To estimate PAF, one needs prevalence data for a specific population and OR and CI for the specific disease (from a c-c or cohort study). In the attached reference Steenland et al discuss some of the specific issues involved in calculating AFs.

Epidemiology 2006 ; 17 (5) 512-9

Saman,

Thanks for the feedback. I agree with you that PAF is the difference between the probability of disease (PD) in exposed population and the probability of disease in non-exposed population divided by the probability in the exposed population. This formula is applicable for rare and chronic diseases. If multiplied by 100 %, the outcome is termed PAF%. This is the excess disease that could have been prevented if exposure did not occur.

I also agree with you that OR is the appropriate measure to estimate the prevalence of the disease. Knowledge about the prevalence of the disease can enable public health practitioners to design and implement interventions to prevent exposure; the ultimate goal being disease prevention.

Thanks,

Ben Olwe