Self-reported Lifestyle Behaviours in Families With an Increased Risk for Type 2 Diabetes Across Six European Countries: the Feel4diabetes-study

Background: The current study aimed to describe self-reported lifestyle behaviours by comparing them to current health guidelines in European Feel4Diabetes-families at risk for developing type 2 diabetes across six countries (Belgium, Finland, Spain, Greece, Hungary and Bulgaria). Methods: Parents and their children were recruited through primary schools located in low socioeconomic status areas. Parents lled out the FINDRISC-questionnaire, which was used for the risk assessment of the family. Sociodemographic factors and several lifestyle behaviours of both adults and children were assessed by parental questionnaires. To investigate families’ lifestyle behaviours, to compare these levels to health guidelines and to assess potential differences between the countries, multilevel regression analyses were conducted. Results: Most Feel4Diabetes-families at risk (parents and their children) did not comply with the guidelines regarding healthy behaviours, set by the WHO, European or national authorities; less than half of parents and children complied with the physical activity guidelines, less than 15% of them complied with the fruit and vegetable guideline, and only 40% of the children met the recommendations of ve glasses of water per day. Clear differences in lifestyle behaviours in Feel4Diabetes-families at risk exist between the countries. Conclusions: Countries are highly recommended to invest in policy initiatives to counter unhealthy lifestyle behaviours in families at risk, taking into account country-specic needs. For future research it is of great importance to focus on families at risk in order to counter the development of type 2 diabetes and reduce health inequity.


Introduction
Diabetes is a serious public health problem in Europe and worldwide. The International Diabetes Federation (IDF) stated that diabetes affects 59.3 million European adults (8.9% of the population) aged 20-79 years -of which type 2 diabetes is the most common type -and the prevalence is estimated to rise to 68.1 million by 2045 (1). Although only a limited amount of information is available about the prevalence of type 2 diabetes in European children, type 2 diabetes is also increasingly diagnosed among this young population (2)(3)(4)(5). The increased prevalence of type 2 diabetes in both children and adults results in a high cost spent on the treatment and management of diabetes and related complications (1). Therefore, efforts are needed to tackle this problem in both children and adults.
Type 2 diabetes is highly preventable through adopting a healthy lifestyle. Several lifestyle behaviours contribute to the development of type 2 diabetes, including insu cient levels of physical activity (PA), high levels of sedentary behaviour and unhealthy dietary behaviours (6). Guidelines and recommendations for these behaviours have been developed for children and adults. PA guidelines set by the World Health Organisation (WHO) recommend children to engage in at least 60 minutes of moderateto-vigorous PA (MVPA) per day (7) and adults to engage in at least 150 min of MVPA per week (corresponding to at least 30 minutes of MVPA during 5 days) (8). For sedentary behaviour, studies and national guidelines recommend for both children and adults to minimise the amount of time spent in prolonged sitting and to break up long periods of sitting as often as possible (9). Further, speci cally for children it is recommended not to exceed two hours of recreational screen-time per day (10)(11)(12). Finally, the WHO recommends to consume at least 400 grams of fruit and vegetables, corresponding to ve portions per day (for both children and adults) (13), and to limit the daily consumption of unhealthy snacks and soft drinks (not exceeding WHO recommendations for free sugar, < 10% of total energy intake) (14). In addition, recommendations based on previous research, suggest to consume breakfast on a daily basis (15) and the European Food Safety Authority recommends a daily water ( uid) intake of 1.6-2.0 litres of water per day for girls and women and boys and men respectively (16).
Several European studies have previously investigated the levels of physical activity (PA), sedentary behaviour and healthy dietary behaviours in primary schoolchildren and adults in a general population.
Results indicated that between 5.0% and 47% of the European children (17) and between 7.0% and 96% of the European adults (18) met the recommended amount of PA per day, depending of assessment methods and country (17,18). Furthermore, European adults reported a sitting time of 309 minutes/day on weekdays (19) and 25% of the girls and 33% of the boys exceeded the recommended daily screen-time limits (20). Regarding dietary behaviour, results of the European Energy-project showed that between 12.0 and 51.7% of children and between 11.8 and 56.1% of adults, depending of country and gender, skip breakfast (21). Furthermore, research has indicated that 76% and 94% of 11-year-old European children (22) and 86% of adults (23) did not reach the recommendations regarding fruit and vegetable intake.
Moreover, a study conducted in six European countries showed that young children consumed on average between 0.25 and 0.85 litres of water per day (depending on gender and country) (24) and French, Italian and Spanish adults consumed on average between 0.60 and 0.72 litres of water per day (25), which is in both children and adults below the recommended amount. Finally, results of the IDEFICS-study revealed high intakes of total sugars and foods and drinks rich in added sugar in European 2-to 10-year-old children (25) and a high percentage (11-49%, depending of country) of European adults was found to exceed the WHO recommendation on free sugars intake, considering solely uids (26). Even more alarming, in vulnerable subpopulations (e.g. children and parents with a low socio-economic status (SES)) higher levels of physical inactivity, high levels of sedentary behaviour and unhealthy dietary patterns have been reported and consequently they have a higher prevalence of overweight, obesity and type 2 diabetes compared to the general population (27,28). In 2019, it was therefore recommended in the International Diabetes Federation Atlas (1) to prioritize families with a higher risk for developing type 2 diabetes in lifestyle interventions. However, up to now to our knowledge no research could be located that investigated the prevalence of the above mentioned lifestyle behaviours speci cally in parents at increased risk for developing type 2 diabetes and their children.
Therefore, the aim of the current study was to assess the levels of several lifestyle behaviours (i.e. PA, sedentary behaviours and dietary behaviours) and compare these levels to health guidelines/recommendations in European families (parents and their primary schoolchildren) at risk for developing type 2 diabetes. As differences in prevalence of type 2 diabetes exist between European countries (1,29), this study also investigated potential differences in lifestyle behaviours in these highrisk families between European countries.

Study background
Within the Feel4Diabetes-project a multi-level intervention was developed aiming to prevent type 2 diabetes in vulnerable families across six European countries (Belgium, Finland, Bulgaria, Hungary, Greece and Spain). These countries represent three socio-economic levels: High-income countries (Belgium and Finland), high-income countries under austerity measures (Greece and Spain) and low-tomiddle-income countries (Bulgaria and Hungary). The Feel4Diabetes-intervention targets three main lifestyle behaviours, namely PA, sedentary behaviour and dietary behaviour. The Feel4Diabetes-study focused on all types of families with children on one hand; and on the other hand, on families with an increased risk for developing type 2 diabetes. This study uses the baseline data from the population at risk for developing type 2 diabetes. More detailed information can be found elsewhere (30).

Procedure
Primary schoolchildren and their parents were recruited via schools located in vulnerable areas. Within low-to-middle income countries (Bulgaria and Hungary), all areas were considered vulnerable; whereas in Belgium, Finland, Greece and Spain recruitment only took place in low SES areas. To select the low SES areas, all municipalities in the selected provinces were divided into tertiles based on socioeconomic indices retrieved from o cial sources and authorities (i.e. literacy or unemployment rates) (31)(32)(33)(34).
Municipalities within the highest tertile (i.e. municipalities with the lowest SES indices) were included in the study. Within the selected areas, schools were randomly contacted and in total 236 primary schools (response rate = 40.2%) con rmed their participation in the Feel4Diabetes-study. In January 2016, children from the rst three grades of primary school and their parents were invited to participate in the study. By signing the informed consent form and lling out two questionnaires (FINnish Diabetes RIsk Score (FINDRISC)-questionnaire and Energy Balance Related Behaviour (EBRB)-questionnaire), families con rmed their participation. The FINDRISC-questionnaire was used to identify families with an increased risk for developing type 2 diabetes (based on the diabetes risk score -see 'Measurements' section). In addition, the high-risk families received two additional questionnaires (i.e. the high-risk questionnairesone for the child and one for the parent) for a more in-depth evaluation of the behaviours. The parents lled out the demographic and lifestyle-related questions for both their children and themselves. In total, 11,396 families con rmed their participation to the project, of which 4,484 families (39.3%) were identi ed as high-risk families for developing type 2 diabetes (de ned as one or both parents having FINDRISC value denoting moderate or high type 2 diabetes risk) (35). Finally, height and weight were measured by researchers in the school (children) and local municipality centers or home setting (parents).
Measurements were conducted by trained researchers, using standardized protocols and calibrated equipment (36).

Measurements
Before applying the questionnaire to the main study, its reliability was assessed in a pilot study using a two-way random effect single measure intra-class correlation coe cient (ICC). Reliability addressed the question of how consistent the answers were from one occasion to the next in the same subject. Parents were asked to complete the questionnaire twice, within a 1-2 week interval. The ICC's were classi ed as excellent (> 0.81), good (0.61 − 0.50), moderate (0.41-0.60) and poor (< 0.40) (37).
Diabetes risk score. The FINDRISC-questionnaire is a validated tool to assess individuals' risk for developing type 2 diabetes. It includes eight questions on age, Body Mass Index (BMI), waist circumference, PA, daily consumption of fruit, berries or vegetables, history of antihypertensive drug treatment, history of high blood glucose and family history of diabetes (38). The Diabetes Risk score is calculated as the sum of these individual scores and varied from 0 to 26 points, with higher scores indicating a higher risk for developing type 2 diabetes (39). A family was selected as a high-risk family if at least one of the parents met the cut-off score of 9 points. Further, the highest FINDRISC-score within the family was assessed based on the highest FINDRISC-score of mother or father, or of both parents.
Physical activity. Moderate-to-vigorous physical activity (MVPA) was measured by the following questions in the EBRB-questionnaire: "On how many days during the last week did you (parent) spent MVPA for a total of at least 30 minutes per day?" and "On how many days during the last week did your child spent in MVPA for a total of at least 1 hour per day?". This question was asked separately for weekdays and weekend days (ICC range 0.367-0.700). To assess the percentage of children and parents meeting the WHO Europe PA guidelines, the number of days meeting the PA guidelines on weekdays and weekend days were summed. The levels of PA in a whole week were dichotomized into not meeting the WHO guideline (< 7 days a week for children and < 5 days a week for parents) and meeting the guideline (7 days a week for children and ≥ 5 days a week for parents).
Sedentary behaviour. Parents' total sitting time on weekdays in parents and screen-time behaviour during the week in children were assessed in respectively the high-risk questionnaire and the EBRBquestionnaire. Total sitting time in parents was assessed using the question: "During the last 7 days, how many hours did you (parent) spend sitting on a weekday?" (ICC = 0.584). Furthermore, children's screentime behaviour was assessed by the following question: "About how many hours per day does your child usually devote to screen-activities (excluding school)?". This question was asked separately for weekdays and weekend days (ICC = 0.647 and 0.694). Answer options were 0 h/d, < 0. 5  time activities on weekdays and weekend days were summed. To determine the percentage of children meeting the recommendation of less than two hours of recreational screen-time per day (based on recent studies and set by national authorities), the total time spend on screen-time activities were dichotomized into not meeting the guideline (≥ 120 minutes of screen-time activities per day) and meeting the guideline (< 120 minutes of screen-time activities per day).
Dietary behaviour. Consumption of water, soft drinks and juices containing sugar, fruit and vegetables, unhealthy snacks (sweets and salty snacks/fast-food) and breakfast were assessed in the EBRBquestionnaire. The general question asked in the questionnaire was: "Please indicate how often you (parent) and your child consume: Water, soft drinks and juices containing sugar, fruit/berries (fresh or frozen), fruit and berries (canned or dried), vegetables, sweets, salty snacks/fast-food" (ICC range: 0.371-0.827). Answer options were: Less than 1 per week, 1-2 per week, 3-4 per week, 5-6 per week, 1 or 2 per day, 3 or 4 per day, 5 or 6 per day, more than 6 per day. Afterwards, these categorical values were recoded into numerical values according to the midpoint method (e.g. 5-6 portions per week was recoded into 0.79 portion per day, 3 or 4 portions per day was recoded into 3.5 portions per day). The portion size, de ned with a household unit, was provided under the question. The consumption of water and soft drinks/juices containing sugar was expressed in glasses, one glass contains a content of 2.5 dl. One portion of fruit and vegetables equals the content of about 1/2 cup (2.5 dl) or the size of a tennis ball. One portion of sweets equals a chocolate bar, half a cup of sweets, cookies or ice-cream and one portion of salty/snacks fast food equals a small hamburger, a small bag of chips or a slice of pizza. To assess the total consumption of fruit and vegetable per day, the daily consumption of fruit/berries (fresh or frozen), fruit and berries (canned or dried) and vegetables were summed. Further, outliers (de ned as values above three standard deviations from the mean) were capped, and reassigned the value of the mean plus three standard deviations, a method conducted within previous research (40). The daily breakfast consumption was measured by the following questions. "On how many days do you/does your child usually eat breakfast?". This question was asked separately for weekdays and weekend days (ICC range 0.088-0.281). The amount of days consuming breakfast on weekdays and weekend days were summed. Within the Feel4Diabetes-study, families were recommended to drink at least 5 glasses of water per day. To assess the percentage of parents and children meeting these recommendations, the amount of water consumption was dichotomized into not meeting the recommendation (< 5 glasses of water per day) and meeting the recommendation (≥ 5 glasses of water per day). Further, to assess the percentage of parents and children meeting the WHO guidelines regarding fruit and vegetable consumption (at least 5 portions per day), the amount of fruit and vegetable intake was dichotomized into not meeting the guideline (< 5 portions of fruit and vegetables per day) and meeting the guideline (≥ 5 portions of fruit and vegetables per day). Finally, to assess the percentage of parents and children consuming breakfast on a daily basis (recommendation based on previous research), the breakfast consumption in a whole week was dichotomized into not meeting the guideline (< 7 days a week) and meeting the guideline (7 days a week).
Sociodemographic variables. Parents reported their birthdate, sex and educational level (years of education), as well as their child's birthdate and sex. Age was calculated based on parents' and children's birthdate and measurement dates. Family SES was categorised as follows: low (both parents having no higher education i.e. ≤14 years of education), medium (at least one of the parents having no higher education),high (both parents having a higher education i.e. >14 years of education) (41). In European education systems more than 14 years of education implies attendance of higher education (e.g., bachelor program).

Statistical analyses
Descriptive statistics for the sample demographics were computed using SPSS statistics 24.0 for Windows (SPSS Inc., Chicago, IL). Besides, in order to have more insight into the strength of the relationship between parents' lifestyle behaviours and their FINDRISC-score and between children's lifestyle behaviours and highest FINDRISC-score within the family a correlation analysis was conducted in SPSS statistics 24.0 for Windows (SPSS Inc., Chicago, IL).
Parents' and children's lifestyle behaviours and the percentages of parents and children meeting the recommendations/guidelines across all countries were investigated. In addition, there was examination of potential differences between countries. Therefore, multilevel regression analyses were conducted using MlwiN 2.28 (Centre for Multilevel Modelling, University of Bristol, UK). Multilevel modelling (threelevel: child; class, school) was used to take clustering of children in classes, in schools into account. For all analyses, statistical signi cance level was set al p < 0.01.
Furthermore, the highest FINDRISC-score within the family ranged between 9 and 24 points, with an average of 12.41 ± 2.94 points. Families were assigned to the high-risk group based the FINDRISC-score of the mother (41.7%), the father (35.3%) or both (22.8%). Descriptive data across all countries, and separately for the six countries can be found in Table 1.
Besides, the correlations were investigated between parents' lifestyle behaviours and their FINDRISCscore (correlation coe cient ranged between 0.037 and 0.098), and between children's lifestyle behaviours and the highest FINDRISC-score within the family (correlation coe cients ranged between 0.009 and 0.079]. The correlation coe cients can be found in Additional File 1.  Consumption of sweets. Descriptive data (means and standard deviations) of the consumption of sweets in children and parents can be found in Table 2. Furthermore, within the table, differences between the countries were indicated (p-values).
Children Consumption of salty snacks and fast food. Descriptive data (means and standard deviations) of salty snacks and fast food in children and parents can be found in Table 2. Furthermore, within the table, differences between the countries were indicated (p-values).
Children. Children across the participating countries consumed on average 0.36 portions of salty snacks/fast food per day and differences exist between the countries (p < 0.001). In Finland and Greece, children consumed the lowest number of salty snacks and fast food (respectively 0.12 and 0.19 portions/day) (p < 0.01), while in Hungary, children consumed 0.97 portions of salty snacks and fast food, which is the highest number compared to the other countries (p < 0.001). From Spanish children, no data are available, due to a question missing in the local survey.
Parents. Parents across the countries consumed on average 0.34 portions of salty snacks and fast food per day and signi cant differences were found between the countries (p < 0.001). Finnish and Greek parents consumed the lowest number of salty snacks and fast food (respectively 0.10 and 0.12 portions/day) compared to the other countries (p < 0.01), while the highest number, 0.61 portions per day, was found in Hungarian parents (p < 0.001) Breakfast consumption. Descriptive data (means and standard deviations) of breakfast consumption in children and parents can be found in Table 2. Furthermore, within the table, differences between the countries were indicated (p-values).
Children. Children across the countries consumed breakfast on 6.58 days per week and in total 86.9% complied with the daily breakfast guideline (consuming breakfast, 7 days/week). Signi cant differences exist between the countries (p < 0.001). The highest number of days in which children consumed breakfast were found in Spain (6.86 days/week) and Finland (6.79 days/week), while the lowest numbers were found in Belgium (6.39 days/week), Greece (6.44 days/week), Hungary (6.41 days/week) and Bulgaria (6.53 days/week) (p < 0.01). The highest percentages of children complying with the daily breakfast guideline were found in Spain (95.9%) and Finland (92.4%), while the lowest percentages were found in Hungary (82.3%), Belgium (82.5%), Greece (82.9%) and Bulgaria (83.1%) (p < 0.001).
Parents. Overall, parents consumed breakfast on average 5.60 days per week and in total 66.5% of all parents complied with the guideline of daily breakfast consumption. Signi cant differences exist between the participating countries (p < 0.001). The highest numbers of days in which parents consumed breakfast were found in Finland (6.71 days/week) and Spain (6.45 days/week) compared to the other countries (p < 0.001). In contrast, Bulgarian parents consumed breakfast on 4.30 days per week, which was signi cantly lower compared to the other countries (p < 0.001). The percentages of parents meeting the breakfast guidelines were the highest in Finland (89.1%) and Spain (85.3%), while the lowest percentage was found in Bulgaria (40.0%) (p < 0.001).
A more detailed description (p-values) of lifestyle behaviours between the participating countries can be found in Additional File 2.

Discussion
The aim of the current study was to assess several lifestyle behaviours (PA, sedentary behaviour and dietary behaviours) in families at risk for developing type 2 diabetes, and to compare the behaviours to health guidelines. Further, potential differences in at risk families' lifestyle behaviours between six Feel4Diabetes European countries were investigated.
Overall, some alarming results were found for both parents and children: less than 30% of the children met the PA guideline, only 40% of the children met the guideline of at least 5 glasses of water per day and less than 10% of the children met the guideline regarding fruit and vegetable intake. In addition, less than 45% of the parents met the PA guideline and only 14% of the parents met the guidelines regarding fruit and vegetable intake. On the other hand, results also showed some more positive results: almost 87% of the children and 68% of the parents met the recommendations regarding daily breakfast consumption, 64% of the children met the guideline of screen-time, and both parents and children managed to keep the consumption of soft drinks or juices containing sugar, sweets, salty snacks/fast food relatively low. These results point out that behaviour change programmes in the selected Feel4Diabetes families at increased risk for type 2 diabetes primarily need to focus on the promotion of physical activity, fruit and vegetable intake and on drinking water. To obtain the adherence to health guidelines, additional efforts to promote daily breakfast consumption and to limit the consumption of unhealthy snacks seem less necessary in this target group.
Despite research showing that people with normal blood glucose levels generally have healthier behaviours compared to people with prediabetes (42,43), results of this European sample of families at increased risk seems to be quite comparable with results of studies conducted in the general population.
More speci cally, similar gures were found for adults and children in meeting the guidelines regarding PA, sedentary behaviour and several dietary behaviours (19,22,23,44,45). Less positive outcomes for health behaviours could be expected in families at increased risk to develop type 2 diabetes, since they have been selected based on the FINDRISC-questionnaire, which includes questions on these behaviours. However, these questions only counts for 3 of 26 points (11.5%) of the total FINDRISC-score, so they only minimally in uence the total FINDRISC-score. Further, results showed that the correlations between parents' lifestyle behaviours and their FINDRISC-score, and the correlations between children's lifestyle behaviours and families' FINDRISC-score are overall weak. This is actually not surprising since the prediction model to assess the probability to develop type 2 diabetes within 10 year (conducted to develop the FINDRISC-questionnaire) showed that questions on physical activity and fruit and vegetable consumption did not add much to the predictive power. Even though they had no impact on the score, these two lifestyle behaviours were included in the FINDRISC-score to emphasize the importance of lifestyle behaviours in preventing type 2 diabetes (38). Despite the comparable lifestyle outcomes between the general population and the at risk population, this study population needs extra attention in future lifestyle interventions in order to prevent the development of type 2 diabetes.
Similar to cross-European health behaviour differences in the general population (41,46,47), results of the present study showed clear cross-European differences in several lifestyle behaviours in high risk families. The cross-European differences, found in all included lifestyle behaviours, are possibly due to political, geographical and cultural differences. In our study, Hungarian children and parents and Finnish children at increased risk had signi cantly higher physical activity levels compared to similar participants in the other countries (Belgium, Spain, Greece and Bulgaria). This might partially be explained by differences in the total hours of mandatory physical education in primary schools: Hungarian schools provide ve hours of physical education per week, whereas schools in the other countries provide 2 to 3 hours of physical education per week (46). Furthermore, countries differ in provided action plans in the health sectors, schools, workplaces to promote physical activity, and in national awareness campaigns on PA (46 2013), conducted in 32 European countries in a general population. These differences might be explained by wealth inequalities, which could in uence several domains of sedentary behaviours (i.e. parents' occupation, transport, leisure-time and household activities), differences in climate, and/or cultural differences; which has also been reported in previous research (19).
Finally, within this study, clear differences were found in water consumption between Northern and Central European countries, compared to Southern European countries, which could also be found in the study of De Craemer et al. (2015), conducted in preschool children in the general population. This might be explained by higher outside temperatures, resulting in a higher thirst to quench (48). In conclusion, clear cross-European differences exist in health behaviours in families at risk for developing type 2 diabetes, which are similar compared to cross-European differences reported in the general population. For future research, it is recommended to deliver effective universal interventions (targeting the general population), with a special focus on targeted interventions in individuals/families with an increased risk for developing type 2 diabetes -known as proportionate universalism -in order to reduce health inequity (50). Second, the FINDRISC-questionnaire, used as a tool to select families at risk for developing type 2 diabetes, is a unique and promising approach in health research to identify families' risk for type 2 diabetes. It is a fast, inexpensive way to raise awareness in parents of their families' risk (38) (54). Yet it is of great importance to involve them in health research as recent research shows a strong correlation between weight status and lifestyle behaviours of fathers and their children (55)(56)(57)(58) and because a growing number of fathers are involved in the health care of their children (59). Therefore, recruitment strategies in which fathers are explicitly invited to participate, or by communicating the bene ts of the research for fathers and their families should be used in further research (60).
A rst strength of the study is the large sample across six European countries, representing high income countries, low-to-middle income countries and countries under austerity measures. Another strength is the focus on a population at risk for developing type 2 diabetes. By using the FINDRISC-questionnaire, at risk families could be detected in a fast, inexpensive and non-invasive way. On the other hand, families' lifestyle behaviours were self-reported, potentially causing response bias (e.g. social-desirability bias) (61), which is a limitation of the current study.

Conclusion
Results revealed that most Feel4Diabetes families at risk (parents and their children) did not comply with the guidelines regarding healthy behaviours set by the WHO, European or national authorities. Despite the fact that no clear differences in lifestyle behaviours exist between the at risk group in the current study and the general population (revealed from previous research), it is still of great importance to focus on this subgroup in future lifestyle interventions in order to counter the development of type 2 diabetes and to reduce health inequity. Countries should investigate in policy initiatives to counter unhealthy lifestyle behaviours in families at risk for developing type 2 diabetes and because of the observed differences between countries, it is recommended to develop interventions taking into account the needs within a speci c country.

Availability of supporting data
All data related to this research are available in the text and tables.

Competing interests
The authors declare that they have no con ict of interest