Review Question - QID 106440

The 95% confidence interval (CI) for a statistically significant outcome cannot include the null value (i.e. RR of 1), making the best answer oF 95% CI 1.01-3.70.

Remember that 95% confidence interval and p-values are related. 95% CI describes the range in which the mean would be expected to fall 95% of the time if the experiment were performed again and again. P value describes the percent probability that a positive result (i.e., showing a difference between 2 study groups) is due to chance. A "statistically significant" p value is always decided upon prior to analysis--this level is known as "alpha", and is frequently (and arbitrarily) 0.05. A 95% CI that does not include 1.0 equates to a p value of 0.05 or less (meaning the study is statistically significant given a desired alpha value of 0.05). In the case of relative risk, the null value is 1 (meaning there is no difference in the probability of occurring vs. not occurring). If the 95% CI for the relative risk includes the null value, then there is insufficient evidence to conclude that the groups are statistically significantly different, meaning the p-value cannot be less than <0.05.

Goodman discusses the p value fallacy in medicine. He notes that this measure may be overused in medicine. Caution must be used when a single number is used to capture the longterm consequences of an experiment and the evidential meaning of a single result.

Goodman and Berlin note that CIs are important when setting sample size and are often interpreted as representing a range of believable values. This is particularly useful in deciding whether a study had adequate numbers of patients to answer a question.

Incorrect Answers:
Answers 2-5: All of these contain the null value, which would not correspond to a statistically significant p-value.