New Research
Simple Identification of Complex ADHD Subtypes Using Current Symptom Counts

https://doi.org/10.1097/CHI.0b013e31819996baGet rights and content

Abstract

Objective

New attention-deficit/hyperactivity disorder (ADHD) subtypes identified through latent class analysis have been recently proposed. Here, we assess the accuracy of simple rules based on symptom counts for the assignment of youths to clinically relevant population-derived ADHD subtypes: severe inattentive (SI) and severe combined (SC).

Method

Data from 9,675 twins and siblings from Missouri and Australia aged 7 to 19 years were analyzed using continuous and categorical models of ADHD symptoms using principal components analysis and subtyping by DSM-IV and by latent class criteria. Cut points were derived for classifying SI and SC subtypes by positive predictive value, negative predictive value, percent positive agreement, and Matthew coefficient of agreement.

Results

Principal components analysis suggested two underlying factors: total number of symptoms and symptom type, with SI and SC latent class subtypes clearly mapping to distinct areas on a plot of these factors. Having six or more total symptoms and fewer than three hyperactive-impulsive symptoms accurately predicts the latent class SI subtype. The latent class SC subtype was best identified by 11 or more total symptoms and 4 or more hyperactive-impulsive. The DSM-IV ADHD subtype criteria accurately identified the SC subtype but only poorly for the SI subtype.

Conclusions

Symptom counts criteria allow the simple and accurate identification of subjects with severe ADHD subtypes defined by latent class analysis. Such simple symptom counts corresponding to screening cut points selected latent class-derived SI subtype subjects with greater precision than DSM-IV criteria.

Section snippets

Description of Samples

Data for this study came from three distinct population-based samples (Table 1): the population-representative Missouri Adolescent Female Twin Study (MOAFTS) examined risk factors for alcoholism in young women and adolescents; Missouri Twin Study (MOTWIN) identified both male and female twins from the Missouri Twin Registry for a genetic epidemiological study of ADHD; and Australian Twin ADHD Project (ATAP) is a large mail survey of child and adolescent twins and siblings participating in the

Statistical Methods

Analyses were conducted using Stata Version 10 (StataCorp, College Station, TX). To validate the use of symptom counts in later analyses and provide a template onto which DSM-IV and latent class subtypes can be compared, we first applied principal components analysis (PCA) to the 18 DSM-IV ADHD symptoms, starting from the polychoric correlation matrix and using only observations from the training sample. Extracted factors were not rotated to avoid placing any preexisting diagnostic conventions

Sample Characteristics

The final analysis sample (Table 1) consisted of 4,079 individual monozygotic twins, 4,532 individual dizygotic twins, and 1,127 nontwin siblings with a mean age of 12.5 ± 3.3 years at assessment. The predominantly female (69%) characteristic is largely due to MOAFTS being 100% female, whereas the ATAP and MOTWIN samples exhibit a more equal sex distribution (50% female in ATAP; 35% in MOTWIN).

Overall, 4% of the subjects were assigned to the SI latent class, and 4.6% of the subjects were

Discussion

These results show that it is possible to identify clinically relevant latent class ADHD subtypes using symptom counts only (Table 4, Table 5). Symptom use estimates from the DSM-IV field trials indicated similar predictive ability for all 18 ADHD symptoms. Inattentive and hyperactive-impulsive items were found to better identify subjects with those named characteristics (e.g., inattentive symptom 1 with primarily inattentive type) than cross-dimension (e.g., inattentive symptom 1 with

References (31)

  • American Psychiatric Association

    Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, (DSM-IV)

    (1994)
  • S Setterberg et al.

    Parent and Interviewer Version of the Children's Global Assessment Scale

    (1992)
  • BB Lahey et al.

    DSM-IV field trials for attention deficit hyperactivity disorder in children and adolescents

    Am J Psychiatry

    (1994)
  • K McBurnett et al.

    Symptom properties as a function of ADHD type: an argument for continued study of sluggish cognitive tempo

    J Abnorm Child Psychol

    (2001)
  • ER Rasmussen et al.

    Replication of the latent class structure of attention-deficit/hyperactivity disorder (ADHD) subtypes in a sample of Australian twins

    J Child Psychol Psychiatry

    (2002)
  • Cited by (32)

    • Altered structural connectivity is related to attention deficit/hyperactivity subtypes: A DTI study

      2016, Psychiatry Research - Neuroimaging
      Citation Excerpt :

      Additionally, children who had been exposed to stimulants and those who had used psychotropic medication within the last six months or had a lifetime history of any substance use disorder were excluded. Patients who sought child psychiatry assessment for ADHD and who were selected as candidates for the study after a screening interview were given the DSM-IV Disruptive Behavior Disorders Rating Scale (parents and teacher form) (Ercan et al., 2001), the Teacher's Report Form (TRF) (Achenbach, 1991), and the Child Behavior Checklist CBCL (parents) (Volk et al., 2009). We invited for psychiatric interview children with ADHD presenting inattention scores that were at least one standard deviation above the mean for the child's age according to both the parent and teacher ratings.

    • Differential oscillatory electroencephalogram between attention-deficit/ hyperactivity disorder subtypes and typically developing adolescents

      2014, Biological Psychiatry
      Citation Excerpt :

      A potential factor that might slightly reduce our capacity to detect a significant difference in beta suppression between the CB and IA group is the cutoff we employed (3 or fewer hyperactive/impulsive symptoms) for inclusion in our IA group. However, some researchers have argued for the use of two or less hyperactive/impulsive symptoms as a conservative inclusion factor for these “pure” IA individuals (56,65). In the current study we used an established cognitive paradigm to identify specific processes involved in the pathology of ADHD.

    • The latent classes of subclinical ADHD symptoms: Convergences of multiple informant reports

      2012, Research in Developmental Disabilities
      Citation Excerpt :

      Previous research using various sources of data (different rating scales, interviews on symptoms of DSM-IV) for LCA provided well-interpretable and useful evidences regarding understanding the structure of ADHD, and regarding the refinement of genetic analyses (e.g., Hudziak et al., 1998, 1999; Neuman et al., 1999; Todd et al., 2001). Some of the identified latent classes can be considered as alternative subtypes, which could serve as a basis for the further refinement of the original ones (for an example on the predominantly inattentive phenotype see Schmitz et al., 2010; Volk, Todorov, Hay, & Todd, 2009). A remarkable observation concerning the LCA studies is that the number of children classified into some latent class other than non-symptomatic varies in a wide range, from about 30 to 39.4% (Althoff et al., 2006) to 63% (Hudziak et al., 1999) in non-referred samples of boys.

    View all citing articles on Scopus

    This article was reviewed under and accepted by Ad Hoc Editor F. Xavier Castellanos, M.D.

    The authors gratefully acknowledge the support of National Institutes of Health grants MH52813 (R. T), MH083823 (A. T), ES013678 (H. V), and MH074272 (H.V.). The Australian work has been supported by the National Health and Medical Research Council (Australia) and the Australian Twin Registry.

    The authors thank Andrew C Heath for access to the Missouri Adolescent Female Twin Study (MOAFTS) data set and the Missouri Department of Vital Statistics for access to Missouri Birth Records. This work was conceived and conducted under the guidance of Dr. Richard D. Todd, who passed away from complications of leukemia.

    View full text