Not fast enough?
================

* Matthew B. Stanbrook

In their report on accelerated publication in medical journals,1 William Ghali and colleagues were apparently not asked to update their citation counts beyond December 1999, although their study was published more than 2 years later. I attempted to replicate their citation analysis for the 12 articles from the* New England Journal of Medicine*, using citation data updated to April 2002. Interestingly, I found that the *Science Citation Index* grossly underestimated the citation count for one paper2 because of a known problem with indexing of articles with group authorship.3

With corrected and updated data, the average citation rate was nearly twice as high for the 6 prereleased articles as for the 6 controls (50 v. 30 citations per year). Importantly, the control articles as well as the prereleased articles had citation rates well above the average for articles published in the journal in the same year (1.6 and 2.7 times higher, respectively). Citation rates for prereleased articles and their matched controls were not independent but were highly correlated (Spearman correlation coefficient = 0.88; *p* = 0.02).

The latter observation illustrates a major flaw underlying Ghali and colleagues' study as a whole: overmatching, which is biased against finding differences between accelerated and control articles. Although the stated objective of the study was to compare accelerated articles to a journal's “usual” output, the control articles were purposely chosen to be as similar as possible in subject and design to the case articles. Use of random controls would have been a fairer and more informative way to meet the study's objective.

For articles published in leading medical journals, the subject of a study appears to have a greater influence on its impact than the particular study itself. Despite this, accelerated articles were found to have higher importance scores in every dimension than nonaccelerated articles on the same topic. Although prerelease of articles is a relatively new phenomenon, journal editors appear to have been generally successful in expediting articles that are more important and will be more widely cited.

**Matthew B. Stanbrook** Editorial Fellow *New England Journal of Medicine* Boston, Mass.

## References

1.  1. Ghali WA, Cornuz J, McAlister FA, Wasserfallen JB, Devereaux PJ, Naylor CD. Accelerated publication versus usual publication in 2 leading medical journals. CMAJ 2002;166(9):1137-43.
    
    [Abstract/FREE Full Text](http://www.cmaj.ca/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiY21haiI7czo1OiJyZXNpZCI7czoxMDoiMTY2LzkvMTEzNyI7czo0OiJhdG9tIjtzOjIyOiIvY21hai8xNjcvNy83MzguMS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

2.  2. International Perinatal HIV Group. The mode of delivery and the risk of vertical transmission of human immunodeficiency virus type 1 — a meta-analysis of 15 prospective cohort studies. N Engl J Med 1999;340:977-87.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1056/NEJM199904013401301&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=10099139&link_type=MED&atom=%2Fcmaj%2F167%2F7%2F738.1.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000079489800001&link_type=ISI) 

3.  3. Dickersin K. Scherer R, Suci EST, Gil-Montero M. Problems with indexing and citation of articles with group authorship. JAMA 2002;287: 2772-4.
    
    [CrossRef](http://www.cmaj.ca/lookup/external-ref?access_num=10.1001/jama.287.21.2772&link_type=DOI) 
    
    [PubMed](http://www.cmaj.ca/lookup/external-ref?access_num=12038908&link_type=MED&atom=%2Fcmaj%2F167%2F7%2F738.1.atom) 
    
    [Web of Science](http://www.cmaj.ca/lookup/external-ref?access_num=000176002200005&link_type=ISI)