Screening for attention-deficit/hyperactivity disorder (ADHD): can high-risk children be identified in first grade?

Authors

  • K. Holmberg,

    Corresponding author
    1. Department of Women's and Children's Health, Section for Paediatrics, Uppsala University, Uppsala
    Search for more papers by this author
  • C. Sundelin,

    1. Department of Women's and Children's Health, Section for Paediatrics, Uppsala University, Uppsala
    Search for more papers by this author
  • A. Hjern

    1. Department of Women's and Children's Health, Section for Paediatrics, Uppsala University, Uppsala
    2. Centre for Health Equity Studies, CHESS, Stockholm University/Karolinska Institute, Stockholm, Sweden
    Search for more papers by this author

Kirsten Holmberg MD PhD, Department of Women's and Children's Health, Section for Paediatrics, Central Unit of Child Healthcare, Uppsala University Children's Hospital, SE-751 85 Uppsala, Sweden. E-mail: kirsten.holmberg@kbh.uu.se

Abstract

Aim  Recent studies have demonstrated the beneficial long-term effects of an indicated parent support programme for acting out behaviour in pre-school children with attention-deficit/hyperactivity disorder (ADHD) traits. In this study we wanted to assess different thresholds for screening with the Conners scale for hyperactive–inattentive behaviours in first grade for ADHD in grade four.

Method  The study population consisted of 422 first graders (6- to 7-year-olds) in one municipality in Stockholm County who were screened with Conners 10-item scale and followed up by ADHD assessment in grade four. Sensitivity, specificity, likelihood ratios, and positive predictive value (PPV) of the screening by parents and teachers in first grade for being diagnosed with ADHD in fourth grade were calculated.

Results  The prevalence of pervasive and situational ADHD was 5.7% and 5.9% respectively. A score ≥10 on the Conners scale in first grade in teachers' reports identified 63% [95% confidence interval (CI): 43–79] of children diagnosed with pervasive ADHD in grade four (P < 0.001) with a PPV of 29% and a positive likelihood ratio (LR+) of 6.72. Parental reports of a score ≥10 yielded a lower sensitivity (29%; 95% CI: 15–49), PPV of 20% and LR+ of 4.24 for pervasive ADHD. The best predictor was a combination of parent and teacher scores ≥10 with a PPV of 50% and LR+ of 16.63. Associations with situational ADHD were weak with LR+ of 1.81 and 2.49, respectively, for teachers' and parental scores ≥10.

Conclusions  This study indicates a strong association between a teacher's report of a score ≥10 on the Conners scale in first grade and pervasive ADHD in grade four, while parental reports were less predictive.

Ancillary