Background: Identifying children with childhood-onset neurodevelopmental problems (NDPs, defined here as autism spectrum disorders [ASDs], attention-deficit/hyperactivity disorder [AD/HD], tic disorders [TDs], learning disorders [LDs] and development coordination disorder), using easily administered screening instruments, is a prerequisite for epidemiological research. Such instruments are also clinically useful to prioritize children for comprehensive assessments, to screen risk groups, and to follow controls.Autism-Tics, ADHD, and other Co-morbidities inventory (A-TAC) was developed to meet these requirements; here the A-TAC's prospective and psychometric properties are examined, when used in a population-based, epidemiological setting.
Methods: Since 2004, parents of all Swedish twins have been asked to take part in an ongoing, nation-wide twin study (The Child and Adolescent Twin Study in Sweden). The study includes the A-TAC, carried out as a telephone interview with parents of twins aged 9 or 12. In the present study, screen-positive twins from three birth year cohorts (1993-1995) were invited to a comprehensive clinical follow-up (blinded for previous screening results) together with their co-twins and randomly selected, healthy controls at age 15 (Total N = 452).
Results: Sensitivity and specificity of A-TAC scores for predicting later clinical diagnoses were good to excellent overall, with values of the area under the receiver operating characteristics curves ranging from 0.77 (AD/HD) to 0.91 (ASDs). Among children who were screen-positive for an ASD, 48% received a clinical diagnosis of ASDs. For AD/HD, the corresponding figure was also 48%, for LDs 16%, and for TDs 60%. Between 4% and 35% of screen-positive children did not receive any diagnosis at the clinical follow-up three years later. Among screen-negative controls, prevalence of ASDs, AD/HD, LDs, and TDs was 0%, 7%, 4%, and 2%, respectively.
Conclusions: The A-TAC appeared to be a valid instrument to assess NDPs in this population-based, longitudinal study. It has good-to-excellent psychometric properties, with an excellent ability to distinguish NDPs (mainly ASDs) from non-NDPs at least three years after the screening evaluations, although specific diagnoses did not correspond closely to actual clinical diagnoses.