Investigating the cross-cultural validity of DSM-5 autism spectrum disorder: Evidence from Finnish and UK samples

Abstract

The recent Diagnostic and Statistical Manual of Mental Disorders–Fifth Edition (DSM-5) reformulation of autism spectrum disorder has received empirical support from North American and UK samples. Autism spectrum disorder is an increasingly global diagnosis, and research is needed to discover how well it generalises beyond North America and the United Kingdom. We tested the applicability of the DSM-5 model to a sample of Finnish young people with autism spectrum disorder (n = 130) or the broader autism phenotype (n = 110). Confirmatory factor analysis tested the DSM-5 model in Finland and compared the fit of this model between Finnish and UK participants (autism spectrum disorder, n = 488; broader autism phenotype, n = 220). In both countries, autistic symptoms were measured using the Developmental, Diagnostic and Dimensional Interview. Replicating findings from English-speaking samples, the DSM-5 model fitted well in Finnish autism spectrum disorder participants, outperforming a Diagnostic and Statistical Manual of Mental Disorders–Fourth Edition (DSM-IV) model. The DSM-5 model fitted equally well in Finnish and UK autism spectrum disorder samples. Among broader autism phenotype participants, this model fitted well in the United Kingdom but poorly in Finland, suggesting that cross-cultural variability may be greatest for milder autistic characteristics. We encourage researchers with data from other cultures to emulate our methodological approach, to map any cultural variability in the manifestation of autism spectrum disorder and the broader autism phenotype. This would be especially valuable given the ongoing revision of the International Classification of Diseases–11th Edition, the most global of the diagnostic manuals.

The underlying symptom structure of autism spectrum disorders: A factor analytic approach using the developmental, dimensional and diagnostic interview

“…In general, these results are partially supportive for the structure of ASD as put forward in DSM-5 (American Psychiatric Association, 2013) and in other factor analytic studies (Frazier et al, 2012(Frazier et al, , 2014(Frazier et al, , 2008Georgiades et al, 2007;Mandy et al, 2012Mandy et al, , 2014Norris et al, 2012;Snow et al, 2009). Though we do not find a unitary social communication and interaction factor, our data support the notion that social communication and interaction cannot be separated (as was the case in DSM-IV-TR and in different studies (Lecavalier et al, 2006;Norris et al, 2012), since two of the factors (Social Communication and Appropriateness of Interaction) comprise both social interaction items and communication items.…”

“…Cultural differences can also influence how disturbing or deviant some behaviours are experienced by parents. However, this was not the case in a study comparing a UK versus Finnish ASD population (Mandy et al, 2014).…”

“…Further analysis revealed no significant differences in fit between the UK and Finnish sample. In a sample of participants with ASD symptoms not meeting diagnostic criteria for ASD, called broader autism phenotype, this model did not fit the data (Mandy, Charman, Puura, & Skuse, 2014).…”

“…This was for instance shown for samples with different levels of verbal functioning (Norris et al, 2012), for populations differing in age range or number of affected family members (Constantino et al, 2004;Frazier et al, 2014), and for participants with ASD compared to participants with a broader autism phenotype (Mandy et al, 2014).…”

Automated Extraction of Diagnostic Criteria From Electronic Health Records for Autism Spectrum Disorders: Development, Evaluation, and Application (Preprint)

“…In addition to case status assignment, more detailed use of the information contained in the free text would be helpful for large-scale analysis, e.g., cultural comparisons as suggested by Mandy et al [12] and for combination with other data. To extract this information automatically, entity extraction algorithms are needed.…”