Population attributable fraction in textbooks: Time to revise

INTRODUCTION: The population attributable fraction is an important measure for assessing the impact of intervention on the disease risk in populations, but it is frequently misused in the research literature. METHODS: In this study, we review the definition, calculation, interpretation and assumptions of PAF in 43 textbooks and highlight important shortcomings. RESULTS: While the Levin formula was proposed as a method of calculation in 29 (67%) textbooks, only in 4 (9%) was the Miettinen formula or its generalization for multilevel exposure recommended to calculate a confounding-adjusted population attributable fraction. Other concepts such as generalized impact fraction and prevented and preventable fractions were briefly discussed in few textbooks. DISCUSSION: We recommend the authors revise the textbooks in light of our proposed framework for teaching the population attributable fraction.