One simple questionnaire could help spot autism in many countries
- A questionnaire predicts autism among 1,400 children in India, Japan, and the UK
- But some questions don’t capture cultural nuances when translated
- The questionnaire may not work universally, but regional versions could help spot autism
Rosa Hoekstra, a psychiatry researcher at King’s College London, and colleagues looked at the Autism Spectrum Quotient (AQ)-Child questionnaire, an open access test that has been shown to effectively predict autism in the United Kingdom. It has already been translated into Japanese and two languages spoken in India.
Hoekstra and her team used these tests to assess more than 1,400 children in India, Japan, and the United Kingdom. The researchers analysed which of the 50 questions was most predictive of an autism diagnosis in children with and without the condition between 4 and 9 years old.
The research, published today (5 November) in Molecular Autism, found that questions that performed well in predicting autism typically did so in more than one country. Five were predictive for autism across all three countries, although four questions indicated potential cultural differences.
“It is important to do an examination to make sure that a questionnaire works in the countries, the cultural settings and contexts that you’re hoping to use it for,”
Hoekstra said she embarked on the research to address the problem that up to 95 per cent of autism cases go undiagnosed in low- and middle-income countries. These countries need globally applicable screening instruments for research and clinical work, she said.
“The vast majority of autism research is done in high-income western countries,” said Hoekstra. “It means that we know relatively little about what autism looks like in other settings.”
Autism is a lifelong developmental disability that can manifest in various symptoms, including speech impairments, repetitive behaviour, and severe learning disabilities. Most cases go undiagnosed in low-income countries.
Hoekstra stressed that the AQ-Child questionnaire, in which parents answer questions about their children’s behaviour, is only a medical aid and cannot lead to diagnosis by itself. Still, it could be used to screen a large number of children to identify those that need further assessment, she says. Her study showed that more than 15 questions from the AQ-Child questionnaire provided excellent diagnostic pointers in each country, a validation that Hoekstra called “encouraging”.
“It is important to do an examination to make sure that a questionnaire works in the countries, the cultural settings, and contexts that you’re hoping to use it for,” Hoekstra said.
One such cultural difference emerged in the phrasing of a question about whether a child “enjoys doing things spontaneously”. In Anglophone countries, parents typically interpreted the question to refer to events without prior planning, which would upset an autistic child. But among Indian and Japanese parents, the question was viewed as referring to a child doing things of their own accord, which autistic children generally enjoy.
Hemamali Perera, who studies mental health in children at the University of Colombo in Sri Lanka, called the attempt to identify questions that are valid across three diverse cultural settings “brave”. But she said producing a universal screening questionnaire “is an impossible task”. Questions relating to behaviour and communication cannot be translated and retain a consistent meaning, due to the vast diversity in expression and vocabularies, she told SciDev.Net.
Hoekstra agreed to an extent and said she would not use the AQ-Child questionnaire for a rural Ethiopian population, where she is currently investigating autism. “Questions like ‘Prefers to go to a birthday party than to a library’ are completely irrelevant because there is no library,” she said.
However, Hoekstra said she remained hopeful that a global diagnosis support tool for autism can still be found.
“I think we can make strides towards developing a measure that can be used in several settings, perhaps with some additional adaptations,” she said. “If we could find a way to create a sub-Saharan Africa screening scale, or if we can find a screening scale that would generally work well in middle-income country settings, then that would be quite helpful.”
This article was produced by SciDev.net's Global edition.