Context: A cardinal feature of the DSM-IV diagnostic criteria for conduct disorder is the distinction between childhood-vs adolescent-onset subtypes. Whether such developmental subtypes exist in the population and have different prognoses should be rigorously tested to inform the DSM-V.Objectives: To evaluate the epidemiological validity of childhood-vs adolescent-onset conduct problems in a prospective birth cohort, and to assess whether life-coursepersistent conduct problems are associated with a greater adult health burden.Design, Setting, and Participants: Our sample includes 526 male study members in the Dunedin Multidisciplinary Health and Development Study, a 1-year birth cohort (April1,1972,throughMarch30,1973.Developmental trajectories were defined using prospective ratings of conduct problems at 7, 9, 11, 13, 15, 18, 21, and 26 years of age.
Main Outcome Measures:Health burden was assessed as mental and physical health problems at 32 years of age measured via diagnostic interviews and physical examinations.
Results:We identified the following 4 developmental subtypes of conduct problems through general growth mixture modeling:(1)childhood-onset/life-course-persistent,(2)adolescent onset, (3) childhood limited, and (4) low. At 32 years of age, study members with the life-course-persistent subtype experienced the worst health burden. To a lesser extent, those with the adolescent-onset subtype also experienced health problems. A childhood-limited subtype not specified by DSM-IV was revealed; its adult health outcomes were within the range of the cohort norm.
Conclusions:Results support the epidemiological validity of the DSM-IV conduct disorder distinction based on age of onset but highlight the need to also consider long-term persistence to refine diagnosis. Preventing and treating conduct problems has the potential to reduce the adult health burden.