934Prevalence by proxy: Estimating youth smoking prevalence using nationally representative household surveys

Abstract

Abstract Background Reducing youth (15-17 years) smoking uptake is critical to tobacco control; accordingly, youth smoking prevalence is a key monitoring and evaluation outcome. Many nationally representative surveys collect youth smoking behaviour data from, or in the presence of, the youth’s parent or caregiver. We aimed to quantify the potential bias conferred by this. Methods We compared youth smoking prevalence when reported by parent proxy, with parent present or by private self-report, in Australian Bureau of Statistics Health Surveys. National youth current smoking prevalence if all data were collected by youth self-report was estimated. Results Smoking behaviour data for over 75% of youth participants in the health survey were collected by proxy or with parent present. Ever-smoking prevalence using private self-report versus report by proxy was 1.29 (95%CI:0.96-1.73) to 1.99 (1.39-2.85) times as high in Aboriginal and Torres Strait Islander youth, and 1.83 (0.92-2.63) to 2.72 (1.68-4.41) times as high in total population youth. Predicted national current smoking prevalence if all youth were to self-report alone was substantively higher than the estimated national prevalence based on actual responses, but still reveals a decline over time. Conclusions Youth smoking estimates drawn from data collected by proxy/with parent present are unlikely to be accurate. Increased youth self-report is crucial to ensure data accuracy to inform effective tobacco control. Key messages The accuracy of youth smoking data collected by proxy/with parent present should be further scrutinised before it is used to inform assessment of national prevalence and trends.