The accuracy of self-reported smoking: a systematic review of the relationship between self-reported and cotinine-assessed smoking status.

INTRODUCTION Smoking is a leading cause of premature mortality and preventable morbidity. Surveillance is most often based on self-reported data, but studies have shown that self-reports tend to underestimate smoking status. METHODS This study systematically reviewed the literature to measure the concordance between self-reported smoking status and smoking status determined through measures of cotinine in biological fluids. Four electronic databases were searched to identify observational and experimental studies on adult populations over the age of 18 years. RESULTS Searching identified 67 studies that met the eligibility criteria and examined the relationship between self-reported smoking and smoking confirmed by cotinine measurement. Overall, the data show trends of underestimation when smoking prevalence is based on self-report and varying sensitivity levels for self-reported estimates depending on the population studied and the medium in which the biological sample is measured. Sensitivity values were consistently higher when cotinine was measured in saliva instead of urine or blood. Meta-analysis was not appropriate because of the substantial heterogeneity among the cutpoints used to define smokers and the poor reporting on outcomes of interest. DISCUSSION Further research in this field would benefit from the standardization of cutpoints to define current smokers and the implementation of standard reporting guidelines to enhance comparability across studies. Accurate estimation of smoking status is important as data from population studies such as those included in this review are used to generate regional and national estimates of smoking status and in turn are used to allocate resources and set health priorities.