View Article PDF

Cardiovascular disease (CVD) is the second most common cause of death after cancer in New Zealand (NZ).1 The prevalence of CVD increases dramatically with age; at 45 years of age only 8% of New Zealanders are diagnosed with CVD compared to 75% of those aged 75 years.2New Zealand has an aging population which will increase the burden of CVD.3 Among the risk factors for CVD, some are partially modifiable, including high blood pressure, high blood cholesterol, overweight and poor glycaemic control.4 Modifiable risk factors are known to be influenced, in part, by dietary intake.4 Consumption of particular heart healthy foods has been shown to improve modifiable risk factors.5,6A heart healthy diet typically includes high intakes of fruit, vegetables, whole grain foods, fish and seafood, lean meat and poultry, low fat dairy products, nuts and seeds and low consumption of high sodium products. Unfortunately, consuming a heart healthy diet and improving one's CVD risk may not be achievable for all sectors of the community particularly for those in lower socioeconomic groups.7 The relationship between lower socioeconomic status (SES) and morbidity and mortality rates from chronic diseases such as cardiovascular disease (CVD) are well established worldwide8 and in New Zealand.9International research has shown that lower socioeconomic status is often associated with higher intakes of dietary fat10–12 and lower intakes of dietary fibre and fruit and vegetables, independently of each other.13,14 Lower socioeconomic status is thought to not only influence the dietary intake of families but, in turn, body weight of children.15 It is not known if these relationships are also present in the NZ population.Socioeconomic status can be measured using a variety of methods, the newest of which is the Economic Living Standard Index short form (ELSISF).16 The ELSI is a measure of individual living standards and has been included in the New Zealand Health Survey since 2006/07 to describe sociodemographics of New Zealanders. The ELSISF has been used to describe dietary intakes of children16 but to our knowledge nutritional intake of adults or body weight status has not been compared using the ELSISF. Nutritional intake of NZ adults has been described according to level of community deprivation (NZDep) but not by individual living standards.7Associations between selected psychosocial factors including standard of living measures, education, knowledge of food and dietary intake, as assessed by investigation of single foods, food groups and nutrients, have been studied internationally11,17–19 and in New Zealand.7,13,15Dietary patterns have been proposed as a solution to investigating relationships between food choice and chronic disease risk as these analyses allow for the multiple factors in the diet to be examined in combination, rather than focusing on intakes of single nutrients or food groups. This is a more powerful approach as foods can have synergistic or inhibitor effects on each other. One such type of analysis is PCA, which has been shown to be successful in producing patterns associated with CVD risk factors in many studies.20 However, most dietary patterns are generated from food frequency questionnaire data, which are not as robust as data from food records.20 This exploratory study uses estimated food records to investigate whether a novel heart healthy dietary pattern could be identified using PCA.A secondary aim of the study was to investigate if such a pattern is associated with a higher level of education and standard of living. Analysis of this type has not been previously conducted using the ELSISF. The aims of this study are in line with the wider aims of the Canterbury Health, Ageing and Life Course (CHALICE) study to build a database of important determinants of health.MethodsStudy design—The data for this pilot study were collected from the first 82 participants who completed baseline assessments in the CHALICE study between August 2010 and February 2011. Data included in this study were collected prior to the Christchurch Earthquake on 22 February 2011.The CHALICE study—CHALICE is a longitudinal, observational study of a random sample of people aged 50 years currently living in the Canterbury District Health Board (CDHB) area.21 CHALICE aims to recruit 1000 participants. Fifty-year olds were randomly recruited from the general and Māori Canterbury electoral rolls from the Kaikoura, Hurunui, Waimakariri, Christchurch city, Selwyn and Ashburton districts. Māori were over-sampled (one in every four participants were selected from the Māori electoral roll). Ethical approval was obtained from the Upper South A Regional Ethics Committee.Data collection—The overall CHALICE methods have been described in detail elsewhere.21 Participants were invited to attend a four to six hour face to face interview at the University of Otago, Christchurch offices where information was collected by trained interviewers on physical and mental health, family and social factors and lifestyle. For this study only body composition, demographic, nutrition related questionnaires and the estimated food record data were included in the analysis. These are described in more detail below.Demographic data collected included ethnicity as assessed by the question used in the NZ census; "Which ethnic group or groups do you belong to?". For the purposes of these analyses results were dichotomised where anyone who selected Māori, even if more than one ethnicity was chosen, were identified as Māori. All other participants were categorised as non-Māori. Participants were asked "What is the total income that your household got from all sources, before tax, or anything that was taken out of it, in the past 12 months" and asked to pick from a range of answers, where the lowest was less than <$5000 and the highest $150,001 or more. For these analyses participant responses were grouped into the following groups: <$40,000, $40,000–$69,999, $70,000–119,999 and $120,000 or over. Highest level of education was classified as either no qualifications, secondary school qualifications, post-secondary (non-degree) qualification, or University degree and for these analyses was dichotomised to Secondary school or less, or Post-secondary qualification.Standard of living was assessed using the ELSISF. This is a 25 item questionnaire that assesses standard of living, based on three components – ownership restriction (e.g. not owning a car), social participation restriction (not doing activities that a participant would like to do because of cost) and economising (e.g. not buying fresh fruit and vegetables due to cost).13 The ELSISF answers were summed to produce a score of between 0 and 31. These scores were then coded into seven categories ranging from "Severe hardship" to "Very good". Due to small participant numbers in these analyses these categories were reduced to three categories: low (score of 0-16), medium (score of 17-24) or high (score of 25-31) standard of living.Weight and height measurements were taken with shoes removed and in light clothing to the nearest 0.1 cm. Weight was measured using an electronic scale and height was measured against a wall using a permanently attached tape measure. Waist circumference was measured against the skin. A tape measure was placed at the mid-point between the lowest rib and the top of the iliac crest. Waist circumference was measured once to the nearest 0.5cm. Body mass index (BMI) was calculated as weight (kg) divided by height (m2).A nutrition literacy score was calculated from 24 questions based on the United Kingdom "The Family Diet Study" questionnaire.22 This questionnaire asked about the fat, sugar, salt and/or fibre content of commonly consumed foods. The foods used in the nutrition literacy questions were modified to make them compatible with the NZ diet e.g. Rice Krispies changed to Rice Bubbles, squash changed to cordial. This questionnaire was pretested in a sample comparable to the CHALICE population before use in this study.23Food records: Participants completed a four day estimated food record in the week after completing the study interview. Participants received a written and verbal explanation of how to complete a detailed record of all food and beverages consumed over three week days and one weekend day. Returned records were checked by a dietitian for completeness. If any information was incomplete, unclear or not detailed enough, the interviewers contacted participants for clarification according to the dietitian's instruction. The food record data were analysed using Diet Cruncher version 1.6.0 (Way Down South Software). For the current analyses mean intakes of total fat, saturated fat, dietary fibre and total energy were obtained and total and saturated fat as a percentage of total energy (% TE) were calculated. Average serves of daily fruit and vegetables were also calculated, to assess whether participants were meeting the NZ Healthy Eating guidelines.24,25 The results of the food record analysis were compared with guidelines for a heart healthy diet. The data from one participant was excluded from analyses because they had a rare health condition controlled by diet and their diet was not representative of the general population.There are many national dietary guidelines available from countries including the United States of America, United Kingdom, Australia and NZ and while some of the individual food and/or nutrients included in these may differ from country to country, the foods and nutrients most consistently included are total fat, saturated fat, fruit, vegetables and fibre.3,4,24,25 Therefore, for the purpose of this study a heart healthy diet was defined as a diet including a healthy intake of all five of these components, based on the current NZ recommendations.24,25 Criteria for this are a total fat contribution of less than 35% of total energy (% TE), saturated fat contribution of than 10% TE, dietary fibre intake of more than 25 g/day, fruit intake of at least two serves per day and vegetable intake of at least three serves (excluding potatoes). Yes/No variables were created to indicate whether each participant met each of the five guidelines.Statistical analyses—Analyses were carried out using R, version 2.13.0 (Vienna, Austria). Median intake and interquartile ranges were calculated for each nutrient and associations with demographic variables and BMI were explored using the Mann Whitney U test and the Kruskal-Wallis test. Associations between standard of living and education were adjusted for income, sex and ethnicity.Heart healthy dietary pattern scores were also generated using PCA. PCA is a commonly used data reduction technique to provide summary measures of diet. As individual foods or food groups are not eaten in isolation, analysis of individual nutrients or food groups does not take into account the complexities of meals eaten, and the possible nutrient interactions occurring when meals are consumed.20Using a PCA-derived pattern means that a greater proportion of the variation within the dietary data can be explained than when using individual food group or nutrient data. The use of PCA means that summary scores can be calculated for each of the five heart healthy diet components entered into the analyses. These scores are based on an equation based on the amount consumed (for nutrients), or the number of servings (for fruit and vegetables), multiplied by the factor loading produced by the PCA.Foods/nutrients with higher factor loadings contribute more to the overall score than those with lower loadings, and in accordance with other similar studies, eigenvalues above 1, the elbow in the scree plot and factor loadings ≤-0.3 or greater than ≥0.3 were considered important when identifying and naming patterns.26Negative factor loadings mean that a particular food or nutrient is contributing negatively to the overall score. A sample size of 62 participants allows for the examination of six variables using PCA, as ten participants per variable are required for robust results.27 Varimax rotation was conducted using z-scores of the five dietary components of a heart healthy diet. As continuous data (not dichotomised data) is needed for PCA data for the amount consumed (for nutrients) or the number of servings (for fruit and vegetables) were used for the PCA. The first principal component was used based on both the elbow in the Scree plot (not shown) and its interpretability based on factor loadings.Multiple linear regression was applied to the first principal component with demographic and nutrition related variables. Non-significant variables that did not contribute to the fit of the model were removed, second order interactions were tested and all but standard of living, education and nutrition literacy were not significant, there was insufficient data in all cells to properly fit higher order interactions. All results were considered statistically significant at the 5% level where reported p values were adjusted for multiple comparisons using Tukey's method. All model assumptions were checked by graphical inspection of residual plots (QQ, Leverage, residual versus fitted values).ResultsOf the 82 CHALICE participants interviewed 62 (75.6%, 32 females and 30 males) returned a completed food record prior to the 22nd of February 2011 and provided information on all variables of interest. Table 1 details the participant characteristics.Table 1. Sample characteristics\r\n \r\n \r\n \r\n Characteristics\r\n \r\n Category\r\n \r\n Number of participants (n=62) (%)\r\n \r\n \r\n \r\n Sex\r\n \r\n Male Female\r\n \r\n 30 (48) 32 (52)\r\n \r\n \r\n \r\n Ethnicity\r\n \r\n Māori Non-Māori\r\n \r\n 8 (13) 54 (87)\r\n \r\n \r\n \r\n Education\r\n \r\n Secondary school or less Post-secondary school\r\n \r\n 21 (34) 41 (66)\r\n \r\n \r\n \r\n Employment (current)\r\n \r\n In paid employment Not in paid employment\r\n \r\n 55 (89) 7 (11)\r\n \r\n \r\n \r\n Household income (per annum)\r\n \r\n <$40,000 $40,000–69,999 $70,000–119,999 ≥$120,000 Don't know\r\n \r\n 12 (19) 13 (21) 19 (31) 15 (24) 3 (5)\r\n \r\n \r\n \r\n Standard of living (ELSISF)\r\n \r\n Low Medium High\r\n \r\n 7 (11) 18 (29) 37 (60)\r\n \r\n \r\n \r\nTable 2 summarises mean body composition of participants. Mean BMI was 28.0±6.0 kg/m2, which is within the WHO overweight range (BMI 25–29.9 kg/m2).4 Twenty-nine percent of participants were classified as normal weight, 40% overweight and 31% obese. Mean waist circumference of participants falls within the range (94–102 cm for men and 80–88 cm for women) which is associated with an increased risk of metabolic complications.Table 2. Body composition of sample\r\n \r\n \r\n \r\n Body composition measures\r\n \r\n Females (n=32) Mean (SD)\r\n \r\n Males (n=30) Mean (SD)\r\n \r\n \r\n \r\n Weight (kg) Height (cm) BMI (kg/m2) Waist circumference (cm)\r\n \r\n 75.2 (18.5) 165.2 (7.0) 27.6 (6.6) 87.6 (15.7)\r\n \r\n 89.5 (17.1) 177.4 (6.5) 28.5 (5.4) 100.0 (13.4)\r\n \r\n \r\n \r\nBMI=body mass index.The median dietary intake of participants is shown in Table 3. Of the five heart healthy dietary components, no participants met all five recommendations, 8% met four recommendations, 24% met three recommendations, 25% met two recommendations, another 25% met one recommendation and 18% met none of the recommendations.Single nutrient analyses—Table 3 also summarises non-parametric and post hoc tests with dietary intake, standard of living and education. The only dietary component which varied significantly with standard of living was total fat intake (P=0.018, P adjusted (for income)=0.024). There was no evidence of a difference between total fat intake for the medium and high ELSI ranges. This result remained unchanged after adjustment for BMI, sex and ethnicity (results not shown).Total fat and vegetable intake did not vary significantly by education. Median saturated fat as a % TE was lower in the higher education group, 12.4% TE compared with 14.8% TE, than in the lower education group (P=0.002). Median intake of both groups was higher than the recommended intake of less of than 10% of TE.Table 3. Median (LQ, UQ) food and nutrient intake compared with heart healthy dietary guidelines and median (LQ, UQ) nutrient and food intake by standard of living category and education\r\n \r\n \r\n \r\n Nutrient\r\n \r\n Total fat (% TE)\r\n \r\n Saturated fat (% TE)\r\n \r\n Dietary fibre (g/day)\r\n \r\n Fruit (portions/day)\r\n \r\n Vegetables (portions/day)\r\n \r\n \r\n \r\n Recommended intake for a heart healthy diet\r\n \r\n <35%\r\n \r\n <10%\r\n \r\n >25g\r\n \r\n >2\r\n \r\n >3\r\n \r\n \r\n \r\n Median (LQ, UQ)\r\n \r\n 33.7 [31.2, 37.0]\r\n \r\n 13.5 [11.5–14.8]\r\n \r\n 22.9 [18.2–27.8]\r\n \r\n 1.8 [0.5–2.5]\r\n \r\n 2.0 [1.3–2.7]\r\n \r\n \r\n \r\n Percentage achieving recommended intake\r\n \r\n 57%\r\n \r\n 11%\r\n \r\n 38%\r\n \r\n 48%\r\n \r\n 14%\r\n \r\n \r\n \r\n Median (LQ, UQ) food and nutrient intake by standard of living\r\n \r\n \r\n \r\n Low\r\n \r\n 40.4 [34.9–41.8]\r\n \r\n 15.0 [14.1–16.4]\r\n \r\n 19.8 [18.9–26.9]\r\n \r\n 0.3 [0.1–1.5]\r\n \r\n 1.5 [1.0–1.6]\r\n \r\n \r\n \r\n Medium\r\n \r\n 32.3 [28.7–36.4]*\r\n \r\n 12.3 [11.2–13.8]\r\n \r\n 24.1 [20.5–31.3]\r\n \r\n 2.0 [0.6–3.0]\r\n \r\n 2.0 [0.5–2.5]\r\n \r\n \r\n \r\n High\r\n \r\n 33.7 [31.1–36.9]*\r\n \r\n 13.2 [11.5–14.6]\r\n \r\n 21.9 [18.0–26.1]\r\n \r\n 1.8 [0.8–2.5]\r\n \r\n 2.0 [1.5–3.0]\r\n \r\n \r\n \r\n P Kruskal-Wallis test\r\n \r\n 0.018\r\n \r\n 0.124\r\n \r\n 0.397\r\n \r\n 0.193\r\n \r\n 0.151\r\n \r\n \r\n \r\n Median (LQ, UQ) food and nutrient intake by education level\r\n \r\n \r\n \r\n Lower education\r\n \r\n 35.8 [32.5–38.2]\r\n \r\n 14.8 [12.3–16.1]\r\n \r\n 19.8 [15.7–25.6]\r\n \r\n 1.0 [0.3–2.0]\r\n \r\n 1.5 [1.0–2.0]\r\n \r\n \r\n \r\n Higher education\r\n \r\n 33.2 [30.5–36.5]*\r\n \r\n 12.4 [11.0–13.8]**\r\n \r\n 23.4 [19.7–29.0]*\r\n \r\n 2.0 [0.5–2.8]*\r\n \r\n 2.0 [1.5–3.0]\r\n \r\n \r\n \r\n P Kruskal-Wallis test\r\n \r\n 0.113\r\n \r\n 0.002\r\n \r\n 0.045\r\n \r\n 0.044\r\n \r\n 0.240\r\n \r\n \r\n \r\nNote: * p<0.05; **p<0.01 compared to low standard of living group or lower education group. % TE=percentage of Total Energy intake, LQ = lower quartile, UQ = upper quartile.Principal components analysis—PCA produced one meaningful pattern (Table 6). The first principal component reflected a "higher CVD risk" dietary pattern: a high saturated fat, and low fruit, vegetable and dietary fibre diet based on factor loadings of 0.42, -0.55, -0.50 and -0.46 respectively. A higher score for an individual for the "higher CVD risk" pattern indicates a higher saturated fat intake and a lower fruit, vegetable and fibre intake, all of which are not recommended as part of a heart healthy diet.Table 6: Principal component analysis loadings and importance\r\n \r\n \r\n \r\n Variables\r\n \r\n "Higher CVD risk" pattern\r\n \r\n \r\n \r\n Total fat Saturated fat Dietary fibre Fruit Vegetables Percentage variation explained\r\n \r\n 0.26 0.42 -0.46 -0.55 -0.50 40%\r\n \r\n \r\n \r\nMultiple linear regression analysis on normalised "higher CVD risk" diet scores showed significantly lower scores for those with a higher education level (0.98, 95% CI [0.27–1.69], P=0.008) while there was evidence for a difference in the impact of nutrition literacy on diet by standard of living (P=0.001). The diet score of the low standard of living group increased by 0.35 (95%CI [0.02–0.67], P=0.039) for each one unit increase in nutrition literacy while the diet score of the high standard of living group decreased by 0.10 (95%CI [0.11–0.79], P=0.011) for each increase in nutrition literacy by one unit. Neither sex nor ethnicity had significant effects at the 0.05% level, nor was there a significant difference between the two higher levels of standard of living, the model explained 28% of the variation.The results show that there were significant interaction effects between nutrition literacy and standard of living on the "higher CVD risk" pattern scores. Those in the medium or high standard of living categories, who had better nutrition literacy, had lower scores for the "higher CVD risk" pattern, indicating consumption of a more heart healthy diet, compared to those of medium or high standard of living who had lower nutrition literacy. Conversely, those in the low standard of living category, who had higher nutrition literacy scores, had higher scores for the "higher CVD risk" pattern, indicating consumption of a less heart healthy diet.Figure 1. Association between nutrition literacy and standard of living interaction with a "higher CVD risk" dietary pattern (multiple linear regression, p=0.001)a,b,c,da All of the data included in the multiple linear regression are plotted in the figure; however the regression lines are only relevant for those participants in the respective lower education group; b Low standard of living β=0.35, 95%CI [0.02, 0.67], p=0.039, c Medium standard of living β=-0.23, 95%CI [-0.97 – -0.17], p=0.006, d High standard of living β = -0.10 – 95%CI [-0.79 – -0.11], p=0.011.DiscussionThe results discussed here are based on preliminary analysis from the first wave of the CHALICE study aiming to recruit 1000 50 year olds from Canterbury over the period 2010–2014. This study describes aspects of the dietary intake and the associations with selected social variables.None of the participants in this study met all five heart healthy guidelines, and the majority of participants met two or less guidelines. These results suggest that, as expected, a higher standard of living, education and nutrition literacy were all associated with consuming a heart healthy diet. Using results of the PCA, those who had a medium or high standard of living and who had higher levels of nutrition literacy had higher scores for a heart healthy diet. However, those with a lower standard of living, and who had higher nutrition literacy had pattern scores that corresponded with a less heart healthy diet.Previous studies have shown associations between socioeconomic status and education with selected foods or nutrients2,7,12,14,22,28,29 but there is limited evidence available from New Zealand populations.30 No research, that we are aware of, has used the ELSISF to describe socioeconomic circumstance or standard of living in relation to adult dietary intake. In this population there was a significant pattern toward decreasing fat (as a % TE) intake as standard of living increased. There are no questions in the ELSISF that assess variables that may influence fat consumption such as frequency of purchasing food from outside of the home or access to affordable healthy foods.Studies from abroad have shown that those living in lower socioeconomic areas may struggle to access affordable, healthy, lower fat foods due to the positioning of supermarkets, fruit and vegetable shops and takeaway outlets30,31 and that eating out more often has been associated with higher fat intake.32 It is not known if this is also the case in NZ. In this population there were no significant patterns with the other components of a heart healthy diet; although this may be due to the wide interquartile ranges of some of the components.In addition to assessing associations between socioeconomic status, education and nutrition literacy and single nutrients this study also investigated associations between social variables and a "higher CVD risk" dietary pattern.Dietary patterns, as with intakes of selected nutrients, have been shown to be less healthy in lower socioeconomic groups2 and deprived sectors of the community.7 Dietary modelling (PCA) showed that standard of living, as well as nutrition literacy and level of education, were inversely associated with a "higher CVD risk" dietary pattern. Interestingly, those with a high standard of living tended to consume a less healthy dietary pattern than those with a medium standard of living, however this did not reach statistical significance (P = 0.28).Conversely there was a significant pattern with the lowest standard of living group towards increasing consumption of saturated fat (as a % TE), and lower consumption of fruit, vegetables and dietary fibre as nutrition literacy increased.The relationship between nutrition literacy and dietary intake is often complex. Previous research that has explored this relationship has shown inconsistent results.11,18,19,22,28,31,33 The lack of a significant association in this group may be due to the small sample size of the low standard of living group (seven participants); further research with a larger sample may show a more definitive pattern.This study was designed as an interim analysis to investigate whether an interpretable heart healthy dietary pattern could be obtained from this small dataset comprising the first participants enrolled into CHALICE. The analysis was also a pre-planned part of our quality assessment strategy, in particular with the dietary instruments, data entry and analysis of the food records.While the study had sufficient sample size to derive a meaningful, robust dietary pattern,27 we acknowledge that a sample size of around 60 participants severely restricts our ability to detect significant differences in individual dietary components. However as the participants are all fifty years of age the overall variability within the dietary data will be smaller than for a similar sample over a larger age range.The Canterbury earthquake sequence began on 4 September 2010 with the second, most destructive, and only fatal major episode on 22 February 2011. The data presented was all collected prior to the 22 February 2011 and is in our view the best representation of diet of 50 year olds unaffected by the earthquakes.Future analyses will attempt to quantify the effect of the earthquake, if any. While we would have preferred more participants the first 62 form a natural grouping. Nevertheless, these findings show that it is possible to generate a heart healthy dietary pattern in smaller samples and we once the final CHALICE recruitment is completed we will explore these relationships further, using more sophisticated analyses.This study provides evidence that the 50 year olds from Canterbury included in these analyses do not consume a heart healthy diet and that socioeconomic status, education and nutrition literacy may influence dietary consumption. If these findings are found to be similar in the whole CHALICE sample then this is particularly pertinent for Canterbury which has an aging population, hence identifying modifiable risk factors in midlife has the potential to better target interventions for this increasing proportion of the population.This study provides further evidence that improving dietary intake through improving the social and financial circumstances of a population may help reduce health inequalities and the burden of disease.This article was corrected on 30 January 2015 as outlined in the Erratum published the same day athttp://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1408/6427\r\n

Summary

Abstract

Aim

Cardiovascular disease is a leading cause of death in New Zealand, but risk factors may be decreased by consuming a heart healthy diet. This pilot study investigated whether participants met the guidelines for a heart healthy diet and whether a novel heart healthy dietary pattern could be identified using principal components analysis (PCA). The second aim of this project was to assess if higher education, standard of living and nutrition literacy are associated with a heart healthy dietary pattern.

Method

This exploratory study was undertaken using data from the first participants enrolled in the Canterbury Health Ageing and Lifecourse study: an observational study of 50 year olds in the Canterbury District Health Board region. Eighty-two people were selected from the General and Mori electoral role and interviewed prior to the 22 February 2011 Christchurch Earthquake. PCA was conducted to identify dietary patterns, based on intake of specific nutrients as indicated by the New Zealand and international heart healthy dietary guidelines.

Results

62 participants completed questionnaires and an estimated food record. No participants met all five of the heart healthy dietary guidelines. One dietary pattern was produced by PCA: a higher CVD risk pattern. Regression analysis indicated that higher standard of living, education and nutrition literacy were inversely associated with a higher CVD risk pattern.

Conclusion

Higher standard of living, education and nutrition literacy were associated with a healthier dietary eating pattern. However, as no participants met all the dietary recommendations more education and support is needed to help people meet these.

Author Information

Acknowledgements

Correspondence

Correspondence Email

Competing Interests

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Cardiovascular disease (CVD) is the second most common cause of death after cancer in New Zealand (NZ).1 The prevalence of CVD increases dramatically with age; at 45 years of age only 8% of New Zealanders are diagnosed with CVD compared to 75% of those aged 75 years.2New Zealand has an aging population which will increase the burden of CVD.3 Among the risk factors for CVD, some are partially modifiable, including high blood pressure, high blood cholesterol, overweight and poor glycaemic control.4 Modifiable risk factors are known to be influenced, in part, by dietary intake.4 Consumption of particular heart healthy foods has been shown to improve modifiable risk factors.5,6A heart healthy diet typically includes high intakes of fruit, vegetables, whole grain foods, fish and seafood, lean meat and poultry, low fat dairy products, nuts and seeds and low consumption of high sodium products. Unfortunately, consuming a heart healthy diet and improving one's CVD risk may not be achievable for all sectors of the community particularly for those in lower socioeconomic groups.7 The relationship between lower socioeconomic status (SES) and morbidity and mortality rates from chronic diseases such as cardiovascular disease (CVD) are well established worldwide8 and in New Zealand.9International research has shown that lower socioeconomic status is often associated with higher intakes of dietary fat10–12 and lower intakes of dietary fibre and fruit and vegetables, independently of each other.13,14 Lower socioeconomic status is thought to not only influence the dietary intake of families but, in turn, body weight of children.15 It is not known if these relationships are also present in the NZ population.Socioeconomic status can be measured using a variety of methods, the newest of which is the Economic Living Standard Index short form (ELSISF).16 The ELSI is a measure of individual living standards and has been included in the New Zealand Health Survey since 2006/07 to describe sociodemographics of New Zealanders. The ELSISF has been used to describe dietary intakes of children16 but to our knowledge nutritional intake of adults or body weight status has not been compared using the ELSISF. Nutritional intake of NZ adults has been described according to level of community deprivation (NZDep) but not by individual living standards.7Associations between selected psychosocial factors including standard of living measures, education, knowledge of food and dietary intake, as assessed by investigation of single foods, food groups and nutrients, have been studied internationally11,17–19 and in New Zealand.7,13,15Dietary patterns have been proposed as a solution to investigating relationships between food choice and chronic disease risk as these analyses allow for the multiple factors in the diet to be examined in combination, rather than focusing on intakes of single nutrients or food groups. This is a more powerful approach as foods can have synergistic or inhibitor effects on each other. One such type of analysis is PCA, which has been shown to be successful in producing patterns associated with CVD risk factors in many studies.20 However, most dietary patterns are generated from food frequency questionnaire data, which are not as robust as data from food records.20 This exploratory study uses estimated food records to investigate whether a novel heart healthy dietary pattern could be identified using PCA.A secondary aim of the study was to investigate if such a pattern is associated with a higher level of education and standard of living. Analysis of this type has not been previously conducted using the ELSISF. The aims of this study are in line with the wider aims of the Canterbury Health, Ageing and Life Course (CHALICE) study to build a database of important determinants of health.MethodsStudy design—The data for this pilot study were collected from the first 82 participants who completed baseline assessments in the CHALICE study between August 2010 and February 2011. Data included in this study were collected prior to the Christchurch Earthquake on 22 February 2011.The CHALICE study—CHALICE is a longitudinal, observational study of a random sample of people aged 50 years currently living in the Canterbury District Health Board (CDHB) area.21 CHALICE aims to recruit 1000 participants. Fifty-year olds were randomly recruited from the general and Māori Canterbury electoral rolls from the Kaikoura, Hurunui, Waimakariri, Christchurch city, Selwyn and Ashburton districts. Māori were over-sampled (one in every four participants were selected from the Māori electoral roll). Ethical approval was obtained from the Upper South A Regional Ethics Committee.Data collection—The overall CHALICE methods have been described in detail elsewhere.21 Participants were invited to attend a four to six hour face to face interview at the University of Otago, Christchurch offices where information was collected by trained interviewers on physical and mental health, family and social factors and lifestyle. For this study only body composition, demographic, nutrition related questionnaires and the estimated food record data were included in the analysis. These are described in more detail below.Demographic data collected included ethnicity as assessed by the question used in the NZ census; "Which ethnic group or groups do you belong to?". For the purposes of these analyses results were dichotomised where anyone who selected Māori, even if more than one ethnicity was chosen, were identified as Māori. All other participants were categorised as non-Māori. Participants were asked "What is the total income that your household got from all sources, before tax, or anything that was taken out of it, in the past 12 months" and asked to pick from a range of answers, where the lowest was less than <$5000 and the highest $150,001 or more. For these analyses participant responses were grouped into the following groups: <$40,000, $40,000–$69,999, $70,000–119,999 and $120,000 or over. Highest level of education was classified as either no qualifications, secondary school qualifications, post-secondary (non-degree) qualification, or University degree and for these analyses was dichotomised to Secondary school or less, or Post-secondary qualification.Standard of living was assessed using the ELSISF. This is a 25 item questionnaire that assesses standard of living, based on three components – ownership restriction (e.g. not owning a car), social participation restriction (not doing activities that a participant would like to do because of cost) and economising (e.g. not buying fresh fruit and vegetables due to cost).13 The ELSISF answers were summed to produce a score of between 0 and 31. These scores were then coded into seven categories ranging from "Severe hardship" to "Very good". Due to small participant numbers in these analyses these categories were reduced to three categories: low (score of 0-16), medium (score of 17-24) or high (score of 25-31) standard of living.Weight and height measurements were taken with shoes removed and in light clothing to the nearest 0.1 cm. Weight was measured using an electronic scale and height was measured against a wall using a permanently attached tape measure. Waist circumference was measured against the skin. A tape measure was placed at the mid-point between the lowest rib and the top of the iliac crest. Waist circumference was measured once to the nearest 0.5cm. Body mass index (BMI) was calculated as weight (kg) divided by height (m2).A nutrition literacy score was calculated from 24 questions based on the United Kingdom "The Family Diet Study" questionnaire.22 This questionnaire asked about the fat, sugar, salt and/or fibre content of commonly consumed foods. The foods used in the nutrition literacy questions were modified to make them compatible with the NZ diet e.g. Rice Krispies changed to Rice Bubbles, squash changed to cordial. This questionnaire was pretested in a sample comparable to the CHALICE population before use in this study.23Food records: Participants completed a four day estimated food record in the week after completing the study interview. Participants received a written and verbal explanation of how to complete a detailed record of all food and beverages consumed over three week days and one weekend day. Returned records were checked by a dietitian for completeness. If any information was incomplete, unclear or not detailed enough, the interviewers contacted participants for clarification according to the dietitian's instruction. The food record data were analysed using Diet Cruncher version 1.6.0 (Way Down South Software). For the current analyses mean intakes of total fat, saturated fat, dietary fibre and total energy were obtained and total and saturated fat as a percentage of total energy (% TE) were calculated. Average serves of daily fruit and vegetables were also calculated, to assess whether participants were meeting the NZ Healthy Eating guidelines.24,25 The results of the food record analysis were compared with guidelines for a heart healthy diet. The data from one participant was excluded from analyses because they had a rare health condition controlled by diet and their diet was not representative of the general population.There are many national dietary guidelines available from countries including the United States of America, United Kingdom, Australia and NZ and while some of the individual food and/or nutrients included in these may differ from country to country, the foods and nutrients most consistently included are total fat, saturated fat, fruit, vegetables and fibre.3,4,24,25 Therefore, for the purpose of this study a heart healthy diet was defined as a diet including a healthy intake of all five of these components, based on the current NZ recommendations.24,25 Criteria for this are a total fat contribution of less than 35% of total energy (% TE), saturated fat contribution of than 10% TE, dietary fibre intake of more than 25 g/day, fruit intake of at least two serves per day and vegetable intake of at least three serves (excluding potatoes). Yes/No variables were created to indicate whether each participant met each of the five guidelines.Statistical analyses—Analyses were carried out using R, version 2.13.0 (Vienna, Austria). Median intake and interquartile ranges were calculated for each nutrient and associations with demographic variables and BMI were explored using the Mann Whitney U test and the Kruskal-Wallis test. Associations between standard of living and education were adjusted for income, sex and ethnicity.Heart healthy dietary pattern scores were also generated using PCA. PCA is a commonly used data reduction technique to provide summary measures of diet. As individual foods or food groups are not eaten in isolation, analysis of individual nutrients or food groups does not take into account the complexities of meals eaten, and the possible nutrient interactions occurring when meals are consumed.20Using a PCA-derived pattern means that a greater proportion of the variation within the dietary data can be explained than when using individual food group or nutrient data. The use of PCA means that summary scores can be calculated for each of the five heart healthy diet components entered into the analyses. These scores are based on an equation based on the amount consumed (for nutrients), or the number of servings (for fruit and vegetables), multiplied by the factor loading produced by the PCA.Foods/nutrients with higher factor loadings contribute more to the overall score than those with lower loadings, and in accordance with other similar studies, eigenvalues above 1, the elbow in the scree plot and factor loadings ≤-0.3 or greater than ≥0.3 were considered important when identifying and naming patterns.26Negative factor loadings mean that a particular food or nutrient is contributing negatively to the overall score. A sample size of 62 participants allows for the examination of six variables using PCA, as ten participants per variable are required for robust results.27 Varimax rotation was conducted using z-scores of the five dietary components of a heart healthy diet. As continuous data (not dichotomised data) is needed for PCA data for the amount consumed (for nutrients) or the number of servings (for fruit and vegetables) were used for the PCA. The first principal component was used based on both the elbow in the Scree plot (not shown) and its interpretability based on factor loadings.Multiple linear regression was applied to the first principal component with demographic and nutrition related variables. Non-significant variables that did not contribute to the fit of the model were removed, second order interactions were tested and all but standard of living, education and nutrition literacy were not significant, there was insufficient data in all cells to properly fit higher order interactions. All results were considered statistically significant at the 5% level where reported p values were adjusted for multiple comparisons using Tukey's method. All model assumptions were checked by graphical inspection of residual plots (QQ, Leverage, residual versus fitted values).ResultsOf the 82 CHALICE participants interviewed 62 (75.6%, 32 females and 30 males) returned a completed food record prior to the 22nd of February 2011 and provided information on all variables of interest. Table 1 details the participant characteristics.Table 1. Sample characteristics\r\n \r\n \r\n \r\n Characteristics\r\n \r\n Category\r\n \r\n Number of participants (n=62) (%)\r\n \r\n \r\n \r\n Sex\r\n \r\n Male Female\r\n \r\n 30 (48) 32 (52)\r\n \r\n \r\n \r\n Ethnicity\r\n \r\n Māori Non-Māori\r\n \r\n 8 (13) 54 (87)\r\n \r\n \r\n \r\n Education\r\n \r\n Secondary school or less Post-secondary school\r\n \r\n 21 (34) 41 (66)\r\n \r\n \r\n \r\n Employment (current)\r\n \r\n In paid employment Not in paid employment\r\n \r\n 55 (89) 7 (11)\r\n \r\n \r\n \r\n Household income (per annum)\r\n \r\n <$40,000 $40,000–69,999 $70,000–119,999 ≥$120,000 Don't know\r\n \r\n 12 (19) 13 (21) 19 (31) 15 (24) 3 (5)\r\n \r\n \r\n \r\n Standard of living (ELSISF)\r\n \r\n Low Medium High\r\n \r\n 7 (11) 18 (29) 37 (60)\r\n \r\n \r\n \r\nTable 2 summarises mean body composition of participants. Mean BMI was 28.0±6.0 kg/m2, which is within the WHO overweight range (BMI 25–29.9 kg/m2).4 Twenty-nine percent of participants were classified as normal weight, 40% overweight and 31% obese. Mean waist circumference of participants falls within the range (94–102 cm for men and 80–88 cm for women) which is associated with an increased risk of metabolic complications.Table 2. Body composition of sample\r\n \r\n \r\n \r\n Body composition measures\r\n \r\n Females (n=32) Mean (SD)\r\n \r\n Males (n=30) Mean (SD)\r\n \r\n \r\n \r\n Weight (kg) Height (cm) BMI (kg/m2) Waist circumference (cm)\r\n \r\n 75.2 (18.5) 165.2 (7.0) 27.6 (6.6) 87.6 (15.7)\r\n \r\n 89.5 (17.1) 177.4 (6.5) 28.5 (5.4) 100.0 (13.4)\r\n \r\n \r\n \r\nBMI=body mass index.The median dietary intake of participants is shown in Table 3. Of the five heart healthy dietary components, no participants met all five recommendations, 8% met four recommendations, 24% met three recommendations, 25% met two recommendations, another 25% met one recommendation and 18% met none of the recommendations.Single nutrient analyses—Table 3 also summarises non-parametric and post hoc tests with dietary intake, standard of living and education. The only dietary component which varied significantly with standard of living was total fat intake (P=0.018, P adjusted (for income)=0.024). There was no evidence of a difference between total fat intake for the medium and high ELSI ranges. This result remained unchanged after adjustment for BMI, sex and ethnicity (results not shown).Total fat and vegetable intake did not vary significantly by education. Median saturated fat as a % TE was lower in the higher education group, 12.4% TE compared with 14.8% TE, than in the lower education group (P=0.002). Median intake of both groups was higher than the recommended intake of less of than 10% of TE.Table 3. Median (LQ, UQ) food and nutrient intake compared with heart healthy dietary guidelines and median (LQ, UQ) nutrient and food intake by standard of living category and education\r\n \r\n \r\n \r\n Nutrient\r\n \r\n Total fat (% TE)\r\n \r\n Saturated fat (% TE)\r\n \r\n Dietary fibre (g/day)\r\n \r\n Fruit (portions/day)\r\n \r\n Vegetables (portions/day)\r\n \r\n \r\n \r\n Recommended intake for a heart healthy diet\r\n \r\n <35%\r\n \r\n <10%\r\n \r\n >25g\r\n \r\n >2\r\n \r\n >3\r\n \r\n \r\n \r\n Median (LQ, UQ)\r\n \r\n 33.7 [31.2, 37.0]\r\n \r\n 13.5 [11.5–14.8]\r\n \r\n 22.9 [18.2–27.8]\r\n \r\n 1.8 [0.5–2.5]\r\n \r\n 2.0 [1.3–2.7]\r\n \r\n \r\n \r\n Percentage achieving recommended intake\r\n \r\n 57%\r\n \r\n 11%\r\n \r\n 38%\r\n \r\n 48%\r\n \r\n 14%\r\n \r\n \r\n \r\n Median (LQ, UQ) food and nutrient intake by standard of living\r\n \r\n \r\n \r\n Low\r\n \r\n 40.4 [34.9–41.8]\r\n \r\n 15.0 [14.1–16.4]\r\n \r\n 19.8 [18.9–26.9]\r\n \r\n 0.3 [0.1–1.5]\r\n \r\n 1.5 [1.0–1.6]\r\n \r\n \r\n \r\n Medium\r\n \r\n 32.3 [28.7–36.4]*\r\n \r\n 12.3 [11.2–13.8]\r\n \r\n 24.1 [20.5–31.3]\r\n \r\n 2.0 [0.6–3.0]\r\n \r\n 2.0 [0.5–2.5]\r\n \r\n \r\n \r\n High\r\n \r\n 33.7 [31.1–36.9]*\r\n \r\n 13.2 [11.5–14.6]\r\n \r\n 21.9 [18.0–26.1]\r\n \r\n 1.8 [0.8–2.5]\r\n \r\n 2.0 [1.5–3.0]\r\n \r\n \r\n \r\n P Kruskal-Wallis test\r\n \r\n 0.018\r\n \r\n 0.124\r\n \r\n 0.397\r\n \r\n 0.193\r\n \r\n 0.151\r\n \r\n \r\n \r\n Median (LQ, UQ) food and nutrient intake by education level\r\n \r\n \r\n \r\n Lower education\r\n \r\n 35.8 [32.5–38.2]\r\n \r\n 14.8 [12.3–16.1]\r\n \r\n 19.8 [15.7–25.6]\r\n \r\n 1.0 [0.3–2.0]\r\n \r\n 1.5 [1.0–2.0]\r\n \r\n \r\n \r\n Higher education\r\n \r\n 33.2 [30.5–36.5]*\r\n \r\n 12.4 [11.0–13.8]**\r\n \r\n 23.4 [19.7–29.0]*\r\n \r\n 2.0 [0.5–2.8]*\r\n \r\n 2.0 [1.5–3.0]\r\n \r\n \r\n \r\n P Kruskal-Wallis test\r\n \r\n 0.113\r\n \r\n 0.002\r\n \r\n 0.045\r\n \r\n 0.044\r\n \r\n 0.240\r\n \r\n \r\n \r\nNote: * p<0.05; **p<0.01 compared to low standard of living group or lower education group. % TE=percentage of Total Energy intake, LQ = lower quartile, UQ = upper quartile.Principal components analysis—PCA produced one meaningful pattern (Table 6). The first principal component reflected a "higher CVD risk" dietary pattern: a high saturated fat, and low fruit, vegetable and dietary fibre diet based on factor loadings of 0.42, -0.55, -0.50 and -0.46 respectively. A higher score for an individual for the "higher CVD risk" pattern indicates a higher saturated fat intake and a lower fruit, vegetable and fibre intake, all of which are not recommended as part of a heart healthy diet.Table 6: Principal component analysis loadings and importance\r\n \r\n \r\n \r\n Variables\r\n \r\n "Higher CVD risk" pattern\r\n \r\n \r\n \r\n Total fat Saturated fat Dietary fibre Fruit Vegetables Percentage variation explained\r\n \r\n 0.26 0.42 -0.46 -0.55 -0.50 40%\r\n \r\n \r\n \r\nMultiple linear regression analysis on normalised "higher CVD risk" diet scores showed significantly lower scores for those with a higher education level (0.98, 95% CI [0.27–1.69], P=0.008) while there was evidence for a difference in the impact of nutrition literacy on diet by standard of living (P=0.001). The diet score of the low standard of living group increased by 0.35 (95%CI [0.02–0.67], P=0.039) for each one unit increase in nutrition literacy while the diet score of the high standard of living group decreased by 0.10 (95%CI [0.11–0.79], P=0.011) for each increase in nutrition literacy by one unit. Neither sex nor ethnicity had significant effects at the 0.05% level, nor was there a significant difference between the two higher levels of standard of living, the model explained 28% of the variation.The results show that there were significant interaction effects between nutrition literacy and standard of living on the "higher CVD risk" pattern scores. Those in the medium or high standard of living categories, who had better nutrition literacy, had lower scores for the "higher CVD risk" pattern, indicating consumption of a more heart healthy diet, compared to those of medium or high standard of living who had lower nutrition literacy. Conversely, those in the low standard of living category, who had higher nutrition literacy scores, had higher scores for the "higher CVD risk" pattern, indicating consumption of a less heart healthy diet.Figure 1. Association between nutrition literacy and standard of living interaction with a "higher CVD risk" dietary pattern (multiple linear regression, p=0.001)a,b,c,da All of the data included in the multiple linear regression are plotted in the figure; however the regression lines are only relevant for those participants in the respective lower education group; b Low standard of living β=0.35, 95%CI [0.02, 0.67], p=0.039, c Medium standard of living β=-0.23, 95%CI [-0.97 – -0.17], p=0.006, d High standard of living β = -0.10 – 95%CI [-0.79 – -0.11], p=0.011.DiscussionThe results discussed here are based on preliminary analysis from the first wave of the CHALICE study aiming to recruit 1000 50 year olds from Canterbury over the period 2010–2014. This study describes aspects of the dietary intake and the associations with selected social variables.None of the participants in this study met all five heart healthy guidelines, and the majority of participants met two or less guidelines. These results suggest that, as expected, a higher standard of living, education and nutrition literacy were all associated with consuming a heart healthy diet. Using results of the PCA, those who had a medium or high standard of living and who had higher levels of nutrition literacy had higher scores for a heart healthy diet. However, those with a lower standard of living, and who had higher nutrition literacy had pattern scores that corresponded with a less heart healthy diet.Previous studies have shown associations between socioeconomic status and education with selected foods or nutrients2,7,12,14,22,28,29 but there is limited evidence available from New Zealand populations.30 No research, that we are aware of, has used the ELSISF to describe socioeconomic circumstance or standard of living in relation to adult dietary intake. In this population there was a significant pattern toward decreasing fat (as a % TE) intake as standard of living increased. There are no questions in the ELSISF that assess variables that may influence fat consumption such as frequency of purchasing food from outside of the home or access to affordable healthy foods.Studies from abroad have shown that those living in lower socioeconomic areas may struggle to access affordable, healthy, lower fat foods due to the positioning of supermarkets, fruit and vegetable shops and takeaway outlets30,31 and that eating out more often has been associated with higher fat intake.32 It is not known if this is also the case in NZ. In this population there were no significant patterns with the other components of a heart healthy diet; although this may be due to the wide interquartile ranges of some of the components.In addition to assessing associations between socioeconomic status, education and nutrition literacy and single nutrients this study also investigated associations between social variables and a "higher CVD risk" dietary pattern.Dietary patterns, as with intakes of selected nutrients, have been shown to be less healthy in lower socioeconomic groups2 and deprived sectors of the community.7 Dietary modelling (PCA) showed that standard of living, as well as nutrition literacy and level of education, were inversely associated with a "higher CVD risk" dietary pattern. Interestingly, those with a high standard of living tended to consume a less healthy dietary pattern than those with a medium standard of living, however this did not reach statistical significance (P = 0.28).Conversely there was a significant pattern with the lowest standard of living group towards increasing consumption of saturated fat (as a % TE), and lower consumption of fruit, vegetables and dietary fibre as nutrition literacy increased.The relationship between nutrition literacy and dietary intake is often complex. Previous research that has explored this relationship has shown inconsistent results.11,18,19,22,28,31,33 The lack of a significant association in this group may be due to the small sample size of the low standard of living group (seven participants); further research with a larger sample may show a more definitive pattern.This study was designed as an interim analysis to investigate whether an interpretable heart healthy dietary pattern could be obtained from this small dataset comprising the first participants enrolled into CHALICE. The analysis was also a pre-planned part of our quality assessment strategy, in particular with the dietary instruments, data entry and analysis of the food records.While the study had sufficient sample size to derive a meaningful, robust dietary pattern,27 we acknowledge that a sample size of around 60 participants severely restricts our ability to detect significant differences in individual dietary components. However as the participants are all fifty years of age the overall variability within the dietary data will be smaller than for a similar sample over a larger age range.The Canterbury earthquake sequence began on 4 September 2010 with the second, most destructive, and only fatal major episode on 22 February 2011. The data presented was all collected prior to the 22 February 2011 and is in our view the best representation of diet of 50 year olds unaffected by the earthquakes.Future analyses will attempt to quantify the effect of the earthquake, if any. While we would have preferred more participants the first 62 form a natural grouping. Nevertheless, these findings show that it is possible to generate a heart healthy dietary pattern in smaller samples and we once the final CHALICE recruitment is completed we will explore these relationships further, using more sophisticated analyses.This study provides evidence that the 50 year olds from Canterbury included in these analyses do not consume a heart healthy diet and that socioeconomic status, education and nutrition literacy may influence dietary consumption. If these findings are found to be similar in the whole CHALICE sample then this is particularly pertinent for Canterbury which has an aging population, hence identifying modifiable risk factors in midlife has the potential to better target interventions for this increasing proportion of the population.This study provides further evidence that improving dietary intake through improving the social and financial circumstances of a population may help reduce health inequalities and the burden of disease.This article was corrected on 30 January 2015 as outlined in the Erratum published the same day athttp://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1408/6427\r\n

Summary

Abstract

Aim

Cardiovascular disease is a leading cause of death in New Zealand, but risk factors may be decreased by consuming a heart healthy diet. This pilot study investigated whether participants met the guidelines for a heart healthy diet and whether a novel heart healthy dietary pattern could be identified using principal components analysis (PCA). The second aim of this project was to assess if higher education, standard of living and nutrition literacy are associated with a heart healthy dietary pattern.

Method

This exploratory study was undertaken using data from the first participants enrolled in the Canterbury Health Ageing and Lifecourse study: an observational study of 50 year olds in the Canterbury District Health Board region. Eighty-two people were selected from the General and Mori electoral role and interviewed prior to the 22 February 2011 Christchurch Earthquake. PCA was conducted to identify dietary patterns, based on intake of specific nutrients as indicated by the New Zealand and international heart healthy dietary guidelines.

Results

62 participants completed questionnaires and an estimated food record. No participants met all five of the heart healthy dietary guidelines. One dietary pattern was produced by PCA: a higher CVD risk pattern. Regression analysis indicated that higher standard of living, education and nutrition literacy were inversely associated with a higher CVD risk pattern.

Conclusion

Higher standard of living, education and nutrition literacy were associated with a healthier dietary eating pattern. However, as no participants met all the dietary recommendations more education and support is needed to help people meet these.

Author Information

Acknowledgements

Correspondence

Correspondence Email

Competing Interests

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Cardiovascular disease (CVD) is the second most common cause of death after cancer in New Zealand (NZ).1 The prevalence of CVD increases dramatically with age; at 45 years of age only 8% of New Zealanders are diagnosed with CVD compared to 75% of those aged 75 years.2New Zealand has an aging population which will increase the burden of CVD.3 Among the risk factors for CVD, some are partially modifiable, including high blood pressure, high blood cholesterol, overweight and poor glycaemic control.4 Modifiable risk factors are known to be influenced, in part, by dietary intake.4 Consumption of particular heart healthy foods has been shown to improve modifiable risk factors.5,6A heart healthy diet typically includes high intakes of fruit, vegetables, whole grain foods, fish and seafood, lean meat and poultry, low fat dairy products, nuts and seeds and low consumption of high sodium products. Unfortunately, consuming a heart healthy diet and improving one's CVD risk may not be achievable for all sectors of the community particularly for those in lower socioeconomic groups.7 The relationship between lower socioeconomic status (SES) and morbidity and mortality rates from chronic diseases such as cardiovascular disease (CVD) are well established worldwide8 and in New Zealand.9International research has shown that lower socioeconomic status is often associated with higher intakes of dietary fat10–12 and lower intakes of dietary fibre and fruit and vegetables, independently of each other.13,14 Lower socioeconomic status is thought to not only influence the dietary intake of families but, in turn, body weight of children.15 It is not known if these relationships are also present in the NZ population.Socioeconomic status can be measured using a variety of methods, the newest of which is the Economic Living Standard Index short form (ELSISF).16 The ELSI is a measure of individual living standards and has been included in the New Zealand Health Survey since 2006/07 to describe sociodemographics of New Zealanders. The ELSISF has been used to describe dietary intakes of children16 but to our knowledge nutritional intake of adults or body weight status has not been compared using the ELSISF. Nutritional intake of NZ adults has been described according to level of community deprivation (NZDep) but not by individual living standards.7Associations between selected psychosocial factors including standard of living measures, education, knowledge of food and dietary intake, as assessed by investigation of single foods, food groups and nutrients, have been studied internationally11,17–19 and in New Zealand.7,13,15Dietary patterns have been proposed as a solution to investigating relationships between food choice and chronic disease risk as these analyses allow for the multiple factors in the diet to be examined in combination, rather than focusing on intakes of single nutrients or food groups. This is a more powerful approach as foods can have synergistic or inhibitor effects on each other. One such type of analysis is PCA, which has been shown to be successful in producing patterns associated with CVD risk factors in many studies.20 However, most dietary patterns are generated from food frequency questionnaire data, which are not as robust as data from food records.20 This exploratory study uses estimated food records to investigate whether a novel heart healthy dietary pattern could be identified using PCA.A secondary aim of the study was to investigate if such a pattern is associated with a higher level of education and standard of living. Analysis of this type has not been previously conducted using the ELSISF. The aims of this study are in line with the wider aims of the Canterbury Health, Ageing and Life Course (CHALICE) study to build a database of important determinants of health.MethodsStudy design—The data for this pilot study were collected from the first 82 participants who completed baseline assessments in the CHALICE study between August 2010 and February 2011. Data included in this study were collected prior to the Christchurch Earthquake on 22 February 2011.The CHALICE study—CHALICE is a longitudinal, observational study of a random sample of people aged 50 years currently living in the Canterbury District Health Board (CDHB) area.21 CHALICE aims to recruit 1000 participants. Fifty-year olds were randomly recruited from the general and Māori Canterbury electoral rolls from the Kaikoura, Hurunui, Waimakariri, Christchurch city, Selwyn and Ashburton districts. Māori were over-sampled (one in every four participants were selected from the Māori electoral roll). Ethical approval was obtained from the Upper South A Regional Ethics Committee.Data collection—The overall CHALICE methods have been described in detail elsewhere.21 Participants were invited to attend a four to six hour face to face interview at the University of Otago, Christchurch offices where information was collected by trained interviewers on physical and mental health, family and social factors and lifestyle. For this study only body composition, demographic, nutrition related questionnaires and the estimated food record data were included in the analysis. These are described in more detail below.Demographic data collected included ethnicity as assessed by the question used in the NZ census; "Which ethnic group or groups do you belong to?". For the purposes of these analyses results were dichotomised where anyone who selected Māori, even if more than one ethnicity was chosen, were identified as Māori. All other participants were categorised as non-Māori. Participants were asked "What is the total income that your household got from all sources, before tax, or anything that was taken out of it, in the past 12 months" and asked to pick from a range of answers, where the lowest was less than <$5000 and the highest $150,001 or more. For these analyses participant responses were grouped into the following groups: <$40,000, $40,000–$69,999, $70,000–119,999 and $120,000 or over. Highest level of education was classified as either no qualifications, secondary school qualifications, post-secondary (non-degree) qualification, or University degree and for these analyses was dichotomised to Secondary school or less, or Post-secondary qualification.Standard of living was assessed using the ELSISF. This is a 25 item questionnaire that assesses standard of living, based on three components – ownership restriction (e.g. not owning a car), social participation restriction (not doing activities that a participant would like to do because of cost) and economising (e.g. not buying fresh fruit and vegetables due to cost).13 The ELSISF answers were summed to produce a score of between 0 and 31. These scores were then coded into seven categories ranging from "Severe hardship" to "Very good". Due to small participant numbers in these analyses these categories were reduced to three categories: low (score of 0-16), medium (score of 17-24) or high (score of 25-31) standard of living.Weight and height measurements were taken with shoes removed and in light clothing to the nearest 0.1 cm. Weight was measured using an electronic scale and height was measured against a wall using a permanently attached tape measure. Waist circumference was measured against the skin. A tape measure was placed at the mid-point between the lowest rib and the top of the iliac crest. Waist circumference was measured once to the nearest 0.5cm. Body mass index (BMI) was calculated as weight (kg) divided by height (m2).A nutrition literacy score was calculated from 24 questions based on the United Kingdom "The Family Diet Study" questionnaire.22 This questionnaire asked about the fat, sugar, salt and/or fibre content of commonly consumed foods. The foods used in the nutrition literacy questions were modified to make them compatible with the NZ diet e.g. Rice Krispies changed to Rice Bubbles, squash changed to cordial. This questionnaire was pretested in a sample comparable to the CHALICE population before use in this study.23Food records: Participants completed a four day estimated food record in the week after completing the study interview. Participants received a written and verbal explanation of how to complete a detailed record of all food and beverages consumed over three week days and one weekend day. Returned records were checked by a dietitian for completeness. If any information was incomplete, unclear or not detailed enough, the interviewers contacted participants for clarification according to the dietitian's instruction. The food record data were analysed using Diet Cruncher version 1.6.0 (Way Down South Software). For the current analyses mean intakes of total fat, saturated fat, dietary fibre and total energy were obtained and total and saturated fat as a percentage of total energy (% TE) were calculated. Average serves of daily fruit and vegetables were also calculated, to assess whether participants were meeting the NZ Healthy Eating guidelines.24,25 The results of the food record analysis were compared with guidelines for a heart healthy diet. The data from one participant was excluded from analyses because they had a rare health condition controlled by diet and their diet was not representative of the general population.There are many national dietary guidelines available from countries including the United States of America, United Kingdom, Australia and NZ and while some of the individual food and/or nutrients included in these may differ from country to country, the foods and nutrients most consistently included are total fat, saturated fat, fruit, vegetables and fibre.3,4,24,25 Therefore, for the purpose of this study a heart healthy diet was defined as a diet including a healthy intake of all five of these components, based on the current NZ recommendations.24,25 Criteria for this are a total fat contribution of less than 35% of total energy (% TE), saturated fat contribution of than 10% TE, dietary fibre intake of more than 25 g/day, fruit intake of at least two serves per day and vegetable intake of at least three serves (excluding potatoes). Yes/No variables were created to indicate whether each participant met each of the five guidelines.Statistical analyses—Analyses were carried out using R, version 2.13.0 (Vienna, Austria). Median intake and interquartile ranges were calculated for each nutrient and associations with demographic variables and BMI were explored using the Mann Whitney U test and the Kruskal-Wallis test. Associations between standard of living and education were adjusted for income, sex and ethnicity.Heart healthy dietary pattern scores were also generated using PCA. PCA is a commonly used data reduction technique to provide summary measures of diet. As individual foods or food groups are not eaten in isolation, analysis of individual nutrients or food groups does not take into account the complexities of meals eaten, and the possible nutrient interactions occurring when meals are consumed.20Using a PCA-derived pattern means that a greater proportion of the variation within the dietary data can be explained than when using individual food group or nutrient data. The use of PCA means that summary scores can be calculated for each of the five heart healthy diet components entered into the analyses. These scores are based on an equation based on the amount consumed (for nutrients), or the number of servings (for fruit and vegetables), multiplied by the factor loading produced by the PCA.Foods/nutrients with higher factor loadings contribute more to the overall score than those with lower loadings, and in accordance with other similar studies, eigenvalues above 1, the elbow in the scree plot and factor loadings ≤-0.3 or greater than ≥0.3 were considered important when identifying and naming patterns.26Negative factor loadings mean that a particular food or nutrient is contributing negatively to the overall score. A sample size of 62 participants allows for the examination of six variables using PCA, as ten participants per variable are required for robust results.27 Varimax rotation was conducted using z-scores of the five dietary components of a heart healthy diet. As continuous data (not dichotomised data) is needed for PCA data for the amount consumed (for nutrients) or the number of servings (for fruit and vegetables) were used for the PCA. The first principal component was used based on both the elbow in the Scree plot (not shown) and its interpretability based on factor loadings.Multiple linear regression was applied to the first principal component with demographic and nutrition related variables. Non-significant variables that did not contribute to the fit of the model were removed, second order interactions were tested and all but standard of living, education and nutrition literacy were not significant, there was insufficient data in all cells to properly fit higher order interactions. All results were considered statistically significant at the 5% level where reported p values were adjusted for multiple comparisons using Tukey's method. All model assumptions were checked by graphical inspection of residual plots (QQ, Leverage, residual versus fitted values).ResultsOf the 82 CHALICE participants interviewed 62 (75.6%, 32 females and 30 males) returned a completed food record prior to the 22nd of February 2011 and provided information on all variables of interest. Table 1 details the participant characteristics.Table 1. Sample characteristics\r\n \r\n \r\n \r\n Characteristics\r\n \r\n Category\r\n \r\n Number of participants (n=62) (%)\r\n \r\n \r\n \r\n Sex\r\n \r\n Male Female\r\n \r\n 30 (48) 32 (52)\r\n \r\n \r\n \r\n Ethnicity\r\n \r\n Māori Non-Māori\r\n \r\n 8 (13) 54 (87)\r\n \r\n \r\n \r\n Education\r\n \r\n Secondary school or less Post-secondary school\r\n \r\n 21 (34) 41 (66)\r\n \r\n \r\n \r\n Employment (current)\r\n \r\n In paid employment Not in paid employment\r\n \r\n 55 (89) 7 (11)\r\n \r\n \r\n \r\n Household income (per annum)\r\n \r\n <$40,000 $40,000–69,999 $70,000–119,999 ≥$120,000 Don't know\r\n \r\n 12 (19) 13 (21) 19 (31) 15 (24) 3 (5)\r\n \r\n \r\n \r\n Standard of living (ELSISF)\r\n \r\n Low Medium High\r\n \r\n 7 (11) 18 (29) 37 (60)\r\n \r\n \r\n \r\nTable 2 summarises mean body composition of participants. Mean BMI was 28.0±6.0 kg/m2, which is within the WHO overweight range (BMI 25–29.9 kg/m2).4 Twenty-nine percent of participants were classified as normal weight, 40% overweight and 31% obese. Mean waist circumference of participants falls within the range (94–102 cm for men and 80–88 cm for women) which is associated with an increased risk of metabolic complications.Table 2. Body composition of sample\r\n \r\n \r\n \r\n Body composition measures\r\n \r\n Females (n=32) Mean (SD)\r\n \r\n Males (n=30) Mean (SD)\r\n \r\n \r\n \r\n Weight (kg) Height (cm) BMI (kg/m2) Waist circumference (cm)\r\n \r\n 75.2 (18.5) 165.2 (7.0) 27.6 (6.6) 87.6 (15.7)\r\n \r\n 89.5 (17.1) 177.4 (6.5) 28.5 (5.4) 100.0 (13.4)\r\n \r\n \r\n \r\nBMI=body mass index.The median dietary intake of participants is shown in Table 3. Of the five heart healthy dietary components, no participants met all five recommendations, 8% met four recommendations, 24% met three recommendations, 25% met two recommendations, another 25% met one recommendation and 18% met none of the recommendations.Single nutrient analyses—Table 3 also summarises non-parametric and post hoc tests with dietary intake, standard of living and education. The only dietary component which varied significantly with standard of living was total fat intake (P=0.018, P adjusted (for income)=0.024). There was no evidence of a difference between total fat intake for the medium and high ELSI ranges. This result remained unchanged after adjustment for BMI, sex and ethnicity (results not shown).Total fat and vegetable intake did not vary significantly by education. Median saturated fat as a % TE was lower in the higher education group, 12.4% TE compared with 14.8% TE, than in the lower education group (P=0.002). Median intake of both groups was higher than the recommended intake of less of than 10% of TE.Table 3. Median (LQ, UQ) food and nutrient intake compared with heart healthy dietary guidelines and median (LQ, UQ) nutrient and food intake by standard of living category and education\r\n \r\n \r\n \r\n Nutrient\r\n \r\n Total fat (% TE)\r\n \r\n Saturated fat (% TE)\r\n \r\n Dietary fibre (g/day)\r\n \r\n Fruit (portions/day)\r\n \r\n Vegetables (portions/day)\r\n \r\n \r\n \r\n Recommended intake for a heart healthy diet\r\n \r\n <35%\r\n \r\n <10%\r\n \r\n >25g\r\n \r\n >2\r\n \r\n >3\r\n \r\n \r\n \r\n Median (LQ, UQ)\r\n \r\n 33.7 [31.2, 37.0]\r\n \r\n 13.5 [11.5–14.8]\r\n \r\n 22.9 [18.2–27.8]\r\n \r\n 1.8 [0.5–2.5]\r\n \r\n 2.0 [1.3–2.7]\r\n \r\n \r\n \r\n Percentage achieving recommended intake\r\n \r\n 57%\r\n \r\n 11%\r\n \r\n 38%\r\n \r\n 48%\r\n \r\n 14%\r\n \r\n \r\n \r\n Median (LQ, UQ) food and nutrient intake by standard of living\r\n \r\n \r\n \r\n Low\r\n \r\n 40.4 [34.9–41.8]\r\n \r\n 15.0 [14.1–16.4]\r\n \r\n 19.8 [18.9–26.9]\r\n \r\n 0.3 [0.1–1.5]\r\n \r\n 1.5 [1.0–1.6]\r\n \r\n \r\n \r\n Medium\r\n \r\n 32.3 [28.7–36.4]*\r\n \r\n 12.3 [11.2–13.8]\r\n \r\n 24.1 [20.5–31.3]\r\n \r\n 2.0 [0.6–3.0]\r\n \r\n 2.0 [0.5–2.5]\r\n \r\n \r\n \r\n High\r\n \r\n 33.7 [31.1–36.9]*\r\n \r\n 13.2 [11.5–14.6]\r\n \r\n 21.9 [18.0–26.1]\r\n \r\n 1.8 [0.8–2.5]\r\n \r\n 2.0 [1.5–3.0]\r\n \r\n \r\n \r\n P Kruskal-Wallis test\r\n \r\n 0.018\r\n \r\n 0.124\r\n \r\n 0.397\r\n \r\n 0.193\r\n \r\n 0.151\r\n \r\n \r\n \r\n Median (LQ, UQ) food and nutrient intake by education level\r\n \r\n \r\n \r\n Lower education\r\n \r\n 35.8 [32.5–38.2]\r\n \r\n 14.8 [12.3–16.1]\r\n \r\n 19.8 [15.7–25.6]\r\n \r\n 1.0 [0.3–2.0]\r\n \r\n 1.5 [1.0–2.0]\r\n \r\n \r\n \r\n Higher education\r\n \r\n 33.2 [30.5–36.5]*\r\n \r\n 12.4 [11.0–13.8]**\r\n \r\n 23.4 [19.7–29.0]*\r\n \r\n 2.0 [0.5–2.8]*\r\n \r\n 2.0 [1.5–3.0]\r\n \r\n \r\n \r\n P Kruskal-Wallis test\r\n \r\n 0.113\r\n \r\n 0.002\r\n \r\n 0.045\r\n \r\n 0.044\r\n \r\n 0.240\r\n \r\n \r\n \r\nNote: * p<0.05; **p<0.01 compared to low standard of living group or lower education group. % TE=percentage of Total Energy intake, LQ = lower quartile, UQ = upper quartile.Principal components analysis—PCA produced one meaningful pattern (Table 6). The first principal component reflected a "higher CVD risk" dietary pattern: a high saturated fat, and low fruit, vegetable and dietary fibre diet based on factor loadings of 0.42, -0.55, -0.50 and -0.46 respectively. A higher score for an individual for the "higher CVD risk" pattern indicates a higher saturated fat intake and a lower fruit, vegetable and fibre intake, all of which are not recommended as part of a heart healthy diet.Table 6: Principal component analysis loadings and importance\r\n \r\n \r\n \r\n Variables\r\n \r\n "Higher CVD risk" pattern\r\n \r\n \r\n \r\n Total fat Saturated fat Dietary fibre Fruit Vegetables Percentage variation explained\r\n \r\n 0.26 0.42 -0.46 -0.55 -0.50 40%\r\n \r\n \r\n \r\nMultiple linear regression analysis on normalised "higher CVD risk" diet scores showed significantly lower scores for those with a higher education level (0.98, 95% CI [0.27–1.69], P=0.008) while there was evidence for a difference in the impact of nutrition literacy on diet by standard of living (P=0.001). The diet score of the low standard of living group increased by 0.35 (95%CI [0.02–0.67], P=0.039) for each one unit increase in nutrition literacy while the diet score of the high standard of living group decreased by 0.10 (95%CI [0.11–0.79], P=0.011) for each increase in nutrition literacy by one unit. Neither sex nor ethnicity had significant effects at the 0.05% level, nor was there a significant difference between the two higher levels of standard of living, the model explained 28% of the variation.The results show that there were significant interaction effects between nutrition literacy and standard of living on the "higher CVD risk" pattern scores. Those in the medium or high standard of living categories, who had better nutrition literacy, had lower scores for the "higher CVD risk" pattern, indicating consumption of a more heart healthy diet, compared to those of medium or high standard of living who had lower nutrition literacy. Conversely, those in the low standard of living category, who had higher nutrition literacy scores, had higher scores for the "higher CVD risk" pattern, indicating consumption of a less heart healthy diet.Figure 1. Association between nutrition literacy and standard of living interaction with a "higher CVD risk" dietary pattern (multiple linear regression, p=0.001)a,b,c,da All of the data included in the multiple linear regression are plotted in the figure; however the regression lines are only relevant for those participants in the respective lower education group; b Low standard of living β=0.35, 95%CI [0.02, 0.67], p=0.039, c Medium standard of living β=-0.23, 95%CI [-0.97 – -0.17], p=0.006, d High standard of living β = -0.10 – 95%CI [-0.79 – -0.11], p=0.011.DiscussionThe results discussed here are based on preliminary analysis from the first wave of the CHALICE study aiming to recruit 1000 50 year olds from Canterbury over the period 2010–2014. This study describes aspects of the dietary intake and the associations with selected social variables.None of the participants in this study met all five heart healthy guidelines, and the majority of participants met two or less guidelines. These results suggest that, as expected, a higher standard of living, education and nutrition literacy were all associated with consuming a heart healthy diet. Using results of the PCA, those who had a medium or high standard of living and who had higher levels of nutrition literacy had higher scores for a heart healthy diet. However, those with a lower standard of living, and who had higher nutrition literacy had pattern scores that corresponded with a less heart healthy diet.Previous studies have shown associations between socioeconomic status and education with selected foods or nutrients2,7,12,14,22,28,29 but there is limited evidence available from New Zealand populations.30 No research, that we are aware of, has used the ELSISF to describe socioeconomic circumstance or standard of living in relation to adult dietary intake. In this population there was a significant pattern toward decreasing fat (as a % TE) intake as standard of living increased. There are no questions in the ELSISF that assess variables that may influence fat consumption such as frequency of purchasing food from outside of the home or access to affordable healthy foods.Studies from abroad have shown that those living in lower socioeconomic areas may struggle to access affordable, healthy, lower fat foods due to the positioning of supermarkets, fruit and vegetable shops and takeaway outlets30,31 and that eating out more often has been associated with higher fat intake.32 It is not known if this is also the case in NZ. In this population there were no significant patterns with the other components of a heart healthy diet; although this may be due to the wide interquartile ranges of some of the components.In addition to assessing associations between socioeconomic status, education and nutrition literacy and single nutrients this study also investigated associations between social variables and a "higher CVD risk" dietary pattern.Dietary patterns, as with intakes of selected nutrients, have been shown to be less healthy in lower socioeconomic groups2 and deprived sectors of the community.7 Dietary modelling (PCA) showed that standard of living, as well as nutrition literacy and level of education, were inversely associated with a "higher CVD risk" dietary pattern. Interestingly, those with a high standard of living tended to consume a less healthy dietary pattern than those with a medium standard of living, however this did not reach statistical significance (P = 0.28).Conversely there was a significant pattern with the lowest standard of living group towards increasing consumption of saturated fat (as a % TE), and lower consumption of fruit, vegetables and dietary fibre as nutrition literacy increased.The relationship between nutrition literacy and dietary intake is often complex. Previous research that has explored this relationship has shown inconsistent results.11,18,19,22,28,31,33 The lack of a significant association in this group may be due to the small sample size of the low standard of living group (seven participants); further research with a larger sample may show a more definitive pattern.This study was designed as an interim analysis to investigate whether an interpretable heart healthy dietary pattern could be obtained from this small dataset comprising the first participants enrolled into CHALICE. The analysis was also a pre-planned part of our quality assessment strategy, in particular with the dietary instruments, data entry and analysis of the food records.While the study had sufficient sample size to derive a meaningful, robust dietary pattern,27 we acknowledge that a sample size of around 60 participants severely restricts our ability to detect significant differences in individual dietary components. However as the participants are all fifty years of age the overall variability within the dietary data will be smaller than for a similar sample over a larger age range.The Canterbury earthquake sequence began on 4 September 2010 with the second, most destructive, and only fatal major episode on 22 February 2011. The data presented was all collected prior to the 22 February 2011 and is in our view the best representation of diet of 50 year olds unaffected by the earthquakes.Future analyses will attempt to quantify the effect of the earthquake, if any. While we would have preferred more participants the first 62 form a natural grouping. Nevertheless, these findings show that it is possible to generate a heart healthy dietary pattern in smaller samples and we once the final CHALICE recruitment is completed we will explore these relationships further, using more sophisticated analyses.This study provides evidence that the 50 year olds from Canterbury included in these analyses do not consume a heart healthy diet and that socioeconomic status, education and nutrition literacy may influence dietary consumption. If these findings are found to be similar in the whole CHALICE sample then this is particularly pertinent for Canterbury which has an aging population, hence identifying modifiable risk factors in midlife has the potential to better target interventions for this increasing proportion of the population.This study provides further evidence that improving dietary intake through improving the social and financial circumstances of a population may help reduce health inequalities and the burden of disease.This article was corrected on 30 January 2015 as outlined in the Erratum published the same day athttp://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1408/6427\r\n

Summary

Abstract

Aim

Cardiovascular disease is a leading cause of death in New Zealand, but risk factors may be decreased by consuming a heart healthy diet. This pilot study investigated whether participants met the guidelines for a heart healthy diet and whether a novel heart healthy dietary pattern could be identified using principal components analysis (PCA). The second aim of this project was to assess if higher education, standard of living and nutrition literacy are associated with a heart healthy dietary pattern.

Method

This exploratory study was undertaken using data from the first participants enrolled in the Canterbury Health Ageing and Lifecourse study: an observational study of 50 year olds in the Canterbury District Health Board region. Eighty-two people were selected from the General and Mori electoral role and interviewed prior to the 22 February 2011 Christchurch Earthquake. PCA was conducted to identify dietary patterns, based on intake of specific nutrients as indicated by the New Zealand and international heart healthy dietary guidelines.

Results

62 participants completed questionnaires and an estimated food record. No participants met all five of the heart healthy dietary guidelines. One dietary pattern was produced by PCA: a higher CVD risk pattern. Regression analysis indicated that higher standard of living, education and nutrition literacy were inversely associated with a higher CVD risk pattern.

Conclusion

Higher standard of living, education and nutrition literacy were associated with a healthier dietary eating pattern. However, as no participants met all the dietary recommendations more education and support is needed to help people meet these.

Author Information

Acknowledgements

Correspondence

Correspondence Email

Competing Interests

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE