Two approaches to school health promotion—a focus on health-related behaviours and general competencies. An ecological study of 25 Swedish municipalities

Abstract

The school is an important potential health-promoting setting for children and adolescents. Two main perspectives on school health promotion have been identified, one addressing health-related behaviours, the other stressing the development of general competencies. From a policy perspective, it is important to establish whether these two approaches are complementary or competing. This question was analysed by examining school administrations in 25 municipalities in Stockholm County, Sweden. An attempt was made to relate level of activity in line with each approach to general adolescent health outcomes. Data related to the two approaches were collected through the administration of a questionnaire. Outcome variables, measured as fraction of students qualified for upper secondary school, fraction of students with high alcohol intake, and fraction of 15–18 year olds suspected of crime, were measured on the basis of registry data. To control for structural patterns, socio-economic, demographic and system characteristics were gathered from public statistics. Spearman rank correlation coefficients were computed. Clusters of determinants assignable to each of the two approaches—promotion of health-related behaviours and supporting the development of student competencies—were identified. The two approaches were found to be negatively correlated. Thus, the perspective focusing on promotion of student health-related behaviours and the one emphasizing general development of student competencies seem to be competing rather than complementary. If this result is confirmed by other studies, it will have important policy implications.

INTRODUCTION

The school is an important potential health-promoting setting for children and adolescents. Two main perspectives on school health promotion can be distinguished: an approach focusing on the promotion of health-related behaviours, and one emphasizing the general development of student competencies.

It is common in school settings to address health-related behaviours (e.g. alcohol and drug abuse, smoking and physical inactivity) by means of health education. Long-term effects, however, are seldom achieved (Peterson et al., 2000; Foxcroft et al., 2003; Wiehe et al., 2005). In accordance with the approach of the health-promoting school, health education is combined with modification of the school environment to support the development of health-promoting behaviours (Denman et al., 2002). Educational and environmental measures are expected to reinforce each other. However, empirical evidence for the effectiveness of the health-promoting school approach is limited (Denman et al., 2002; Mũkoma and Flisher, 2004).

The second perspective on school health promotion places an emphasis on students' competencies. Education is presumed to enhance a wide range of competencies in children and adolescents and, by that means, improve students' health and social adaptation (Keating and Herzman, 1999). Several studies have shown that, if children develop linguistic and mathematical skills, the risk of school failure decreases, which, in turn, may diminish the risk of mental ill health (Sylva, 1994; Case et al., 1999; Willms, 1999). Poor school achievements have also been associated with negative health-related behaviours, such as smoking (Bryant et al., 2000) and physical inactivity (Tammelin et al., 2003). In a recent WHO study of young people's health, academic achievements were found to be associated with fewer health complaints, good general self-rated health, greater overall life satisfaction, and a lower risk of smoking (Ravens-Sieberer et al., 2004). In a theoretical paper, Markham and Aveyard suggested that education facilitates the ability to function well, and that this ability is the basis for good health (Markham and Aveyard, 2003). According to Markham and Aveyard, schools do not need health education classes or health education staff to be health promoting. On this line of thinking, health promotion at school means development of the school in its core task. Over the past few decades, the ‘effective school’ concept has been central to understanding the characteristics of schools that are effective in producing socially well-adjusted students with good academic results (Rutter et al., 1979). Effective schools are characterized by high expectations of student achievement, learning and behaviour, an emphasis on student responsibilities and rights, monitoring progress at all levels, effective leadership, a positive school culture, staff development and parental involvement (Reynolds and Teddlie, 2000). Because ‘effective schools’ promote student competencies, these schools might also be expected to promote student health.

To conclude, two main perspectives on school health promotion have been identified, one that addresses health-related behaviours, the other that stresses the development of general competencies. From a policy perspective, it is important to establish whether these two approaches are complementary or competing. This question was analysed by examining school administrations in 25 municipalities in Stockholm County, Sweden. A further issue considered was the relation between level of activity within each approach to adolescent health outcomes.

STUDY POPULATION AND METHODS

The study encompassed 25 municipalities in Stockholm County, Sweden, with an average of 42 000 inhabitants. Its design was ecological and cross-sectional. An ecological study is one in which both exposure and outcome variables are measured at group level, as either derived variables (based on aggregated individual data) or integral variables (based on contextual characteristics without individual level analogues).

The Swedish school system

In 1989, the Swedish Government resolved that municipalities should be responsible for local schools. Consequently, the Swedish school system is based on policies at three levels of governance: school law and the curriculum at national level, municipal school plans and the operational plans of individual schools. Because the municipalities govern the schools, the focus of the study was on municipal school administrations rather than particular schools. The Swedish perspective is of special interest because of the relatively homogeneity of its school system, in which all pupils are supposed to have equivalent access to high-quality education. Indeed, differences in academic results between schools are lower in Sweden than other OECD countries (The Swedish Ministry of Education and Science, 2004). Moreover, social variation between Swedish schools is comparatively small.

School activity variables

School activity variables, designed to capture municipal intentions to improve health-related behaviours and the general competencies of students, were collected through the administration of a questionnaire. Item areas, response alternatives and coding are shown in Tables 1 and 2.

To our knowledge, there are no general valid measures of the relevant constructs at central municipal level. Hence, a questionnaire was elaborated and pre-tested in six municipalities (all of which were located outside Stockholm County). The activities selected for questionnaire items—to the greatest possible extent—have been shown by earlier research to be effective, present in Swedish municipalities, and accessible by the method used in this study. It was made clear that the questions pertained to activities that the municipality was responsible for, even if contractors other than the municipality actually ran the schools in question. The questionnaires were sent to public officials in municipal school administrations in February 2002. One written reminder was sent, and several telephone calls were made, until responses had been received from all the municipalities (July 2002).

The questionnaire comprised 35 items, formulated to capture health- and competence-promoting activities aimed at school children governed by municipal school administrations. All activities were weighted independent of quality of evidence, or level of reach into the municipalities. First, responses to the items were given a preliminary score ranging from 0 to 1—based on percentages, dichotomous variables or rating scales. Next, standardized scores were computed with a mean of 1.00. A high level of activity corresponds to a high score.

Items with low variance were removed from the data set before correlation analyses were performed. The following were removed from the ‘health-related behaviour variable group’: municipal aims and follow-up methods concerning sports and health, municipal aims concerning students' way to school, and local policies for preventing bullying and gathering local bullying statistics. The items with the following values were removed from the ‘general competence variable group’: schools with a local plan of action, schools producing quality accounts, head-teachers on time-restricted employment contracts, municipal school policies, further education for school managers, municipal aims concerning the influence of parents and students and the objectives of pedagogic activities.

Four determinant groups were constructed in the final data set: physical health promotion, bullying prevention, municipal control and school development, and the influence of students and parents. The two former groups were related to the perspective that focuses on promotion of the health-related behaviours of students, whereas the latter two groups related to the perspective emphasising the general development of student competencies.

Socio-economic, demographic and system variables

To control for structural patterns, socio-economic, demographic and system characteristics of the 25 municipalities were collected from the Swedish National Agency for Education and Statistics Sweden.

Outcome variables

Registry data were collected on fraction of students qualified for upper secondary school (the Swedish National Agency for Education), fraction of students with high alcohol intake (the Swedish Council for Information on Alcohol and Other Drugs), and fraction of 15–18 year olds under suspicion of crime (the Police Authority in Stockholm County).

Statistical analysis

Spearman rank correlation coefficients (Bland, 2000) were computed to assess the relationship between health- and competence promoting measures, socio-economic variables and outcomes. Statistical significance was set at p < 0.05. A tendency level was set at p < 0.10.

RESULTS

Descriptive data

Socio-economic, demographic and system characteristics of the 25 municipalities, collected from public statistics, are shown in Table 3. The variable ‘adults with more than 12 years of education’ served as a socio-economic status control variable.

Questionnaires were used to collect information about municipal intentions to promote positive health-related behaviours and general competencies on the part of students. Responses were obtained from all 25 municipalities. The municipalities generally seemed to be rather well endowed with such activities, although—as can be seen from Tables 1 and 2—variation was considerable. There were a lot of missing responses to the question about health-promoting schools, and the item was therefore removed from the data set.

To relate level of health- and competence-promoting activities in school to adolescent health outcomes, the following three variables were assessed: fraction of students qualified for upper secondary school, fraction of students with high alcohol intake and fraction of 15–18 year olds under suspicion of crime (Table 4).

Correlation analyses

After removing questions with low variance from the data set, four determinant groups were created according to levels of physical health promotion, bullying prevention, municipal control and school development and the influence of students and parents. Spearman rank correlation analyses were performed both within and between the four groups.

A significant positive correlation was detected between the two groups of determinants that reflected the perspective focusing on the promotion of student health-related behaviours. The same condition existed for the two determinant groups that reflected the other perspective, i.e. the general development of student competencies. Thus, clusters of determinants assignable to each of the two approaches to school health promotion were discerned. A significant negative correlation was identified between one determinant group linked to the health-related behaviour perspective, and one determinant group linked to the general competence perspective (Table 5).

No statistically significant results were obtained regarding outcomes. However, one determinant group—physical health promotion—did show a tendency to a positive relationship with the outcome variable ‘15–18 year olds under suspicion of crime’. Another determinant group—municipal control and school development—tended to be negatively correlated with this outcome variable (Table 6).

Fraction of adults with more than 12 years of education served as a socio-economic status control variable. No significant correlations were detected between the determinant groups and the socio-economic status control variable (Table 5). As expected, the two outcome variables ‘students qualified for upper secondary school’ and ‘15–18 year olds suspected of crime’ were found to be significantly correlated with the socio-economic status control variable (Table 6).

In sum, the health- and competence-promoting activities measured by the variables were found to be quite common in the municipalities covered by the study. Clusters of determinants assignable to the two perspectives on school health promotion were distinguished, and a negative relation was revealed between them. Only tendencies were detected concerning the outcome variables related to the different approaches.

DISCUSSION

The variables connected with the health-related behaviour perspective were found to cluster, as did the variables connected with the general competence perspective. The two perspectives were negatively correlated.

The very fact that there was clustering among the variables connected with the two perspectives on school health promotion indicates that these variables reflect meaningful constructs. The negative correlation between the determinants linked to the health-related behaviour perspective and the determinants linked to the general competence perspective suggests a competing relation between them. This finding indicates that municipalities may largely concentrate on either one of the two approaches. In other words, if decision makers in a municipality focus on, for example, the control and follow-up of academic achievement in school, they will tend to focus less on, for example, the promotion of physical activity.

The responses given in the questionnaires were not validated by information from any independent source. Activities aimed at supporting student competencies generally occurred more frequently than activities aimed at promoting student health-related behaviours. This finding was expected because ‘effective school characteristics’ are very close to the core business of any ordinary school. Yet, this focus might also be related to the interests of particular respondents. In a similar way, a focus on health-promoting activities might reflect other interests on the part of respondents. However, because all respondents belonged to a homogenous group of municipal school administrators, it is unlikely that our results are attributable to response bias.

There were no significant effects regarding the outcome variables. Such lack of effects might be because the school activity variables or the outcome variables, or both, were too indistinct to detect potential differences. However, in light of the clustering of variables within each of the two perspectives, we believe that the school activity variables are valid. Further, the outcome variables correlated with the socio-economic status control variable as expected, which supports the validity of the outcome variables. However, despite assumed validity of variables, we could not demonstrate any significant outcomes. This reflects a general problem in public health research, where long causal chains in public health interventions are typical (Victora et al., 2004). In this study, the causal chain extends from municipal policy, through local school plan and classroom activity, to student health-related behaviours and general competencies, and finally to health outcomes. Thus, the distance between measures and effects is long, with several intermediate links, which naturally dilutes potential effects.

An ecological study design implies that both exposure and outcome variables are measured at group level, as either derived or integral variables. The exposure variables in this study are integral, and the outcome and socio-economic status variables derived. No conclusions are drawn at individual level. Because inference is at group level, the ecological fallacy cannot apply (Diez Roux, 2002).

Socio-economic status often influences outcomes in public health (Marmot et al., 1997; Power et al., 1998; Cubbin et al., 2000). In this study, we used the fraction of adults with more than 12 years of education as a municipal socio-economic status control variable. No significant correlations were found between this control variable and the school activity variables. Thus, intentions to promote student health-related behaviours and intentions to support the general development of student competencies appear in municipalities of both high and low socio-economic status.

The study describes intentions to promote student health-related behaviours and to support the general development of student competencies in 25 Swedish municipalities. Clusters of determinants were detected within each perspective, and a negative relation was found between them. Thus, the two perspectives (student health-related behaviours and the general development of student competencies) seem to be competing rather than complementary. If this result is confirmed in other studies, it will have important policy implications.

The authors would like to acknowledge the municipal officials who meticulously responded to the questionnaire. The study was funded by the Swedish National Institute of Public Health.

REFERENCES