In the article by Lachs and colleagues (1), Table 1 (column 1) gives incorrect percents for patients in the high-probability subgroup. Because of a programming error, four patients with missing previous probabilities of urinary tract infection were mistakenly classified as being in the high-probability subgroup. Thus, the percents should be based on a denominator of 103. The values for sensitivity and specificity and their associated CIs in Table 3 are all based on n = 103 and remain correct.
Correction: Spectrum Bias in Evaluation of Diagnostic Tests. Ann Intern Med. ;117:975. doi: 10.7326/0003-4819-117-11-975_2
Download citation file:
Published: Ann Intern Med. 1992;117(11):975.
Copyright © 2018 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use