Transcript SOBER - RCEC

Transsexualism in a European context
The value of NIRT in assessing the utility of DSM-IV-TR
diagnostic criteria: a multi-site study on Gender Identity Disorder
Muirne Paap
Baudewijntje Kreukels
Peggy Cohen-Kettenis
Hertha Richter-Appelt
Griet De Cuypere
Ira Haraldsen
Warning!

The topic is applied (N)IRT…

…no formulae/equations!
About me

Currently taking PhD at the clinic for
Gender Identity Disorder (GID) in
Oslo, Rikshospitalet

Studied psychology with a major in
clinical psychology and minor in
statistics
Collaboration

Our clinic has a research collaboration
with the clinics in Amsterdam,
Hamburg and Ghent

Goal: standardize diagnostics and
gather questionnaire data
Background



IRT has been gaining ground in
psychiatric research
investigating properties of clinical
diagnoses or instruments
DIF analyses
Aim of this study…
…is to use the DSM-IV-TR criteria for the
diagnosis Gender Identity Disorder
(GID) as an example to illustrate how
the utility and generality of different
aspects of diagnostic criteria for any
DSM diagnosis can be investigated,
using Nonparametric Item Response
Theory (NIRT).
GID Diagnosis: 4 criteria




Strong and persistent cross-gender
identification
Persistent discomfort about one's assigned
sex or a sense of inappropriateness in the
gender-role of that sex
The diagnosis is not made if the individual
has a concurrent physical intersex condition
Clinically significant distress or impairment
in social, occupational, or other important
areas of functioning
Criterion A

Criterion A:




Stated desire to be the other sex
Frequent passing as the other sex
Desire to live or be treated as other sex
Conviction that he or she has the typical
feelings of the other sex
Criterion B

Criterion B:


Preoccupation with getting rid of sex
characteristics
Belief to be born the wrong sex
Scoring

Criteria from DSM were scored 0 or 1 on:
 severity
 duration
 onset
 persistance
 frequency

Item = subcriterion x aspect
Screenshot of scoring sheet
Sample

N=214 applicants (mean age = 32.3, SD =
12.2), seen between jan 07 – march 09

42% were biological females ( mean age =
28.4, SD = 10.4) and 58% were biological
males (mean age = 35.11, SD = 12.7).

82% were diagnosed with GID (mean age =
32.8, SD = 12.2). FtMs: 90%, MtFs: 77%.
Method
Comparing the centers:
 On item (symptom) level
 On scale level

Done by using Nonparametric Item
Response Theory (Mokken-scale
analysis) to construct scales and
examine items
NIRT
Two models:

Monotone Homogeneity Model (MHM)

Double Monotonicity Model (DMM)
Main research question current project
Are there any differences between centers
in the way the GID-criteria are used to
reach a diagnosis?


Are the symptoms (items) interpreted in the
same way in the four centers?
Is the ordering with respect to popularity
comparable?
NIRT
Some advantages in a clinical setting:
1. Any functional form of the IRF is
allowed, as long as it is monotonely
nondecreasing  higher chance of
good model-data fit
2. Can be used for relatively small datasets
Results – ’international scale’
 When all data was analyzed together,
only 1 scale emerged, combining
criterion A & B!
Results – per center
For three of the four clinics, a onescale sollution was found, similar to
the international one
 In Amsterdam, a two-scale sollution
was found
 however, this was not a two-scale
sollution congruent with the A and B
criteria in the DSM-IV-R!

Results - Amsterdam


Scale one: ‘onset’ and ‘duration’ items
(‘Amst 1’)
Scale two: ‘severity’ and ‘persistence’
items (‘Amst 2’)
 Dutch clinicians might have a different
conception of GID
Results – IIO in subgroups

The rank-order of the items, according to their
’difficulty’, was similar over the four centers

With the exception of the ’persistence’ and
’severity’ of ”Conviction that he or she has the
typical feelings of the other sex”

These were relatively ‘difficult’ items in Gent. To
the contrary, in Hamburg the items are relatively
‘easy’
IRF persistence conviction typical
feelings of the other sex
1= Gent
2= Hamburg
3= Amsterdam
4=Oslo
Summary



With exception of 1 item, all items were
used in a similar fashion in the four clinics
when reaching a diagnosis (rank-order)
Criterion A & B ended up in one scale in our
analysis (international scale)
For Amsterdam, a two-scale sollution was
found  international differences in
diagnostic decisions?
Implications


GID: we would suggest that the
severity and duration of symptoms
should be taken into account in the
next version of the DSM
Generally: we urge more researchers
to consider to use NIRT to scrutinize
diagnostic criteria (as listed in the
DSM)
Current developments

At the moment the DSM-V is under
development and it is being considered to
enhance the DSM by adding a dimensional
adjunct to each of the traditional categorical
diagnoses in the DSM
 IRT!
Kraemer HC: DSM categories and dimensions in clinical and research contexts.
International Journal of Methods in Psychiatric Research 2007; 16: S8-S15
Thank you!

Thank you for your attention!

Any questions?

Email: [email protected]