Clinical utility of dimensional models for personality pathology

1
Clinical utility of dimensional models for
personality pathology

• Roel Verheul

Viersprong Institute for Studies on Personality
Disorders (VISPD) Center of Psychotherapy De
Viersprong Department of Psychology University
of Amsterdam
2
Criteria for an optimal classification system (1)

• It is not easy to arrive at a diagnostically
  valid classification system, i.e. a system
• providing sufficient coverage of the domain of
  personality pathology, and
• being consistent with up-to-date empirical
  findings regarding the
• dimensionality versus taxonicity,
• pathogenesis, and
• changeability of personality disorder,.

3
Criteria for an optimal classification system (2)

• . yet it is perhaps even more difficult to
  design a clinically useful classification system!
• Clinical utility should be the driving force
  behind future revisions!

4
Clinical utility

• Definition
• The systems clinical utility is the extent to
  which it assists clinical decision makers in
  fulfilling the various functions of a psychiatric
  classification (cf. First et al., 2004)

5
Clinical utility

• Definition
• Diagnostic validity is a requirement for clinical
  utility

6
Clinical utility

• Definition
• Diagnostic validity required
• Functions of psychiatric classification
• Covering a certain domain of psychopathology
• Helping clinicians to understand patients
• Providing reliable diagnoses
• Enabling professional communication
• Assisting clinical decision making

7
Components of clinical utility

• User acceptability and accuracy
• Professional communication
• Interrater reliability
• Subtlety of information
• Clinical decision making

8
1. User acceptability and accuracy

• Classification is worthless if not used at all,
  or if not used correctly

9
1. User acceptability and accuracy

• Classification is worthless if not used at all,
  and if not used correctly
• Users have
• Limited time
• Limited resources
• Limited education / understanding
• Different language(s)

10
1. User acceptability and accuracy

• Classification is worthless if not used at all,
  and if not used correctly
• Users have
• We need a system that is
• Fast, simple, cheap, feasible, and comprehensible
  !!

11
Complexity of current system structure for each
criterion
Probing
If denial no reason to doubt the answer
Score 0
yes
Does answer fit in with criterion?
No
Score 0
yes
Are the general criteria met?
Axis-I-related or lt past 5 year
Score 0
yes
frequency Depending on severity
consequences
Score 1,2(,3)
12
Recommendations

• This element is critically important, and should
  be covered by the field trials, e.g.
• Surveying users reactions to proposed changes
  reasons behind judgments (!)
• Measuring the impact of changes on ease of use
  (e.g., by timing duration of assessment)
• Comparing accuracy of clinician-based with
  expert/interview/consensus-based diagnoses

13
2. Professional communication

• Ease in communication has often been referred to
  as one of the advantages of categorical system
• However
• Only true for prototypic cases ( 20)
• Not true for 80 non-prototypic subthreshold
  cases
• Clinicians like detailed information
• Hierarchical systems might be simpler to
  communicate
• Supposed advantage might actually reflect
  familiarity

14
Recommendations

• This element is likely to be a consequence rather
  than a prerequisite of clinical utility
• Nevertheless, DSM-V should pay considerable
  attention to communicative aspects

15
3. Interrater reliability

• DSM system is developed for practitioners
• Reliable DSM-IV diagnoses in realworld practice
  not feasible, especially in non-prototypic cases!
• Inter-instrument disagreement!
• Aim should be to develop a system that can be
  accurately (reliably) applied

16
Recommendations

• Important issue is the acceptability of
  structured and standardized measures!
• Perhaps, it is recommendable
• (a) to obtain expert consensus about a standard
  measure or even develop one,
• (b) to develop guidelines for assessment
• The possible introduction of a completely new
  system would create a unique chance in this
  regard

17
4. Subtlety of diagnosis

• Dimensional systems gt categorical systems
• Subtlety depends on of dimensions, e.g. FFM gt
  spectra models.
• However, models differ with respect to
• Clinical relevance of dimensions
• Coverage of adaptive versus maladaptive range
• Inclusion of strengths in addition to pathology

18
Recommendations

• Sufficient comprehensiveness is critically
  important, although purpose is to simplify
  complexity
• Research possibilities include
• Clinician surveys to identify clinically relevant
  concepts
• Comparison of models with respect to level of
  richness/relevance of descriptions of case
  vignettes
• Psychometric analyses to study
• coverage (IRT) of (mal)adaptive range
• incremental coverage of an integrated model
  (EFA/CFA)

19
5. Clinical decision making

• Misunderstanding yes/no decisions are the
  exception, not the rule
• Clinical decision making typically involves
  determining the appropriate degree of various
  treatment characteristics and therapist behaviors
• Perhaps the most important determinant of
  clinical utility ? what can we learn from the
  current situation?

20
Situation in The Netherlands

• DSM-IV is viewed as an administrative system
• Most frequent diagnoses include PDNOS and
  diagnosis deferred (799.9)
• Polarization between descriptive and structural
  diagnosis
• Kernbergian thinking is popular severity
  dimension
• For the DSM-V to have more impact on clinical
  practice, it is important not to ignore these
  signals

21
Three major domains of decision making

• Determining the
• Necessity and benefit of treatment
• Macrotreatment level general treatment model
  most likely to be effective and efficient
• Microtreatment level type of interventions most
  likely to be helpful within the treatment model

22
Three major domains of decision making

• Necessity and benefit
• Spontaneous recovery
• e.g., isolated facet elevations within N, while
  normal A/C
• Patients not likely to respond
• e.g., FFM misery triad high N, low E, low C ?
  supportive, palliative, focusing on
  rehabilitation, practical issues and symptom
  relief
• Patients at risk for negative response
• e.g., strong psychopathic traits

23
Three major domains of decision making

• Necessity and benefit of treatment
• Macrotreatment decisions higher-order level
• Setting severity/rigidity, work/children
• Format interpersonal and system problems
• Major strategies extraversion / openness (?)
• Duration severity/rigidity, strengths
• Medication personality symptoms, comorbidity

24
Theoretical orientation and personalityMiller,
1991
25
Three major domains of decision making

• Necessity and benefit of treatment
• Macrotreatment decisions
• Microtreatment decisions lower-order level
• Goal setting, e.g. self-harm, trust
• Matching patient characteristics to therapists,
  e.g. dependency, grandiosity
• Determining degree of support, e.g. vulnerability
• Determining degree of structure / limit setting,
  e.g. conduct problems, compulsivity

26
Recommendations

• In general, dimensional system more consistent
  with clinical decision making
• It is tempting to demand hard evidence (RCTs?),
  but
• Bridge too far for PD no practice guidelines
  yet, so sole reliance on outcome
• Such a requirement would be unprecedented
• Revisions typically yield gradual and delayed
  changes
• I would suggest case vignette studies comparing
  clinician and expert decisions across various
  models
• Review of predictor studies!

27
Consumers report
28
Tentative comparison of models

• Overall clinical utility
• Categorical lt Hybrid lt Purely dimensional models
• Dimensional models
• DAPP, SNAP, NEO best candidates
• Higher-order level similar!!!
• Integrated or collapsed model might be optimal!
• e.g. coverage both adaptive and maladaptive
  range, inclusion of positive and negative traits,
  support in the field

29
Strategies for case identification

• Necessity
• Legal purposes
• Medical purposes
• Administrative purposes
• Credibility of public health issue
• Funding purposes

30
Strategies for case identification

• Necessity
• Cutoffs do not solve the problem!
• Statistical deviance is neither a necessary, nor
  a sufficient criterion for disorder

31
Strategies for case identification

• Necessity
• Cutoffs do not solve the problem!
• Several strategies have been proposed
• Cloninger (2000)
• Livesley Jang (2000)
• Westen Shedler (2000)
• Widiger et al. (2002)
• Tyrer (1996)

32
Some consensus

• No redundancy
• one disorder (e.g., Tyrer, Livesley, Cloninger)
• Definition
• impairments, dysfunctioning, adaptive failure
• Severity dimension
• with several cutoffs

33
Conclusion

• Clinical utility should be the driving force
  behind future revisions
• Dimensional gt hybrid gt categorical models
• Integrated dimensional model might be optimal
• Strategy for case identification one category,
  impairment definition, severity dimension

34
Research priorities
35
Research priorities