SMOKING PREVALENCE, MORTALITY, AND TOBACCO CONTROL: A DATA-DRIVEN EVALUATION OF PORTUGAL IN REGIONAL CONTEXT

As a Costa Rican student living in Lisbon, one of the initial cultural differences that stood out to me was the open and frequent usage of tobacco. In Costa Rica, stringent and frequent policies of tobacco control, particularly in public spaces, have made smoking a less visible habit in everyday life. Witnessing its prevalence in Portugal not only surprised me but also intrigued me about the effectiveness of public health policy in shaping this behavior. This contrast provided the start of an exploration of a more nuanced nature of the way that organized policies of tobacco can control direct smoking behavior and its health effects. This dissertation examines the evolution and impact of anti-smoking policies in Portugal with a focus on the MPOWER approach of the World Health Organization. With WHO and IHME data, the work evaluates trends in mortality attributed to smoking and prevalence of smoking between 2000 and 2035. By using lag-adjusted correlation models and forecasting techniques, I estimate the time-lagged impact of public policy in six major domains. We find a significant disparity: whereas mortality continually declined, the prevalence of smoking remains generally unchanged, a finding invoking concern about the regularity, scope, and popular salience of measures implemented. Among MPOWER components, tax (R) and cessation services (O) have the greatest correlation with reductions in long-term mortality, while media (A) and population protection (P) have irregular application and less substantial impacts on prevalence trends. Regional peer-country analysis based on a cluster of Southern and Western European countries provides the context to Portugal’s policy performance. Countries like Greece and Malta have achieved significant prevalence reductions with the support of higher and more stable policy scores. Portugal, however, ranks in the middle of this cluster, which suggests the presence of a framework but the absence of appropriate leverage. A PCA and K-means-based model of clustering was performed for further analytical richness to categorize countries by MPOWER implementation profiles. This categorizes Portugal among its peers, performing on a moderate note, with policy intensity trailing regional leaders like Spain. The results show a critical test for Portugal: unless its policy framework is implemented more vigorously and consistently, particularly in the areas of fiscal control, public communications, and population-tailored interventions, progress against the decline of prevalence could be restrained again. This dissertation not only quantifies these trends but views them, as a data science and public health interested student, with a sense of contributing to a more data-informed and responsive policy for tobacco control.