 |
|
Analytic Errors Undermine Conclusions of Cardiovascular Study
|
|
| Since this article does not have an abstract, we have provided the first 150 words of the full text and any section headings. |
|
|
|
See Retracted Article and Notice of Retraction.
In their study of hypertension as a predictor of cardiovascular mortality, we believe that Panagiotakos and colleagues1 have made serious analytic errors, undermining their reported conclusions.
We question two specific sets of results. First, the authors report inferences based on "log-likelihood differences."1(p2144) These are usually represented by  2 statistics, as produced by the authors' software (version 11.0; SPSS Inc, Chicago, Ill). However, they report values ranging from 15 640 to 15 978, which are inconceivable 2 values given the sample size. More straightforwardly, these "–2 log-likelihood" values are really minus two times the optimized values of the log-likelihood for each model. Hence, the lowest reported value, using mid blood pressure, is actually the best fitting model considered, making mid blood pressure the best predictor. Pulse pressure is the worst of the single predictors, the opposite of the authors’ stated conclusions. Further evidence comes from Table 31(p2145). . . [Full Text of this Article]
AUTHOR INFORMATION
Kenneth Rice, PhD;
Richard Kronmal, PhD;
Thomas Lumley, PhD
RELATED ARTICLE
Notice of Retraction: "The Relation Between Pulse Pressure and Cardiovascular Mortality in 12 763 Middle-aged Men From Various Parts of the World: A 25-Year Follow-up of the Seven Countries Study" (Arch Intern Med. 2005;165:2142-2147)
Philip Greenland
Arch Intern Med. 2006;166(16):1787.
EXTRACT
| FULL TEXT
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES
Notice of Retraction: "The Relation Between Pulse Pressure and Cardiovascular Mortality in 12 763 Middle-aged Men From Various Parts of the World: A 25-Year Follow-up of the Seven Countries Study" (Arch Intern Med. 2005;165:2142-2147).
Greenland
Arch Intern Med 2006;166:1787-1787.
FULL TEXT
|
