![]() ![]() ![]() 2009) Their approach can be adapted for many regression models. (Greenland and Drescher 1993, Ruckinger, von Kries et al. Greenland and Drescher first introduced methods to calculate adjusted AFs directly from regression models. Notice that the MH approach can be viewed as a special instance of the weighted sums approach.Īdjusted Attributable Fractions Through Modeling However, the weighted sums approach should not be used when analyzing sparse data. The weighted sums approach makes no assumptions about interaction. However, the MH variance estimates are more cumbersome to calculate. MH estimators have the advantage of consistent variance even under conditions of sparse data. The Mantel–Haenszel approach to adjusted attributable fractions assumes no interaction between the exposure and the third variable. 1985)īenichou describes two main methods for calculating adjusted AFs using stratified analysis: the Mantel–Haenszel (MH) approach and the weighted sums approach. Thus, there is a need for methods to calculate adjusted AF estimates. This approach requires the rare-disease assumption.Īs with other traditional measures of association, the AF will be biased if there are confounding factors. the general population can result in less generalizable AFs.Ĭase-control study: In a case-control study, the AF is traditionally calculated using the odds ratio to estimate the risk ratio. The potential difference in prevalence of exposure in a cohort study vs. However, cohort studies are often used when the exposure of interest is rare to allow for over-sampling of the exposure. The interpretation of the AF changes with the type of study data used to calculate the measure.Ĭross-sectional study: In a cross sectional study, the AF represents the excess of prevalent cases of disease that can be attributed to the exposure of interest.Ĭohort study: In a cohort study, the AF represents the excess of incident disease cases that can be attributed to exposure of interest. Interpretation of Attributable Fraction (Benichou 2001) Authors need to be careful in wording the interpretation of AFs so as not to confuse the excess and etiologic fractions. Etiologic fractions are never directly estimable because the disease risk in the exposed takes into account both those individuals who got the disease due the exposure (causal) and individuals who got the disease through a mechanism not including the exposure (doomed).Įxcess fractions will be less than etiologic fractions even if the disease is rare, and Rothman states that they can be thought of as “lower bounds” for the etiologic fraction. The distinction may be framed within a discussion of timing (with regards to classifying cases that would have still occurred in the study period, but at a later moment, had they not been exposed) or with reference to assumptions of monotonicity (with regards to whether the population includes preventive types that may couterbalance some of the causal types), When calculating AFs researchers often hope to interpret the measure as an etiologic fraction, however, they have likely calculated an excess fraction. This distinction between excess and etiologic fractions causes a tension. 2008)Įxcess fractions measure the number of excess cases of disease due to a specific exposure of interest.Įtiologic fractions measure the number of cases of disease due to a specific exposure of interest. Rothman explains that there are two major classes of attributable fractions. More recently, “attributable fraction” appears to have matched the popularity of “population attributable risk” Literature reviews suggest that “attributable risk” and “population attributable risk” are the most commonly used terms (Uter and Pfahlberg 1999). It is important to note that Levin’s attributable fraction is referred to by various terms in the literature. The proposed formula assumed that the proportion with disease would be equal among the exposed and the unexposed under the null hypothesis of no influence of exposure on disease, and used the relative risk (RR) and the prevalence of the exposure of interest to estimate the number of exposed cases in excess of those expected. Attributable fractions (AFs) are measures of association between a disease and a specific exposure that attempt to assess the public health impact of that exposure.
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |