Socio–economic differences in health risk behavior in adolescence: Do they exist?

https://doi.org/10.1016/S0277-9536(98)00034-3Get rights and content

Abstract

Socio–economic differences in risk behaviors in adolescence can be seen as a prelude to the re-emergence of socio–economic health differences in adulthood. We studied whether or not socio–economic differences in health risk behaviors are present in male and female adolescents in The Netherlands. The relation between socio–economic status (SES) and health risk behaviors was examined, by testing both the main and interaction effects of SES and gender on separate health risk behaviors on one hand, and on the behaviors cumulatively on the other. The data were derived from 1984 adolescents in the four northern provinces of The Netherlands. SES was measured by means of the educational level and the occupational status of both parents. Four health risk behaviors were included in this study: smoking, alcohol consumption, soft drug use, and (no) physical exercise. We found that the relationships between SES and health risk behaviors are not as linear as is often found in adulthood. Our findings can be characterised overall by an absence of relationship between SES and health risk behaviors. The only exception applies to sport, which is linearly related to SES. Adolescents in the lower SES groups engage in sport less than adolescents in the higher SES groups. There was an irregular relationship between the father's occupational status and the adolescents' smoking and drinking. Adolescents in the highest, lowest and middle of the six SES groups have the highest rates of health risk behaviors. All observed relationships are similar for both male and female adolescents. A relationship between gender and the separate health risk behaviors was found only for alcohol consumption and drug use. For both male adolescents showed higher rates of risk behavior. Males also scored higher on the cumulative health risk behaviors than their female counterparts. The findings of this study do not support the hypothesis of latent differences in adolescence.

Introduction

Socio–economic health differences are age-specific. In contrast to adults and children, most studies have found little variation by social class in morbidity and mortality among adolescents (West et al., 1990; Macintyre and West, 1991; Rahkonen, 1993; Rahkonen et al., 1995). These studies conclude that adolescence is characterised, unlike other stages in the course of life, more by the absence than the presence of socio–economic health differences. Recently, however, some suggestions have been made that there may be socio–economic health differences in adolescence (Tuinstra et al.). However, these differences are relatively small and are also found only for a limited number of health measures.

Even if there are no consistent social class differences in health among adolescents, there might be socio–economic differences in health risk behaviors during this period of life. These health risk behaviors have an adverse effect on health in the long term, which has often been established (Pietilä et al., 1995). Therefore, socio–economic differences in health risk behaviors in adolescence are one of the predictors of socio–economic health differences in adulthood. We consider this hypothesis as the hypothesis of latent differences: while there are few differences in health in adolescence, health behaviors are class differentiated. We refer specifically to health risk behavior, in contrast to West, who used the same concept for, for example, height (West, 1988).

In addition, in the inequality debate, lifestyle or behavioral factors are seen as providing an important area of explanation for socio–economic health differences (Van der Lucht, 1992; Ranchor, 1994; Stronks et al., 1996). Studies on adults have established that these behavioral factors often show a pattern which is disadvantageous to the lowest socio–economic status (SES) groups (Ranchor et al., 1990; Graham, 1994; Ranchor, 1994; Kooiker and Christiansen, 1995; Van der Lucht and Groothoff, 1995; Stronks et al., 1996; Sanderman et al., submitted). People with a low SES exhibit more risk behaviors, such as smoking, poor diet and physical inactivity, than high-SES people. Alcohol consumption, however, is a common exception. The relation between SES and alcohol consumption is usually weak or reversed (Mackenbach, 1992; Mulder et al., in press).

Evidence about the relationship between SES and health risk behaviors in adolescence, however, is often inconsistent or even contradictory. Some studies support the hypothesis that behavior like smoking, drinking, drug use and physical inactivity is more frequent among adolescents with a low family SES (Green et al., 1991; Greenland et al., 1995; Pietilä et al., 1995; Lowry et al., 1996). Other studies reject this hypothesis and show that SES has little or no relationship to health risk behaviors (O'Malley et al., 1993; Donato et al., 1994; Glendinning et al., 1994; Donato et al., 1995). The picture gets even more complex with the introduction of differences between male and female adolescents.

Recently, the importance of the combination of health risk behaviors has been acknowledged (Osler, 1993), mainly supported by the very strong associations between the behaviors (Lytle et al., 1995). Also, the cumulative effect of health risk behaviors is to make people vulnerable to health complaints, so that in combination they have a greater power to predict the occurrence of ill health. Recent studies support the idea that an accumulation of health risk behaviors is associated with SES, both for adults (Sanderman et al.; Mulder et al., in press) and for adolescents (Raitakari et al., 1995; Lowry et al., 1996). These socio–economic differences in adolescence might be one of the preludes to the re-emergence of socio–economic health differences in adulthood.

Three questions are asked: firstly, whether a relationship could be found between SES and health risk behaviors, for the separate health risk behaviors on the one hand, and for the combination of these behaviors, on the other. Secondly and thirdly, we examined the main and the interacting influences of gender in the relationship between SES and health risk behavior.

Section snippets

Procedure and sample

Data were collected in 1994 and 1995 with the cooperation of four Public Health Services in four provinces in the north of the Netherlands, one from each province. In these areas, a random sample of 18 secondary schools, vocational and high schools, was drawn. Only two of the 18 schools approached refused to cooperate. To acquire the desired sample number of adolescents, an additional sample was drawn from two areas, one school per area. Eventually, 1984 of the total sample (N=2090) of fourth

Separate health risk behaviors

The first set of analyses examined the effect of the SES of the parents, of gender and of the interaction between these two variables on the separate health risk behaviors (Table 1).

As can be seen from Table 1, the relations of SES and gender vary for the different behaviors. With respect to SES, we found no overall relationship with smoking and alcohol consumption. For drug use and physical exercise, the relationship with SES was statistically significant.

Taking a closer look at the overall

Conclusion and discussion

The most important conclusion of our study is that the idea of latent differences in health in adolescence is in general not supported by the data. Latent differences are defined as a combination of a lack of socio–economic health differences with the presence of differences in health risk behaviors, which can be seen as one of the forerunners of the re-emergence of socio–economic health differences in adulthood.

The overall conclusion is that no relationships exist between SES and smoking,

References (33)

  • P. West et al.

    Social class and health in youth: Findings from the West of Scotland Twenty-07 Study

    Social Science and Medicine

    (1990)
  • Berkel-van Schaik, A. B. and Van Tax, B. (1990) Naar een standaardoperationalisatie van sociaal–economische status voor...
  • CBS, Centraal Bureau voor de Statistiek (1985) Beroepenclassificatie 1984. CBS,...
  • F. Donato et al.

    Smoking among high school students in 10 Italian towns: Patterns and covariates

    The International Journal of the Addictions

    (1994)
  • G. Green et al.

    Like parent like child? Associations between drinking and smoking behaviour of parents and their children

    British Journal of Addiction

    (1991)
  • K.J. Greenland et al.

    Impact of father's education and parental smoking status on smoking behavior in young adults

    American Journal of Epidemiology

    (1995)
  • Cited by (0)

    View full text