SINCE I SPENT considerable time working in the Rhode Island juvenile detention center (the “Training School”) in the past year, I’ve developed an interest in how one becomes a delinquent or anti-social adolescent. I don’t mean to imply that youth in the juvenile-justice system constitute a homogeneous group — far from it!

In my experience, detained youth run the gamut from normal teenagers who’ve made a bad mistake (usually under the influence of one or more drugs), to psychiatric patients whose symptoms are primarily behavioral, to a group who are already well launched on a path of criminality. All that these three groups have in common is that they’ve done something they shouldn’t have and gotten arrested for it. However, the determinants, implications and treatment of their behavior are unique for each group.

It’s the third group, the ones who already have a pattern of fighting, stealing, lying, truancy, drug use, running away and incorrigibility, that interests me the most. Understanding how they get that way is a prerequisite to intervening effectively — something we now don’t do very well with this group of adolescents.

We do know a reasonable amount about risk factors for anti-social behavior. Child abuse and/or neglect clearly put a child at risk for conduct disorder, and subsequent anti-social behavior. To treat others with empathy and to view rules with respect, children must experience adults as loving and trustworthy, and their immediate environment as safe. It thus makes intuitive and, the evidence shows, empirical sense that abused children are at greater risk of mistreating others or breaking society’s rules as they age.

But not all abused children exhibit anti-social behavior during adolescence. Recent genetic work may offer an explanation, as a number of genes linked to anti-social behavior have been identified. Most interesting is a gene on the X chromosome that controls the enzyme that breaks down important neurotransmitters. The short version of this gene leads to high levels of the neurotransmitters because its code is to produce little of the enzyme.

In a large and elegant study of 1,000 boys, a strong link was found between childhood abuse and the presence of the short version of the gene in relation to predicting later anti-social behavior. In the group of boys who suffered severe abuse and who had the short gene variant, 85 percent engaged in antisocial behavior. Thus an abusive environment combined with the particular genotype appears to be a powerful determinant of anti-social behavior — a classic example of gene-environment interaction.

Another important longitudinal study found that the clinical picture of ADHD plus conduct disorder in boys 6-12 years of age predicted criminal behavior 30 years later: 44 percent of the subjects with these two diagnoses had been arrested compared with 15 percent of controls.

Less extensive studies have associated a number of disparate factors with anti-social behavior during adolescence, including maternal smoking during pregnancy, neighborhood violence, poverty, childhood head injury resulting in a loss of consciousness and inconsistent discipline. The discouraging element common to most of these factors — with inconsistent discipline being the prominent exception — is how difficult they are to prevent or change. One’s genetic makeup is unalterable, at least for the present; prevention of child abuse has proven as difficult as reducing poverty or violence on a large scale.

There are few studies demonstrating powerful prevention or treatment effects for interventions intended to reduce antisocial behavior. The longitudinal study (above) that identified ADHD and conduct disorder as determinants also evaluated the impact of up to three years of individualized treatment on antisocial behaviors in adolescents. A significant effect was found only for treatment longer than two years, and this benefit was not maintained into adulthood. Perhaps the most encouraging early intervention has been the Nurse-Family Partnership, developed by David Olds, Ph.D., in which nurses provided at-risk families a variety of advice and assistance from the prenatal period through the first two years of the child’s life. The impact on adolescent antisocial behavior has been so substantial that the program is receiving dissemination to various sites worldwide.

Until we can better delineate more determinants of antisocial behavior that are modifiable, or develop better interventions for risk factors that now appear immutable, our success will be modest at interrupting an adolescent’s downward trajectory toward criminality once it is established.

Gregory K. Fritz, M.D., an occasional contributor, is academic director at the E.P. Bradley Hospital, in East Providence, and editor of the Brown University Child and Adolescent Behavior Letter.