This is the text extract for Heart Protection Study, browse documents here.

---

CORRESPONDENCE

Although a change to expedite publication of correspondence is welcome, we feel that reducing the time allowed for submission of letters from 8 to 2 weeks might be too short. This reduction could deter or prevent important correspondences to The Lancet. Instead, a deadline of 3–4 weeks after publication of an article might be more accommodating for both the editorial staff and authors. This length will allow authors the time to adequately research and compose a letter and communicate with colleagues, if needed, while fulfilling other professional and family duties. We strongly recommend that the editorial staff of The Lancet seek the opinion of contributing authors, if it has not already, to determine whether the deadline we propose is more appropriate than the newly instituted 2-week deadline.

*James Bradley Summers, Joseph Kaminski

Department of Diagnostic Radiology, University of South Alabama, Mobile, AL 36616, USA (JBS); and Department of Radiology, Medical College of Georgia, Augusta, GA 30912, USA (JK) (e-mail: Orotic@scubadiving.com) 1 Mullan Z. Lancet Correspondence: old letters, new rules. Lancet 2003; 361: 12.

Sir—I applaud your new rules about old letters.1 The example you cited of a letter to the editor about a paper published in April appearing in the last issue of October is almost as bad as one of mine, which I submitted on July 3 about an article published on June 22 but which was not published until Dec 14, 2002. However, the 2-week limit between publication of the article and submission of letters seems a bit harsh and impractical. It sometimes takes more than 2 weeks for my library to receive overseas journals. Furthermore, you penalise old-fashioned physicians who might not have access to e-mail. As long as The Lancet still publishes the journal in paper form rather than exclusively electronically, the process of submission of letters to the editor should not be too restrictive. Many pearls of medical wisdom appeared first in the form of letters to the editor, so let us not discourage them.

Tsung O Cheng

George Washington University Medical Center, Washington, DC 20037, USA (e-mail: tcheng@mfa.gwu.edu) 1 Mullan Z. Lancet Correspondence: old letters, new rules. Lancet 2003; 361: 12.

ence.1 But restricting submission to within 2 weeks of publication seems too strict, and can only lessen the quality of postpublication peer review. Just 7 months ago, in an entire issue of JAMA relating to peer review in biomedical publication, the short time limits for submitting letters commenting on published articles were strongly criticised. “Time limits discourage potential postpublication peer review, deterring potential correspondents by the unambiguous cut-off. . . Time limitation on correspondence denies readers the opportunity to draw attention to methodological deficiencies. In effect, there is a statute of limitations by which authors of articles in these journals are immune to disclosure of methodological weaknesses once some arbitrary (short) period has elapsed”.2 And yet the time limits cited there for the six top medical journals were 4–8 weeks. Richard Horton, in the same JAMA issue, concluded that failure to recognise the critical footprint of primary research weakens the validity of guidelines and distorts clinical knowledge.3 Just how does restricting to 2 weeks help stop potentially incorrect conclusions, based on faulty analysis, remaining in the literature to be cited uncritically by others?4

Scott Metcalfe

c/o Pharmaceutical Management Agency (PHARMAC), Wellington, New Zealand (external contractor) (e-mail: scott.metcalfe@pharmac.govt.nz) 1 2 Mullan Z. Lancet Correspondence: old letters, new rules. Lancet 2003; 361: 12. Altman DG. Poor-quality medical research: what can journals do? JAMA 2002; 287: 2765–67. Horton R. Postpublication criticism and the shaping of clinical knowledge. JAMA 2002; 287: 2843–47. Bland M. Fatigue and psychological distress: statistics are improbable. BMJ 2000; 320: 515–16.

authorship within days of the issue being posted. However, not everyone’s circumstances permit access to the online version of the journal, and we recognise that delivery times for the paper issue are far from ideal in some countries. We will certainly not dismiss outright anyone who is not able to write in within the 2-week deadline because of geographical or delivery reasons. We sincerely hope that our “tough” stance on submission times merely makes our correspondence section more up-to-date, and does not detract from the quality, openness, and worldwide applicability of the debate that characterises it.

Zoë Mullan

The Lancet, 32 Jamestown Road, London NW1 1BJ, UK (e-mail: zoe.mullan@lancet.com)

Sir—Yesterday (Jan 17), all deadlines seemed so far away

my

Now I need an internet connection on which to bang away My beloved Lancet (dated Jan 4) came in the mail But what I read on page 121 made me pale So I’ve express-written this missive And hope that, for its publication, you’ll be permissive

Colin Butler

4 Queen Street, Campbell Town, Tasmania 7210, Australia (e-mail: csbutler@iprimus.com.au) 1 Mullan Z. Lancet Correspondence: old letters, new rules. Lancet 2003; 361: 12.

3

4

DEPARTMENT OF ERROR

Paclitaxel plus carboplatin versus standard chemotherapy with either single-agent carboplatin or cyclophosphamide, doxorubicin, and cisplatin in women with ovarian cancer: the ICON3 randomised trial—In this Article by the International Collaborative Ovarian Neoplasm (ICON) Group (Aug 17, p 505), the last sentence of the “Procedures” section (p 508) should have read: “Patients allocated the combination of paclitaxel and carboplatin were to receive paclitaxel at a dose of 175 mg/m2 given in a 3-h infusion followed by carboplatin at the same dose as the control group set out above”. The number of patients in the “CAP as control” group whose non-protocol treatment was defined as “not known” (table 3) should have been 34 for those assigned CAP and 26 for those assigned paclitaxel plus carboplatin. The number of patients at risk in the progression-free survival curve for carboplatin control (figure 3) should have been 478, 393, 293, 223, 184, 150, 129, 111, 85, 52, 34, and 13, respectively, for those assigned paclitaxel plus carboplatin, and 943, 745, 537, 424, 340, 281, 238, 201, 143, 96, 55, and 26, respectively, for those assigned control.

Editor’s reply

We at The Lancet believe in shaking things up a little, and we certainly seem to have done so with our new rules for submission of letters to the editor. Our intention was to galvanise potential correspondents into writing to us without delay if they feel moved to do so. We would then reward such alacrity with prompt dispatch to the article’s author, and timely publication of what hopefully provides clarification or explanation to a wide audience. All our content is available the day before the publication date on our website (http://www.thelancet.com), and many correspondents can and do compile detailed critiques with multiple

Sir—The Lancet is worried about undue delays between publication of original papers and consequent correspond-

706

THE LANCET • Vol 361 • February 22, 2003 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.