Reliability of self-reported behavioural health risk factors in a South Australian telephone survey

Objective: To test the reliability of telephone health survey questions. Method: A telephone survey on mental health of South Australians in 1997 was re-administered to a random sub-sample of 102 respondents between 32 and 79 days after the original survey. Results: Demographic questions (age, gender, number of adults and children in the household) showed the highest reproducibility and were almost perfect. Questions regarding health risk factors, such as smoking and drinking behaviour, showed substantial to almost perfect agreement. Co-morbidity variables were substantially reproducible where prevalence estimates were not close to zero. Conclusions: The results were comparable to findings from similar studies associated with the Behavioral Risk Factor Surveillance System (BRFSS) in the United States. The study suggests that the telephone health survey instrument used in South Australia is reliable for estimating health conditions and behaviours in the population.