My Research Papers
My Research Papers
net/publication/313246269
CITATIONS READS
0 1,019
2 authors:
All content following this page was uploaded by Peter Donald Griffiths on 07 February 2017.
One of the least agreeable aspects of editing an previously published paper to answer the same or similar
academic journal is dealing with alleged breaches of research question or test the same hypothesis is a
publication ethics. In this month’s issue of the Interna- duplicate publication.
tional Journal of Nursing Studies (IJNS) one of the Usually the authors of the duplicate manuscript are
dilemmas and difficulties that arise all too frequently is the same as the original authors, but not necessarily, and
the subject of two Commentaries (Moon, 2008; there are cases where another author has published the
Gøtzsche, 2008). This case refers to a paper by Moon same material elsewhere, without knowledge of the
and his colleagues (Lee et al., 2006) which was accepted original author. In some cases duplicate publication may
for publication in this journal after the usual peer review be perfectly legitimate. Publication of a previously
process. The article appeared in the Article-in-Press published paper in a different journal, perhaps in a
section of the journal’s website. ‘‘Articles-in-Press’’ are different language or for a different readership, with the
peer-reviewed papers which are made available online knowledge and agreement of the editors of both journals
(through ScienceDirect) but which have not yet been is not unwarranted duplicate publication. For example
published in the printed journal. The article has now the publication of very similar versions of Cochrane
been withdrawn from the Article-in-Press service be- Systematic Reviews in both the Cochrane Library and a
cause we judged it to be an unwarranted duplicate traditional journal is a frequent occurrence with the
publication; this was brought to our attention by intention of ensuring wide dissemination of the review.
someone conducting a systematic review who noticed Indeed the IJNS has published a version of a Cochrane
that most of the data in the paper was from a previously review (Griffiths et al., 2004, 2005) after its publication
published study. in the Cochrane Library and would gladly do so again.
In this age of computerised healthcare databases and Prior publication of the abstract of a paper or a report
Internet search engines there is good chance that such to funders with limited circulation is not generally
papers will be identified when they are screened by the problematic and would not be considered duplicate
editors prior to being sent for peer review, but some- publication at all. However in all cases we, as editors,
times, as happened in this case, a paper slips through the would want to know about this so we can make a
net. We do not propose to say any more about this judgement of the novelty of the full paper and how far
particular paper here, but to comment more generally its publication will contribute to what is already known.
about the problem of duplicate publication and explain Crucially when legitimate duplication occurs it must be
why we are concerned about it. fully and explicitly acknowledged in the publication.
Firstly, what is duplicate publication? Duplicate Duplicate publication which is covert is clear scientific
publication, in any form (on line or in print), has been misconduct. A passing reference is not sufficient.
defined as ‘publication of a paper that overlaps It is worth considering how duplicate publication
substantially with one already published’ (URMSBJ, differs from the practice of ‘salami slicing’. Salami
1997). But what constitutes ‘substantial’? We are slicing or salami publication usually refers to submitting
reminded of Judge Potter Stewart’s famous definition different manuscripts drawn from data collected from a
of pornography: ‘I know it when I see it’ (Jacobellis v. single research study or a single data collection period.
Ohio, 378 U.S. 184; 1964). For practical purposes, In some cases studies may draw on data that has been
however, a paper which reports the same data as another reported and analysed previously and largely repeat the
0020-7489/$ - see front matter r 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijnurstu.2008.07.003
ARTICLE IN PRESS
1258 Editorial / International Journal of Nursing Studies 45 (2008) 1257–1260
analysis with the addition of some new data or analysis those working away to index articles for databases; the
not previously reported. There is, in our opinion, problem is so serious that MEDLINE has duplicate
considerable similarity between some ‘salami’ publica- publication as one of its publication types. If, as
tion and duplicate publication and no clear demarcation suggested by Tramer et al.’s (1997) paper, around 17%
between the two practices. This is a world more often of the papers on MEDLINE are duplicates, just think
comprising shades of grey than black and white but in of the time saved by indexers and by those searching
general salami publication is not acceptable when the the database, had they not been published in the first
aim is simply to get as many individual papers from a place.
single project as possible. So, why do authors risk it? The academic and research
However, this said, there may be cases where community must take at least some responsibility for the
reporting the results of a study in one paper is not problem of duplicate publication through providing
possible; for example the results of large clinical or perverse incentives. The old adage ‘publish or perish’ is
epidemiological studies which address a series of distinct true, in some countries and academic communities more
research questions. A more debateable example is of than in others, and curriculum vitae’s inflated by a large
studies usually involving collection of substantial inter- number of publications is seen by many as the fast-track
view, observational or survey data, which researchers route for promotion and academic advancement.
split up and report in separate papers, perhaps under the Quantity is emphasised at the expense of the quality,
different themes that have emerged from the analysis. with the result that the mean contribution to knowledge
What’s wrong with duplicate publication and why are of papers published in academic journals is diminished.
we bothered about it? There are two major ethical In the UK the Research Assessment Exercise, went some
objections to duplicate publication. First, particularly in way to countering this trend with its requirement that
this era of systematic reviews, duplicate publications can each member of faculty who were returned submitted up
lead to exaggeration of research findings and so threaten to four of their highest quality publications. The effects
the evidence base of treatments. A convincing example is of its replacement with the Research Excellence Frame-
a systematic review of the effects of the drug ondanse- work (Nolan et al., 2008), with its emphasis on citation
tron on post-operative sickness (Tramer et al., 1997) counts as a measure of research quality, is, as yet,
which showed that positive trials are more likely to be unknown.
subject to duplicate publication than those with negative Authors are central to this whole issue. They need to
results, so exacerbating the existing problem of publica- respect the rights of academic journals to receive original
tion bias. In this case the authors found that data from manuscripts and failure to do this violates the trust
nine of the 84 trials included in the review, appeared in which should exist between authors, editors and readers.
23 separate publications; this included four pairs of Our advice for authors is to publish the results from a
identical papers, but written by different authors. In single study together in one substantial paper wherever
only one paper was prior publication of the same data possible. The generous word length of the IJNS
acknowledged. Failure to exclude duplicate publications compared to that of many other journals offers this
would have overestimated ondansetron’s efficacy by possibility. Where a single publication is not possible full
23% (improving the number needed to treat from 6.4 to and open disclosure is required to ensure that there is
4.9). Whilst most readers trust the veracity of healthcare full transparency. If in doubt the best advice to authors
publications Tramer et al.’s (1997) paper and others is to protect themselves through disclosure to the editor
suggests that the proportion of duplicate publications is of any prior publications which draw on the same data
higher than one would suppose. For example, Bailey, set or address the same research questions. The IJNS
Editor of Archives of Otolaryngology—Head and Neck requires authors to identify what is already known about
Surgery reports that of 1000 randomly selected authors the subject and specify the contribution of this paper to
from his journal, 201 published 644 papers which could knowledge; if authors can provide a convincing answer
be regarded as duplicates (Sun, 1989). Even in cases of to this question then it may be that publishing different
salami publication (short of actual duplication) a papers drawing from the same data set is justified.
misleading impression of the weight of independent Crucially, just as with cases where legitimate duplica-
evidence can be created. tion occurs, full and explicit disclosure is important.
The second main ethical objection to duplicate This is not a matter of inserting a reference to prior
publication is that it can displace valuable articles from publications in the paper. It is not uncommon for us to
publication, since publication space, whilst plentiful, is a return manuscripts to authors for revision because a
finite resource and ought to be used judiciously. As close reading suggests that the sources that they draw on
editors we are particularly conscious of demands on our are, in fact, other reports of the same study.
excellent international panel of peer reviewers, and the We would not wish to impugn the integrity of authors
additional load duplicate publication imposes on them in such cases although we have no doubt that on
to no purpose. And we need also to spare a thought for occasion there is a deliberate attempt to increase the
ARTICLE IN PRESS
Editorial / International Journal of Nursing Studies 45 (2008) 1257–1260 1259
appearance of novelty. Sometimes no reference to prior Importantly these guides apply not just to publica-
publication is made but we find other studies while tions with the IJNS but to subsequent papers that might
searching for reviewers. In these cases it is hard not to be published elsewhere. Different editors may have
suspect that there is some attempt to deceive us. What is different views about how studies should be cross
required is sufficient detail that the reader knows that referenced, where duplicate publication merits space
aspects of a study have been published previously or, and when multiple publications from the same study are
where future publications will address the same study, genuinely warranted. However what is crucial is that
the confidence that those studies will be reported in such neither the editors nor the subsequent readership of a
a way that makes things clear. paper are left in any doubt about the relationship
We recognise, of course, that delays in reviewing and between a series of publications and the sources of data
publication can cause even the most assiduous of upon which it draws.
authors problems in cross referencing publications as
additional manuscripts are prepared and submitted
prior to the acceptance or publication of preceding References
ones. This is one reason we encourage single publication
where possible. For randomised controlled trials pro- Camerino, D., Estryn-Behar, M., Conway, P.M., van Der
spective registration of the trial gives a unique trial Heijden, B.I.J.M., Hasselhorn, H.-M., 2008. Work-related
identification number which can be used to report a factors and violence among nursing staff in the European
study and links all publications from the study and we NEXT study: a longitudinal cohort study. International
now ask authors to provide this detail where they have Journal of Nursing Studies 45 (1), 35–50.
Estryn-Behar, M., van der Heijden, B., Camerino, D., Fry, C.,
registered. For other study designs this is not so readily
Le Nezet, O., Conway, P.M., Hasselhorn, H.-M., and the
available but the principle can apply. NEXT Study group, 2008. Violence risks in nursing—results
An excellent example of good practice is the European from the European ‘NEXT’ Study. Occupational Medicine
NEXT study which explored factors related to early exit (London) 58 (2), 107–114.
from the profession. The scale of this study and the large Flinkman, M., Laine, M., Leino-Kilpi, H., Hasselhorn, H.M.,
number of factors probably warranted several publica- Salanterä, S., 2008. Explaining young registered Finnish
tions and indeed the authors have reported them in nurses’ intention to leave the profession: a questionnaire
several journals including the IJNS (Camerino et al., survey. International Journal of Nursing Studies 45 (5),
2008; Estryn-Behar et al., 2008; Flinkman et al., 2008; 727–739.
Simon et al., 2008). Where (on occasion) the authors Gøtzsche, P., 2008. Covert duplicate publication and mislead-
ing sample size calculation: commentary on Lee et al
have not cross referenced clearly related papers the
(2006). International Journal of Nursing Studies 45 (9),
reader can be left in no doubt that they are reading 1398.
reports of the same study (e.g. Camerino et al., 2008; Griffiths, P., Edwards, M., Forbes, A., Harris, R., 2005. Post-
Estryn-Behar et al., 2008) and while we would be acute intermediate care in nursing-led units: a systematic
happier as editors if the later paper on work related review of effectiveness. International Journal of Nursing
violence referred to the earlier one it is clear that there is Studies 42 (1), 107–116.
no attempt to deceive. Griffiths, P., Edwards, M., Forbes, A., Harris, R., Ritchie, G.,
What we advise is that those submitting a paper from 2004. Intermediate care in nursing-led in-patient units:
a study where there will be several publications do the effects on health care outcomes and resources. Cochrane
Database of Systematic Reviews 2004 (4), Art. No.:
following:
CD002214.pub002212.
Lee, I.S., Moon, J.S., Yoo, Y.S., 2006. Effectiveness of
Make sure that any prior publications are both
bedding control instruction for patients with respiratory
properly cited and referred to as being part of the allergies: a randomized controlled trial. International
same study. Journal of Nursing Studies, doi:10.1016/j.ijnurstu.
Give the fullest possible reference to ‘in press’ papers 2006.08.009 (Withdrawn).
but not those currently under review. Moon, J.S., 2008. Response to Gøtzsche (2008). International
Update the citations in all papers that are being Journal of Nursing Studies 45 (9), 1399.
prepared if the status of a paper changes while there Nolan, M., Ingleton, C., Hayter, M., 2008. The research
is still opportunity during the publication processes. excellence framework: a major impediment to free and
Submit copies of other manuscripts (including those informed debate?. International Journal of Nursing Studies
under review or in press) so that we can assess the 45 (4), 487–488.
Simon, M., Tackenberg, P., Nienhaus, A., Estryn-Behar, M.,
degree of overlap and the novel contribution.
Maurice Conway, P., Hasselhorn, H.M., 2008. Back or
Identify your study in a distinctive way and refer to it neck-pain-related disability of nursing staff in hospitals,
in this way in all publications so that a reader can nursing homes and home care in seven countries—results
clearly and easily identify that all papers emanate from the European NEXT-Study. International Journal of
from the same study. Nursing Studies 45 (1), 24–34.
ARTICLE IN PRESS
1260 Editorial / International Journal of Nursing Studies 45 (2008) 1257–1260
Sun, M., 1989. Peer review comes under peer review. Science Ian Norman, Peter Griffiths
244, 910–912. Florence Nightingale School of Nursing and Midwifery,
Tramer, M.R., Reynolds, D.J.M., Moore, R.A., McQuay, H.J., King’s College London, James Clerk Maxwell Building,
1997. Impact of covert publication on meta-analysis: a case 57 Waterloo Road, London SE1 8WA, UK
study. British Medical Journal 315, 635–640. E-mail addresses: ian.j.norman@kcl.ac.uk (I. Norman),
Uniform Requirements for Manuscripts Submitted to Biome-
peter.griffiths@kcl.ac.uk (P. Griffiths)
dical Journals (URMSBJ), 1997. Annals of Internal
Medicine 126, 36–47.