Preschool psychopathology: lessons for the lifespan

Abstract

One of the clearest lessons emerging from psychiatric research over the last decade or so is that what were once thought of as typically ‘adult’ disorders are more often than not reported to have had their onsets in childhood and adolescence (Insel & Fenton, 2005; Kessler et al., 2005). In the subspecialty of child and adolescent psychiatry there has also been a move towards identifying disorders in ever younger groups of children. For instance, 15 years ago, when our group began the Great Smoky Mountains Study (Costello et al., 1996), the decision to include only children aged nine and older was driven by the lack of diagnostic measures of psychopathology for younger children. Similarly, the leading birth cohort studies of the time did not begin to diagnose psychiatric disorders until around the same age, or later (e.g., Kashani et al., 1983). The briefest perusal of the papers included in this special section devoted to preschool psychopathology shows how dramatically the situation has changed. Parentand teacherreport questionnaires led the way (Achenbach, Edelbrock, & Howell, 1987; Behar, 1977; Richman et al., 1974), but now we see represented here a wide range of reliable assessment technologies ranging from such questionnaires to DSM-IV-based symptom checklists and more detailed assessments of preschool behavior, full DSM-IV (American Psychiatric Association, 2000), ICD-10 (World Health Organization, 1987) and DC:0–3-based (Zero to Three, 2005) diagnostic interviews, child selfreports, standardized observational assessments, and neuropsychological batteries. We can now confidently assert that we have wherewithal to assess the psychiatric status of children down to age two. Of course, that does not mean that we have perfected everything we could ever want – that would not be true of psychiatric assessment at any age – but it does mean that there is no general methodological reason to exclude young children from studies of specific psychiatric disorders. The child and adolescent psychiatric field has become used to using the ‘or rule’ in diagnosing disorders in older children – if either the child, the teacher, or the parent reports a symptom as being present, then we regard it as being present. Kerr, Lunkenheimer, and Olsen (this issue) found that mothers’, fathers’, teachers’, and examiners’ ratings of child behavior at age three all predicted problems at age 5, and confirmed that fathers were a valuable independent source of information. Examiner information proved to be of more limited utility, but the use of the examiners (to complete the Teacher’s Report Form (TRF) on a child they did not know prior to the examination) seems likely to have been non-optimal. Wakschlag et al. (this issue) used their examiners to implement a series of presses for disruptive behavior incorporated in the DB-DOS, and found that they provided significant prediction of impairment over and above that derived from maternal reports of the frequency of disruptive behaviors. Coplan, Clossen, and Arbeau (this issue) conducted a short direct interview with five-yearolds about loneliness, and their data suggest that such interviews yield meaningful information. Ialongo and colleagues showed some time ago that questionnaire-based self-reports from first-graders predicted anxiety and depressive problems and diagnoses at ages 10 and 14 (Ialongo, Edelsohn, Werthamer-Larsson, Crockett, & Kellam, 1995; Ialongo, Edelsohn, & Kellam, 2001), but it remains an open question how young we can go with selfreport assessments of this sort. Codings of responses to story stems, doll-house scenarios or puppet-based assessments are viable down to age four or five (see, e.g., Stadelmann et al., this issue), but shorter attention spans, fear of large doggy puppets, and difficulties in inhibiting proponent play responses suggest that more work needs to be done to develop scorable assessments for even younger children. Stadelmann et al.’s paper also adds to the growing evidence that behavioral problems in preschoolers/ kindergarteners are as stable over time as they are in older children and adolescents (Briggs-Gowan, Carter, Bosson-Heenan, Guyer, & Horwitz, 2006). There is also a more general problem with incorporating child ‘self-reports’ into the diagnostic process for younger children. By the age of nine, most children can complete an ‘adult style’ diagnostic interview and, although it has been shown that they have problems with time-based constructs, such as dates of onset (Breton et al., 1995), we can collect information that is formally equivalent to that obtained from parents. This information feeds easily into DSM/ICD-style diagnostic algorithms using the ‘or rule’. In younger children, such adult-style interviews simply do not work, and we do not, as yet, have an agreed-upon set of methods for incorporating other sorts of information into our Conflict of interest statement: No conflicts declared. Journal of Child Psychology and Psychiatry 48:10 (2007), pp 961–966 doi:10.1111/j.1469-7610.2007.01832.x