Date of Award

12-15-2020

Degree Type

Thesis

Degree Name

MS in Epidemiology & Biostatistics

First Advisor

Dr. Sarah Saleem

Second Advisor

Dr. Atif Rahman

Third Advisor

Mr. Iqbal Azam

Department

Community Health Sciences

Abstract

Background: Early detection of Autism in children is critical to creating an autism-friendly environment leading to a better prognosis. In a low middle-income country like Pakistan, there is a dearth of trained professionals and a lack of infrastructure to screen Autism in children. This study aimed to develop and validate a technology-mediated culturally relevant diagnostic tool, shifting the task of a detailed assessment of gold standard by a trained professional to a layman by giving an equally robust Screening of Autism in children in low resource settings, enhancing chances of early diagnosis of Autism hence improving prognosis.
Methods: The study was conducted in two phases; during Phase 1 the autism screening tool, by the name of ‘Currim Autism Digital Evaluation Tool (CADET)’ was developed, In Phase 2 CADET was validated on children aged 31 months to 10 years . Phase 1: CADET was developed in 6 steps. The tool was conceptualized, based on existing validated tools and discussed with an expert group who selected, and prioritized six items on their relevance for autism diagnosis in children. These 6 items were illustrated on the storyboard and animations were developed in consultation with Dr Saleem Sayani Director Technology Innovation Support Center and his group at Aga Khan University CADET was pretested on 5 mothers. Based on the feedback tool was modified and was rolled out for validation. During validation, user feedback was taken to improve the tool further. Phase 2: In this phase the validation study of the CADET was carried out. The validation study of the tool consisted of content validity, construct validity, and criteria validity. Content validity was part of the phase one as well where experts selected the items and gave their consensus on the content of the six items. Construct validity and criteria validity was assessed by screening and comparing children using CADET in 3 groups who have been earlier diagnosed as autistic, atypically developing, and typically developing children by a standard tool “Childhood Autism Rating Scale)’: A total of 84 children, 28 in each group were randomly selected from the data records of Out Patient Department of Children’s Hospital, Aga Khan University Hospital. The construct validity and criteria of CADET were assessed by checking the tool's internal consistency and ROC analysis using Stata version.14 and Medcalc 19.6.
Results: The CADET is a web-based application consisting of 6 items. Each item had two animations showing autistic and normal behavior. The construct validity of CADET was assessed by using Cronbach's α with an estimated value of 0.93. Considering typically developed children as reference, CADET showed sensitivity of 96.4% (95% CI (81.6% - 99.9%) and specificity of 100% (95% CI (87.6% - 100.0%) The mean score difference between an autistic and typically developing child was 5.35 (p-value < 0.001).
Conclusion: The tool's initial testing shows it works at par with the Gold standard and can be easily used by a layperson.

First Page

1

Last Page

75

Download

Share

COinS