%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Poor diet in early life is a major modifiable risk factor with a multitude of health outcomes. Although numerous studies have focused on the association between defined health outcomes and diet, many of these studies are frequently plagued by methodological discrepancies and often feature a cross-sectional design which represents a limitation when attempting to interpret long-term trends(Reference Shim, Oh and Kim1). The geographical location of a given country is also an important factor influencing the quality of data, as in some areas data collection had been hindered due to political reasons for a long time. With some exceptions, such as the HAPIEE study (Health, Alcohol and Psychosocial factors In Eastern Europe), which focused primarily on the elderly(Reference Boylan, Welch and Pikhart2), this long-term lack of quality follow-up on the dietary characteristics of country-wide populations as well as specific subpopulations also applies to Central Europe. Even though a recent large-scale study performed on data from nineteen countries focused on clustering energy balance-related behaviours and on associating identified clusters with weight status using data on the fruit and vegetable intake of Eastern European children and adolescents(Reference Bel-Serrat, Ojeda-Rodríguez and Heinen3), complex evaluation of the diet and especially its longitudinal outcomes in children is largely missing(Reference Northstone and Emmett4–Reference Borland, Robinson and Crozier8).
Some information on the dietary behaviour of Central European children comes from a Czech study entitled Health Behaviour in School-aged Children. While the Health Behaviour in School-aged Children study focused on the consumption of fruit, vegetables, soft drinks and sweets by Czech children and adolescents from 2002 to 2014(Reference Voráčová, Sigmund and Sigmundová9), it did not focus specifically on dietary patterns. Recently, the dietary patterns of pregnant women enrolled in the European Longitudinal Study of Pregnancy and Childhood (ELSPAC–CZ) study were analysed by Bienertová-Vašků et al. (Reference Bienertová-Vašků, Zlámal and Pruša10). However, longitudinal studies describing the dietary patterns of children are currently missing. Moreover, this lack of data must also be viewed in the context of major socio-economic and political changes which took place in Central Europe beginning with the early 1990s: these extensive and profound changes may be expected to have resulted in substantial dietary behaviour changes within the space of the past 20 years.
In the 1990s, Czech society underwent a unique transition from a planned economy to a market-based system; this historical occurrence may be viewed as a suitable basis for studying the impact of socio-economical changes on the dietary behaviour of entire families(Reference Koyame-Marsh11). Like most Central and Eastern European countries, Czechoslovakia (subsequently split into the Czech Republic and Slovakia) experienced a transitional recession period in the early 1990s as a result of an inflation rate jump. While unemployment rose steeply in Slovakia during the recession of 1991–1992 (14·7 % in 1993), it remained stable in the Czech Republic (4 % in 1993). Although the Czech Republic and Slovakia experienced a rise in income inequality during the transition period(Reference Koyame-Marsh11), second-generation reform policies adopted by the Czech government to correct the shortcomings of first-generation reforms subsequently led to sustained economic growth beginning with the mid-1990s.
In order to address the issue of dietary pattern stability in a profoundly changing population undergoing a major socio-economic transformation, this study aims to investigate the longitudinal aspects of dietary behaviour – assessed in the form of dietary patterns – between 1991 and 2006 in the ELSPAC–CZ. This study specifically aims to investigate: (i) the dietary patterns of children at 7, 11 and 15 years of age, (ii) the stability (level of adherence) of the observed patterns throughout a follow-up period of 8 years and (iii) the predisposing factors for identified dietary patterns in the observed population.
Methods
Study design
The ELSPAC is a population-based study examining the effects of biological, psychosocial, economic and environmental factors; the study focuses on how these factors affect the health and development of children from the prenatal period to adulthood. The study was initiated by the World Health Organization Regional Office for Europe in 1985 with the aim of collecting data from 40 000 children across Europe(12). The project was originally coordinated by Bristol University – Avon Longitudinal Study of Pregnancy and Childhood (ALSPAC)(Reference Golding, Pembrey and Jones13) – and operated by eight independent centres based in the UK, Czechoslovakia, Greece, Ukraine and Russia (the latter two were initially part of the USSR). Protocol development, including follow-up planning and questionnaire design, was coordinated by ALSPAC(Reference Piler, Kandrnal and Kukla14). The Czech portion of the study (ELSPAC-CZ) included eligible mothers from the Brno (the second largest city in the Czech Republic) that was home to 389 000 inhabitants in 1991–1992(15). All participating mothers were expected to deliver between 1 April 1991 and 30 June 1992. Self-reported questionnaires filled out by participants included records of demographics, lifestyle, dietary habits, partnership, life attitudes, life events, social factors and environmental exposure. In addition to its primary aims, the Czech ELSPAC–CZ study also examined socio-economic changes related to the societal transformation after the fall of Communism in 1989(Reference Bienertová-Vašků, Zlámal and Pruša10,Reference Piler, Kandrnal and Kukla14) .
Exclusion criteria and final sample size
The analysis presented in this study is based on data collected from the mothers in the study who reported on their children at two time points: 7 and 11 years. The children themselves reported on their dietary behaviour at 15 years of age. For ease of reporting, we will refer to these time points as 7, 11 and 15 years throughout the paper. The cases (children of reporting mothers or the reporting children themselves at 15 years) were excluded from the analysis in case the absolute values of their factor scores exceeded 3·5. A total of 0·9 % of all subjects were excluded from the study as a result of extreme factor scores. At 7 years, the analysis thus included 3230 children, at 11 years the analysis included 2509 children and at 15 years the analysis included 1589 children.
Dietary assessment
Dietary data were obtained from a self-reported semi-quantitative FFQ (SFFQ) collected according to the ELSPAC–CZ protocol from all study participants. The SFFQ was based on ALSPAC SFFQ, with several local food items added. Data were collected from children at three time points. Mothers were asked to fill in SFFQ about each child’s diet around the seventh and eleventh birthday of their child. The children were asked to fill out the SFFQ themselves at 15 years of age.
Participants were asked to indicate consumption frequency for 40–60 food questions using the following criteria: (1) never or rarely; (2) twice a month; (3) one to three times per week; (4) four to seven times per week and (5) several times a day. Participants were also asked to report portion sizes. For factor analysis purposes, a score for each component was created by multiplying the factor loadings by a standardised value for each food item and summing through all food items. All food items were divided into thirty-nine food groups. Food groups with no detectable effects were removed for greater clarity. This method has been used before, for example, by the HAPIEE study(Reference Boylan, Welch and Pikhart2,Reference Bienertová-Vašků, Zlámal and Pruša10) . A detailed description of food item grouping across investigated time periods is provided in Supplemental Table 1.
Statistical analysis
Children’s dietary patterns were estimated by exploratory factor analysis with varimax rotation for thirty-nine food groups. The Kaiser–Olkin–Meyer test was used to identify whether factor analysis results were appropriate. When a suitable number of factors were identified for a given group and period, factor scores were extracted. To determine the longitudinal stability of individual dietary patterns, factor analysis was performed again, this time with extracted factor scores from the previous analysis. The Pearson correlation coefficient was used to assess the association between extracted factor scores for a given pair of dietary patterns across observed time periods. This method was selected due to the high dropout rate of subjects when considering analyses of full cases from cross-sectional and longitudinal point of view. Food groups with absolute factor loadings greater than 0·20 for a given dietary pattern were considered to significantly contribute to the pattern, and conversely, negative loadings lower than −0·2 were considered to contribute inversely to the dietary pattern. All analyses were performed using R software, version 3.3.3. P-values of less than 0·05 were considered statistically significant.
Results
Descriptive results
Description of the biological characteristics of the children in the study is given in Table 1. The percentage of total variance the dietary patterns account for stood at 14·8 % at 7 years, 20·4 % at 11 years and 25·5 % at 15 years.
Non-dietary characteristics of children
Height z-score, height-for-age z-score; BMI z-score, BMI for-age z-score.
* Statistically significant difference, P < 0.001.
Children’s dietary patterns
Three dietary patterns were identified at the three examined time points (7, 11 and 15 years). The pattern scores at each time point are provided in Table 2. Consumption at 7 years was characterised by two dietary patterns, labelled ‘prudent’ and ‘junk food’. Consumption at 11 years was characterised by three dietary patterns, labelled ‘junk food’, ‘prudent 1’ and ‘prudent 2’. Consumption at 15 years was characterised by two dietary patterns, labelled ‘junk food’ and ‘prudent’. Although three separate patterns were thus identified, two dietary patterns were found to be consistently present in children across all three time points: the ‘junk food’ and the ‘prudent’ pattern. The trajectories of children’s dietary patterns across each time period and between individual time periods are provided in Fig. 1. A description of the full cases represented by all 1364 children with dietary patterns available across all three time points is provided in Fig. 2.
Factor loading scores for relationships between food groups and children’s dietary patterns*
PP, prudent pattern; JF, junk food; PP1, prudent pattern 1; PP2, prudent pattern 2.
* Factor loadings with absolute value <0·2 not shown.
Stability and changes in children’s dietary patterns across three time periods. The reported values are the Pearson pairwise correlation coefficient for a pair of factor scores corresponding to given dietary patterns across the investigated time points. Only statistically significant correlations are shown. PP1, prudent pattern 1; JF, junk food; PP2, prudent pattern 2; c7, children’s 7 years time period; c11, children’s 11 years time period; c15, children’s 15 years time period. Line thickness is directly proportional to the correlation coefficient value. Solid lines connect two adjoining time periods (7–11 and 11–15). Dashed lines connect other time periods (7–15)
Stability and changes of children’s dietary patterns from the total of 1364 children across all three time points. The reported values are the Pearson correlation coefficients for a pair of factor scores in full case analysis (where DP is known in all three ages) corresponding to given dietary patterns. Only statistically significant correlations are shown. PP1, prudent pattern 1; JF, junk food; PP2, prudent pattern 2; c7, children’s 7 years time period; c11, children’s 11 years time period; c15, children’s 15 years time period. Line thickness is directly proportional to the correlation coefficient value. Solid lines connect two adjoining time periods (7–11 and 11–15). Dashed lines connect other time periods (7–15)
An analysis of correlations across investigated time points showed the consistent presence of the ‘prudent’ pattern across all three time points. The ‘junk food’ dietary pattern was consistently characterised by typical high-fat and high-sugar foods. The correlation obtained using the Pearson correlation coefficient (0·85) showed the robust presence of the ‘junk food’ dietary pattern across investigated time points.
Furthermore, the calculated factor scores were evaluated using the univariate regression models in relation to socio-economic and maternal biological factors (Supplemental Table 2). Maternal age at birth, maternal education, maternal BMI before pregnancy, substance abuse, alcohol consumption, smoking and relationship status were analysed. Although no significant associations were observed for the first two time points in the study, some of the evaluated variables were significantly correlated with the presence of the ‘junk food’ pattern at 15 years: maternal education (6·82 %), maternal age at birth (6·24 %), maternal BMI before pregnancy (6·08 %), substance abuse (5·31 %), maternal alcohol intake during pregnancy (5·48 %), maternal smoking during pregnancy (5·80%) and maternal relationship status (5·05 %) (multiple R 2 values in the linear regression model for the variable determining the factor score in dependency on gender, evaluated variable and their interaction).
We also investigated the relationship between the dropout of children between investigated time points and other above-mentioned variables. Dropout between 7 years and 11 years was significantly associated with maternal age (non-missings were older than missings by approximately 0·4 years), ‘junk food’ pattern score at 7 years (missings had higher scores than non-missings), sex of the child (non-missings: men and women were approximately equally likely to continue, missings: men were more likely to drop out, OR = 1·23) and maternal education (children of parents with higher education were more likely to continue in the study). Dropout between 11 and 15 years was significantly associated with maternal age (non-missings were older than missings by approximately 0·9 years), maternal height (non-missings were approximately 0·7 cm higher than missings), maternal BMI before pregnancy (missings had higher BMI than non-missings by approximately 0·3 kg/m2), ‘junk food’ pattern score at 11 years (missings had higher scores than non-missings), sex of the child (women were more likely to be non-missings, whereas men were more likely to be missings, OR = 1·18), maternal prenatal smoking status (more never smokers were found in the non-missing group) and maternal and paternal education (children of parents with higher education level were more likely to continue in the study).
Discussion
This study is the first to identify stable long-term trends in the dietary behaviour of children in a Central European longitudinal cohort in times of socio-economic transition. This study utilises data from the ELSPAC-CZ cohort, collected in 1998, 2002 and 2006, to analyse children’s dietary patterns as well as their stability and changes at ages 7, 11 and 15 years. We analysed dietary data obtained using self-reported SFFQ and derived dietary patterns using factor analysis. We identified two stable dietary patterns in our study population of children (adolescents) which corresponded to patterns established in an ALSPAC population using the same methodology.
Study limitations include a significant dropout number among study subjects which may have biased the remaining data (discussed in detail in the Results). A second major limitation concerns the possible differences between data reported by proxies (mothers) at ages 7 and 11 and data reported by the study participants themselves at age 15. A third limitation is posed by the distribution of SFFQ, submitted once per time point, and thus potentially incapable of recording all seasonal variations in food consumption. The last limitation of the study, that is, the unstable setting of a country transitioning from a planned economy to a market-based system, is also its greatest strength. It could be expected that a transition associated with rising social inequality and liquidity should have a profound effect on both nutritional composition and eating behaviours. However, it seems that within the observed period of almost two decades after the Velvet Revolution, the composition of children’s diets remained relatively very stable, corresponding to patterns observed in the UK using the same methodology.
We observed two main stable patterns – a health-conscious one, labelled as ‘prudent’, and a ‘junk food’ pattern – which was consistently present across all three time points. While the ‘junk food’ pattern was consistently associated with higher consumption of biscuits, pies and cakes, sausages, cola and other soft drinks, and salty snacks, the ‘prudent’ pattern was associated with the consumption of breakfast cereals, yoghurts, fish, cheese, vegetables and fresh fruit and milk. Except for the absence of other types of meat than the red meat, the ‘prudent’ pattern is consistent with food pyramid-based recommendations valid at that time. It could be speculated that as the range of food items gradually expanded and the availability of various snacking options increased, pronounced changes in dietary behaviour of the study subjects would take place as a result of increased marketing pressure. However, no such occurrence was observed in this study, with observed dietary patterns remaining quite stable. It may thus be speculated that family dietary patterns have a more pronounced influence on children’s diets than food item availability or massive marketing. However, it could also be suggested that the specific dietary of the family was due to other confounding variables.
In view of the above-mentioned profound socio-economic transformation, it is rather surprising that the results of this study bear a striking similarity to outcomes reported in studies focusing on ALSPAC(Reference Northstone, Smith and Newby5). The results of such comparisons are even more compelling taking into account that the ELSPAC study generally used the same methodology in terms of the wording of the questionnaires as well as the frequency of administration to both the probands of the study and their proxies. Existing ALSPAC studies(Reference Northstone and Emmett4,Reference Northstone, Smith and Newby5,Reference Smith, Emmett and Newby16–Reference Northstone, Smith and Cribb18) reported very similar dietary pattern ranges, which further supports the observation that patterns observed in the UK and in the Czech Republic were quite conserved. The greatest degree of similarity between ELSPAC and ALSPAC was observed between children’s dietary patterns labelled ‘prudent’, that is, ‘healthy’(Reference Northstone, Smith and Newby5,Reference Bull and Northstone17) , ‘health-conscious’(Reference Northstone and Emmett4), ‘health aware’(Reference Smith, Emmett and Newby16) and ‘traditional/health-conscious’(Reference Northstone and Emmett4) and between children’s dietary patterns labelled ‘junk food’ in this study, that is, ‘processed’(Reference Northstone and Emmett4,Reference Northstone, Smith and Newby5,Reference Bull and Northstone17) . Taking into account the entirely different socio-economic environments of early 1990s UK and Czechoslovakia (and subsequently the Czech Republic), we may conclude that dietary patterns are relatively less susceptible to socio-economic trends and are probably much more conserved than originally expected. However, the reasons behind this are likely highly complex; a deeper understanding of the nature of this phenomenon thus requires further investigation.
Socio-economic transition generally indicates changes at the macro level and is expected to be a result of globalisation, urbanisation and industrialisation. It is expected that better improved economic situation status allows individuals to learn and live in safer and more friendly environments while achieving higher education levels(Reference Lee and Sobal19). Socio-economic changes are also inextricably linked to nutrition transition; however, this process has not been sufficiently documented. In a recent study which empirically examined the system of transitions using data from three countries (Indonesia, South Korea and the USA) at different stages of development, only vegetable oil consumption was found to have significantly increased in all three nations, with no particular significant trends established with respect to nutrition transition(Reference Linando and Lee20). However, this publication assessed nutrition transition using either a percentage of energy obtained from carbohydrates, protein and fat(Reference Lee and Sobal19) or based on a few items selected from FFQ: rice, sugar and sweeteners, vegetable oil, animal fat, milk and meat(Reference Linando and Lee20). A study focusing on a society in transition was conducted in South Africa, evaluating the nutritional records of 5-year-old urban black South African children from 1984 to 1995(Reference MacKeown, Cleaton-Jones and Edwards21). The study revealed a robust shift towards a more ‘western’ diet: while the 1984 diet consisted of 30 % fat and 61 % carbohydrates, fat intake had increased to 41 % of total energy, while carbohydrate intake had decreased to 52 % in 1995. However, FFQ were only used for the evaluation of 1995 records, and the overall results are thus difficult to interpret.
In general, the effects of socio-economic transitions on nutritional determinants are vast, complex and far from adequately understood. No unifying trend capable of explaining the effects of socio-economic circumstances on food item choices and overall eating behaviour of populations or subpopulations has thus far been identified.
When analysing the influence of selected variables on dietary pattern distribution, we observed a strong influence of some investigated variables (maternal age, maternal education, substance abuse, maternal alcohol intake during pregnancy and maternal smoking during pregnancy) on pattern distribution, with the influence of maternal education and smoking most pronounced in the earlier time points after birth. This is unsurprising as plenty of evidence indicates that maternal education and smoking are important socio-economic parameters influencing dietary behaviour. In a study by Kiefte de Jong et al., low paternal education, low household income, parental smoking, multiparity, maternal BMI, maternal carbohydrate intake and television watching of the child were determinants of a ‘Western-like’ diet equivalent to the ‘junk food’ pattern established in our study(Reference Kiefte-de Jong, de Vries and Bleeker22). However, while Kiefte de Jong et al. primarily focused on toddlers, our study followed study individuals for a significantly longer period of time.
To conclude, this is the first study to identify the dietary patterns in a Central European population during a major socio-economic transition, thereby providing important information regarding the speed and quality of the dietary transition in this population.
Acknowledgements
Acknowledgements: The authors of this study wish to thank the participating families as well as the gynaecologists, paediatricians, school heads and class teachers who took part. Our thanks also go to Dr Lubomír Kukla, PhD, ELSPAC national coordinator 1990–2012, and the entire ELSPAC team. The authors are grateful to Pavel Piler and Vít Kandrnál for valuable comments and cohort description assistance. Financial support: The ELSPAC study is currently supported by the RECETOX Research Infrastructure of Masaryk University CETOCOEN PLUS (CZ.02.1.01/0.0/0.0/15_003/0000469), the National Feasibility Programme of the Ministry of Education of the Czech Republic (LO1214) and by the CELSPAC: Central European Longitudinal Study of Pregnancy and Childhood (MUNI/M/1075/2013) support programme. Conflicts of interest: None. Authorship: M.G. analysed data, drafted and revised the paper. F.Z., L.A. and J.Š. assisted in data analysis. H.P. designed the research, formulated the research question and revised the manuscript draft. J.B.-V. edited and reviewed the manuscript and has primary responsibility for the design of the paper and its final content. Ethics of human subject participation: An ethics committee was established at the international level to support the international legal framework and ethics of the study design. The ELSPAC–CZ study was approved by local ethics committees at the national level. Informed consent was obtained from all study participants during each data collection period. This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving study participants were approved by the local ethics committees at the national level.
Availability of data and materials
The data that support the findings of this study are available from www.elspac.cz by Research Centre for Toxic Compounds in the Environment (RECETOX) but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the Research Centre for Toxic Compounds in the Environment.
Supplementary material
For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980020001056
'/%3e%3cpath%20d='M17%207.971L15.711%206.677%2010.912%2011.495%2010.912%200%209.088%200%209.088%2011.495%204.289%206.677%203%207.971%2010%2015z'%20transform='translate(8%208)'/%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
