Corrections
BMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7056.550 (Published 31 August 1996) Cite this as: BMJ 1996;313:550Grand Rounds—Hammersmith Hospital: Tuberculous enteritis
An editorial error led to authorship of this paper being attributed to the wrong person on the front cover (27 July, pp 215-7). The author is Farah B Ahmed [not S D Taylor Robinson, who is the series editor].
Trials to assess equivalence: the importance of rigorous methods
We regret that an editorial error occurred in this article by Jones et al (6 July, p 36). In the appendix two of the formulas for power calculations had brackets in the wrong places and two had a -1 wrongly inserted. The power calculations for normally distributed data should have appeared as: Two sided case
Power = 2(Phi) (δ/(square root)s2(2/n) - z(1-(alpha))-1
One sided case
Power = (Phi) (δ/(square root)s2(2/n) - z(1-(alpha))
The power calculation for binary data, one sided case, should have appeared as:
Power = (Phi) (δ/(square root)p(100 - p) (2/n) - z1-(alpha))
In addition, in the boxed example on p 37 the Greek delta symbol ((delta)) should have appeared as a Greek sigma to give the phrase: the prior estimate of (sigma)2.