Sandbox

Flu Shots

The Wall Street Journal of September 24, 2009 reports on a study in the current issue of the New England Journal of Medicine dealing with the efficacy of flu shots during the “2007-2008 flu season.” There were “1952 healthy adults who received either a placebo or one of two [types of] vaccinations,” either a shot or a spray.

“Of the 813 volunteers who received the flu shot, 28 of them or 3.4%, later developed a confirmed case of influenza. In the FluMist group; 56 of the 814 volunteers, or 6.9% developed the flu, while 10.8% of the placebo group had confirmed flu cases.”

Discussion

1. Here is the Minitab output for the comparison between the two forms of vaccination:

Test and CI for Two Proportions

Sample -- X -- N -- Sample p
1 - 56 -- 814-- 0.068796
2 - 28 -- 813 --0.034440<br?

Difference = p (1) - p (2)
Estimate for difference: 0.0343557
95% CI for difference: (0.0129208, 0.0557907)
Test for difference = 0 (vs not = 0): Z = 3.14 P-Value = 0.002

Fisher's exact test: P-Value = 0.002

How does this output support the conclusion that the shot works better than the spray?

2. Because the number of volunteers is 1952 and then total number in the control arms is 1627 (814 + 813), the number of volunteers is 325. Below it’s the Minitab output comparing the controls vs. the placebo:

Test and CI for Two Proportions

 Sample X N Sample p 1 35 325 0.107692 2 84 1627 0.051629