Examining Nutrition Knowledge, Skills, and Eating Behaviours in People with Severe Mental Illness: A Cross-Sectional Comparison among Psychiatric Inpatients, Outpatients, and Healthy Adults

Compared to the general population, people with severe mental illness (SMI) have an increased risk of weight gain and metabolic syndrome, but also of malnutrition, in part due to unhealthy lifestyle behaviours. The aim of this cross-sectional study was to identify barriers to healthy eating, including nutrition knowledge and skills in people with SMI. For this purpose, we compared the means of anthropometric data such as body mass index, waist-to-hip ratio, and interview data on nutrition knowledge and skills, health-related variables, eating behaviours, personality, motivation, and attitudes in 65 inpatients and 67 outpatients of the Psychiatric Hospital of the University of Zurich and 64 healthy adults using ANOVA and chi-squared tests. The results showed that patients with SMI had worse nutritional status and lifestyle compared to the healthy controls, including disordered (e.g., night eating) and unhealthy (e.g., high intake of sugary foods) eating habits. However, levels of nutrition knowledge, cooking and food skills, and motivation to eat healthily were not significantly lower in the psychiatric patients than in the healthy adults and were not associated with weight change. Based on our findings, nutritional support for people with SMI is urgently needed and should include not only educational but also behavioural and long-term approaches.


Introduction
Compared with the general population, people with severe mental illness (SMI), including schizophrenia, major depression, and bipolar disorder [1], often have poorer physical health. They are at higher risk of weight gain and metabolic syndrome, which is associated with an increased risk of diabetes, cancer, and coronary heart disease, as well as a reduced life expectancy [2][3][4][5]. In addition to the side effects of psychiatric medications, such as increased appetite [6][7][8][9], lifestyle factors such as physical inactivity and dietary habits play a crucial role in the pathogenesis of metabolic syndrome in people with SMI [10][11][12]. In general, people with SMI have a poorer diet (higher calorie intake; more processed foods with higher salt and sugar content; less fruit, vegetables, and fibre) compared to the general population [12][13][14][15]. They also have a higher risk of malnutrition with deficiencies in vitamins, minerals, and micronutrients due to inadequate diet [16,17].
People with SMI face several barriers that can lead to unfavourable eating habits and a reduced diet quality [14]. Some of these barriers, such as daily habits, financial constraints, reduced motivation to eat healthily, price, taste, low social support, and the unavailability eating disorder. In addition, the healthy controls were required to have no mental disorders. For the psychiatric samples, clinical staff screened all eligible individuals with psychotic or depressive disorders (classified as F2 or F3 according to the ICD-10 diagnostic tool) upon admission to one of the acute wards of the PUK. At the beginning of their treatment, n = 161 inpatients (response rate = 40.4%) and n = 94 outpatients (response rate = 71.3%) were consecutively asked to participate in the study. In the case of agreement and the provision of written informed consent, the participant's contact details were given to the research team. The recruitment process of the psychiatric patients was based on stratified sampling to minimise selection bias and has been fully described and illustrated in a recent publication [17]. For the group of healthy controls, a convenient sample was recruited through various information channels in Zurich, such as flyers in family practice centres, shops, and city swimming pools; university mailing lists; and word-of-mouth advertising among mental health professionals and acquaintances. We used stratified sampling for age and gender for all groups [29,30]. Each interview lasted an average of one hour and was conducted by a specially trained research assistant with background in psychology or medicine. The psychiatric patients were interviewed an average of 14 days after admission. Participants were offered a drink and they were allowed to take breaks according to their needs, but they received no financial compensation. All data were entered into the LimeSurvey tool [31] using case identification codes.

Measures
In addition to nutrition knowledge and skills, a selection of socio-demographic variables, health-related variables, and variables related to the participants' eating behaviours were included in this study. An overview of all study variables can be found in reference [29].

Personal Variables
The participants' gender, age, education, nationality, living situation, source of income, and financial situation for food were assessed (see Table 1). Education (high school diploma and above = 1, other = 0) and financial situation ([much] too little = 0, rest = 1) were dichotomised as control variables.

Health Status and Behaviours
Mental health was assessed using the nine-item Symptom Checklist (SCL-K-9) and the nine-item Patient Health Questionnaire (PHQ-9).
The SCL-K-9 is a reliable, efficient, and validated scale for assessing psychopathological symptomatology which asks participants to rate their mental and physical health problems of the past seven days on a five-point Likert-type scale [32]. Higher scores indicate more severe symptoms. The Cronbach's alpha in the present study was 0.88 (n = 196).
The PHQ-9 is a validated and reliable screening tool for assessing the severity of depression in the preceding two weeks [33,34]. Higher scores indicate more depressive symptoms (10-14 = mild depression, 15-19 = moderate depression, 20-27 = severe depression). The Cronbach's alpha in the present study was 0.89 (n = 196).
To quantify the relevance of psychiatric medication on potential weight gain, a sum score of the prescribed psychiatric medications was calculated, with each individual score derived from the clinical information system (higher scores = negative effect on weight gain). Details of the score calculation procedure have been described elsewhere [17].
Participants were asked whether they had received any inpatient or outpatient treatment because of poor physical or mental health in the previous three to six months. Participants were also asked about their smoking status and to report the number of physically active days in a typical week (at least 30 min of moderate-to-vigorous physical activity).
To assess body mass index (BMI; kg/m 2 ) and waist-to-hip ratio (WHR; abdominal girth/hip girth), weight (kg), height (m), abdominal girth (cm), and hip girth (cm) were measured using a digital scale (Soehnle 63850 PWD Style Sense Compact 100) and a SECA measuring tape. Each measurement was repeated twice, and the mean was calculated.
Participants wore light clothes and no shoes. In addition, participants were asked whether they had lost or gained weight (kg) in the last three to six months and to subjectively rate their nutritional status (supply of energy and nutrients) on a scale ranging from 1 = very poor to 10 = very good. Furthermore, participants were also asked whether they suffered from nutrition-related diseases such as obesity, diabetes, coeliac disease, food allergies, and (subjective) food intolerances, as well as whether they had ever received nutrition counselling. The validated short form of the Big Five Inventory (BFI-K) was used to assess five personality factors of the participants by rating 21 statements on a Likert-type scale from 1 = strongly disagree to 5 = strongly agree [35]. The Cronbach's alphas (n = 196) were as follows: 0.82 for Extraversion, 0.56 for Agreeableness, 0.70 for Conscientiousness, 0.81 for Neuroticism, and 0.74 for Openness to experience.

Dietary Behaviours and Attitudes
As there are currently no specific dietary assessment tools available for mental health settings [15], we used a selection of simple food frequency questions that have been validated for the general Swiss population and have been shown to be indicators of healthy or unhealthy eating behaviours [36,37]. Participants were asked how often they eat/drink vegetables and salad, fruit, juice, sweets, and sugary drinks (daily, 4-6 times/week, 1-3 times/week, 1-3 times/month, or rarely) and how many portions they usually eat (1-6 handfuls). To compare the intake of sugary foods among the groups, we added up the portions of juice, sweets, and sugary drinks consumed per week.
Participants were also asked to report their usual types of daily meals (e.g., breakfast, morning snack, lunch, afternoon snack, dinner, other snacks between meals), whether they followed a special diet (e.g., gluten-free, lactose-free, low in sugar, high in protein), and whether they intentionally restricted their food intake to regulate weight or health.
Selected items from the Social Functioning Scale [38], specifically developed and validated for use in people with SMI, were used to assess participants' eating habits. Thereby, participants were asked to indicate, on a scale from 1 = never to 4 = often, how often they performed a particular activity, such as planning or buying food, storing food, preparing meals (food-related activities), eating a healthy and balanced diet, eating prepared meals (healthy eating), eating too little, eating too much (overeating), cooking together with friends, and eating together with friends (social eating).
Participants' feelings about nutrition were assessed using the technique of spontaneous associations with the words "nutrition" and "diet" [39,40]. In the first step, the participants were asked to name the first words or images that came to mind when they thought of the word "nutrition" or "diet". In the second step, the participants rated their associations on a scale ranging from 1 (very negative) to 7 (very positive). We then calculated the mean of the ratings. In addition, the participants rated the importance of a healthy and balanced diet to them on a scale from 1 to 10, with higher scores indicating greater importance.

Nutrition Knowledge and Skills
Levels of nutrition knowledge (objective measure) as well as cooking and food skills (subjective measures) were assessed using the sum scores of two different questionnaires.
Nutrition knowledge was assessed using the "practical nutrition knowledge about balanced meals scale" (PKB-7 scale) which has been validated for the Swiss population and used in several countries [37,41,42]. The scale consists of seven multiple-choice items (with only one correct answer) of different levels of difficulty. It aims to measure knowledge relevant to making healthy food choices.
Cooking and food skills were assessed using the Cooking and Food Skills questionnaire [43], previously validated for the Swiss population [44]. The questionnaire includes ratings of confidence in 14 cooking tasks, such as "preparing and cooking vegetables", and confidence in 19 food tasks, such as "planning how much food to buy". Confidence ratings were made on a scale from 1 = very poor to 7 = very good and 0 = never/rarely done. The Cronbach's alpha was 0.89 for the cooking scale and 0.86 for the food scale (n = 196).

Statistical Analyses
A sample size calculation suggested the inclusion of at least 192 participants (n = 64 per group); for details, see the previously published study protocol [29]. Sociodemographic variables are presented as frequencies by group in Table 1. Chi-squared tests or ANOVA were performed for the variables of participants' health status and behaviours, dietary behaviours, attitudes, and nutrition knowledge and skills among the inpatients, outpatients, and healthy adults. Where appropriate, post-hoc tests were performed using the Bonferroni correction. In addition, differences in participants' dietary behaviours, attitudes, and nutrition knowledge and skills were controlled for using ANCOVA with education and financial situation as control variables. Relationships between variables were assessed using Spearman correlation coefficients. In addition, an exploratory extreme-group analysis was performed between psychiatric patients who had experienced weight gain and those who had experienced weight loss using independent samples t-tests. The significance level was set at 0.05 for all analyses.

Characteristics of Participants
A total of 196 individuals took part in the interviews. Table 1 shows the sociodemographic characteristics of the participants, differentiated according to the study groups. The mean age of the study participants was 38.7 (SD = 11.9) years. The group of healthy adults were more highly educated and had better financial situations than the groups of psychiatric patients.

Health Status and Behaviours
The participants' health status and behaviours differed significantly between psychiatric patients and healthy controls (see Tables 2 and 3). By definition, the patients had more psychiatric symptoms, had used more medications, and had used medical treatments more frequently in the past few months. There were also significant differences in four of the five personality factors. The psychiatric patients had a higher BMI and WHR and suffered more from nutrition-related diseases and food intolerances compared to the healthy controls. In addition, most patients with SMI had experienced weight changes in the previous three to six months, whereas the healthy controls had more stable weights. In addition to poorer nutritional status, psychiatric patients were less likely to engage in healthy behaviours such as physical activity and not smoking. Note. M = mean, SD = standard deviation. For continuous data, ANOVA was calculated; difference in letters (a,b,c) indicate significant Bonferroni-adjusted post-hoc tests, p = 0.05. SCL-K-9: higher scores = more psychiatric symptoms; PHQ-D: higher scores = more depressive symptoms; sum score of prescribed psychiatric medication: higher scores = negative effect on weight gain; personality factors: higher scores = higher expression; subjective evaluation of nutritional status: 1 = very bad, 10 = very good. Table 4 shows the differences in dietary behaviours, attitudes, and nutrition knowledge and skills among the three groups. The psychiatric patients ate more sugary foods and tended to skip main meals when eating late snacks or at night. Higher intake of sugary foods was positively associated with unhealthy behaviours such as late snacking (rs = 0.24, p = 0.006, n = 132), night eating (rs = 0.17, p = 0.046, n = 132), and overeating (rs = 0.19, p = 0.033, n = 132); moreover, this behaviour was negatively associated with healthy behaviours such as eating an individualised diet (rs = −0.23, p = 0.008, n = 132) and a healthy diet (rs = −0.18, p = 0.038, n = 132).

Dietary Behaviours, Attitudes, and Nutrition Knowledge and Skills
The psychiatric patients also performed fewer nutrition-related activities and ate with others less often than the healthy controls. They also had more negative feelings about nutrition, which correlated with the SCL-K-9 scores (rs = −0.32, p < 0.001, n = 132), the PHQ-9 scores (rs = −0.34, p < 0.001, n = 132), and social eating (rs = 0.20, p = 0.020, n = 132).
Although the psychiatric patients had a higher BMI and more nutrition-related diseases and food intolerances (especially outpatients), the proportion of people who ate an individualised diet based on their needs or restricted calories did not differ among the three groups. In contrast to the unhealthier eating behaviours, motivation for a healthy and balanced diet as well as nutrition knowledge and skills were not significantly lower in the psychiatric patients compared to the healthy controls. More precisely, levels of nutrition knowledge and skills were highest in the outpatient group. In addition, participants living in institutions or who were homeless (n = 7) reported lower levels of cooking skills (Md = 29.0) and food skills (Md = 54) than those living independently; however, they had similar levels of nutrition knowledge (Md = 5.0).

Differences between Psychiatric Patients with Weight Gain and Loss
In an exploratory extreme-group analysis, we compared study variables between psychiatric patients who had experienced weight gain (n = 57) in the previous three to six months and those who had experienced weight loss (n = 48) during this time to explore differences in health behaviours and attitudes. Sociodemographic variables (e.g., gender, age, education), illness-related variables (e.g., diagnosis, mental state, medication), motivation for healthy eating, and nutrition knowledge and skills were not associated with weight gain or loss. Participants with weight gain had a BMI = 28.7, and those with weight loss had a BMI = 24.7 (t [103] = 4.18, p < 0.001), while there was no significant difference in WHR. Weight gain was associated with less physical activity ( Figure 1 as percentages of the scale scores (100% = highest possible scale score). cording to their needs, compared to 33% of those with weight gain (X [1] = 7.70, p = 0.006). Weight loss was also associated with higher levels of the following items:  Figure 1 as percentages of the scale scores (100% = highest possible scale score).

Discussion
Patients with SMI are known to be at risk for nutritional problems, mostly associated with increased weight but also malnutrition, leading to reduced life expectancy and quality of life [2][3][4][5]14,16,17]. In the present study, the levels of nutrition knowledge and skills for healthy eating were investigated in individuals with SMI compared to healthy controls, along with the participants physical and mental health status and eating behaviours. The results showed that patients with SMI had poorer nutritional status and lifestyle compared to healthy adults. In particular, a substantial proportion of the patients reported disordered and unhealthy eating habits, reduced activities related to food preparation and eating, and negative feelings about nutrition. Despite various nutritional problems, less

Discussion
Patients with SMI are known to be at risk for nutritional problems, mostly associated with increased weight but also malnutrition, leading to reduced life expectancy and quality of life [2][3][4][5]14,16,17]. In the present study, the levels of nutrition knowledge and skills for healthy eating were investigated in individuals with SMI compared to healthy controls, along with the participants' physical and mental health status and eating behaviours. The results showed that patients with SMI had poorer nutritional status and lifestyle compared to healthy adults. In particular, a substantial proportion of the patients reported disordered and unhealthy eating habits, reduced activities related to food preparation and eating, and negative feelings about nutrition. Despite various nutritional problems, less than half of the participants with SMI followed an individualised diet. However, levels of nutrition knowledge and skills and motivation to eat healthily were not lower in the psychiatric patients than in the healthy controls.
Some previous studies have discussed a lack of nutrition knowledge, cooking and food skills, and motivation to follow a healthy diet as possible causes of unhealthy eating behaviours in people with SMI [14,[19][20][21][22]. However, to the best of our knowledge, our study is the first to systematically investigate levels of nutrition knowledge and skills in people with SMI. Contrary to previous assumptions, and despite the higher levels of education of the healthy controls (and likely greater interest in nutrition due to convenience sampling), nutrition knowledge was not higher in the healthy controls. In addition to objectively assessed practical nutrition knowledge, subjective levels of cooking and food skills were not lower in the psychiatric patients than in the healthy controls. More specifically, the inpatients had similar scores to the healthy controls, while the outpatients had the highest scores. One reason for this may be that half of the psychiatric patients (49%) had already received dietary advice. There may be two reasons for the differences between the inpatients and outpatients. First, because of their higher BMI and poorer nutritional status, the outpatients may have been more interested in gaining knowledge about a healthy diet. Second, some of the inpatients who lived in residential care homes (including one homeless person) were unable to prepare food by themselves and therefore had very low cooking and food skills. In addition to nutrition knowledge and skills, motivation to eat healthily was not significantly lower in the psychiatric patients than in the healthy controls. Moreover, levels of knowledge, skills, and motivation were not related to weight gain or weight loss in the psychiatric patients.
Nevertheless, the psychiatric patients seemed to have difficulty translating their knowledge into action. In line with previous findings [10][11][12]24], they had overall unhealthier lifestyles compared to the healthy controls, such as higher smoking rates, lower physical activity levels, and unhealthy and disordered eating habits, such as higher consumption of sugary foods, skipping main meals, night eating, less social eating, and generally less nutrition-related activities such as planning meals or shopping for food. Patients with SMI often have disturbed circadian rhythms and a lack of daily structure (e.g., unemployment). Both diagnoses included in the study (psychosis and depression) can be associated with disordered sleep patterns [45]. The combination of SMI, sleeplessness, and social deprivation may lead to night eating and reduced social eating and activities during the day [46]. This is associated with higher calorie intake, problematic aspects of body metabolism, and a higher risk of metabolic syndrome [47]. Furthermore, social aspects of eating are associated with feelings of inclusion and psychological well-being and may lead to preferable food choices and healthier eating patterns [48,49]. Indeed, in an implicit association task, the patients expressed more negative feelings towards nutrition than the healthy controls.
Aside from lifestyle factors, increased appetite and weight gain in patients with SMI are often discussed as being caused by side effects of psychiatric medication [23]. For instance, second-generation antipsychotics and some antidepressants are known to lead to increased appetite. In our study, most patients had medications with a known risk of metabolic side effects. The inpatients with 62% affective diagnoses (F3) had a higher sum score, which entails a higher risk for weight gain than outpatients with 87% affective diagnoses [17]. However, despite a higher medication-induced risk for weight gain in the inpatients and the need to prevent metabolic side effects of psychiatric medication [6,7,9], BMI was significantly higher in the outpatients, which indicates the importance of lifestyle factors. Similarly, the results of the exploratory extreme-group analysis showed positive associations between an unhealthy lifestyle including low physical activity, unhealthy eating patterns (high sugar as well as low vegetable and fruit consumption), and weight gain. In contrast, patients who reported weight loss during the past three to six months were more physically active, followed a healthy and individualised diet, and were characterised by more conscientiousness, extraversion, and openness. Some personality traits seem to facilitate or hinder the adoption of a healthy lifestyle in patients with SMI, and strategies to increase levels of conscientiousness have been discussed to achieve desirable behaviours and outcomes in psychotherapy [50]. However, their specific role in eating behaviours must be explored in future studies.
A strength of this study is the use of a broad range of different objective (e.g., BMI, nutrition knowledge) and subjective measures (e.g., weight change, cooking and food skills) related to nutrition being applied in patients with SMI. Thus, we can assume that biases based on subjective assessments and social desirability were not substantial. As well, we used a complete stratified sampling frame for the recruitment of the patients, meaning that each patient meeting the inclusion criteria had the chance to participate during the recruitment period, thus minimising selection biases. However, this was not the case for the convenience sample of the healthy controls. The patient groups and healthy controls differed in terms of their education and socioeconomic status. Nevertheless, these gaps have been frequently described [51] and might also represent the naturalistic setting. Another limitation of this study is the use of a cross-sectional design, which was appropriate for our research aims but did not allow us to examine changes and causal relationships.

Conclusions
In summary, the results of this study showed that the participants with SMI had worse mental and physical health conditions compared to the healthy adults. The patients with SMI were frequently affected by nutrition-related diseases such as overweight and diabetes and more often had disordered eating habits (e.g., night eating) and unhealthy lifestyle and eating patterns (e.g., low physical activity, smoking, high intake of sugary foods). Thereby, nutrition knowledge, cooking and food skills, as well as motivation for a healthy diet did not seem to be relevant barriers for these behaviours, although they might be important prerequisites for healthy eating. Rather, the results indicated that people with SMI would benefit from nutritional support that aims at improving their daily structure and social inclusion, for instance via behavioural approaches related to meal planning and social eating. Therefore, the development of low-threshold, long-term support services are needed in addition to individual nutrition counselling. Based on the study's results, there might also be different needs for support services in inpatient and outpatient settings as well as in terms of living situation (e.g., living alone or in a residential care home). In addition, the results indicate a need for more knowledge about how food intolerances and personality factors affect the eating behaviours of people with SMI. Funding: This research was partially funded by the Foundation for the Encouragement of Nutrition Research in Switzerland (SFEFS; grant number: 551_Rev). The SFEFS was not involved in the chosen study design, data collection, analysis, or interpretation of the data, nor in the preparation, review, and approval of the manuscript.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki and approved by the Swiss Ethics Committee, Switzerland (2019-01485).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The dataset generated for the current study is available from the first author (S.M.) upon written request.