, Fisheries and Food shows that the prevalence of processed food consumption increased by 2.2% from 2011/12 to 2017. In addition, since 2015, there has been a reduction in the number of households with children and in the average family size. According to a previous study [62], people living alone are less likely to follow a varied diet and have lower fruit, vegetable, and fish consumption than those living with others.
There is a broad consensus in the scientific literature regarding the association between the adherence to a poor diet quality and the presence of CMD [63][64][65][66]. The lack of energy or enthusiasm for preparing or enjoying food, as well as appetite loss, may influence diet quality among people with CMD symptoms [41,67]. Furthermore, emerging evidence suggests that diet may influence the onset of mood disorders and specifically depression. For instance, recent systematic reviews have demonstrated associations between measurements of diet quality and the probability and risk of depression [15,63]. Thus, diet may impact on in mental health via several pathways, including those related to oxidative stress, inflammation, and mitochondrial dysfunction, which are disrupted in people with mental disorders [68], and unhealthy diets contain certain compounds that may negatively affect these pathways. For example, elements commonly found in processed foods such as saturated fatty acids, artificial sweeteners, and emulsifiers may alter the gut microbiome and activate inflammatory pathways [43], which is associated with a significantly higher incidence of depressive symptoms, even among those without diagnosed mental disorders [17,20,69].
Our findings show that the prevalence of a moderate diet quality was higher in individuals without CMD. However, previous studies have pointed out that a history of CMD may stimulate these subjects to improve their diet quality in the long term [65,67].
Emerging adulthood (individuals aged 18 to 24 years old) is characterized by a continuing process of self-definition and increasing autonomy and it is a time when major role transitions take place [56,70,71]. Emerging adults tend to have a poor quality diet, often accompanied by an excessive intake of soft drinks with sugar and fast food [72]. Similarly, young adulthood is a period when healthy patterns are established, such as a good quality diet, which will carry over into later adulthood.
In our analysis, both emerging and young adults obtained poor diet quality index scores, regardless of their mental status. It is well-known that young people have an excessive intake of soft drinks with sugar, cold meats and cuts and sweets, while their intake of vegetables and fruit is lower in comparison with older age groups [73][74][75][76][77]. In fact, our findings showed that 'never or hardly ever' consuming soft drinks, sweets and cold meats and cuts was higher as age increased among people with CMD in comparison with those without CMD, except in older adults. Several possible mechanisms linking food with sugar intake and anxiety/depression symptoms are assumed, including inflammation markers [78]. Previous studies reported other influences on an unhealthy diet such as lack of motivation, time constraints and cost [79][80][81]. Meanwhile, a few intervention studies have correlated improvements in mental health status with an increase in fruit and vegetable intake [82][83][84][85]. For example, Mujcic and Oswald [86] showed that fruit and vegetables consumption increased happiness, wellbeing and life satisfaction. Moreover, middle-aged and especially older adults had a better diet quality in comparison with the two first life stages, although other studies in these advanced age groups with depressive symptoms found a poor diet quality [87,88], perhaps influenced by social factors, such as marital status and social contacts [89]. Our results show that a higher frequency of daily consumption of bread/grains, vegetables or fruit was observed among middle-aged and older adults with CMD in comparison with those without CMD. A higher intake of dietary fiber helps the nervous system, which has a beneficial influence on mental health [90]. Other food groups, such as legumes, contain B vitamins, magnesium, folic acid and potassium, among others, which could act as protection against the risk of psychological disorders [91]. Our findings show that 1-2 weekly consumption of legumes was higher among emerging and young adults without CMD compared to those with CMD, and among middle-aged and older adults with CMD in comparison with those without CMD. A randomized clinical trial suggests that a dietary rich in legume and nuts had beneficial effects on depression [92]. As regards the 1-2 weekly consumption of meat, it was higher among middle-aged and older adults with CMD, in line with another study [93].
Educational level may be one of most important social factors explaining differences in food habits [94][95][96]. Here, it was found that the probability of having a poor or moderate diet quality was greater in adults with CMD with a lower educational level compared with those without CMD with higher education. This is not the first time that a higher educational level has been associated with good dietary habits [97][98][99], although some studies have failed to find any association [100] or even found an inverse link [101,102], probably due to people with a higher educational level having access to better knowledge about food [103,104]. In addition, some studies have shown that a favorable attitude towards a healthy diet was associated with a higher educational level [105][106][107]. Our results contrast with another study, which found that university students whose diet was in need of improvement or had a poor diet had twice or triple the risk of CMD, respectively, than those who had a healthy diet [108].
Several studies have found an association between poor or moderate diet quality; the latter indicates a medium adherence to the recommendations proposed by the Spanish Society of Community Nutrition [45] and high BMI [109][110][111]. Nevertheless, our results have shown that overweight and obese were protective factors for poor or moderate diet quality in adults with CMD. Recently, it has also been shown that obese and overweight adults are motivated to lose weight by taking up a healthier lifestyle, such as healthy eating or doing physical activity [112]. Moreover, the sensory pleasure derived from eating palatable meals could reduce stress and produce a positive mood [113]. In contrast, being underweight was a risk factor for poor or moderate diet quality in adults without CMD, in line with other studies [114][115][116], probably due to a reduced intake of nutritious food [117].
As regards lifestyle behavior, the multivariate analysis showed a link between lower rates of poor or moderate diet quality and those who engaged in physical activity during their leisure time among adults with CMD. Moreover, not engaging in physical activity was a risk factor for poor or moderate diet quality in adults without CMD. It is well-documented that leisure-time physical activity has considerable mental health benefits throughout a patient's lifespan [118][119][120][121]. In fact, engaging in leisure-time physical activity is negatively associated with CMD symptoms [122]. In addition, several studies have found many cases of individuals who had a good diet quality, and took part in healthy behaviors associated with engaging in physical activity and smoking less [123][124][125].

Strengths and Limitations
This study has certain limitations. First, the present study is a cross-sectional design, so it is currently not possible to assign causality. Second, diet quality was based exclusively on the frequency of food consumption and not on the quantity of consumption; this does not allow us to judge if the p-values given for the small percentages obtained for CMD versus non-CMD are clinically significant despite the fact that the sample size is very large. Third, mental status was constructed from self-reported questionnaires, and the responses could therefore have been biased by social desirability and/or memory. Fourth, there was a lack of data on institutionalized populations, since the personal interviews were conducted at the participants' homes. Fifth, the BMI data were taken from the participants' own responses about their height and weight in the self-reported surveys. Finally, the SNHS surveys were carried out with different samples, which means that the different characteristics of the participants could have affected our results.
On the other hand, one strength of the present study is that data from the SNHS surveys have been obtained using carefully planned methodology, including sampling, well-designed forms, preparation of the survey participants, supervision of the survey and filtering of the data, all of which guarantee a representative sample of the population and enable comparisons to be made. Moreover, the validity of the Health Eating Index (HEI) has been demonstrated with plasma biomarkers [126,127], where a higher score in the SHEI was associated with blood concentrations of certain markers with a protective effect against particular diseases. Finally, this study included people aged ≥ 65 years old, and the total sample was therefore representative of all adults living in Spain.

Implications for Research and Practice
The findings of this work offer useful insights for further studies. Regardless of mental status, a high percentage of people had a moderate diet quality, which suggests that more effort needs to be made to improve healthy dietary habits. In addition, our findings could help to identify which age groups had a poor and/or moderate diet quality. In this case, it would be desirable for public health policies to promote healthy eating habits, especially among young people [128,129]. In particular, it is crucial to target favorable opinions of family, friends and health professionals in relation to the consumption of fruit and vegetables [130]. One way could be by promoting short cooking programs to improve young people's skills and confidence in certain areas of food literacy, especially those related to the consumption of fruit and vegetables [131,132]. Although middle-aged adults and older individuals had higher diet quality index scores in comparison with other age groups, further intervention studies are required to promote healthy diet quality in these age groups, considering the psychological and social factors [133]. In addition, the results of the multivariate analysis may be used as a guide to improve Spanish public health policies and health promotion guidelines about diet quality [1]. Lastly, it would be of great interest to carry out further longitudinal studies to evaluate the impact of CMD on diet quality in the stages of life highlighted in this study.

Conclusions
Regardless of mental status, a better diet quality could be observed in older adults than emerging and young adults. In this regard, older adults with and without CMD meet the existing dietary guidelines in terms of daily consumption of bread/grains, fruit and dairy products. Nevertheless, a higher frequency of daily consumption of bread/grains is observed among middle-aged and older adults with CMD compared to those without CMD. In contrast, the consumption 'never or hardly ever' of soft drinks, sweets and cold meats and cuts increases with age among people with CMD in comparison to those without CMD, except in older adults. Among people with CMD, not having primary studies is considered a risk factor for having a poor or moderate diet quality. In contrast, in adults, overweight or obese and doing leisure-time physical activity are considered protective factors against having a poor or moderate diet quality. Among individuals without CMD, the likelihood of a having poor or moderate diet quality decreases in those who have university studies; however, it increases in those who are underweight or do not do any physical activity in their leisure time.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/nu13051727/s1, Table S1: criteria to define the score for each item of the Spanish Health Eating Index (SHEI), File S1: research data.
Funding: This research received no specific grant from any funding agency.
Institutional Review Board Statement: Ethical review and approval were waived for this study, due to according with Spanish legislation, when secondary data are used, there is no need for approval from an ethics committee.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: The data are available as Supplementary Material (File S1: research data).