Evaluation of the effects of an interdisciplinary lifestyle and health program for adolescents and its benets for maintaining health.

109 interdisciplinary school-based study. The aim was to assess various implementation outcomes based on the effect of health education, including the promotion of healthy eating patterns and a limitation of sugar-containing beverage consumption via the feedback of analysed anthropometric data, biomedical parameters, physical tness tests and an eating and physical activity behaviour questionnaire. improvement, signicant decrease/decline).

Obesity is associated with comorbidities such as metabolic syndrome (11) diabetes mellitus (12), cardiovascular diseases (13,14), musculoskeletal disorders and cancer (15). Longitudinal studies have shown that prenatal and postnatal factors can already be decisive to predict whether a child becomes obese (16)(17)(18). Therefore, information on a balanced diet and physical activity is important and prevention programs should start as early as possible. As we face large socioeconomic problems caused by an unhealthy lifestyle, any age class should be considered for prevention and if necessary for intervention.
To prevent obesity, different programs for pre-school children (19)(20)(21), primary school children (22,20) and adolescents (23,22) have been initiated. They differ in study design and the way of intervention. The IDEFICS-study was a comprehensive study investigating the effect of dietary-and lifestyle-induced health effects in children and infants in several different countries (24). The participants of this study were at the age of 2 to 9.9 years and the recruitment took place at pre-schools and primary schools. The key messages for intervention were focused on the reduction of soft drinks consumption, the daily intake of more fruits and vegetables, less media-time and more daily physical activity, increased family time and an adequate sleep duration.
To increase the outcome and effectiveness of such interventions in terms of public health decision making regarding health policies, programmes and practices, a framework for the design, execution and interpretation of health-related research has been initiated. This framework, summarised under the term implementation research, provides a number of key principles and implementation outcome variables for the design, execution and analysis of health related research (25).
The present study aimed to assess the feasibility of studying the implementation outcomes of an interdisciplinary school health project with regards to healthy nutrition and physical activity of adolescents aged 14 to 17 years. The aims were twofold, rst to de ne the potential and adoption of this implementation project in schools and second to assess the acceptability, appropriateness and coverage of the planned intervention via the measurement of anthropometric data, biomedical parameters (26) and physical tness parameters (health parameters). Moreover, a psychologically approved questionnaire about food preferences, physical activity and other individual factors was included. As intervention tool, hands on workshops increasing the awareness for nutrition and beverages and associated health parameters were used.

Methods
The study design meets the Helsinki Declaration (World Medical 2013) and was approved by the local ethics committee of the county of Salzburg, Austria. The participants of the study were informed collectively about the project and the planned measurements. In addition, they had the possibility for personal interviews, if there were any additional questions. Only students who provided their informed consent (by signing the information sheets themselves and by a parent or legal guardian) were included in the study. All participants could stop the participation in this study at any given moment without the need to provide a reason. The different analyses were performed at T0 (start of the study), T1 (1 month) and T2 (6 months). Initially, it was thought that certain parameters might change after weeks, but this was not the case. Therefore, time point T1 was not included in the analysis presented here, since the time frame of 1 month was too short to determine measurable effects of the dietary-and lifestyle-program. The students were divided in a group obtaining health-related information (intervention-group: N=56) and a control-group (N=51). Data were collected in a pseudo-anonymous manner. The students obtained a code and only the paediatrician knew the key between the codes and the respective students. In case of abnormal laboratory values, the student and his/her parents were informed and invited to an additional medical assessment by the paediatrician.
At the end of the school year, the head of the school and the teachers involved (natural sciences and sports) were asked for permission to perform the study in the following school year. At the beginning of the next school year, an informative meeting was organized for students and parents and they were asked to sign the informed consent form for the study. Based on the number of valid informed consent forms, the school visits of the teams of the different disciplines were planned in close cooperation with the teachers and administrators of the school. Venipuncture and anthropometric measurements were always performed during the rst lesson and the students were instructed to fasten, i.e. not to eat or drink anything except water, ten hours before drawing the blood samples. To ensure protocol adherence, students were asked if they were able to follow the instructions before venipuncture. The time frame to draw blood and measure the values from 25 to 30 participants was calculated with approximately 30 to 40 min and a healthy breakfast was served right afterwards for all participants. The physical tness tests were conducted by sports scientists trained for the Physical Fitness Test battery (PFTB) within the same week (27). Encoded questionnaires were handed out within the same week and were lled out by the students during one of their biology courses. The participants were randomized into a control group and an intervention group. The intervention group had workshops on nutrition, beverages and associated health parameters including anthropometric data, biomedical parameters, physical tness parameters as well as eating behaviour directly after measuring baseline levels (T0) and the control group attended this workshop after the third measurement (T2). The workshop included theoretical information on various types of food and its components, a hands-on workshop to estimate and determine fat, sugar and protein content of speci c foods and beverages and a direct link to the analysis of certain health parameters, e.g.
glucose. Moreover, a crossword puzzle and a quiz were included to test the acquired knowledge. The puzzle and the quiz were only performed to manifest the acquired knowledge; however, the analysis was not included in this study.
The different measurement tools were used at three time points. To exclude changes due to external in uences, the participants were randomly assigned to an intervention or control group matched for age and gender.

Anthropometric data
Anthropometric measurements for evaluating body mass index (BMI) and waist-to-hip ratio (WHR) were performed with a medically approved stadiometer (Seca, Seca Austria, Wien, Austria) and reference tape and S (coe cient of variation of BMI) are parameters from the Cole`s LMS method and are available on the CDC website. Z-scores (or standard deviation (SD) scores) are widely used in anthropometry to quantify a measurement's distance from the mean (28).

Biomedical data
The biomedical data were obtained as described previously (26). In brief, samples for biomedical analyses were withdrawn using the BD vacutainer® Push Button Blood collection set in combination with tubes for serum, Ethylendiamintetraacetat-and Li-Heparin-plasma (BD Becton Dickinson Austria GmbH, Vienna, Austria). The samples were transported and stored according to pre-analytical guidelines and good laboratory practise. A total of 50 laboratory parameters were measured. Care was taken to comply with good practice in pre-analytics, analytics and post-analytics. Laboratory parameters were analysed as previously reported (26).

Physical tness test battery
Physical tness was assessed by a standardized age-and percentile-matched physical tness test battery (PFTB) (27). The test includes seven exercises for determining motor skills, strength, endurance and coordination. A detailed test description can be found in Table 1. The percentile rank for the 20MR, SLJ, LBFJ and 6MR tests were calculated from the raw data using the normalised values of the German motor function tests (29). The percentile ranks for SHJ were calculated using the 'Munich Fitness Tests (30). Raw data was used for PU and OBR, since no normalised values are available for the age group tested.

AD-EVA Questionnaire
AD-EVA is an interdisciplinary test system for the diagnosis and evaluation of obesity and other diseases that can be in uenced by eating and movement life-style changes (31). This validated questionnaire covers nine main questionnaires (Scales) consisting of several sub-questionnaires (sub-scales) of which seven were analysed for this study. A detailed description of the scales is shown in Table 2. FEV-Path (pathogenic eating behaviour) was not included in the study, since the assumption was made that the students were per se healthy. The FBEB (handling of food including carving and addiction) scale was excluded from analysis due to low response rates.
In order to achieve a standardisation, the evaluation tool of the standardized questionnaire (31)    calculations and descriptive statistical analysis were performed. Data is presented as percentile (25,50 and 75) to provide a better view of the distribution of the data. Statistical analyses between the control group and the intervention group were performed using the Independent Samples Mann-Whitney U Test with a signi cance level of p < 0.05 (Anthropometric, biomedical and PFTB) and independent sample test including Levene's test for equality of variances and t-test for equality of means (questionnaire).
Statistical analyses between time points T0 and T2 were performed using the Paired-Sample Wilcoxon signed rank test including a Bonferroni correction (Anthropometric, biomedical and PFTB) and paired sample t-test (questionnaire).

Participants and Compliance
In total, 109 students participated in the study (42 females and 67 males both aged 15.0 ± 0.7 years).
Dropout rates from the time point T0 to T2 differed for the different tests. For the anthropometric and PFTB data the dropout rate was 2.5% and 2.4% respectively for females and 4.7% for both data sets in males. For the laboratory measurements, the dropout rate was 7.5% and 6.3% for females and males, respectively. In case of missing values in the data range, the whole pair was excluded in the analysis comparing T0 and T2.
For the AD-EVA questionnaire we had a total of 38 females and 65 males, however, the groups were not split between control and intervention, since only one subscale (evaluation of fun and satisfaction (FBM-F&B)) resulted in a signi cant difference at T2 between the control-and intervention-group for the BMInorm group in the male population out of all measurements. Anthropometric data In Table 4 anthropometric data is shown for the start of the study at T0 for females and males, presented as 25, 50 (median) and 75 percentiles (control and intervention group combined). Detailed information including data for females and males at both time points can be found in the additional le 1. The analysis of the anthropometric data showed that the BMI is distributed according to the age-and gendermatched reference range. No statistical difference was observed between the control group and the intervention group at T0 and T2. To determine the change over time for the anthropometric data, statistical analysis was performed between the time point for both groups (Crtl and Int) in females and males ( Table 4, Table S1). Over the period of 6 months, the male population showed a small, but signi cant increase in weight, height, BMI, waist-and hip circumference, whereas the female population showed only an increase in height, waist-and hip-circumference and hip/waist ratio (see Table 4 for group details). Interestingly, whole body bio-impedance as well as right and left arm bio-impedance was signi cantly lower at T2 in males (Table 4). This could be explained as a result of the increase in physical tness shown in Table 5.  To determine the change over time for the PFTB data, statistical analysis was performed between the time point T0 and T2 for both groups (Crtl and Int) in females and males (Table 5, Table S2). Over the period of 6 months, the female population showed a signi cant improvement at the exercises PU, LBFJ and OBR and the male population at the exercises 20MR, PU, LBFJ and OBR (for group details see Table   6).    (Table 6).     (Table 8). Detailed description of the 'Body image' questionnaire can be found in Table 2 and Table 8. All values were between 2.5 and 3.8, which is considered an average response indicating no disorder in terms of body image. Interestingly, the results for item 5 ('who is the least attractive') had a substantial spread ranging from 1-6 with a standard deviation close to the mean (Table   8). No signi cant differences were observed between T0 and T2, therefore only T0 is shown in Table 8.  presented as mean and standard deviation as well as the corresponding percentage range and T-value for each for the BMI groups (BMI-norm, BMI-low, BMI-high) in both gender.

Discussion
The study was designed to assess the potential of an interdisciplinary life-style program on the health status of adolescents over a 6-month period. The focus was placed on the changes observed over time including anthropometric, biomedical, and physical tness parameters as well as the results of the AD-EVA questionnaire. The results are reported as a comparison of the measured parameters at the start and the end of the study. Control and intervention group are presented separately in order to identify a potential impact of the intervention tool "hand-on workshops".
In terms of the adoption variable, the intention to employ a new intervention in this speci c school setting was met with great acceptance from the head of school and teachers, participants and their parents as well as the ethics committee and policies involved. This positive adaption was also linked to a high feasibility of the program.
In terms of the study aims acceptability and appropriateness, the program was well accepted by the participants, with a dropout rate of maximally 7.5% and a general positive feedback of the participants was received. The study population was selected to show effects of the intervention rather than selecting bene ciaries, since such a life-style program is aimed at the wide variety of young adolescents without particular comorbidities or speci c social, cultural or economic environment.
An extensive part of the program was linked to the degree to which participants might bene t from this intervention. The results of the anthropometric data revealed a signi cant increase in weight, height and BMI, as well as hip-, waist-and neck-circumference over the course of 6 months within the normal reference range. This can be explained by the continuous growth in adolescence. Interestingly, the increase was more pronounced in the male population in both the control and intervention group, compared to the female population. Moreover, a signi cant decrease in whole body-as well as right-and left-arm-bio-impedance measurement results was observed after 6 months, implying an increase in muscle mass paralleled by a decrease in fat mass. This can be linked to a signi cant increase in physical tness related to pull ups, coordination and motor skills, caused by the repeated analysis of these measures in both groups. In addition, the muscle metabolite creatinine was signi cantly increased in the male intervention group, indicating a potential increase in muscle turnover and physical activity (32).
Moreover, the earlier ndings focussing on the calculation of reference intervals of haematological and biochemical markers in an Austrian adolescent study cohort were extended in the present report (26).
Various clinical laboratory parameters were measured at different time points. The comparison of the parameters at T0 and T2 showed signi cant changes in the haematological values and parameters for the determination of carbohydrate-and fat-metabolism as well as renal function. The number of erythrocytes, as well as the haemoglobin, haematocrit and mean corpuscular haemoglobin values were signi cantly decreased, nevertheless all values were still in the normal reference range ( Table 6 and Table   S3). These changes might be due to daily uctuations, especially for the female population (33). Some of the values indicated, that changes might be due to an enhanced physical activity, a direct connection cannot be drawn at this stage and this should be studied in more detail. Nevertheless, a study investigating the effects of exercise and training on the oxygen supply of red blood cells showed, that trained athletes have a decreased haematocrit which is brought about by an increased plasma volume (34). Changes in the carbohydrate metabolism were observed for glucose and insulin values in only a number of female and male participants, which could be due to a speci c diet before the day of the measurement. However, no long-term conclusion can be drawn from these observations. A connection between dietary changes and a long-term modulation of the carbohydrate metabolism can only be drawn if there would have been a signi cant change in glucose-, insulin-and HbA1c values within one of the groups (35,36). Interestingly a number of fat metabolism-related parameters decreased signi cantly over the course of the study, mainly in the intervention group of the male population, whereby all changes remained in the normal reference range. Parameters included LDL-cholesterol, LDL/HDL ratio, ApoA1 and Adiponectin. A decrease in LDL-cholesterol concentration and LDL/HDL ratio indicates a general positive trend towards a healthier lifestyle, whereby a decrease in adiponectin, a hormone which is involved in regulating glucose levels as well as fatty acid breakdown, indicates an increase in fat cell depot and fat cell size (37). This is in contrast to the observed results for whole body-as well as right-and left-arm bio impedance, which implied an increase in muscle mass paralleled by a decrease in fat mass. This could be explained by daily uctuations of the adiponectin level in contrast to a more consistent change of bioimpedance over the 6 month time point (38). A study investigating the lifestyle and dietary determinants of serum apoA1 and apoB concentrations showed that high intake of products containing added sugar, such as pastries, sweets, chocolate, jam/sugar and sugar-sweetened beverages, was associated with lower ApoA1 concentrations but higher ApoB concentrations and a higher ApoB/ApoA1 ratio (39). In our study, the concentration of ApoB did not change signi cantly over time.
In order to further evaluate the psychological state of the adolescents in relation to their eating habits, a psychologically approved questionnaire was included in this study (31). Statistical analysis showed, that only a very limited effect was seen between the control group and the intervention group after the 6 months period, indicating a low impact of the nutrition workshop on the eating habits. Therefore, results were further analysed as control-and intervention-group combined. Some of the questionnaires scales and subscales signi cantly changed over the course of the study. A signi cant decrease was observed in 'Nutrition preference list-snacks, hearty and SUM' (EPL-S, EPL-D and EPL-SUM), which indicates a decreased consumption of snacks, hearty food and an overall decrease in the ratings of snacks, hearty and healthy food, respectively. A signi cant increase was observed for 'Salutogenic behaviour-Ability of implementing suggestions' (also known as FEV-EU), which indicates an increased ability of implementing suggestions. In general, the changes indicate a trend towards slightly healthier eating habits by the reduction of snacks and hearty food. However, no signi cant increase in healthy food was observed and all changes happened only in a subgroup of participants. In addition, it should be mentioned that the results of the 'Nutrition preference list' (EPL) scale, presented as T-values, are classi ed as 'extremely below the average' compared to the reference group de ned by (31). This is due to the fact, that the participants rated the list of products including snacks, hearty and healthy food with an average of 'neither…nor' (3) or 'not really' (2) which resulted in the low overall value.
The T-value results of the 'Quality-of-Life' (KINDL) questionnaire were classi ed as below average for the BMI-low and the BMI-norm group which would indicate a reduced quality of life compared to the reference values of adolescents with a similar weight range. Regarding the 'Body image' (SKB) scale, the results are within the normal range. Interestingly, for item 5-who is the least attractive-the answer was very diverse ranging from the skinniest person to the chunkiest person.
By the use of this questionnaire, we aimed to evaluate the eating behaviour or change of diet over the course of the study, however we did not expect a substantial change in the outcome of the questionnaire comparing the start and the end of the study since our study participants were not considered to have an eating disorder in any of the scales tested.
A number of limitations have to be considered when interpreting the results. Especially the biomedical parameters have to be judged carefully, since these values are mainly linked to a healthier lifestyle in terms of intake of healthy food. In this study, we could not control for the availability of healthy food including the school buffet or the consumption of fruits and vegetables as snacks provided by the parents (40). In addition, the supportiveness of the school environment and at home to live a healthier lifestyle should be given and this factor could also not be in uenced in this study (21).
Nevertheless, the results demonstrate a positive effect of the performed dietary-and lifestyle-program in terms of body fat, muscle mass and physical activity. A Trend was shown towards a healthier lifestyle regarding muscle metabolism through the creatinine pathway and fat metabolism via a decrease in LDL levels and the LDL/HDL ratio. In addition, the measurement of a set of biomedical parameters facilitate an early detection of physical de ciency symptoms, e.g. anaemia or hyperlipidaemia, which allows for an early therapeutic intervention.
Some of these positive changes were more pronounced in the intervention group and some were present in both the control and the intervention group. These ndings demonstrate that measuring the health status and confronting young people with their well-being might have a positive effect on their well-being and self-empowerment.

Conclusion
Early information and prevention programs in schools, in order to increase the awareness of a healthy lifestyle including dietary health as well as physical health, have a number of important implications. The implementation of such projects as permanent topics in the school curriculum over an extended period of time could improve the health of students, leading to less days of absence from classes, an increased ability for self-perception and self-re ection as well as self-empowerment.
Even though this study lasted only 6 months, we could already detect a positive impact on the health status of the students by an increase in physical tness and increased awareness for nutrition and beverages as well as health-associated parameters. In addition, the fact that the students had regular interrogations regarding their health status and tness might per se lead to an increased awareness and subsequently positive effect.