News and Events

Dr. Stephen Hinshaw discussed several principles of developmental psychopathology: (a) there is a mutual interplay of normal and atypical—perspectives on normal and atypical development are mutually enhancing and informative; (b) psychopathology is a deviation from normal developmental processes; (c) context matters—it is important to adopt a transactional approach; (d) development may be thought of in terms of continuities and discontinuities of two types—across time and in terms of normal vs. abnormal (e) development may reflect equifinality (differential pathways lead to a common endpoint) or multifinality (the same initial state leads to diverging outcomes); and (f) longitudinal investigations are necessary (e.g., in revealing information about the temporal course of events), but are not sufficient, in terms of drawing causal inferences. Causal risk factors are best ascertained via experimental means. The field has underutilized “experiments of nature” and developmentally informed prevention and intervention trials, as ways of improving our ability to determine causal inferences about the nature of psychopathology from longitudinal data. As Rutter and colleagues have stated, alternative strategies for disentangling various kinds of environmental effects include adoption, blended family, and twin study designs, as well as migration designs, studies looking at time trends, and studies looking at genes as moderators. "Hybrid" designs are recommended (see Howe, Reiss,, & Yuh) , in which the best aspects of clinical or prevention trials (e.g., random assignment) are coupled with the best aspects of longitudinal correlational designs (e.g., multiple data collection points during interim periods, careful attention to mediational processes).

There are a growing number of experimental trials. One study discussed in greater detail was a multi-site clinical trial of 579 children with ADHD, Combined Type. Children (ages 7-9.9 years) were randomly assigned to 14 months of (a) medication management, (b) intensive behavioral treatment, (c) a combination of medication and behavioral treatment, or (d) a community-treated comparison. Across a 14-month period, medication alone fared slightly better than combined treatment, and both were preferable to behavior-only or community treatment. One-third of children in the sample had a comorbid anxiety disorder; for these children, combined treatment fared best (followed by medication-only, then behavior-only, and lastly the community treatment). For the two-thirds of the sample without an anxiety disorder at baseline, over a 450-day period (i.e., 1415 months), those receiving combined treatment or medication- only fared better, compared to those receiving behavior-only or community treatment. In addition, negative/ineffective discipline was found to be a mediator of improvement for school outcomes. This parenting variables inappropriately harsh discipline and inconsistent discipline (e.g., yelling, slapping; promising rewards without following through). For children with parents showing the greatest improvments with respect to this category of harsh discipline, after 14 months, combined treatment proved to be strongly effective, bringing children into the normal range with respect to school-based aggression and disruptive behavior.

Four key conclusions were discussed: (1) For developmental psychopathology, prospective longitudinal investigations are mandatory, but not sufficient; (2) we must build on the science of “basic” processes (e.g., biological, cognitive, emotional, family, peer); (3) there are limits of passive longitudinal studies regarding inferring causality, but this can be attempted by strategies such as controlling key variables, or pulling apart variables that usually go together; (4) it is important to design clinical trials with moderator and mediator variables, which can help to specify important subgroups showing differential response to intervention and important processes responsible for change.

Stephen Hinshaw's presentation "The Role of Longitudinal Investigations in Developmental Psychopathology: Necessary but Not Sufficient" can be viewed in PDF format, using Adobe® Acrobat® Reader®.