Development of a reliable questionnaire to assist in the diagnosis of fetal alcohol spectrum disorders (FASD)

Background: A battery of clinical assessments was used in the Lililwan* Project, Australia's first population-based Fetal Alcohol Spectrum Disorders (FASD) prevalence study, conducted in the remote Fitzroy Valley, Western Australia. One objective was to develop and assess test-retest reliability of an acceptable questionnaire for collecting health information in remote Aboriginal communities feasible for use in the Lililwan Project. Methods: A questionnaire was developed by paediatricians to assist in diagnosis of FASD. Content was based on a literature review of FASD diagnostic criteria, existing questionnaires and risk factors for FASD and birth defects. Aboriginal community members, including qualified Aboriginal language interpreters, adapted the questionnaire to ensure language and cultural components were appropriate for use in the Fitzroy Valley. Locally developed pictorial aids were used for gathering accurate information on alcohol use. Aboriginal 'community navigators' assisted researchers to translate the questions into Kimberley Kriol or local Aboriginal languages depending on participant preference. A subset of 14 questions was assessed for test-retest reliability in 30 parents/carers of children in the Lililwan Project cohort, who were interviewed by one rater using the entire questionnaire, then by a second rater who repeated 14 critical questions at least 6 hours later. Results: The full questionnaire contained 112 items and took 50 minutes to administer. For a subset of 14 items from the full questionnaire percent exact agreement between raters ranged from 59-100%, and was below 70% for only 1 question. Test-retest reliability was excellent (Kappa 0.81-1.00) for 5 items, substantial (Kappa 0.61-0.80) for 5 items, and moderate, fair or slight (Kappa <= 0.60) for the remaining 4 items tested. Test-retest reliability for questions relating to alcohol use in pregnancy was excellent. When questions had moderate, fair or slight agreement, information was obtained from alternate sources e. g. medical records. Qualitative feedback from parents/carers confirmed acceptability of the questionnaire. Conclusions: This questionnaire had acceptable test-retest reliability and could be used to collect demographic, socio-cultural and biomedical information relevant to the diagnosis of FASD in Aboriginal communities throughout Australia and elsewhere. Community input is crucial when developing and administering questionnaires for use in cross-cultural contexts. *Lililwan is a Kimberley Kriol word meaning 'all the little ones'. Kimberley Kriol is the main language spoken by Aboriginal people in the Fitzroy Valley.