Tobacco Cost of Illness Studies: A Systematic Review

Introduction: To identify studies reporting costs arising from tobacco use and detail their (1) economic approaches, (2) health outcomes, and (3) other cost areas included. Methods: We searched PubMed, Scopus, Cochrane Library, EconLit, and Google Scholar for studies published between 2008 and April 2018 in English. Eligible articles reported tobacco-related costs and included all tobacco-using populations (multinational, national, subpopulations, and involuntary smokers). All economic approaches that resulted in monetary outcomes were included. We reported USD or converted local currencies to USD. Two health economists extracted and two researchers independently reviewed the data. Results: From 4083 articles, we reviewed 361 abstracts and examined 79 full-texts, with 63 (1.6%) deemed eligible. There were three multinational, thirty-four national, twenty-one subpopulation or condition(s)-specific analyses, and five evaluating involuntary smoking. The diverse approaches and outcomes precluded integrating costs, but these were substantial in all studies. For instance, about USD 1436 billion in global health expenditures and productivity losses in 2012 and USD 9 billion in lost productivity in China, Brazil, and South Africa in 2012. At the national level, costs ranged from USD 4665 in annual per respondent health expenses (Germany 2006-2008) to USD 289-332.5 billion in medical expenses (United States 1964-2014). Conclusions: Despite wide variations in the methods used, the identified costs of tobacco are substantial. Studies on tobacco cost-of-illness use diverse methods and hence produce data that are not readily comparable across populations, time, and studies, precluding a consistent evidence-base for action and measurement of progress. Recommendations are made to improve comparability. Implications: In addition to the health and financial costs to individual smokers, smoking imposes costs on the broader community. Production of comparable estimates of the societal cost of tobacco use is impaired by a plethora of economic models and inconsistently included costs and conditions. These inconsistencies also cause difficulties in comparing relative impacts caused by differing factors. The review systematically documents the post-2007 literature on tobacco cost-of-illness estimations and details conditions and costs included. We hope this will encourage replication of models across settings to provide more consistent data, able to be integrated across populations, over time, and across risk factors.