Abstract
BACKGROUND: Population studies in mostly high-income countries have shown that pictorial health warnings (PHWs) are much more effective than text-only warnings. This is the first quasi-experimental evaluation of the introduction of PHWs in Africa, comparing the change from text-only to PHWs in Kenya to the unchanged text-only health warning in Zambia. METHODS: Data were from International Tobacco Control (ITC) Surveys in Kenya (n=1495), and Zambia (n=1628), cohort surveys of nationally representative samples of adult smokers in each country. The ITC Kenya Survey was conducted in 2012 and 2018 (2 years after the 2016 introduction of three PHWs). The ITC Zambia Survey was conducted in 2012 and 2014 with no change to the single text-only warning. Validated indicators of health warning effectiveness (HWIs) (salience: noticing, reading; cognitive reactions: thinking about health risks, thinking about quitting; and behavioural reactions: avoiding warnings; forgoing a cigarette because of the warnings), and a summary measure-the Labels Impact Index (LII)-measured changes in warning impact between the two countries. RESULTS: PHWs implemented in Kenya led to a significant increase in all HWIs and the LII, compared with the text-only warning in Zambia. The failure to implement PHWs in Zambia led to a substantial missed opportunity to increase warning effectiveness (eg, an estimated additional 168 392 smokers in Zambia would have noticed the warnings). CONCLUSIONS: The introduction of PHWs in Kenya substantially increased the effectiveness of warnings. These results provide strong empirical support for 34 African countries that still have text-only warnings, of which 31 are Parties of the Framework Convention on Tobacco Control and are thus obligated to implement PHWs.