Adolescents in high-risk trajectory: clustering of risky behavior and the origins of socioeconomic health differentials

Abstract

Background: We have evaluated high-risk behavior of adolescents 12 to 17 years of age on the basis of seven binomial psychosocial variables in order to assess whether there is a tendency of these variables to cluster in the same individuals and to identify socioeconomic covariates of risky behavior.

Methods: Study participants were 547 adolescents from four high schools in Greece: two in rural areas, one in an upper-medium socioeconomic class areas, and one in a low-to-medium socioeconomic class area of Athens. Clustering was assessed by evaluating concordance of high-risk attributes examined in pairs, and was expressed as a series of odds ratios (ORs) as well as by factor analysis.

Results: All but one OR were higher than the null value, but they were particularly high with respect to smoking and nonuse of safety belts (OR = 3.2, P < 10(-4)), smoking and binge drinking (OR = 3.3, P < 10(-4)), smoking and riding with a drunk driver (OR = 5.3, P = 10(-4)), smoking and driving under the influence of alcohol (OR = 9.7, P < 10(-4)), nonuse of oral contraceptives and riding a car with a drunk driver (OR = 15.4, P = 0.002), and driving under the influence of alcohol and riding with a drunk driver (OR = 18.6, P < 10(-4)). Factor analysis indicated that risky behavior could be explained in terms of two component factors, namely carelessness in the context of self interest and irresponsible sexual behavior. A composite index integrating information of all seven high-risk indicators regressed on sociodemographic characteristics showed that risky behavior increased sharply with age and was concentrated strongly in the low-education families and the lower income areas.

Conclusions: Several aspects of high-risk behavior tend to aggregate in the same individuals, and the clustering pattern has already been developed by late adolescence, mostly among the less privileged families and population groups. It appears that socioeconomic class health differentials may have strong roots in late adolescence.