Question

Systematic reviews are not free of biases. A meta-analysis was recently published that examined the association both red and processed) meat consumption and colorectal cancer. In their results the authors reported that the publication bias was in the studies on processed meat and colon cancer, noting that small studies with inverse missing. When present publication bias with these characteristics is most likely to lead to:

a over-estimation of the association between an exposure and an outcome.

b.confounding the association between an exposure and an outcome.

c.under-estimation of the association between an exposure and an outcome.

d.lack  of the association between an exposure and an outcome.

Answer 1

The meta-analysis showed that there is a publication bias. This publication bias has favored the reporting of studies that relate processed meat to colorectal cancer. However, there are fewer studies on red meat and colorectal cancer.

Hence, the results that will be obtained by these studies will be very specific to the exposure to processing meat and the outcome of the same on the prevalence of colorectal cancer.

Therefore, this is an instance of confounding bias where the alternative scenario is being missed by the researchers due to the under-reporting of the complementary studies.

Over-estimation or under-estimation depends on the nature of the estimation as to whether there is overfitting of data due to the studies on processed meat or underfitting of data. Since this is not known as the studies on the inverse are absent, therefore, this is an instance of confounding.

Thus, the answer is B.