Abstract
Objective:
School-based nutrition education can promote the development of skills, such as food label reading, that can contribute to making healthier food choices. The purpose of this study was to assess the reliability of a Russian language version of the previously validated Food Label Literacy for Applied Nutrition Knowledge (FLLANK) questionnaire and also to evaluate the impact of school-based nutrition education programmes on nutrition knowledge and food label–reading skills in Russian adolescents and young adults.
Methods:
Two nutrition education study programmes were administered to two separate groups of students – at a middle school (n = 23) and a university (n = 17). The study programmes consisted of a single 2-hour classroom session and emphasised basic nutrition and food components. The FLLANK questionnaire was administered to both groups of students before and after the intervention to assess change in food label–reading skills. Cronbach’s α was calculated for the post-test scores to assess the instrument’s internal consistency.
Results:
Middle school students increased their mean FLLANK scores from 54% to 95% (p < .01). University students increased their scores from 46% to 91% (p < .01). In the school group only, there was an interaction between time and gender, with boys improving significantly more than girls (p < .05). The Cronbach’s α estimates were low for both groups: −.15 and .03 for school and university, respectively.
Conclusion:
Food label literacy improved among Russian adolescents and young adults after the implementation of nutrition education curricula. The translated version of the FLLANK retained its utility for capturing change in food label literacy, but performed very poorly with regard to internal consistency. Additional research with larger sample size and variance of scores to assess reliability is needed.
Introduction
Overweight and obesity are more common than underweight in all income groups in Russia, including food insecure people in the lowest income decile (Liefert, 2004). Among adult men and women in Russia, the prevalence of overweight increased from 34.3% in 1994 to 36.1% in 2004, while the prevalence of obesity increased from 20.3% to 28.0% (Huffman and Rizov, 2007). It is hypothesised that these changes occurred in response to changing food consumption practices in Russia, related to economic adjustments made by households in the recent free market economy (Huffman and Rizov, 2007; Liefert, 2004).
In Russia and in other countries that have recently undergone economic transition with associated significant dietary changes, there is a dual burden of underweight and overweight within the population (Doak et al., 2005; Wang et al., 2002; Wang and Lobstein, 2006). Therefore, it may be preferable to implement public health programmes that emphasise making nutritious food choices rather than weight loss per se.
Food label reading has been associated with higher quality diets (Graham and Laska, 2012; Kreuter et al., 1997; Neuhouser et al., 1999). Although reported use of food labels is high in many countries, understanding of food labels is comparatively low (Besler et al., 2012; Grunert and Wills, 2007; Lubman et al., 2012). In the Russian Federation (RF), food labels are mandatory by Federal Law, but the font used to label items tends to be very small. School-based nutrition education may improve understanding and use of food labels, and overall diet quality among adolescents and young adults. There are currently no food label education programmes in Russia.
It is also important to evaluate the effectiveness of educational programmes with appropriate instruments (Alsaffar, 2012; Sapp and Jensen, 1997). The purpose of this study was to assess the reliability of a previously validated instrument that has been adapted and translated for use in a Russian population. In addition, we evaluated the impact of school-based nutrition education programmes on nutrition knowledge and food label–reading skills in Russian adolescents and young adults. The instrument was a Russian language adaptation of the Food Label Literacy for Applied Nutrition Knowledge (FLLANK) questionnaire (Reynolds et al., 2012).
Methods
Participants
At the beginning of the school year in 2009, first-year medical students at the Moscow State University of Medicine and Dentistry chose one of three elective courses, which included a course on the basic principles of nutrition. Information about these elective courses was posted on the website of the Dean’s office. Students who elected to take the nutrition course were provided study programme details during the course by the course lecturer. In middle school, students and parents were informed about the programme in the beginning of the school year, and information about the programme was also posted on the school website. The programme was offered in addition to their course work. The students who participated were in 9th and 10th grades (corresponding to 10th and 11th grades in the US school system).
Intervention
Two programme curricula (middle school and university) were developed by the first author (KG). The programmes were designed to teach the basic principles of nutrition, with an emphasis on the contributions of major and minor components of food to health. The classes covered topics such as calories, vitamins and minerals, food fibre and food additives (Gurevich et al., 2005; Leshkevich et al., 2006). In addition, course content was adapted from the International Life Sciences Institute (ILSI, 1998) Europe Concise Monograph Series Healthy Lifestyles Nutrition and Physical Activity. All materials were age-adapted.
The middle school curriculum was based on the published work ‘Teaching a Healthy Lifestyle’ ’ (Gurevich et al., 2005). The university curriculum was based on the published work ‘Forming of Healthy Life-style in School’ (Leshkevich et al., 2006). Both study programmes were delivered by KG to students over a 2-hour classroom session that took place in one single day.
Students completed the programmes in the 2009–2010 academic year. The middle school curriculum was delivered to students at a Moscow school. The university curriculum was delivered at Moscow State University of Medicine and Dentistry as an elective course. All first-year students in medical school at the University were eligible to participate in the study. All participants were tested twice: immediately before the education course and 1 week after the programme was administered.
Change in nutrition knowledge was assessed using a Russian language version of the FLLANK questionnaire (Reynolds et al., 2012). The FLLANK has been shown to be a valid and reliable tool to assess food label literacy in elementary school age children (Reynolds et al., 2012). The English language version of the tool was designed to measure the ability to differentiate between more nutritious and less nutritious foods based on the information on Nutrition Facts panels (US Food and Drug Administration, 2004) and ingredient lists on packaged food products. It was first used to assess change in food label literacy in children exposed to the Nutrition Detectives™ (ND) school-based education programme (Katz, 2005), which aims to improve food label–reading skills. Each questionnaire item presents a pair of Nutrition Facts panels and ingredient lists representing one of six types of packaged food products (breads, crackers, cereals, cereal bars and cookies). Students are asked to identify the more healthful choice. Each item has three response options: label A, label B or ‘don’t know’. The total score is the percentage of products identified correctly. The original FLLANK and scoring instructions are available online (Katz and Katz, 2009). The questionnaire was administered before and after the intervention. Even though FLLANK has been previously validated for primary school children, its utility is applicable to the general public.
The Russian language FLLANK was translated by the Russian study team with consultation provided by the Russian National Research Institute of Nutrition (RNRIN). The FLLANK was translated from English to Russian by two independent translators. The two resulting translated FLLANK were reviewed by RNRIN, and a final version was developed. This final version was then sent to two independent translators who translated the FLLANK back to English. The food items used to produce the food labels on the FLLANK were not changed when translated from English to Russian. Only several minor components not registered in Russia were omitted from the translated version of the FLLANK.
The study was approved by local ethical committee at the Moscow State University and Dentistry. All participants provided oral inform consent in order to participate in the study. 1
Statistical analyses
To assess the internal consistency of the translated version of the FLLANK in each sample, Cronbach’s α was determined from post-test item responses. Demographic characteristics of each sample were summarised with frequencies and proportions or means and standard deviations, as appropriate. A paired t-test was used to determine the changes in overall FLLANK score from pre-test to post-test in each sample. Post hoc tests were conducted to assess the influence of within-group factors such as gender and age through the use of repeated measures analysis of covariance (ANCOVA). All analyses were based on intent-to-treat principle. The statistical software package SPSS version 15.0 was used to conduct all analyses. All tests were conducted with a two-tailed alpha level set at .05.
Results
Study population characteristics
A convenience sample of 23 students in the middle school group participated. The average age of students was 16.2 (± 0.8) years, and the majority of school students were female (70%). 2 A total of 17 students in the university group received the nutrition programme. The average age of university participants was 17.6 (±0.7) years. The majority of the university students were female (59%) (Table 1).
Baseline characteristics of school and university participants.
SD: standard deviation.
FLLANK internal consistency
At post-test, Cronbach’s α for the administration of the Russian FLLANK in the ‘school’ group was −.15. The ‘university’ groups’ item responses at post-test had a Cronbach’s α of .03 (Table 2).
Participants’ correct FLLANK item responses by administration.
FLLANK: Food Label Literacy for Applied Nutrition Knowledge.
Data in Table 2 are presented as the number (and percentage) of participants who identified the correct label for each FLLANK test item.
Nutrition knowledge
At pre-test, participants were able to identify 54% of the more nutritious choices on the FLLANK. At post-test, participants were able to identify 95% of the more nutritious choices. The mean within-subject increase in FLLANK score was 41% and was statistically significant (p < .01) (Table 3). In ANCOVA, a significant interaction between intervention and gender was observed (p < .05). There was no significant interaction between intervention and age (p = .65).
FLLANK total scores by administration.
FLLANK: Food Label Literacy for Applied Nutrition Knowledge; SD: standard deviation.
p<0.01.
At pre-test, university students were able to identify 46% of the more nutritious choices on the FLLANK. At post-test, participants were able to identify 91% of the more nutritious choices. There was significant interaction observed between intervention and gender in the ‘school’ students, which was a reflection of significant differences between males and females in the magnitude of improvement in nutrition knowledge from baseline (males: 54.3% ± 17.2 vs females: 35.6% ± 13.1; p < .01). The mean within-subject increase in FLLANK score was 45% and was statistically significant (p < .01) (Table 3). There were no significant interactions between intervention and gender (p = .92) or intervention and age (p = .16) in university students.
Discussion
This study demonstrated that the Russian FLLANK questionnaire could be used to evaluate programmes taught in Russian school and university settings by measuring the construct of ‘food label literacy’. Low scores at the pre-test in both the school (54% correct) and university (46% correct) groups demonstrated that students could not readily identify more nutritious food choices in a majority of the item pairs presented. This is fairly consistent with the findings of a study published in 2012, which assessed food label use among former Soviet Union adult immigrants in the USA (Lubman et al., 2012). In that study, Lubman et al. (2012) found that 41% of the sample indicated that they read food labels ‘not often/never’ and that 67% of the sample had ‘low label-reading skills’.
At follow up, students in both groups correctly identified more than 90% of the healthier choices on the FLLANK. Students at both the school and university levels significantly increased their performance on the FLLANK by greater than 40%. Not only do these results demonstrate an increase in food label literacy among students in both groups following a nutrition education intervention but they also confirm that the Russian FLLANK was able to retain its validity as a measure of the construct ‘food label literacy’ (Reynolds et al., 2012). Finally, the significant interaction observed between intervention and gender in the school students was a reflection of significant differences between young men and young women in the magnitude of improvement in nutrition knowledge from baseline (males: 54.3% ± 17.2 vs females: 35.6% ± 13.1; p < .01). At pre-test, female students had significantly higher scores than male students in the school group (10.5 ± 31.7, p = .033). So while young men showed greater improvement overall, young women had greater baseline knowledge when tested. Most previous studies of nutrition education interventions in children and young people have not assessed gender differences in nutrition knowledge (Anderson et al., 2005; Warren et al., 2003). In the Pathways Study, which evaluated a school-based obesity prevention programme in US American Indian children, there were improvements in nutrition and physical activity knowledge in both boys and girls in the intervention group, with no significant differences between genders (Stevens et al., 2003). In contrast to our findings, a study conducted among Indian school children observed greater improvement in nutrition-related knowledge in girls compared with boys following a health and nutrition education programme (Shah et al., 2010).
In our study samples, the FLLANK was found to have low internal consistency, as evidenced by values of Cronbach’s α (−.15 in the ‘school’ group and .03 in the ‘university’ group). Internal consistency of an instrument is usually established when an instrument’s test items have a Cronbach’s α of greater than .70 and less than .90 (Nunnally and Bernstein, 1994). A negative Cronbach’s alpha may be a result of using a test with few items and small sample size (Hays, 1981; Lord and Novick, 1968). The original English version of the FLLANK was found to have a Cronbach’s α = .77 (Reynolds et al., 2012), and so the estimates obtained in the Russian populations were unexpected. Helms et al. (2006) have suggested that reliability is driven by variance, with greater score variance leading to greater score reliability. The small sample sizes used in this study and the overwhelming lack of variance in post-test scores likely contributed to the low Cronbach’s α estimates. The sample size and variance in test scores were both substantially greater in the previous validation testing of the English language FLLANK (Reynolds et al., 2012), thus potentially explaining some of the discrepancy between the results of this study and those of the prior study.
There are several important limitations to this study. We did not have control groups, and therefore cannot exclude the possibility that the effect of repeated testing spuriously increased scores on the post-test. Because convenience samples were used, it is also possible that selection bias influenced our results. If study participants were more interested in learning about nutrition or otherwise more receptive to the programmes than non-participants, this would limit our ability to generalise our findings to other school and university students. Although construct validity has been demonstrated for the FLLANK, it is not known whether scores on this assessment correspond to increased ability to differentiate between foods of high and low nutritional quality among the vast array of foods that are typically available in a grocery store. Finally, food label–reading skills, although important for identifying more nutritious foods, may not necessarily lead to choosing more nutritious foods. Because we did not assess changes in behaviour, we cannot know whether the nutrition programmes we evaluated influenced participants’ dietary choices.
Conclusion
This study’s findings suggest that the nutrition curricula tested may have a positive impact on food label literacy in Russian adolescents and young adults. The translated version of the FLLANK retained its utility for capturing change in food label literacy following a nutrition education programme, but it did not perform as well with regard to internal consistency (Reynolds et al., 2012). This may be the result of small sample size or low variance in scores. Additional studies in larger groups of Russian-speaking participants are needed to confirm the reliability of the FLLANK in this population. The validation of this tool in Russian will enable its use with these populations. A controlled trial is also necessary to demonstrate the effectiveness of the nutrition education programmes.
Footnotes
Funding
This research project was funded by the US Centers for Disease Control and Prevention (grant #5 U48 DP001945-05).