Translation and Validation of Autism Screening and Diagnostic Tools in to Hindi and Bengali

Background: There is a serious dearth of epidemiological research on Autism Spectrum Conditions (ASC) in South Asia. The availability of standardized screening and diagnostic instruments in principal regional languages constitutes a necessary first step to addressing this gap. Hindi and Bengali are two of the most widely spoken languages of this region (number of speakers > 371 million). Standardised screening tools in these languages would be crucial for further ASC research in this region, and can lead to better diagnostic facilities for ASC children.

Objectives: To translate and validate screening and diagnostic tools for ASC in Hindi and Bengali in two Indian cities (Delhi and Kolkata).

Methods:  

TRANSLATION: Five widely used and standardized instruments for screening and diagnosis were translated into Hindi and Bengali by individuals familiar with ASC: Ten Questions (TQ) (Durkin et al 1995), Social Communication Disorders Checklist (Skuse et al 1995), Social Communications Questionnaire (SCQ) (Berument, Rutter et al 1999), Autism Observation Schedule (ADOS) (all 4 modules). Blind back translation was carried out by a language expert, and this cycle was repeated until the backtranslation was approved by the creator/ copyright holder of the respective instruments. VALIDATION: 45 children with ASC and 43 control children (between 4-7years of age) were recruited in Kolkata for validation of Bengali questionnaires. 40 ASC and 42 Control children in the same age range were recruited in Delhi for validation of Hindi questionnaires. All cases had a ICD-10/DSM-IV based diagnosis from a recognized clinician, and this was confirmed by a child psychiatrist where necessary. The translated ADOS modules 1 and 3 were administered on 20 cases and 20 controls each by a trained ADOS administrator in both cities.All data was analysed using SPSS.

Results:  

Hindi : 86.67% of the children with ASC and 41.86 % of controls met the cut off of 9 on the SCDC and 73.8% of children with ASC  & none  of the controls met the cut off of 15 on the SCQ. On the ADOS 77% of cases, and none of the  controls met the cut off score of 12. AQ-C scores were significantly different between cases (Mean = 81.53) and controls (Mean=45.54) (p<0.001).TQ scores were significantly higher for cases than controls (p<0.001).

Bengali:  All of the children with ASC and 0.67 % of controls met the cut off of 9 on the SCDC and  85.1 % of children with ASC and none of the controls met the cut off of 15 on the SCQ. On the ADOS all of the cases, and none of the  controls met the cut off score of 12. AQ-C scores were significantly different for cases (Mean = 85.86) and controls (Mean=39.93) (p<0.001). TQ scores were significantly higher for cases than controls (p<0.05).

Conclusions:  

Screening and diagnostic tools translated into regional languages and validated in a case-control sample were found to show similar properties to the original instruments. This has direct implications for improving diagnosis of ASC by clinicians in both rural and urban areas in South Asia.