for treatment selection in personality disorders

1
Towards a Clinical Decision Support System

• for treatment selection in personality disorders
• Janine van Manen, Roel Verheul
• SPR 2005 Montreal
• Janine van Manen
• Viersprong Institute for Studies on Personality
  Disorders (VISPD)
• Halsteren, the Netherlands
• janine.van.manen_at_deviersprong.nl, www.vispd.nl

2
Why a Clinical Decision Support System?

• Different dosages psychotherapeutic treatment
• setting outpatient, dayhospital, inpatient
• duration shorter than half a year, longer than
  half year
• intensity confronting, supportive
• theoretical orientation cognitive behavioural,
  psychodynamic
• Rational assignment optimizes the
  (cost)effectiveness of treatment.......

3
Why a Clinical Decision Support System?

• ...... but treatment selection is largely based
  on
• availability of treatment facility (Chiesa et
  al., 2002)
• personal experience and strong belief (or faith)
  (Beutler, 2000)
• socio-demographically variables (income,
  insurance) (Rudolf, Manz, 1993)
• patient and therapist preferences (Bleyen et al.,
  1999)
• There is little empirical knowledge
• Clinical Decision Support System

4
Objective of the study

• Development of evidence based Clinical Decision
  Support System
• Evidence based means based on an optimal
  integration of
• clinical knowledge and expertise
• empirical data
• patient preferences
• User-friendly instrument
• administration within 10-30 minutes
• transparent ? maybe decision tree?

5
Interviews with intake cliniciansobjective

• Objective of the interview
• Investigate systematically clinical knowledge
  about treatment selection strategies
• We are looking for patient characteristics that
  discriminate between different treatments.
• identification of patient characteristics
  important for treatment selection
• identification of treatment parameters
• Which are the most relevant patient
  characteristics for treatment selection?

6
Interviews with intake clinicians objective

• Which patient characteristic is correlated to
  which treatment parameter ?
• Selection of patient characteristic clinical and
  empirical literature, face validity, opinion of
  intake clinicians ? 12 characteristics
• Treatment parameters are defined in setting,
  duration, intensity and theoretical orientation

7
Interviews with intake cliniciansinterviewees

• 27 interviewees characteristics
• Profession
• 19 psychotherapists, 6 psychiatrists, 2
  psychotherapists in training
• Location
• 5 different psychotherapeutic centres in the
  Netherlands
• Setting
• 13 day hospital, 8 inpatient setting, 4
  outpatient setting
• Duration
• treatment program 21 longer than half a year, 4
  shorter, 2 longer and shorter

8
Interviews with intake cliniciansinterviewees

• Theoretical orientation
• 18 psychodynamic, 6 integrative, 2 cognitive
  behavioural, 1 client-centered
• Work experience

9
Interviews with intake clinicianshow did we
analyse the data?

• We are looking for patient characteristics that
  discriminate between different treatments
• Or more specific different levels of the
  indicated patient characteristics match different
  levels of dosage

10
Results 1
11
Results 1what does a percentage mean?

• E.g. the relation between ego-strength and
  intensity

12
Results 1
13
Results 1level of consensus

• Level of consensus
• 27/48 matching relationships
• 13 relations with 40-49 consensus
• 4 relations with 50-59 consensus
• 10 relations with gt 60 consensus
• Consensus per treatment parameter appears to be
• Modest for setting
• Moderate for duration and theoretical orientation
• Strong for intensity

14
Interviews with intake clinicianshow did we
analyse the data?

• Objective of the interview
• Investigate systematically clinical knowledge
  about treatment selection strategies
• We are looking for patient characteristics that
  discriminate between different treatments
• Which are the most relevant patient
  characteristics for treatment selection?
• Question in the interview
• Which are the 4 most important patient
  characteristics for treatment selection? And
  which are 4 least important characteristics?

15
Results 2

• Importance of the patient characteristics in the
  intake

16
Summary of the findings

• We found a reasonable level of consensus 27/48
  matching relationships
• Setting severity and social role
• Duration severity and strength
• Intensity severity and strength
• Theoretical Orientation strength
• In general, patient characteristics are not
  specific for certain dosage parameters, except
  for
• Setting treatment history, work, care
• Duration focality
• Intensity capacity for a therapeutic relation
• Patient characteristics differ widely in terms of
  observed importance

17
Conclusions

• This interview strategy does seem to detect
• important patient characteristics that match with
  dosage psychotherapeutic treatment
• hierarchy of importance among the patient
  characteristics
• Since consensus moderate, additional strategies
  are necessary to develop prototype decision model
• What are the next steps in developing a
  Clinical Decision Support System?

18
Further directions

• Phase 1 Development prototype decision tree
• interviews with intake clinicians
• clinical and empirical literature search
• development of a prototype Clinical Decision
  Support System
• Phase 2 Optimizing and validating the prototype
• Optimizing adjusting the prototype
• retrospective analysis
• prospective analysis
• Validating empirical testing of the prototype
• test of matching hypothesis
• (cost)effectiveness of Clinical Decision Support
  System