Examining the Structure of the Repetitive Behavior Scale-Revised (RBS-R) in Young Children with Autism Spectrum Disorder Peter Szatmari IMFAR 2008.

Examining the Structure of the Repetitive Behavior Scale-Revised (RBS-R) in Young Children with

Autism Spectrum Disorder

Peter Szatmari IMFAR 2008

RBS-R background • The Repetitive Behavior Scale-Revised (RBS-R;

Bodfish et al. 2000) is a questionnaire (usually completed by parent) designed to examine restricted/repetitive behaviors in ASD

• The RBS-R has 43 items organized in 6 subscales:

• Compulsive• Ritualistic• Sameness• Self-Injurious• Stereotyped• Restricted Behavior

Recent factor analytic study • Lam and Aman (2007), using a sample of children

and adults (mean CA=184 months):

• only five factors with a merged Ritualistic/Sameness factor

• explained 47.5 % of the variance

• Goodness of fit - RMSEA .06 (reasonable)

• Some factors highly correlated

Why do we use factor analysis ?• The two main objectives of factor analysis are:

• to reduce a large number of variables to a smaller number of factors for modeling purposes

• to create a set of factors to be treated as uncorrelated variables as one approach to handling multicollinearity in procedures as multiple regression (i.e. as predictors)

Other studies• Repetitive sensory motor behaviours are associated

with IQ, age, adaptive functioning, and other autistic symptoms (Szatmari et al. 2006, Hus et al 2007)

• Insistence on sameness behaviours are independent of IQ, and other autistic symptoms (Hus et al 2007), BUT

• Szatmari et al (2006): Insistence on sameness behaviours associated with autistic symptoms related to communication

Objectives• To examine the factor structure of the RBS-R in a

sample of preschool children with ASD

• To examine the correlates of derived RBS-R factors in a sample of preschool children with ASD

• General idea: use fewer, independent, empirical, dimensional variables (i.e. factors), to model human behaviours/symptoms/traits

Pathways in ASD Study (N=400)

• Describe how children with ASD change and develop over time and to identify factors associated with optimal outcomes.

• Ecological

• Multi-level

• Longitudinal

Child

Family

Services

Community

Social CompetenceBehaviour AdjustmentAdaptive Functioning

Communication

Family well-being

n % Mean SD

Male 193 85.80

Female 32 14.20

Child’s Age (months) 39.81 9.09

Child’s Age group:

2-year olds 95 42.20

3-year olds 93 41.30

4-year olds 37 16.40

MP-R Dev. Index St. Score

56.2422.7

0

MP-R Age Equiv. (months)

25.9554.1

1

Caucasian 157 69.80

Sample: 225 newly-diagnosed children

Analyses

• Principal axis factor analysis with a Quartimax rotation with Kaiser Normalization was used to examine the structure of the RBS-R.

• All 43 items of the RBS-R were used in the analysis

• Pearson correlations were calculated to examine relationships between derived factors and other ASD symptoms, as well as cognitive and adaptive

function.

Results

Stop at 3 factors

Variance explained 40.20 %Confirmatory analyses RMSEA .047 (generally <.05 good fit)

Results• A 3-factor solution was selected, using scree plot,

goodness-of-fit criteria, and interpretability

• The 3 factors were labeled:• Compulsive Ritualistic Sameness Behaviour (CRSB)• Self Injurious Behaviour (SIB)• Stereotyped Restricted Sensory Motor Behaviour (SRSMB)

• The 3 factors were independent of each other:

CRSB SIB CRSB - - SIB .03 - SRSMB .03 .07

Results – items with highest loadingsCompulsive Ritualistic Sameness Behaviour (CRSB)

38. Insists on routine 29. Placement of objects 39. Insists on time 37. Difficult transitions 27. Play/leisure routine 34. Appearance/behavior of others31. No interruption 26. Transportation routine

Results – items with highest loadingsSelf Injurious Behaviour (SIB)

7. Hits w/body 11. Pulls hair/skin 10. Bites self 8. Hits against surface 9. Hits w/object 13. Inserts finger/object 12. Rubs/scratches 14. Picks skin

Results – items with highest loadingsStereotyped Restricted Sensory Motor Behaviour (SRSMB)

42. Preoccupied with part of object 43. Preoccupation with movement 3. Finger movements 5. Object usage 40. Preoccupation with subject 4. Locomotion 6. Sensory 2. Head movements 41. Attached to object 1. Body movements

Proposed RBS-R structure model for preschool children with ASD

Stereotyped Behavior Subscale

Self-Injurious Behavior Subscale

Compulsive Behavior Subscale

Ritualistic Behavior Subscale

Sameness Behavior Subscale

Restricted Behavior Subscale

Compulsive Ritualistic Sameness Behavior (CRSB)

Self-Injurious Behavior (SIB)

Stereotyped Restricted Sensory Motor Behaviour (SRSMB)

ORIGINAL SUBSCALES EMPIRICAL FACTORS

CRSB

SIB

SRSMB

Age at assessment 0.220** -0.145* ADI-R Social domain 0.231** 0.107* ADI-R Communication domain nonverbal/verbal

0.152*

ADI-R Behaviours domain 0.353** 0.182* ADOS all modules Communication ADOS all modules Social 0.222** ADOS all modules Play 0.172* ADOS all modules Behaviours 0.182* ADOS all modules Communication + Social 0.188** ADOS all modules Behaviours + Play 0.234** Merrill-Palmer R dev. index growth score -0.327** SRS Total raw score 0.583** 0.184* 0.257** VABS II adaptive behaviour composite -0.289** -0.169**

Correlations between factors and other variables

Results• CRSB scores are positively associated with age,

ADI-R scores, and SRS scores.

• CRSB scores are negatively associated with Vineland adaptive behaviour composite

• SRSMB scores are positively associated with ADOS scores, ADI-R scores (weak), and SRS scores

• SRSMB scores are negatively associated with Merrill Palmer scores, and Vineland adaptive behaviour composite

• SIB appears to be independent of other variables

• Males score higher on the SRB factor (p<.01)

Conclusions• The structure of the RBS-R can also be captured

using fewer, more inclusive factors in a population of preschool children, compared with a previous factor analysis in an older sample

• Our model satisfies both main objectives of factor analysis:

• reduces a large number of variables (43) to a smaller number of factors (3) for modeling purposes (still explaining 40 % variance)

• creates a set of factors to be treated as uncorrelated variables as one approach to handling multicollinearity in procedures like multiple regression (i.e. as predictors)

Conclusions• If we exclude the Self-Injurious subscale

• CRSB and SRSMB are similar to the empirical ADI-R Insistence on Sameness (IS) and Repetitive Sensory and Motor Behaviours (RSMB) factors (Szatmari et al. 2006, Cuccaro et al. 2003, Shao et al. 2003)

• So more evidence for the High vs. Low level repetitive behaviours (even in young children)

Future directions• This ongoing longitudinal study may reveal useful

information on increasing differentiation of repetitive behaviours, as well as potential factors associated with their development

• Examination of the stability/plasticity of the repetitive behaviours structure across time may provide useful information for interventions

• Based on empirical factor loadings, one could develop algorithms with ‘weighted’ items (not simply adding items) that are more developmentally relevant for describing repetitive behaviours in children with ASD.

• It will be interesting to see if these distinct empirical factors predict different behavioural Pathways of children with ASD

Team effort for this presentation

• Stelios Georgiades• Eric Duku• Isabel Smith• Pat Mirenda• Susan Bryson• Eric Fombonne• Wendy Roberts• Tracy Vaillancourt• Joanne Volden• Charlotte Waddell• Lonnie Zwaigenbaum• Pathways in ASD Study Team*

*Pathways in ASD Study TeamBrady E. Dua V. Fraboni T. Hawkins A. Kalynchuk K. Kawchuk J.McLean JShea S. Simon R. Steiman M. Tidmarsh L. Tsonis M. Tuff L. Vanier M. Wellington S. Abruzzese L.

Bacchus S. Bauld S.Bennett T. Bopp K. Brosseau-Lapré

F.Brown J. Chalupka M. Clarke S. Colli L. Couture M.DaSilva B. Elfert M. El-Khatib L. Fleming L. Fossum K. Green J.

Jull S. Kleeberger V. Lazoff T. Lynch S. Magill C. McCartney J. Miodrag N.Murphy M. Noseworthy J.O’Connor I. Olson J. Peacock S. Phillips T. Quirke S. Saracino J. Sepas-Khah M.A.

Shepherd C.Stock R. Thompson A.Trottier N. Vaccarella L.Yun I.

Acknowledgements • Children and families participating

• Our sponsors:

• Canadian Institutes of Health Research (CIHR)• Autism Speaks• British Columbia Government• Alberta Heritage Foundation for Medical Research