Nutrition “fat facts” are not common knowledge
Abstract
Issue addressed
Individuals who are knowledgeable about nutrition are more likely to eat healthily. Yet, few studies have investigated levels of nutrition knowledge using a validated tool. The present study measured nutrition knowledge using the Re-examined General Nutrition Knowledge Questionnaire (R-GNKQ) to confirm influencing demographic characteristics.
Methods
Adults aged 18-60 years were recruited. Nutrition knowledge was assessed using the R-GNKQ, examining four domains (dietary guidelines, sources of nutrients, choosing everyday foods, and diet-disease relationships) with 96 questions.
Results
Of 606 respondents (mean age 38.8 ± 11.8 years), 506 completed all questions. R-GNKQ score was positively associated with education (p<0.001) and age (p<0.001). Those with the highest education levels scored higher across 89% of the R-GNKQ and the oldest (≥50 years) respondents scored higher than younger respondents. Other characteristics that were associated with higher levels of knowledge were being female, and having a healthy BMI. Lowest knowledge pertained to questions about diet-disease relationships and fatty acids in foods.
Conclusions
The majority of individuals had a good understanding about the Australian Dietary Guidelines, however the health benefits of adhering to the dietary guidelines was less well understood. Gaps in knowledge pertained to the specific details of how to adhere to the guidelines, particularly knowledge about the types of beneficial fats and their everyday food sources.
So what?
Those with lower educational attainment, younger, males and those with higher BMI's would benefit from nutrition communications that identify foods with beneficial fatty acids to assist with adherence to the Australian Dietary Guidelines.
1 INTRODUCTION
Diet is a behavioural risk factor that influences the health of every Australian.1 Poor nutrition now accounts for the greatest proportion of our national burden of disease.2, 3 It has been estimated that over a third of premature deaths in Australia are the result of preventable, diet-related chronic diseases, including cardiovascular disease, type 2 diabetes and specific cancers.1 Consequently, the World Health Organisation advocates the promotion of diversified, balanced, and healthy dietary intakes4 and the Australian Dietary Guidelines provide nutrition recommendations to optimise intakes.1 Yet the latest Australians national nutrition survey identified that over a third of mean daily energy intakes come from energy-dense, nutrition-poor foods.5 This is a major public health concern. However, improving eating patterns is challenging as food choices are influenced by a multitude of factors such as taste, cost, convenience, availability, culture, religious beliefs and nutrition knowledge.6 Therefore, it is important to understand the factors that positively influence dietary intake to help improve the health of Australians.
Nutrition knowledge plays a valuable role in influencing dietary intake.7 Individuals with the highest level of general nutrition knowledge are up to 25 times more likely to eat healthily, compared to those with the lowest level of knowledge.7 This finding is among the few studies to have used valid, reliable tools to assess both nutrition knowledge and dietary intake.8-10 Due to development of the “General Nutrition Knowledge Questionnaire” (GNKQ) a psychometrically valid and reliable instrument (Cronbach's alpha 0.97, test–retest reliability 0.98)11 nutrition knowledge has been measured in selected populations.12-16 Results from studies, using the GNKQ across several countries have identified that knowledge does vary between individuals, tending to be lower among those who are younger, male and/or of lower educational attainment.17 However, limited attention has been given to describing how overall knowledge varies and what knowledge gaps exist both within and across population groups. Identifying knowledge gaps is important, as addressing differences through targeted nutrition education initiatives can support individuals in adopting healthier dietary habits.18, 19
Knowledge contributes to the rationale, motivation and capability for behaviour change.20 Hence national nutrition education campaigns are designed to improve nutrition knowledge in order to help persuade individuals to adopt healthier dietary habits.19 However, it is important to recognise that an individual's existing knowledge and understanding influences their interpretation of new nutrition information. For example, the Australian ‘Go for 2&5’ fruit and vegetable campaign addressed Australians’ lack of awareness that most fruit and vegetable intakes were lower than recommended, as indicated by national dietary surveys.21 Despite the simplicity of the ‘Go for 2&5’ message, individuals who believed that eating one serve of fruit and three serves of vegetables a day was sufficient for good health perceived ‘Go for 2&5’ as aspirational rather than a realistic target message, which diminished its effectiveness.22 Gaining greater insight into current nutrition knowledge may assist in developing tailored and effective nutrition communications.
Today the popularity of social media exposes the community to a diversity of nutrition communications from sources with variable credibility. It is unclear whether this has influenced nutrition knowledge within the general community. Given the importance of improving dietary intake, it appears timely to revisit current nutrition knowledge. The GNKQ has been adapted, updated and psychometrically validated (Cronbach's alpha 0.94, test-retest reliability 0.92) for an Australian population.23, 24 However, the use of the revised version (R-GNKQ) in Australia has been limited to investigating nutrition knowledge in a narrow demographic group where respondents were predominately elite athletes (50.9%), and younger than 25 years (83%).24 The aim of the current study was to use the R-GNKQ to evaluate general nutrition knowledge in a broader demographic sample of Australians that includes a wider age, level of education and socioeconomic status range than previously studied. A secondary aim was to examine current demographic characteristics associated with nutrition knowledge.
2 METHODS
2.1 Study design
Men and women aged 18 to 60 years were invited to complete an anonymous, on-line self-administered survey collecting general demographic information including sex, year of birth, postcode, education attainment, employment status, income rating and health status. Other questions included weight, height, weight-monitoring habits, knowledge of nutrition, general well-being, physical activity and eating behaviours. The respondent's post code data were cross-referenced with the Socio-economic indexes for Areas (SEIFA) database to obtain a measure of socio-economic status.25 The effects of demographic characteristics on nutrition knowledge were then analysed.
2.2 Recruitment
Recruitment methods included distribution of information flyers at medical clinics and social media strategies such as Facebook posts, university blog page, radio interviews and tweets. Eligibility criteria included proficiency in English and access to the Internet. Respondents could elect to receive healthy lifestyle information and go in a draw to win 1 of 10 $50 grocery vouchers. The University of Newcastle Human Research Ethics Committee approved the survey and completion implied consent.
2.3 Data collection
The survey was administered, using an online survey tool (SurveyMonkey Inc 1999-2016). Nutrition knowledge was assessed using the R-GNKQ. Across 96 items, four domains of nutrition knowledge were measured: dietary guidelines (11 items), sources of nutrients (62 items), choosing everyday foods (5 items) and diet-disease relationships (18 items).24 The instrument questions were predominately multiple-choice format, with some, using open-ended comment boxes. The R-GNKQ (Appendix S1) was scored using methods developed previously (Appendix S2), with one point allocated for each correct item and no deductions for incorrect or ‘not sure’ answers, with a maximum score of 96. Within the diet-disease relationships section, five items allowed no response.
2.4 Statistical analysis
Demographic characteristics are presented as number (%): differences between those respondents who completed the R-GNKQ (completers) and those who did not (non-completers) were examined, using chi-square tests. Internal consistency for the overall nutrition knowledge score (R-GNKQ) was measured with Cronbach's Alpha (0.86). Adjusted associations between demographic characteristics and overall nutrition knowledge score (R-GNKQ), and each domain, were examined using linear regression. Assumptions for linear regression were checked (unusual and influential data, normality of residuals, heteroscedasticity, multicollinerarity, using Variation Inflation Factor (VIF), and non-linearity). For examining differences in knowledge about nutrients in foods, this domain was sub-divided into eight categories by nutrient: carbohydrate, added sugar, total fat, salt, protein, fibre, fat types and miscellaneous (limited to two questions: calcium in milk; vitamins and minerals in bread). The statistical significance threshold was set at P < .05 and analysis was conducted, using Stata software, version 13.1 (StataCorp. 2013 College Station, TX: StataCorp LP).
3 RESULTS
3.1 Recruitment
Recruitment was conducted from January 2014 to May 2015, with 606 attempting the survey and 506 completed all sections (termed “completers”) and included in the current analysis. Social media channels were the most effective method of recruitment (57%) across all age groups, with the remaining participants recruited using traditional approaches. Across recruitment methods, two sex differences were identified with a higher percentage of females (37.5%) responding to Facebook compared to males (26%)(P = .028), while a higher percentage of males (12.5%) responded to a radio interviewer's tweet, compared to females (5%)(P = .024).
3.2 Participant characteristics
Demographic characteristics of survey respondents (mean age 38.8 ± 11.8 years) are summarised in Table 1. R-GNKQ “completers” had a mean (SD) age 39.3 (11.6) years, range 18-60 years. Completers were predominately female (82%), had a range of BMI levels (Table 1), were well educated (51% university degree) and employed, had higher income levels and little socio-economic disadvantage. More males (P = .03) and younger respondents (P = .03) were non-completers.
| Demographic characteristic | Group | Completers (n = 506) | Non-completers (n = 100) | P-value |
|---|---|---|---|---|
| Age (years) | <35 | 182 (36%) | 50 (50%) | .025b |
| ≥35 - ≤49 | 204 (40%) | 29 (29%) | ||
| ≥50 | 120 (24%) | 21 (21%) | ||
| Sex | Female | 417 (82%) | 73 (73%) | .029b |
| Male | 89 (18%) | 27 (27%) | ||
| Body mass index (kg.m−2) | 18.5 - < 25 | 220 (43%) | 28 (28%) | .310 |
| ≥ 25 - < 30 | 123 (24%) | 26 (26%) | ||
| ≥ 30 | 142 (28%) | 23 (23%) | ||
| Not reported | 21 (4%) | 23 (23%) | ||
| Education | High school not completed | 27 (5%) | 8 (8%) | .672 |
| High school completed | 80 (16%) | 15 (15%) | ||
| Vocational training | 140 (28%) | 26 (26%) | ||
| University degree | 259 (51%) | 45 (45%) | ||
| Not reported | 0 (0%) | 6 (6%) | ||
| Employment | Employed | 395 (79%) | 73 (78%) | .913 |
| Other | 105 (21%) | 20 (22%) | ||
| Not reported | 6 (1%) | 7 (7%) | ||
| Income | Low | 179 (35%) | 36 (38%) | .587 |
| High | 327 (65%) | 58 (62%) | ||
| Not reported | 0 (0%) | 6 (6%) | ||
| SEIFAa percentile | Low (1-33) | 124 (25%) | 30 (32%) | .430 |
| Medium (34-66) | 181 (37%) | 33 (35%) | ||
| High (67-100) | 186 (38%) | 32 (34%) | ||
| Not reported | 15 (3%) | 5 (5%) |
- a Socio-economic Indexes For Areas.25
- b significant P-values at the P < .05 threshold using Chi squared test.
3.3 Nutrition knowledge
Overall nutrition knowledge was found to be high with a mean of 73% of questions answered correctly, R-GNKQ mean (SD) 70.5, (11.1) and almost all completers (97%) answered over half of the questionnaire correctly. Significantly, more of the highest scorers (R-GNKQ score 73-96) were university educated (56%) compared to those who completed high school (33%), had vocational training (31%) or had not completed high school (7%)(P < .001). This equated to university-educated respondents correctly answering a mean of nine questions more than those who had not completed high school (P < .001). Age was a factor associated with nutrition knowledge, with older respondents (≥ 50 years) scoring higher than the youngest group (P < .001). Weaker associations for differences in total knowledge scores were found for sex (P = .049) and BMI status (P = .045). Males scored lower than females, and those with a BMI ≥30 had lower scores than those with a healthy BMI (18.5-25). Employment, income and SEIFA percentile were not associated with total R-GNKQ scores (Table 2).
| Demographic characteristic | Group | β (95% CI) | P-value |
|---|---|---|---|
| Age (years) | <.001a | ||
| ≥35 - ≤49 | 2.38 (0.04, 4.7) | .046a | |
| ≥50 | 5.28 (2.6, 8.0) | <.001a | |
| Sex | Male | −2.48 (−4.9, −0.0) | .049a |
| Body mass index (kg.m−2) | .045a | ||
| 25-<30 | −2.02 (−4.4, 0.4) | .099 | |
| ≥30 | −3.59 (−6.1, −1.1) | .005a | |
| Not reported | −1.73 (−6.5, 3.1) | .479 | |
| Education | <.001a | ||
| High school completed | 1.68 (−3.3, 6.7) | .510 | |
| Vocational training | 2.68 (−1.8, 7.2) | .244 | |
| University degree | 9.02 (4.5, 13.6) | <.001a | |
| Employment | Other | 0.70 (−1.7, 3.2) | .573 |
| Income | High | 0.31 (−1.7, 2.3) | .762 |
| SEIFAb percentile | .935 | ||
| Medium (34-66) | −0.33 (−2.8, 2.1) | .786 | |
| High (67-100) | −0.45 (−2.9, 2.0) | .717 |
- Reference categories: Age < 35; Female; BMI 18.5 - <25; Education - High school not completed; SEIFA percentile low (1-33); Employed; Low income.
- a Significant P-values at the P < .05 threshold.
- b Socio-economic Indexes for Areas.25
3.4 Domain-specific knowledge
Of the four domains “dietary guidelines” had a high proportion of questions answered correctly (79%) and was the only domain not to be associated with any demographic characteristic. In contrast, the diet-disease relationships domain had the lowest proportion of questions answered correctly (58%) and had the most factors associated with it (age, BMI, education). Higher knowledge of diet-disease relationships was found among those who were oldest (≥50 years)(P = .009), of a healthy BMI (18.5-25)(P = .021) and university-educated (P < .001). Employment and SEIFA percentile were the only factors not to be associated with any of the four knowledge domains (Table 3). All domain scores were positively correlated with each other (P < .001).
| Domaina | Dietary guidelines | Sources of nutrients | Choosing everyday foods | Diet-disease relationships | |||||
|---|---|---|---|---|---|---|---|---|---|
| R-GNKQ domain scoreb: mean (SD) | 8.7 (1.7) | 47.0 (7.9) | 4.4 (0.8) | 10.4 (3.2) | |||||
| Demographic characteristic | Group | β (95% CI) | P-value | β (95% CI) | P-value | β (95% CI) | P-value | β (95% CI) | P-value |
| Age (years) | .088 | .001d | .096 | .009d | |||||
| ≥35 - ≤49 | −0.15 (−0.5, 0.2) | 2.20 (0.6, 3.8) | <.009d | 0.04 (−0.2, 0.2) | 0.29 (−0.4, 1.0) | .419 | |||
| ≥50 | 0.30 (−0.2, 0.8) | 3.54 (1.6, 5.5) | <.001d | 0.21 (0.0, 0.4) | 1.22 (0.4, 2.0) | .004d | |||
| Sex | Male | 0.04 (−0.4, 0.4) | .858 | −1.7 (−3.4, 0.0) | .052 | −0.29 (−0.5, -0.1) | .003d | −0.50 (−1.2, 0.2) | .185 |
| Body mass index (kg.m−2) | .220 | .193 | .339 | .021d | |||||
| 25- <30 | −0.11 (−0.5, 0.3) | −1.2 (−2.9, 0.5) | −0.09 (−0.3, 0.1) | −0.58 (−1.3, 0.1) | .117 | ||||
| ≥30 | −0.39 (−0.8, 0.0) | −1.9 (−3.7, −0.1) | −0.14 (−0.3, 0.1) | −1.2 (−1.9, −0.4) | .003d | ||||
| Not reported | 0.21 (−0.6, 1.0) | −1.0 (−4.4, 2.4) | 0.13 (−0.2, 0.5) | −1.1 (−2.5, 0.4) | .142 | ||||
| Education | .281 | <.001d | .034d | <.001d | |||||
| High school completed | −0.32 (−1.1, 0.5) | 0.8 (−2.7, 4.3) | .665 | 0.10 (−0.3, 0.5) | .628 | 1.1 (−0.4, 2.6) | .145 | ||
| Vocational training | −0.23 (−1.0, 0.5) | 1.5 (−1.7, 4.7) | .355 | 0.18 (−0.2, 0.5) | .318 | 1.2 (−0.1, 2.6) | .074 | ||
| University degree | 0.05 (−0.7, 0.8) | 6.1 (3.0, 9.3) | <.001d | 0.35 (−0.0, 0.7) | .053 | 2.5 (1.1, 3.8) | <.001d | ||
| Employment | Other | 0.06 (−0.3, 0.5) | .763 | 0.2 (−1.5, 2.0) | .784 | −0.07 (−0.3, 0.1) | .451 | 0.5 (−0.3, 1.2) | .207 |
| Income | High | −0.01 (−0.3, 0.3) | .931 | 6.1 (−1.2, 1.7) | .750 | 0.18 (0.0, 0.3) | .032d | −0.1 (−0.7, 0.5) | .802 |
| SEIFAcpercentile | .438 | .808 | .695 | .541 | |||||
| Medium (34-66) | −0.24 (−0.6, 0.2) | −0.5 (−2.2, 1.2) | 0.03 (−0.2, 0.2) | 0.4 (−0.3, 1.1) | |||||
| High (67-100) | −0.23 (−0.6, 0.2) | −0.5 (−2.2, 1.2) | −0.04 (−0.2, 0.2) | 0.3 (−0.4, 1.0) | |||||
- a Each domain is a subset of questions related to general nutrition knowledge about this topic.
- b Domain (maximum score): Dietary Guidelines (11), Sources of Nutrients (62), Choosing Everyday Foods (5), Diet-disease Relationships (18).
- Reference categories:
- Age < 35; Female; BMI 18.5 - <25; Education - High school not completed; SEIFA percentile low (1-33); Employed; Low income.
- c Socio-economic Indexes for Areas.25
- d Significant P-values at the P < .05 threshold.
3.5 Specific nutrient knowledge
The “sources of nutrients” domain makes up 65% of the questionnaire and examining responses for the eight “sources of nutrients” sub-scores identified that the highest percentage of correct answers were for questions relating to carbohydrates (87%), protein (85%), added sugar (83%), salt (83%), fibre (80%) and two miscellaneous questions (70%). The lowest scores were for questions about total and types of fatty acids in foods, with correct response rates decreasing to 69% and 63% respectively. Education was the only factor related to all eight “sources of nutrients” sub-scores with those university-educated scoring higher than the least educated group. Age and sex were associated with scores for fibre and fat. Both older groups scored better than the youngest group (< 35 years) for the fibre questions (P < .01). The oldest group also correctly answered more questions on types of fatty acids in food (P < .001). In contrast, males scored slightly lower than females on knowledge about the fat content of foods (P = .005).
The R-GNKQ targets dietary fat knowledge over 34% of the questionnaire and lower correct response rates were identified as more specific fatty acid knowledge was required. For example, 93% of respondents correctly answered dietary recommendations advocate for a lower intake of fatty foods, 87% knew to reduce saturated fat intake with 66% answering saturated fat raises serum cholesterol levels, whereas 61% knew to consume reduced fat dairy foods. Fewer identified red meat (57%) and dairy products (40%) as sources of saturated fat. Only half knew (50%) sunflower margarine is a low source of saturated fat and fewer respondents correctly identified olive oil as a high source of monounsaturated fat (36%). Only 40% were aware that fat is the macronutrient with the highest kilojoules/gram, and that butter has a comparable total fat content to regular margarine (35%) and polyunsaturated margarine (44%) (Appendix S3).
4 DISCUSSION
The current study provides a contemporary evaluation of nutrition knowledge in Australia across a more broad population sample of adults than previously studied. It draws attention to current levels of nutrition knowledge to describe differences and identify gaps in understanding. The majority had a good understanding of the Australian Dietary Guidelines, particularly of the broad concepts of eating more healthful foods such as vegetables and less fatty, sugary and salty foods. However, only those with the highest education levels had a high level of knowledge across all nutrition domains. Other characteristics that were associated with higher levels of knowledge were older age, being female, and having a healthy BMI. Lowest knowledge scores pertained to questions about diet-disease relationships, which suggests that the rational for the Australian Dietary Guidelines being to optimise health and prevent disease is less well understood, particularly amongst those who are younger, less educated and for individuals with obesity. Other specific knowledge gaps were evident, especially regarding fatty acids in foods. These results suggest that general important nutrition messages are reaching the community, especially related to the key concepts of the dietary guidelines. However, there remains scope to improve nutrition knowledge amongst those with lower educational attainment, younger, males and those with higher BMIs.
The results from the current study are consistent with the findings of others.8, 13 In Canada, confusion about dietary fat recommendations has been identified with limited understanding of the role of fatty acids in promoting health.26, 27 These authors argued that there is a role for specific, quantifiable and tailored nutrition guidelines, consistent with consumer requests for practical advice to improve eating habits.28 In the current study, individuals with obesity had significantly lower total knowledge scores and the lowest level of knowledge about diet-disease relationships. Similarly, an Italian study (n = 744, 52.1 years ± 9.44, 50.3% male) found associations between nutrition knowledge and obesity, reporting the odds ratio of being obese decreased with higher levels of nutrition knowledge (OR = 0.61, 95% CI: 0.40-0.95 for highest versus lowest nutrition knowledge group).13 Importantly, higher knowledge was also associated with consuming a higher quality diet.13 It is plausible that lower knowledge about diet-disease relationships and fatty acids, identified in the current study, could impact on both capability and motivation for adherence to dietary guidelines. Given that obesity is associated with a higher risk for and prevalence of chronic disease29 this is an area of research that merits further attention.
Limitations of the current study are acknowledged. Due to potential recruitment bias caution is warranted in extrapolating this data to other populations. Findings are likely to indicate “best-case” results due to the survey population being older, relatively well educated and predominately female. Other researchers have experienced similar recruitment bias.7, 8, 30 However, non-representative samples have been found to contribute reliable estimates of relative risk when examining within group comparisons,31 with survey non-response previously shown to not seriously bias analysis even when more women of higher social class responded.32 The method of recruitment unintentionally favoured women, with Facebook posts contributing to a higher response rate. It is recommended that future nutrition knowledge surveys target recruitment to increase participation of males, younger individuals and those with lower levels of education. For example, in the current study, males responded more favourably to tweets from a radio interviewer, hence it may have been beneficial to conduct more radio interviews to increase male response rates. Although the level of literacy required to complete the knowledge survey has not been assessed, its length, as well as the on-line nature of the survey requires at least reasonable literacy, and computer and Internet access. This may explain the higher participation of university-educated, higher-income respondents in the current study.
It was unexpected that socio-economic status did not explain any of the variation in nutrition knowledge, as identified in other studies.7, 8, 13 This lack of association may be explained by the method used to determine socio-economic status. It was inferred from postcode data, and so relied on the collective socio-economic characteristic of a given area, instead of that of the individual.25 Furthermore, the scoring method for the R-GNKQ requires updating to acknowledge the updated evidence base regarding the relationship between higher intakes of free sugars and greater risk of obesity,33 and higher salt intake with increased risk of coronary heart disease and stroke.29
5 CONCLUSION
Knowledge is an essential predisposing component for change34 and there is growing evidence to demonstrate that being informed about nutrition influences food choices and changes in dietary intake.13, 35 This study supports the need to move beyond promoting general nutrition recommendations to clarifying the health benefits of adhering to the dietary guidelines, the need to choose reduced fat dairy foods, explicitly stating the weekly red meat intake recommendations in terms of amounts, and the types of beneficial fats and their everyday food sources. For example, rather than simply promoting a reduction in saturated fat intake, also including specific information about appropriate food swaps, such as using avocado instead of butter. It is plausible that tailored nutrition communications aimed at addressing these knowledge gaps and targeted to those most in need may contribute to adherence to the Australian Dietary Guidelines.
ACKNOWLEDGEMENTS
We wish to thank Gillian Hendrie for providing permission to use the R-GNKQ and Susan Heaney for providing unpublished methods for instrument scoring. A special acknowledgement to the late Professor Jane Wardle, who designed the original UK version of the GNKQ and previously approved its use in our research. Thanks to Kerrin Palazzi for her contribution to the statistical analysis.
CONFLICT OF INTEREST
The authors declare that there are no funding or conflicts of interest in connection with this article.
AUTHOR CONTRIBUTIONS
CH was the principle author, responsible for conception, completion and drafted the initial manuscript. All authors revised and approved the final manuscript.
