A Quantitative Hierarchical Model for DSMV

1
A Quantitative Hierarchical Model for DSM-V

• David Watson
• University of Iowa
• October 29, 2005

2
The Basic DSM Structure

• Symptoms organized into disorders
• Disorders organized into diagnostic classes
• Thus, DSM organization implies an underlying
  structure
• Related symptoms define disorders
• Related disorders define classes

3
DSM-IV A Rational Taxonomy

• Diagnostic classes rationally based on shared
  phenomenological features
• Mood disorders disturbance of mood as the
  predominant feature
• Anxiety disorders symptoms of anxiety and
  avoidance behavior

4
DSM-IV Mood Disorders
5
DSM-IV Mood Disorders
6
DSM-IV Anxiety Disorders
7
DSM-IV Anxiety Disorders
8
Basic Structural Problems with Current DSM Scheme

• DSM-IV taxonomy fails to model strong mood and
  anxiety disorder comorbidity
• Current depressed-anxious mood distinction does
  not represent optimal subdivision of these
  disorders

9
Correlations between PANAS-X Fear Sadness Scales

• _____________________________________
• Rating Type Overall N Mean r
• _____________________________________
• Self-Ratings 8,685 .58
• Other-Ratings 978 .54
• _____________________________________

10
Correlations between MASQ GD Depression
Anxiety Scales

• _____________________________________
• Sample Type Overall N Mean
  r
• _____________________________________
• Non-distressed 4,272 .68
• Distressed 1,589 .74
• _____________________________________

11
Correlations between IDAS Depressed Anxious
Mood Scales

• __________________________________________
• Sample N
  r
• __________________________________________
• High School Students 247 .78
• College Students 673 .76
• Community Adults 362 .78
• Psychiatric Patients 353 .77
• __________________________________________

12
Tetrachoric Correlations between Major
Depression GAD

• __________________________________________
• Sample N
  r
• __________________________________________
• NCS (U.S.) 8,098 .59
• NEMESIS (Wave 1) 7,076 .68
• NEMESIS (Wave 2) 5,618 .70
• Australian NSMHWB 10,641 .66
• __________________________________________

13
Tetrachoric Correlations between Dysthymia GAD

• __________________________________________
• Sample N
  r
• __________________________________________
• NCS (U.S.) 8,098 .64
• NEMESIS (Wave 1) 7,076 .67
• NEMESIS (Wave 2) 5,618 .70
• Australian NSMHWB 10,641 .69
• __________________________________________

14
Rethinking the DSM

• Problems indicate need for alternative approach
• Quantitative structural analyses --gt
• better, more accurate taxonomy
• Current rational system reflects hypothesized
  similarities
• Replace with quantitative scheme which captures
  actual similarities between disorders

15
Factor Loadings from CFA of NCS Data

• _________________________________________________
• Disorder Distress
  Fear
• _________________________________________________
• Major Depression .62
• Dysthymia .53
• GAD .47
• PTSD .40
• Simple Phobia .54
• Agoraphobia .51
• Social Phobia .49
• Panic Disorder .41
• _________________________________________________

16
Factor Loadings from CFA of NEMESIS Data (Wave 1)

• _________________________________________________
• Disorder Distress
  Fear
• _________________________________________________
• Dysthymia .93
• GAD .84
• Major Depression .83
• Panic Disorder .94
• Social Phobia .86
• Agoraphobia .81
• Simple Phobia .75
• _________________________________________________
• Adapted from Vollebergh et al. (2001)

17
Factor Loadings from CFA of Australian NSMHWB

• _________________________________________________
• Disorder Distress
  Fear
• _________________________________________________
• GAD .85
• PTSD .83
• Dysthymic Disorder .82
• Major Depression .81
• Panic Disorder .83
• Agoraphobia .83
• Social Phobia .82
• OCD .73
• _________________________________________________
• Adapted from Slade Watson (2005).

18
Toward an EtiologicallyBased System

• Phenotypic and genotypic structures parallel one
  another
• Kendler et al. (AGP, 2003) concluded
• the structure of these genetic risk factors
  bears a conspicuous resemblance to the phenotypic
  structure (p. 935)
• Thus, phenotypic structural analyses point toward
  etiologically based classification system

19
Revised Structural Model of Mood Anxiety
Disorders
20
Revised Structural Model of Mood Anxiety
Disorders
21
Revised Structural Model Basic Features

• Distress Disorders subgroup characterized by
• pervasive subjective distress
• Fear Disorders subgroup characterized by
• behavioral avoidance
• more limited distress

22
Placement of OCD (I) Factor Correlations from a
CFA

• _______________________________________
• Factor 1 2
  3
• _______________________________________
• 1 OCD .--
• 2 Dissociation .56 .--
• 3 Schizotypy .58 .90 .--
• _______________________________________
• N 455.

23
Placement of OCD (II)Factor Correlations from a
CFA

• __________________________________________
• Factor 1 2
  3 4 5
• __________________________________________
• OCD .--
• Dissociation .76 .--
• Mistrust .69 .66 .--
• Oddity .54 .84 .54 .--
• Social Anhedonia .33 .45 .64 .42 .--
• __________________________________________
• N 1,286.

24
Placement of OCD (III) Factor Correlations from
a CFA

• __________________________________________
• Factor 1 2
  3 4 5
• __________________________________________
• OCD .--
• Social Phobia .46 .--
• BII Phobia .42 .56 .--
• Depression .48 .41 .27 .--
• Dissociation .60 .23 .14 .46 .--
• __________________________________________
• N 359.

25
OCD Conclusions

• Strongly related to dissociation and schizotypal
  PD symptoms
• Correlates more strongly with dissociation/schizot
  ypy than other anxiety disorders
• May define separate higher order factor in more
  comprehensive structural model

26
A Three Superclass Taxonomy
27
A Three Superclass Taxonomy
28
Advantages of a Fully Quantitative Hierarchical
Model

• Directly models patterns of comorbidity
• Solves heterogeneity problem
• Captures information related to severity of
  dysfunction
• Balances parsimony and precision

29
Thanks to

• NIMH Grant 1-R01-MH068472-1
• Gavin Andrews, Michael Chmielewski, Ron de Graaf,
  Wakiza Gamez, Roman Kotov, Elizabeth
  McDade-Montez, Michael OHara, Jennifer Gringer
  Richards, Leonard Simms, Tim Slade Wilma
  Vollebergh