Tobacco Induced Diseases (Mar 2018)
Trends in smoking prevalence in South Africa for 1998 - 2014 - overcoming the challenge of multiple data sources
Abstract
Background Monitoring prevalence of smoking is essential for evaluating the impact of national tobacco control policies. This study aims to estimate national pooled smoking prevalence trends over a 16 year period for adult South Africans. Methods Smoking prevalence by age, sex and race from 14 South African population and national household surveys conducted between 1998 - 2014 were used to generate trends over time by 1) performing a series of transformations of the aggregated prevalence data, 2) investigating interactions with race/time, and 3) using a posthumous regression. Bias within surveys were adjusted for by using a risk-of-bias weight. Results The overall observed smoking prevalence decreased from 23.7 (22.4; 24.9) in 1998 to 20.3 (19.3; 21.4) in 2014. Downward trends were observed in both males and females, although it was steeper in males. There were significant differences between race groups, as well as within race groups across age, most notably Asian males, which also seemed to drive the steep decline among males. Whilst the downward trend was evident in older ages (25 + years), the youngest group (15 - 24 years) had a non-linear trend differing from the linear trend in the older group. Conclusions Pooling estimates from different national surveys over time, gave an effective picture of trends in smoking in South Africa. Although not directly observable, the trends reported indicate that the South African Tobacco Control Act of 1993 had influenced and maintained decreasing smoking trends in the last couple of decades, albeit to a lesser extent among females.
Keywords