LETTER
Calculation Errors in Meta-Analysis
Andrea Messori, PharmD;
Giovanna Scroccaro, PharmD; and
Nello Martini, PharmD
1 January 1993 | Volume 118 Issue 1 | Pages 77-78
TO THE EDITOR:
Meta-analysis is increasingly being used in clinical research to address controversial issues where data from clinical trials are inconclusive [1, 2].
The Mantel-Haenszel test [3, 4], the most commonly used statistical technique for determining the overall odds ratio (or relative risk) of a meta-analysis, is calculated by pooling appropriately the individual odds ratios in published studies. Logistic regression [4] can also be used even though the technique is considerably more complex.
The recent paper by Hommes and colleagues [5] reports a meta-analysis of six randomized trials comparing subcutaneous heparin with continuous intravenous heparin for the initial treatment of deep vein thrombosis. To recalculate the odds ratio (with 95% CI), we used a computer program (Messori A. Unpublished observations) developed at our institution that is based on equations 7.18, 7.20, 7.21 (including the correction for zero values), 7.22, and 7.23 [4]. The result of our calculation was an odds ratio of 0.61 (95% CI, 0.298 to 1.251; P > 0.05); this figure differs greatly from the value reported by Hommes and associates (odds ratio, 0.62; 95% CI, 0.39 to 0.98; P < 0.05). When our recalculations involved the odds ratios (with 95% CI) of the six individual studies, we obtained the same identical results as those reported in Table 2 of Hommes' article.
Based on our recalculation of the overall odds ratio, we concluded that subcutaneous heparin is not more effective than intravenous heparin, exactly opposite to that of Hommes and colleagues. This could have resulted from a calculation error or because our technique differed from theirs. However, it would be surprising for two different, although well-accepted, statistical techniques to produce strikingly different results when performing the same calculation.
|
Author and Article Information
|
|---|
Hospital Pharmacy; Ospedale Policlinico Borgo Roma; Verona, Italy
1. Sacks HS, Berrier J, Reitman D, Ancona-Berk VA, Chalmers TC. Meta-analyses of randomized controlled trials. N Engl J Med. 1987; 316:450-5.
2. Lau J, Antman EM, Jimenez-Silva J, Kupelink B, Mosteller SF, Chalmers TC. Cumulative meta-analysis of therapeutic trials for myocardial infarction. N Engl J Med. 1992; 327:248-54.
3. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959; 22:719-48.
4. Schlesselman JJ. Basic methods of analysis. In: Schlesselman JJ; ed. Case-Control Studies. Design, Conduct, Analysis. New York: Oxford University Press; 1982:171-226.
5. Hommes DW, Bura A, Mazzolai L, Bueller HR, tenCate JW. Subcutaneous heparin compared with continuous intravenous heparin administration in the initial treatment of deep vein thrombosis. Ann Intern Med. 1992; 116:279-84.
About Letters
The Editors welcome submissions for possible publication in the Letters section. Authors of letters should:
•Include no more than 300 words of text, three authors, and five references
•Type with double-spacing
•Send three copies of the letter, an authors' form signed by all authors, and a cover letter describing any conflicts of interest related to the contents of the letter.
Letters commenting on an Annals article will be considered if they are received within 6 weeks of the time the article was published. Only some of the letters received can be published. Published letters are edited and may be shortened; tables and figures are included only selectively. Authors will be notified that the letter has been received. If the letter is selected for publication, the author will be notified about 3 weeks before the publication date. Unpublished letters cannot be returned.
Annals welcomes electronically submitted letters.
Top
Author & Article Info
References
Article
