Waarop letten beoordelaars van 'oorspronkelijke stukken' die ter publicatie zijn aangeboden aan het Nederlands Tijdschrift voor Geneeskunde?

To determine to what extent reviewers of the Nederlands Tijdschrift voor Geneeskunde (Dutch Journal of Medicine, NTvG) use specific criteria when reviewing original articles.

Design

Descriptive.

Setting

Editorial office of the NTvG.

Methods

From 21 October 1993 to 1 March 1994, 89 original research manuscripts were received. Of these, 60 (67) were submitted to reviewers. Each manuscript was sent to a second reviewer, with the same expertise, as well. On the basis of two checklists with a total of 16 quality elements, the manuscripts were analysed and compared with the chief editor's judgement. The results of each individual referee were correlated with his year of graduation. The duration of the peer review procedure was established.

Results

The reviewers of the NTvG responded to 46 of the explicit questions from the editors (list one), while they commented upon 43 of the criteria in the second implicit list. There was no relation with the year of graduation. Approximately 12 of the flaws in the manuscripts, as assessed by the editor, were missed. About 42 of the correct elements in the manuscript were not mentioned in the evaluation. The interrater agreement between paired reviewers was 0.26 (kappa); on average 10.4 of the 16 checkpoints were mentioned or not by both. Peer review took 41 days on average.

Conclusion

Only a small fraction of the flaws in a manuscript were missed by the reviewers. Peer review is reasonably uniform at the NTvG, but would probably benefit from standardization.