Sharmila Banerjee Mukherjee1, Manoj Kumar Malhotra, Satinder Aneja, Satabdi Chakraborty, Smita Deshpande. 1. Department of Pediatrics, LHMC and Associated Hospitals, New Delhi; Department of Psychiatry, Pt. BD Sharma PGIMS, Rohtak, Haryana; and Department of Psychiatry, PGIMER, Dr. Ram Manohar Lohia Hospital, New Delhi, India. Correspondence to: Dr Sharmila B Mukherjee, Department of Pediatrics, Kalawati Saran Childrens Hospital, Bangla Sahib Road, New Delhi 110 001, India. theshormi@gmail.com.
Abstract
OBJECTIVE: To determine the diagnostic accuracy of Indian Scale for Assessment of Autism (ISAA) in children aged 2-9 year at high risk of autism, and to ascertain the level of agreement with Childhood Autism Rating Scale (CARS). DESIGN: Diagnostic Accuracy study. SETTING: Tertiary-level hospital. PARTICIPANTS: Children aged between 2 and 9 year and considered to be at a high risk for autism (delayed development, and age-inappropriate cognition, speech, social interaction, behavior or play) were recruited. Those with diagnosed Hearing impairment, Cerebral palsy, Attention deficit hyperactivity disorder or Pervasive developmental disorders (PDD) were excluded. METHODS: Eligible children underwent a comprehensive assessment by an expert. The study group comprising of PDD, Global developmental delay (GDD) or Intellectual disability was administered ISAA by an investigator after one week. Both evaluators were blinded. ISAA results were compared to the Experts diagnosis and CARS scores. RESULTS: Out of 102 eligible children, 90 formed the study group (63 males, mean age 4.5y). ISAA had a sensitivity 93.3, specificity of 97.4, positive and negative likelihood ratios 85.7 and 98.7 and positive and negative predictive values of 35.5 and 0.08, respectively. Reliability was good and validity sub-optimal (r low, in 4/6 domains). The optimal threshold point demarcating Autism from No autism according to Receiver Operating Characteristic curve was ISAA score of 70. Level of agreement with CARS measured by Kappa coefficient was low (0.14). CONCLUSIONS: The role of ISAA in 3-9 year old children at high risk for Autism is limited to identifying and certifying Autism at ISAA score of 70. It requires re-examination in 2-3 year olds.
OBJECTIVE: To determine the diagnostic accuracy of Indian Scale for Assessment of Autism (ISAA) in children aged 2-9 year at high risk of autism, and to ascertain the level of agreement with Childhood Autism Rating Scale (CARS). DESIGN: Diagnostic Accuracy study. SETTING: Tertiary-level hospital. PARTICIPANTS: Children aged between 2 and 9 year and considered to be at a high risk for autism (delayed development, and age-inappropriate cognition, speech, social interaction, behavior or play) were recruited. Those with diagnosed Hearing impairment, Cerebral palsy, Attention deficit hyperactivity disorder or Pervasive developmental disorders (PDD) were excluded. METHODS: Eligible children underwent a comprehensive assessment by an expert. The study group comprising of PDD, Global developmental delay (GDD) or Intellectual disability was administered ISAA by an investigator after one week. Both evaluators were blinded. ISAA results were compared to the Experts diagnosis and CARS scores. RESULTS: Out of 102 eligible children, 90 formed the study group (63 males, mean age 4.5y). ISAA had a sensitivity 93.3, specificity of 97.4, positive and negative likelihood ratios 85.7 and 98.7 and positive and negative predictive values of 35.5 and 0.08, respectively. Reliability was good and validity sub-optimal (r low, in 4/6 domains). The optimal threshold point demarcating Autism from No autism according to Receiver Operating Characteristic curve was ISAA score of 70. Level of agreement with CARS measured by Kappa coefficient was low (0.14). CONCLUSIONS: The role of ISAA in 3-9 year old children at high risk for Autism is limited to identifying and certifying Autism at ISAA score of 70. It requires re-examination in 2-3 year olds.