Effects of statewide health promotion in primary

0 downloads 0 Views 560KB Size Report
children's sick days, visits to a physician, and days parents had to stay off work to care for a sick child. .... (BMI ≥ 30), and abdominally obese (WHtR ≥ 0.5).

Kesztyüs et al. BMC Public Health (2016) 16:1244 DOI 10.1186/s12889-016-3903-2


Open Access

Effects of statewide health promotion in primary schools on children’s sick days, visits to a physician and parental absence from work: a cluster-randomized trial Dorothea Kesztyüs1,2* , Romy Lauer1, Meike Traub1, Tibor Kesztyüs3 and Jürgen Michael Steinacker1

Abstract Background: Based on the World Health Organization’s global school health initiative we investigate intervention effects of statewide health promotion in schools on the numbers of children’s sick days and visits to a physician, and parental days off work due to child illness. Methods: Cluster-randomized trial with 1-year follow-up in primary schools in the state of Baden-Württemberg, Germany. Anthropometric measurements of first and second grade school children were taken by trained staff. Parents filled in questionnaires for information about socio-demographics, health-related variables, numbers of children’s sick days, visits to a physician, and days parents had to stay off work to care for a sick child. Longitudinal differences in the outcome variables were calculated between baseline and follow-up. Intraclass correlation coefficients were determined to quantify a possible clustering of data in schools. Accordingly, linear models and linear mixed models were applied to identify relationships and ascertain significances. Results: Data from 1943 children (1st grade n = 1024, 6.6 ± 0.4 years old; 2nd grade n = 919, 7.6 ± 0.4 years old) were available at baseline. Unadjusted differences regarding both grades were found between mean longitudinal changes in intervention and control group in children’s sick days (−3.2 ± 7.1 vs. -2.3 ± 5.6, p = 0.013), and maternal days off work (−0.9 ± 2.4 vs. -0.5 ± 2.8, p = 0.019). The intervention effect on sick days was adjusted in a linear regression for baseline values, gender and migration background and confirmed for first grade children (B = −0.83, p = 0.003). The intervention effect on maternal days off work lost its significance after adjusting for baseline values. No significant differences were detected in the numbers of children’s visits to a physician and paternal days off work. Conclusions: School-based health promotion slightly reduces sick days in first grade children. Subsequently, parents may not need to stay off work themselves. Small individual effects add up to larger benefits in a statewide implementation of health promotion. Additionally, health promotion may also positively contribute to school success. Trial Registration: The study was registered on the German Clinical Trials Register (DRKS), Freiburg University, Germany, under the DRKS-ID: DRKS00000494. Registered: 25 August 2010. Keywords: Sick leave, Child, Schools, Health promotion, Physicians, Utilization, Parents

* Correspondence: [email protected] 1 Division of Sport and Rehabilitation Medicine, Ulm University Medical Center, 89075 Ulm, Germany 2 Ulm University, Institute of General Medicine, Helmholtzstraße 20, D-89081 Ulm, Germany Full list of author information is available at the end of the article © The Author(s). 2016 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Kesztyüs et al. BMC Public Health (2016) 16:1244

Background In 1995 the World Health Organisation (WHO) launched the global school health initiative to mobilise and strengthen health promotion and education activities at the local, national, regional and global levels [1]. In their subsequent report on the effectiveness of health promotion in schools in 2006, the WHO identified programmes that promote mental health, healthy eating, and physical activity as being amongst the most effective [2]. They detected that most interventions used classroom-only approaches, and some combined a curriculum approach with environmental changes or family involvement [2]. Combined approaches were more likely to be successful, but many interventions appeared to be ineffective [2]. None of the reviews included in the WHO report provided information on the costs, or cost-effectiveness, of health promotion programmes in schools [2]. However, due to a large amount of existing interventions but limited (financial) resources, cost-effectiveness analyses are essential. For such analyses, direct and indirect costs have to be assessed, and the numbers of visits to a physician or productivity losses due to parental absence from work because they have to care for a sick child are important parts of it. Only recently, we determined the cost-effectiveness of school-based health promotion in terms of averted cases of incidental abdominal obesity as a result of the BadenWürttemberg Study (unpublished data). A crosssectional analysis of the baseline data of this study also showed that abdominal obesity in primary schoolchildren was associated with higher rates of children’s sick days and more visits to a physician [3]. We use the term sick days for the findings, because absenteeism may have a broader meaning beyond sickness; even though absenteeism and sick leave in the context of school children is used interchangeably by many authors. While school absenteeism has been studied elsewhere from different perspectives [4], information about health-related absenteeism in primary school children is limited. Some authors found associations of body mass index and obesity with school absenteeism, but results are not consistent [5–10]. In a Dutch population sample of 3960 8-year-old children, obesity was associated with significantly more school absenteeism [5]. In contrast, data from 920 fourth grade children in the United States (US) did not show a significant relationship between the number of school days absent for each child and the BMI percentile category [6]. Another study from the US examined more than 165,000 students in grades 1–12 and found only weak associations between obesity and increased school absences [7]. A national survey in the US of 3470 adolescents reported overweight and obese pupils aged 12–17 years of having 36 and 37% more sick days, respectively, than their normalweight peers [8]. Among 1387 US children (6–11 years) and 2185 adolescents (12–18 years) a relationship between

Page 2 of 9

severe absenteeism and weight status was found only in children but not adolescents [9]. Overweight children were significantly more absent than their normal-weight peers in a US study of 1069 forth to sixth graders [10]. Chronic health conditions like asthma, diabetes or attention deficit hyperactivity disorder (ADHD) lead to higher rates of school absenteeism [11]. The association between children’s health and performance in school is undisputable and school absenteeism poses a threat to academic achievement [6]. Additionally, visits to a physician, and productivity losses due to parental absence from the workplace to care for a sick child, mean higher costs for health services and national economies. The aim of this study is to investigate intervention effects of the school-based health promotion program “Join the Healthy Boat” on parameters of direct and indirect medical costs, namely the number of days children missed school because they were sick, their visits to a physician and the number of days a working parent had to stay at home to care for a sick child.

Methods Study description

As the outcome evaluation of the school-based health promotion program, “Join the Healthy Boat”, the BadenWürttemberg Study (BW Study) was conducted all over the state of Baden-Württemberg in southern Germany. The cluster-randomized prospective trial included an intervention group and a wait-list control group. The participating primary schools were used as units for the randomization process. According to the number of classes and grades, six different school types were identified and used as strata for randomization [12]. Randomization was performed on 164 teachers in 91 schools, and resulted in 45 schools in the intervention group and 46 schools in the control group. The study period covered the academic year 2010/11. Written informed consent was obtained from the parents. The study was registered on the German Clinical Trials Register (DRKS), Freiburg University, Germany, under the DRKS-ID: DRKS00000494. Detailed information on the study protocol has already been published elsewhere [12]. Intervention

Based on the social cognitive theory [13] and the socioecological model [14], scientists from several disciplines and dedicated practitioners, developed the materials for a health promotion program for all four grades of primary school (age group 6–10), following the intervention mapping approach [15]. The “Join the Healthy Boat” program provides teachers in primary school with various materials to integrate into the regular curriculum without the need for additional lessons. Materials include suggestions for

Kesztyüs et al. BMC Public Health (2016) 16:1244

classroom and homework activities, active breaks, and material for parental involvement. The program aims at a reduction of screen media use and intake of sugar sweetened beverages on the one hand, and at an increase in physical activity on the other hand. Significant results were found for subgroups of the intervention group for screen media use and tendencies but no significant differences for physical activity and the consumption of sugar sweetened beverages [16]. To ensure a state-wide implementation of the program, 32 experienced teachers from all over Baden-Württemberg were trained in two seminars in the concept and materials (“train-the-trainer”). In turn, they trained teachers in their region (“peer-to-peer”) in three vocational training sessions, to secure an appropriate application of the program in their classes.

Page 3 of 9

stages of the study. Data for baseline characteristics were available for 1943 pupils. Datasets with longitudinal information about any of the outcome variables sick days of children, visits to a physician and parental days off work were available for 1379 children (71%). Data collection

Baseline data collection was conducted at the beginning of the academic year 2010/11, follow-up measurements took place after a one year intervention period in fall 2011. Four teams of scientific researchers and specially trained students visited each participating class in schools all over Baden-Württemberg, and among others carried out anthropometric measurements and fitness testing of children. Parental questionnaires were handed out by teachers and returned to the data centre.


All pupils within first and second grade classes of teachers who registered for the training on the program in the academic year 2010/11 were eligible for participation. Parents of 1968 pupils gave their written informed consent. Figure 1 shows a flow chart with the respective available numbers of pupils and datasets at the different

Anthropometrics and fitness testing

The weight and height of children were taken by trained staff according to the International Society for the Advancement of Kinanthropometry (ISAK) standards [17]. Waist circumference (WC) was measured midway between children’s ileac crest and lower costal arch to the nearest 0.1 cm. Body Mass Index (BMI) was calculated through dividing weight in kilogram by height in square meter (kg/m2), and was converted to BMI percentiles (BMIPERC) using German reference data for children [18]. Excess weight and obesity were accordingly defined at or above the 90th and 97th age- and gender-adjusted percentile for children, respectively. Waist-to-height ratio (WHtR) was calculated as the ratio of WC and height, each of them in centimetres. Abdominal obesity was defined as a WHtR ≥ 0.5 following the recommendation of McCarthy and Ashwell [19]. A six minute run test was conducted and used as a proxy for cardiorespiratory fitness [20]. Parental BMI and WHtR were calculated according to self-reported weight, height and WC in the questionnaires, and categorized as overweight (BMI ≥ 25), obese (BMI ≥ 30), and abdominally obese (WHtR ≥ 0.5). Questionnaires and derived variables

Fig. 1 Flowchart of participants/datasets in the BW Study 2010–2011. Number of respective written informed consents and available data sets from anthropometric measurements and parental questionnaires at several stages of the outcome evaluation process

Parents were asked to recall the sick days of their children in the last year of school or kindergarten and their children’s numbers of visits to a physician during this period (see Table 1). They gave information about their employment status and their weekly working hours, the monthly household income and the number of days they could not go to work because they had to care for their sick child. Furthermore, both parents gave information on their educational level. They were asked to report their height, weight and WC, and to state their current smoker status. Additionally, single parenthood was assessed.

Kesztyüs et al. BMC Public Health (2016) 16:1244

Page 4 of 9

Table 1 Items in the parental questionnaires for the outcome variables in the BW Study 2010 Outcome variable


Response options (free text)

Child’s sick days

On how many days during the last year of school or kindergarten was your child unable to go to school/kindergarten because they were sick?

Number of days

Visits to a physician

How often during the last year of school or kindergarten did you have to visit a physician because your child was sick?

Number of visits to a physician

Days off work mother/father

If you are employed: On how many days during the last year of school or kindergarten did you have to stay off work because your child was sick?

Father: number of days Mother: number of days

A migration background of the child was assumed if at least one parent was born abroad or at least one parent mainly spoke a foreign language during the child’s first years of life. As parameters of children’s health behaviour, their daily time of outdoor play, the number of days per week they met the WHO guideline of moderate to vigorous physical activity (MVPA) of more than 60 min per day [21], their amount of screen media consumption, their consumption of sugar sweetened beverages, and their breakfast habits were assessed. Family educational level was ranked in accordance with the CASMIN classification system as the highest level of two parents or the level of a single parent [22]. It was dichotomized for analysis into tertiary level, versus primary and secondary level. Monthly household income was grouped into a low (< €1750) and a high (≥ €1750) category. Outdoor playing was dichotomized at above 60 min/day. Reaching the WHO Guideline for MVPA on four days and more a week was compared to reaching it on three days or less. The use of screen media was divided into more than 60 min/day or less, and the consumption of sugar sweetened beverages into more than one time per week or less. Having breakfast was divided into “never”, and “rarely” vs. “often”, and “every day”. Losses to follow-up, missing data

Losses to follow-up and missing data are common problems in observational trials and may bias the results [23]. Therefore, baseline differences between participants who took part in both measurements and those who were lost to follow-up or had missing outcome variables were examined with the appropriate statistical tests as described below. Statistical analysis

Differences between groups were tested according to the scale level and the underlying distribution and variance of the respective variable with Fisher’s exact test for categorical data, the Mann–Whitney U-test, t-test or Welch test for continuous data. Significance level was set to α < .05 for two-sided tests. Sample sizes in the analyses differ because of missing data.

Intraclass correlation coefficients (ICC) were calculated for all outcome variables to quantify a possible clustering of data in schools. Intervention effects on the outcome variables were examined in linear regression models or linear mixed models, the latter with respect to a clustering effect. All variables from Table 2, except for the four outcome variables themselves, were considered as possible influential factors on the outcome variables and therefore included in the process of stepwise linear regression analyses. Baseline-differences were tested for their influence on the outcome in regression analyses, and all reported results from regression analyses were adjusted for the baseline values of the outcome. Descriptive and bivariate statistics were conducted with IBM SPSS Release 21.0 for Windows (SPSS Inc, Chicago, IL, USA). Linear regression models as well as linear mixed models were calculated with the statistical software package R Release 3.1.2 for Windows (http://cran.r-project.org).

Results Baseline characteristics

Baseline characteristics for both groups are displayed in Table 2. Differences between intervention and control group were found in the number of children with a migration background, which was significantly higher in the intervention group. Both parents of children in the intervention group had more days off work to care for a sick child than those in the control groups. Furthermore, children in the intervention group played outside less often and reached a slightly lower distance in the six minute run test. Intervention effects

Data for the longitudinal changes in the outcome variables are shown in Table 3. In total, a decrease appeared in all four variables over the study period of one year. Significant differences occurred in the number of days children missed at school because of sickness and in the number of maternal days off work in favour of the intervention group. Figure 2 visualises these changes. Children in the intervention group had a significantly higher reduction in sick days. This effect remains stable for children in grade 1 after adjustment for sex, migration

Kesztyüs et al. BMC Public Health (2016) 16:1244

Page 5 of 9

Table 2 Baseline characteristics of participants in the BW Study 2010 Missing Values

Intervention (n = 1072)

Control (n = 871)

Total (n = 1943)

Boys, n (%)

536 (50.0)

459 (52.7)

995 (51.2)

Age, years [m (sd)]

7.09 (0.64)

7.06 (0.64)

7.08 (0.64)

Grade 1, n (%)

550 (51.3)

474 (54.4)

1024 (52.7)

207 (28.0)

525 (31.9)

Migration background, n (%)



318 (35.1)

Anthropometry BMIPERC, [m (sd)]


49.13 (28.07)

48.75 (27.67)

48.96 (27.88)

Waist circumference, cm [m (sd)]


55.72 (5.87)

55.41 (5.89)

55.58 (5.88)

WHtR,[m (sd)]


0.45 (0.04)

0.45 (0.04)

0.45 (0.04)

Overweight/obesity, n (%)


107 (10.3)

83 (9.7)

190 (10.0)

Abdominal obesity, n (%)


90 (8.7)

68 (8.0)

158 (8.4)

Single parent, n (%)


101 (10.9)

76 (10.1)

177 (10.5)

Tertiary family educational level, n (%)


293 (32.7)

229 (31.6)

522 (32.2)

Parental characteristics

Household income < 1750 €, n (%)


113 (14.0)

94 (13.8)

207 (13.9)

Overweight/obesity (mother), n (%)


279 (32.1)

217 (30.5)

496 (31.4)

Overweight/obesity (father), n (%)


512 (62.3)

386 (59.1)

898 (60.9)

Abdominal obesity (mother), n (%)


275 (49.0)

206 (44.2)

481 (46.8)

Abdominal obesity (father), n (%)


386 (75.2)

306 (73.0)

692 (74.2)

Smoking (mother), n (%)


172 (20.7)

175 (21.2)

347 (21.0)

Smoking (father), n (%)


234 (29.4)

238 (30.1)

472 (29.7)

Days off work (mother), median [m (sd)]


2 [3.09 (4.62)]2

0 [2.36 (4.23)]

1 [2.77 (4.47)]

Employed (mother), n (%)


657 (72.3)

519 (70.9)

1176 (71.7)

Working hours/week (mother), [m (sd)]


20.83 (10.68)

21.18 (11.79)

20.99 (11.18)



Days off work (father), median [m (sd)]


0 [0.78 (2.31)]

0 [0.33 (1.19)]

0 [0.56 (1.87)]

Employed (father), n (%)


833 (95.6)

694 (96.7)

1527 (96.1)

Working hours/week (father), [m (sd)]


43.60 (10.65)

43.99 (10.35)

43.78 (10.51)

Sick days, median [m (sd)]


5 [7.50 (7.70)]

5 [6.73 (5.97)]

5 [7.15 (6.97)]

Visits to a physician, median [m (sd)]


2 [2.99 (2.76)]

2 [2.95 (3.23)]

2 [2.97 (2.98)]

Health and lifestyle characteristics


Playing outside > 60 min/day, n (%)


607 (66.6)

523 (71.2)

1130 (68.6)

Physical active ≥ 4 days/week ≥ 60 min/day, n (%)


238 (26.6)

199 (27.3)

437 (26.9)

6 min run test, meter [m (sd)]


839.67 (122.88)4

855.28 (120.10)

846.69 (121.85)

Screen media > 1 h/day, n (%)


144 (15.4)

100 (13.2)

244 (14.4)

Sugar sweetened beverages > 1 time/week, n (%)


236 (25.1)

180 (23.6)

416 (24.4)


97 (12.7)

126 (13.4)

223 (13.1)

Skipping breakfast, n (%)

m mean, sd standard deviation, WHtR Waist-to-height ratio, a with regard to those who stated an employment 1 p = 0.002, 2p < 0.001, 3p = 0.048, 4p = 0.006

and baseline values for sick days. First grade children in the intervention and control group had a change of −5.15 vs. -3.64, respectively, in the numbers of sick days between baseline and follow up, indicating a group difference of 1.51 days (p = 0.02). ICC for the differences in sick days was 0.045 (95% CI [0.012; 0.078]), thus indicating that 4.5% of the overall variance was due to the clustering of data in schools. The subsequent analysis in a linear mixed

model controlling for data clustering showed no differences in regression coefficients compared to the simple linear regression model; the latter is shown in Table 4. The difference for maternal days off work lost its significance after adjusting for its baseline values and a random school effect. No intervention effects were found for the paternal days off work and the number of the child’s visits to a physician. These results were confirmed

Kesztyüs et al. BMC Public Health (2016) 16:1244

Page 6 of 9

Table 3 Differences between baseline (2010) and follow-up (2011) in the BW Study

Sick days, m (sd)

Missing values

Intervention (n = 745)

Control (n = 634)

Total (n = 1379)


−3.18 (7.08)1

−2.31 (5.56)

−2.78 (6.44)

Visits to a physician, m (sd)


−0.85 (2.41)

−0.93 (2.70)

−0.88 (2.54)

Days off work (mother), m (sd)


−1.11 (3.89)2

−0.52 (2.84)

−0.85 (3.48)

Days off work (father), m (sd)


−0.25 (2.21)

0.06 (1.56)

−0.10 (1.94)

m mean, sd standard deviation 1 p = 0.013, 2p = 0.019

in linear mixed models adjusted for the respective baseline values and a random school effect. Losses to follow-up, missing data

Datasets with any information on the outcome variables (n = 1379) were compared to those with completely missing outcome variables (n = 564). Those with completely missing outcome variables showed manifold significant baseline differences compared to the others. Lost participants were more likely to be male and to have a migration background; they had larger WC and WHtR and were more often overweight and obese. Their parents were more often single parents, had less often tertiary family education level, and more often a lower household income. Their mothers were more often overweight or obese, both parents were more likely to smoke and mothers were more often unemployed. Children had more sick days in the preceding year of kindergarten or school, they reached shorter distances in the 6 min run test, were more likely to use screen media for more than one hour per day, consumed more sugar sweetened beverages and were more likely to skip breakfast.

Discussion The effectiveness of multicomponent health promotion in a low-threshold approach is not easily determined in

Fig. 2 Longitudinal changes in outcome parameters between baseline (2010) and follow-up (2011) in the BW Study. Numerical changes during the period under observation regarding children’s sick days (n = 1310), children’s visits to a physician (n = 1253), maternal days off work (n = 473), and paternal days off work (n = 732) in order to care for a sick child

a limited time period. The intervention group shows a stronger downward trend than the control group in all outcome variables except the number of visits to a physician. Only the reduction of sick days in first grade pupils remains statistically significant. This reduction in the number of sick days in the intervention group may result from behavioral changes in physical activity, nutrition, and media consumption, as described by Kobel et al. [16], but also from a reduced incidence in abdominal obesity (unpublished data), which is, as mentioned in the background section, correlated with higher rates of sick days [3]. Although the mean group difference of one sick day seems small for the individual child, this adds up to 550 days for all first grade children in the intervention group, which is a significant number. Small individual effects which benefit the whole group are the typical results of a public health intervention [24]. There is, assumedly, a natural decline in the number of sick days from kindergarten through the first years of primary school, which may at least partly be comprehensible with regard to the natural maturing of the immune system. Furthermore, parents may be more reluctant to keep their children at home as soon as they attend school, rather than they would have been during the time in kindergarten. Nonetheless, a small but statistically significant difference between intervention and control group was detected for the decline in sick days in first grade children. Despite being small, this is an important result for various reasons. Firstly, it shows the positive impact of health promotion on health itself. Secondly, the first year of school may be regarded as crucial with respect to the complete education life-cycle and therefore missing fewer days at school may affect school success. The missing significance of longitudinal changes in sick days between intervention and control group in grade 2 may have several reasons. First, second grade children and parents are meanwhile familiar with school and the initial enthusiasm and attention to new information has settled and come to a normal extend. This means, school is now day-to-day life for both, parents and children, and they have developed a certain routine. Hence, first grade children (and their parents) may be more susceptible and eager to learn and implement new, alternative behaviours. Second, the decrease of sick days

Kesztyüs et al. BMC Public Health (2016) 16:1244

Page 7 of 9

Table 4 Results of the linear regression analysis for the change in children’s sick days between baseline (2010) and follow-up (2011) in the BW Study All (n = 1268)

Grade 1 (n = 654)


B (SE)



B (SE)




−0.30 (0.22)



−0.83 (0.28)



Grade 1

−1.15 (0.23)


Suggest Documents

The Steam Engines of Oz | Animation | DJRoom VR Online Radio Disco apk