Dietary Patterns of Persons with Chronic Conditions in a Multi-Ethnic Asian Population

Background: Chronic conditions are a leading cause of death and disability worldwide and respective data on dietary patterns remain scant. The present study aimed to investigate dietary patterns and identify sociodemographic factors associated with Dietary Approach to Stop Hypertension (DASH) scores within a multi-ethnic population with various chronic conditions. Method: The present study utilised data from the Knowledge, Attitudes, and Practices on diabetes study in Singapore – a nationwide survey conducted to track the knowledge, attitudes, and practices pertaining to diabetes. The study analysed data collected from a sample of 2,895 Singapore residents, with information from the sociodemographic section, DASH diet screener, and the modied version of the World Mental Health Composite International Diagnostic Interview (CIDI) version 3.0 checklist of chronic medical conditions. Results: Respondents with no chronic condition had a mean DASH score of 18.5 (±4.6), those with one chronic condition had a mean DASH score of 19.2 (±4.8), and those with two or more chronic conditions had a mean DASH score of 19.8 (±5.2). Overall, the older age groups [35– 49 years (B = 1.78, 95% CI: 1.23 – 2.33, p <0.001), 50–64 years (B = 2.86, 95% CI: 22.24 – 3.47, p <0.001) and 65 years and above (B = 3.45, 95% CI: 2.73 – 4.17, p <0.001)], Indians (B = 2.54, 95% CI: 2.09 – 2.98, p <0.001) reported better diet quality, while males (B = -1.50, 95% CI: -1.87 – -1.14, p <0.001) reported poorer diet quality versus females. Conclusion: Overall, respondents with two more reported quality diet while the sociodemographic gender

The etiology of chronic conditions is complex and multifactorial. Risk factors include age, family history, genetic predisposition, current and lifetime weight and physical activity, smoking, alcohol, and diet (12)(13).
Of these risk factors, dietary choices may be the most amenable to modi cation and with the greatest public health impact (14). Of note, So et al. (2010) highlighted that individuals reporting a greater degree of adherence to healthier dietary intake such as the Mediterranean diet showed a signi cant protection against the development of chronic conditions; with approximately 6-13% reduction in deaths and/or incidence of neurodegenerative disease, cardiovascular disease and cancer. Evidently, such a dietary pattern is recognised as a major contributing factor towards the prevention, development and treatment of chronic conditions (15).
Correspondingly, a growing and evolving body of scienti c inquiry on dietary intake with health and chronic conditions has led to recommendations emphasizing a variety of plant-based foods (e.g., vegetables, fruits, legumes, whole grains, nuts, and seeds) and deemphasizing processed food consumption with added sugar or a diet that is rich in meat content (14,16).
Studies in Asian populations have identi ed various dietary patterns labelled "traditional," "meat," "Western," and "prudent". Dietary patterns observed in China characterised by a high intake of meat and dairy products have been associated with obesity (17)(18). In Thailand, a more traditional carbohydrate-rich pattern was associated with metabolic syndrome (19). In other populations such as Japan and Korea, a more traditional dietary pattern was inversely associated with risk factors such as high blood pressure (20)(21). In Pakistan where 33% of the adult population suffers from hypertension, the high intake of sh, prawns and yoghurt was found to be inversely associated with hypertension (22). Within the Singapore Chinese population, a "fruit-vegetable-soy" pattern was inversely associated, and a "meat-dim-sum" pattern was directly associated with cardiovascular disease mortality (23). Evidently, the association between various dietary patterns and chronic conditions has been established within singular ethnic populations across the considerable corpus of research in this area. Yet, much less is known regarding the association between healthy dietary patterns and chronic conditions across a multi-ethnic population.
Singapore is a multi-ethnic city-state situated in Southeast Asia with a population of approximately 5.6 million of which 4.1 million are Singapore residents (Singapore citizens or permanent residents) (24). The population is largely comprised of inhabitants from 3 major Asian ethnic groups: Chinese (76.0%), Malay (15.0%) and Indian (7.5%) (25). Based on the results from the Singapore Mental Health Study, it was found that a total of 25.4% individuals reported having one chronic condition, and 16.3% had MCC (26). Given the relatively substantial gures of individuals living with chronic conditions, it underscores the need to address the paucity of dietary data in this population. Additionally, a study in this setting provides a unique opportunity to elucidate the dietary patterns of a multi-ethnic population, the results of which can be extrapolated to other countries with a similar ethnic composition.
Given the diverse ethnic composition in Singapore, it follows that a culturally relevant diet screener is required for use in this multi-ethnic population.  developed a reliable and validated short diet screener designed to assess the intake of selected food groups that is representative of the overall dietary patterns across a multi-ethnic Asian population. Speci cally, the short 37-item diet screener assesses the intake of selected food groups representative of a multi-ethnic Asian population via a priori dietary quality indices such as the Dietary Approaches to Stop Hypertension (DASH). In the extant literature, DASH has been demonstrated to be the most sensitive diet score to examine associations between diet and various health-related outcomes (28). Adopting DASH dietary patterns has several bene ts including but not limited to lowered mortality from cardiovascular diseases and diabetes (28-29), lowered blood pressure (30), decreased body weight and waist circumference in obesity related weight management (31).
Taken together, the aims of the present study were to 1) characterise and compare the dietary patterns of a multi-ethnic population between persons with no chronic condition, one chronic condition and MCC through scoring their dietary intake according to the DASH score, and 2) identify socio-demographic correlates of DASH scores amongst persons with no chronic condition, one chronic condition and MCC.

Participants and procedures
The data for this research comes from a population based, cross-sectional study aimed at evaluating the Knowledge, Practice and Attitudes towards Diabetes Mellitus (DM) amongst residents of Singapore aged 18 years and above. A more detailed methodology of the study can be found in an earlier paper (32). The sample was randomly selected via a disproportionate strati ed sampling design according to ethnicity (Chinese, Malay, Indian, Others) and age groups (18-34, 35-49, 50-64, 65 and above) from a national population registry database of all citizens and permanent residents within Singapore. The study oversampled certain minority populations, such as Malay and Indian ethnicity, as well as those above 65 years of age, in order to ensure su cient sample size and to improve the reliability of the parameter estimates for these subgroups.
Citizens and permanent residents who were randomly selected were sent noti cation letters followed by home visits by a trained interviewer from a survey research company to obtain their informed consent to participate in the study. Face-to-face interviews with those who were agreeable to participate were conducted in their preferred language (English, Mandarin, Malay, or Tamil). Responses were captured using computer assisted personal interviewing. Individuals who were unable to be contacted due to incomplete or incorrect addresses, were living outside of the country, or were incapable of attending the interview due to severe physical or mental conditions, language barriers, or were institutionalized or hospitalized at the time of the survey were excluded from the study. Written informed consent was obtained from all respondents prior to the survey, and for those aged 18 to 20 years, parental consent was sought as the o cial age of majority in Singapore is 21 years and above.

Diet screener
The diet screener comprises a list of 30 food/beverage items, that respondents' rate on a 10-point scale ranging from 'never/rarely' to '6 or more times per day', the frequency at which they consumed a particular food/beverage within the last one year ). The diet screener was interviewer-administered.
Standard serving sizes were indicated for each food/beverage item to facilitate this process. Intake frequencies were standardised to a number of servings per day for each food/beverage item. DASH scores were calculated to account for seven intake components: fruit, vegetables, nuts/legumes, whole grains, red and processed meat, low fat dairy, and sweetened beverages. For each of these seven components, participants received a score between 1 and 5 corresponding to the quintile of the intake they fall in, with reverse scoring utilized for meat and sweetened beverages, and these seven quintile scores were summed to form the overall DASH score.

Chronic physical conditions
A modi ed version of the World Mental Health Composite International Diagnostic Interview (CIDI) version 3.0 (34) checklist of chronic medical conditions was used and the respondents were asked to report any of the conditions listed in the checklist. The question was read as, "I am going to read to you a list of health problems some people have. Has a doctor ever told you that you have any of the following chronic medical conditions?" This was followed by a list of 18 chronic physical conditions (such as asthma, high blood sugar, hypertension, arthritis, cancer, neurological condition, Parkinson's disease, stroke, congestive heart failure, heart disease, back problems, stomach ulcer, chronic in amed bowel, thyroid disease, kidney failure, migraine headaches, chronic lung disease, and hyperlipidaemia) which were considered to be prevalent among Singapore's population. If the participant gave a positive response for any of the conditions listed, they were then asked, "How old were you when you were diagnosed with the medical condition?" and, "Did you receive any treatment for it at any time during the past 12 months?".

Statistical analysis
Analyses in the present study were conducted with Stata version 15. In order to ensure representativeness of the data to the general population, survey weights were used to account for complex survey design. Means and standard deviations are provided for continuous variables, while frequencies and percentages are presented for categorical variables. In order to examine the variables associated with the total DASH score, four linear regressions were conducted (within the total sample, and the three subgroups of no chronic condition, one chronic condition and multimorbidity) with the following predictor variables: age, gender, ethnicity, education, marital status, employment status, personal income, BMI and chronic conditions. Statistical signi cance was set at the conventional alpha level of p < 0.05, using two-tailed tests.

Results
Socio-demographics distribution of the sample Table 1 summarised the socio-demographic characteristics for the sample of 2,895 respondents. 46.2% had no chronic medical condition, 26.3% had one chronic medical condition, and 27.2% had two or more chronic medical conditions. Chinese respondents made up 75.8% of the sample distribution, Malays 12.7%, Indians 8.6%, and others 2.9%. 51.6% of the respondents were female, and BMI scores indicated that 53.4% of the respondents were in the normal range based on WHO BMI classi cation.

DASH components score distribution and comparisons between chronic medical condition groups
The means and standard deviations of each of the seven DASH components (fruit, vegetables, nuts/legumes, low fat dairy, whole grains, red and processed meat, and sweetened beverages) and the overall DASH scores of the total sample, those with no chronic medical condition, one chronic medical condition and two or more chronic medical conditions are displayed in Table 2.
Socio-demographic correlates of DASH score Table 3 shows the socio-demographic correlates associated with DASH scores within the full sample and across the three chronic medical condition groups. Within the full sample, the older age group [35-

Discussion
The current study found that 42.9% of the population had no chronic medical/physical condition, 26.3% had one chronic medical/physical condition, and 30.5% had MCC. Overall, people with MCC had demonstrated better dietary practices in terms of number of servings taken per day for the DASH components. In the context of Singapore, this could be attributed to various initiatives and support received from primary care providers. Firstly, the focus upon nudging and facilitating healthier dietary choices is seen across the multitude of nationwide health promotion campaigns. One such example relates to the "Healthier Dining Programme" launched in 2014 which provides incentives to restaurants offering 500-calorie meals. Essentially, consumers are "nudged" towards choosing such healthier meal options that are identi ed with a "Healthier Choice Symbol" on menus in these restaurants (35) 2016), this is consistent with the notion that younger individuals demonstrate certain dietary habits that reduce overall diet quality. Food and beverages with high saturated fat, sugar and sodium contents such as those purchased at quick service restaurants feature prominently in the younger population's diet across many countries such as USA, UK and Australia (39).
Gender differences in terms of dietary patterns was also demonstrated to be consistent with prior studies; where women tended to report better diet quality in comparison to men. For example, in Montreal, women's diets were closer to recommendations for vegetables, fruits, and sodium intake as compared to men (41). As demonstrated in prior literature, women generally reported being more invested in relation to food-related matters and having better knowledge in terms of food and nutrition (42)(43). Additionally, women reported consuming higher intakes of fruits and vegetables, dietary bre, and lower intakes of fat and salt (42,44). Taken together, it follows that greater importance attributed by women to their diet correspondingly translates into better dietary practices.
Across the sample population, ethnicity was also identi ed to be signi cantly associated with diet quality.
Among the major ethnic groups in Singapore, Indians reported having healthier diet based on respective DASH scores. This is consistent with the healthy dietary pattern as outlined in the National Nutrition Survey conducted in 2010 (45), with Indians consuming the most bread and breakfast cereals, vegetable dishes, fruit, milk and dairy products and fewer eggs, poultry and meat dishes.
Diet and nutrition represent important factors in both promotion and maintenance of good health throughout the entire life course. We identi ed several important factors associated with diet quality among persons with no chronic conditions and one chronic condition. Within these subgroups, education level was signi cantly associated with diet quality. Speci cally, those with less than a degree reported having poorer diet quality. Similar ndings regarding educational level have been reported elsewhere; adults with a college diploma in USA demonstrated having a better overall diet quality as compared to all other education levels (46). Therein, it has been posited that education might be associated not only with increased nutritional knowledge, but could also be an indicator of ability to translate such nutritional knowledge into better dietary practices throughout the person's lifetime (46). Lastly, it was interesting to observe that BMI was not a factor associated with diet quality in the present study. In current literature, it has been evidenced that individuals with chronic conditions generally reported higher-than-normal BMI (47). Accordingly, multiple studies have also demonstrated that healthier dietary patterns are associated with lower BMI (48-49). Nonetheless, it should be noted that the use of self-reported chronic conditions may plausibly have resulted in reporting bias -underestimating the true prevalence of chronic conditions and therein, an underestimation of the strength of the association between BMI and diet quality.
The present study is a nationwide survey conducted in four different languages (English, Chinese, Malay and Tamil) to address potential language barriers for participation in a multi-ethnic population. Additionally, the use of a large sample size, randomised design and survey weighted analysis improves the overall reliability of the present results. Nonetheless, several limitations of the present study warrant comment and conclusions drawn should be considered in light of these limitations. Firstly, given that the present study adopts a cross-sectional design, we are not able to establish any causal relationship between dietary pattern and chronic conditions. Secondly, the diet screener utilised is fundamentally based on the past-year selfreported diet recall of the respondent, with no correlation to any blood or urinary parameters. In that regard, we are not able to rule out the likelihood of recall bias in this self-reported format.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing Interests
The authors declare that they have no competing interests.  Tables   Table 1 Socio-demographic distribution of the sample (n = 2895)