Lack of Statistical Power
Keyword(s):
To the Editor.— Dr Donowitz described an interesting research project in which the efficacy of overgown use in the prevention of pediatric intensive care unit nosocomial infection was examined.1 He concludes that "overgowns were an expensive and ineffective method of preventing or decreasing nosocomial infections."1 Although Dr Donowitz may be correct, his inattention to sample size and statistical power in this negative trial makes his argument less than compelling. Given the nosocomial "base rate" of about 10% without gowns, the detection of an infection rate difference of 50% between the nongown and gown subgroups (α = .05, β = .2) would require a sample size of about 500 in each subgroup.2
1989 ◽
Vol 10
(11)
◽
pp. 515-520
◽
1989 ◽
Vol 10
(11)
◽
pp. 515-520
◽
2016 ◽
Vol 5
(1)
◽
pp. 1570-1576
1986 ◽
Vol 41
(10)
◽
pp. 636-637
◽
1991 ◽
Vol 19
(2)
◽
pp. 160-165
◽
2012 ◽
Vol 45
(4)
◽
pp. 475-479
◽
2013 ◽
Vol 12
(1)
◽
pp. 32
◽