PREPARING MANUSCRIPTS AND RESPONDING TO REFEREES’
REPORTS: INSIDE THE EDITORIAL BLACK BOX
Ian Stolerman
INTRODUCTION
This chapter describes how the peer review process
works and presents suggestions to authors of manuscripts.
It is based upon the experiences of scientists and clinicians
who have many years of experience as editors of prominent
addiction journals. The task of the editor is to publish manuscripts
appropriate to the journal and to assist would-be authors
in the production of suitable material. Many of the problems
facing authors writing for scholarly peer-reviewed journals
in the addiction field are similar to those in other fields.
Therefore, it is recommended that readers consult one or more
of the full-length books that have already been published
in this general area. Literally hundreds of books have been
published, as can be seen by searching amazon.com or PubMed
for 'scientific writing'. For example, a search of amazon.com
on 3 October 2003 produced 188 results, many of which were
relevant. A short list of recent books is provided at the
end of this chapter in Appendix 6.1.
TRIAGE: THE FIRST SELECTION
The author's quest to find a suitable publication
outlet ends with a letter stating 'I am pleased to inform
you that your manuscript is acceptable for publication …'.
The first step is to get the manuscript into
the peer-review system. This is not the inevitable consequence
of submitting to a peer-reviewed journal. Some journals state
formally that they operate a system of 'triage' whereby the
editor or his or her assistants decide which submitted articles
will be entered into the peer review process. In practice,
it is likely that all journals have such a system to protect
the profile of the journal and to avoid bothering authors
and peer reviewers with a long and laborious process where
a negative result can be predicted (see Box 6.1). Thus, if
something is received that clearly has no hope of acceptance
it may be rejected without review. Here, the difference between
journals is quantitative rather than qualitative: in the journals
of highest impact in science and medicine generally, it may
be more than half of the submissions that are rejected at
this stage. In contrast, in very specialised journals such
as those in the addiction field, it may be a fairly rare occurrence
for submissions to be rejected without review. Nevertheless,
this rate can be as high as 10% for some journals.
BOX 6.1 REASONS FOR REJECTION BY TRIAGE
• The submission is outside
the scope of the journal (e.g., it is about an abused drug
but it is not relevant in any discernible way to abuse of
or dependence on it)
• The manuscript type is not appropriate
(e.g., a case report is submitted to a journal that does not
publish case reports) • Clear ethical problems such as
apparent violation of current generally accepted standards
for the treatment of human or animal subjects
• The article is poorly organized
• The report is purely descriptive,
has no hypotheses, or reaches no conclusions
• Major methodological weakness
• There appears to be nothing new
in it
• Instructions to Authors are flagrantly
ignored in some way not covered by the above
Note: the editor has a duty to referees
as well as to authors, and tries not to waste their time by
requesting reviews of work that has no chance of acceptance
for one or more of the reasons above.
There are some aspects of manuscript preparation
that are so easy that everyone should get them right. To ensure
your manuscript has the best chance at triage (and beyond),
make sure you do all these things as set out in the Instructions
to Authors that every journal provides. Follow all advice
and recommendations to the letter, format your submission
precisely as requested, check to make sure that all sections
are complete, and that no tables, figures, or figure legends
are missing. Check the reference list to ensure that all cited
references are in it, and no others. Check the accuracy of
each citation. Look in the journal to see exactly how they
are styled. Then check them over again, after you have made
the corrections, until no more errors can be found. This sort
of work is tedious but does not need expensive resources,
profound knowledge of the subject or outstanding intellectual
ability.
If the editor sees at a glance that you do not
even get these straightforward, mechanical things right, he
or she may well develop a jaundiced view about your capability
to deal with more complex matters. Try to look at the manuscript
as an editor might. If you do not bother to do the easier
things required of an author, the editor might reasonably
conclude that you will not be able to do complex revisions
either and you may not be given the opportunity to revise
and resubmit.
COMMUNICATION WITH MORE EXPERIENCED WRITERS
Would-be authors may seek the advice of more
experienced colleagues at almost any stage of the publication
process. When planning a publication, discussion with colleagues
after a presentation of the work at a seminar in the home
institution may yield some tips as to the type of journal
that may be interested in the study. Subsequently, during
preparation of the manuscript, it may be appropriate to seek
the advice of local colleagues on technical aspects such as
statistical analyses. When a manuscript exists in a complete
form, it can often be immensely helpful to ask somebody to
read through it and make comments and suggestions. People
are very often willing to help if authors make clear they
value an expert opinion on aspects such as coverage of the
literature, clarity, style, language and validity of conclusions.
If there is no person in the author's own institution, it
is possible to approach outsiders and ask if they would be
willing to comment. Both people whom you know personally and
others who have published in the area are likely to feel flattered
and pleased that you value their opinion and may well provide
advice. The manuscript that cannot be improved has yet to
be written, and even experienced authors often seek the opinions
of colleagues because after working on a manuscript for months,
through revision after revision, it can be difficult to spot
the little problems that spring to the attention of a new
reader.
WRITING IN A FOREIGN LANGUAGE
It is an unavoidable fact that many authors
have to write in a language other than their own, and to convey
complex scientific ideas with clarity and precision can be
a difficult task even in one's native language. Authors may
therefore find it worthwhile to seek the assistance of colleagues
with more experience of writing in the chosen language and,
if possible, enlist a native speaker of the language to correct
the manuscript. If that is not possible, it may be necessary
to obtain the assistance of a professional translator to suggest
corrections. Journals provide varying amounts of assistance
in the correction of errors after accepting a manuscript for
publication, but they cannot do anything to assist reviewers
of poorly written manuscripts. A fuller consideration of language
issues my be found in Chapter 3.
THE PEER REVIEW PROCESS: SELECTION OF REFEREES
The next step for the editor is selection of
referees who will advise him or her of the strengths and weaknesses
of the work and recommend whether or not it should be accepted,
reconsidered after revision, or rejected outright. Referees
are also expected to make constructive suggestions for improving
the manuscript in a report that can be sent to the authors.
The criteria used for selecting referees are diverse and probably
few if any journals have tightly defined procedures. Box 6.2
shows the main criteria used by editors to identify referees.
The number of referees varies within and between journals,
but most commonly there are two referees. The editors of some
journals may work with only one referee but this seems to
be increasingly rare. Occasionally, three or more are used,
depending upon the journal and the editor's perception of
the complexity and significance of the work. For example,
multi-disciplinary manuscripts may require more than two referees
to ensure sufficient expertise. Similarly, if a study seems
likely to have major practical impact, for example upon policy
or treatment, the editor may wish to be especially certain
that it is assessed thoroughly. If the two referees selected
initially disagree about the paper it is a common practice
to turn to a third person for additional advice.
Box 6.2 SOME METHODS USED BY EDITORS TO IDENTIFY
REFEREES
• Recognised expertise
in the specific field of the manuscript as noted in the journal's
database of previous referees and authors
• Previous invitations to referee
for the journal have resulted in thorough, wellwritten, polite
reviews submitted in a timely manner
• Record of recent publications in
the field as determined by searches of databases such as MEDLINE
and PsycINFO
The following criteria are used to exclude
reviewers from consideration:
• Persons who are known to have a
very close connection to the authors or to have a conflict
of interest with the authors will be avoided
• People who are currently refereeing
another manuscript for the same journal or have refereed one
within a set period (e.g., three months) will be avoided
• Excessive praise or criticism of
the reviewer's work in the manuscript to be assessed
Note: different bulleted points will be
used in combination to reach a decision on whom to invite
and there will inevitably be appreciable variations between
journals with respect to the use of these different methods
of selection.
For all reports, regardless of whether they
are quantitative or qualitative, each journal has its own
set of instructions for referees; journals differ with respect
to the attributes of their 'ideal' manuscripts. There will
sometimes be a requirement for referees to complete a questionnaire
as part of the review, with ratings of the manuscript according
to criteria such as importance and likely impact on the field,
as well as technical competence. The referees are usually
also asked to make a recommendation on the fate of the manuscript,
and to justify it in confidential comments to the editor.
Finally, referees are in all cases expected to produce a report
that the editor will forward to the authors. The main purposes
of this report are (1) to make suggestions enabling the author
to improve the manuscript and (2) to list criticisms that
the referee believes need to be addressed if the report is
to be published. The report to the authors should not include
specific recommendations for acceptance or rejection of the
manuscript because that decision is in the hands of the editor.
The task of the editor is to reconcile sometimes conflicting
reports from different referees and to make a personal judgement
based on a variety of other considerations. The task is made
more difficult if the reviews contain conflicting recommendations
for publication. Referees are also asked to act according
to ethical guidelines that are presented and discussed elsewhere
in this book (see Chapters 7 and 8 and ISAJE's ethical guidelines
at Appendix B).
CRITERIA FOR EVALUATION OF MANUSCRIPTS
If the journal has published its Instructions
to Referees, or put them on a website, they will give you
an idea about the features that both editors and referees
will look at. Some desirable features of highly-rated studies
are probably shared among many journals.
If it is a quantitative study, the criteria
include the use of a sufficiently large and suitably representative
sample of the population under study, a high response rate
among invited participants, valid measures, absence of procedural
biases and minimal confounding of one independent variable
with another, and the use of appropriate controls. Similarly,
referees will look for as full a description of the methods
as allowed by the space available, with referencing of earlier
publications that provide more detail and establish the validity
of the methods and measuring instruments (where applicable).
Results must be described in a clear and logical sequence,
with all necessary information presented. No more detail should
be provided than can be covered in the discussion section.
The discussion should bring out the importance of all the
main findings, indicate how the work advances the state of
knowledge and understanding in the relevant sub-field. Additionally,
alternative interpretations of the data may be given, thus
acknowledging limitations of the study. Referees pay attention
to all the preceding points, and to many others.
In quantitative research, one of the common
problem areas is the data analysis where several mistakes
occur repeatedly. These include:
- failure to deal adequately with confounding
variables;
- claims to have shown something without performing
a (statistical) test that supports it directly and unequivocally;
- failure to control for multiple comparisons;
- drawing inappropriate conclusions from non-significant
associations or differences: we probably all realise that
lack of significance means only that we have failed to find
an effect and does not prove that no effect exists, but
we don't always remember this in our enthusiasm to explain
how our results fail to support the ideas of a scientist
whose theory we dislike.
Authors developing reports of randomised controlled
trials may wish to follow the CONSORT (Consolidated Standards
of Reporting Trials) checklist, which includes 22 items considered
essential to judge the reliability or relevance of the findings
(presented in Appendix 6.2 to this chapter in slightly abbreviated
form).
The criteria for the evaluation of qualitative
reports vary depending on the type of data and methods of
analysis (e.g., participant observation and ethnography, qualitative
interviews, content analysis, textual analysis, discourse
analysis, ethnography and conversation analysis). The following
issues should be addressed in most types of qualitative report
(see also Cuba and Cooking 1994; Mäkelä 1990; Medical
Sociology Group of the British Sociological Association 1994).
- Clear criteria for the selection of data
or subjects should be given. The material should therefore
be carefully positioned in the social and cultural space.
For example, different genres of fiction represent different
segments of the culture.
- A detailed account should be presented of
where and when data were collected or which existing data
sets were used. In studies based on fieldwork, the relation
between the fieldworkers and the subjects should be described,
and possible influence of the data collection on the phenomenon
under study should be discussed. Ideally, the researcher
should keep careful records of the data and offer them for
independent examination.
- The way the analysis was done should be stated
clearly with indication of whether reliability was assessed,
for instance by replicating the analysis.
- Themes, concepts and categories derived
from the data should be described. The interpretation process
should be divided into short steps, with rules of classification
and interpretation specified.
- Steps taken to guard against selectivity
in the use of data should be outlined; exceptions and deviant
cases should be discussed. Ideally, the reader should be
able to apply the same classifications, take the same analytic
steps and reach the same kind of results with another dataset.
- Data must be presented systematically so
that quotations, field notes and so on are easily identifiable.
- There should be enough primary evidence to
show a relation between evidence and conclusions, but presentation
of too many illustrations is to be avoided; the focus should
be on the most representative ones.
COMMON PROBLEMS WITH MANUSCRIPTS
All parts of a manuscript are open to criticism
from the title onwards. The first requirement for gaining
the confidence of an editor or a referee is to describe the
findings objectively and in a sober style without the use
of hyperbolic language. If your data are good, they will speak
for themselves. It is always better for the reader to find
that the results themselves are stronger than you claim, instead
of being obscure and unconvincing in relation to your description.
Every 'data not shown' statement may raise the suspicions
of referees that the authors are trying to hide something.
If there really is not enough space to show important data
graphically or in tabular form, then give some examples of
the more important of such results in the text (with means,
standard errors or other indicator of variance and numbers
of subjects, if it is a quantitative study).
The discussion section is the most difficult
part of a manuscript to write, and it often shows. Sometimes
the opening paragraph is only a summary of the results, which
is not satisfactory. One approach is to decide which are the
main new findings, mention only them, and summarise two or
three important conclusions that follow from them. It is also
common to find that the discussion does not focus on the aims
as stated in the introduction, and sometimes discusses issues
on which no background was given. A failure to place findings
in the context of previous knowledge means that the case for
publication is not made. Referees and editor want to know
what is new, what is confirmatory, and of any important failures
to confirm previous findings. Instead, authors may attempt
to extract too much from their data, in the sense of trying
to address too many different issues. The effect of this error
is to dilute strong conclusions with weakly supported ones,
giving an overall unfavourable impression leading to rejection.
The discussion should also consider alternative
interpretations of the study and acknowledge major limitations.
These may arise from methodological weaknesses; or unexpected
findings that could not be pursued to a firm conclusion, due
to practical limitations such as the project period coming
to an end, or a financial constraint (these reasons do not
need to be stated). If the referees discover these weaknesses
they will consider themselves smart and are likely to make
sure you know it; if you show that you are aware of the limitations
and understand the implications, they will perceive you as
smart and honest, which counts for a lot.
Don't waste time and space discussing 'trends'
that are not statistically significant; if the effect is not
there, its implications do not need discussing. Remember there
are more than enough 'significant' effects that do not replicate
and there is no need to create new myths. If it is believed
that a real and important difference was undetected due to
lack of statistical power, what needs to be done is a new
experiment, and then there may be something to discuss. Particular
but somewhat different problems are associated with reviews
and theoretical articles. If a review claims to be comprehensive,
it is now common practice to state in it the way that the
literature search was carried out, and to define criteria
for including articles. Articles that do not claim to be comprehensive
but rather argue for a particular theoretical viewpoint or
set of ideas are frequently less comprehensive. In such cases
authors often cite publications that support their own position
in a rather uncritical manner, and they may refer to few or
no articles that oppose it. Editors may then firmly but politely
ask the author to state the assumptions made and ensure that
the article makes clear to readers where issues are controversial.
Alternatively, where the intention is to let a distinguished
writer express a personal view based on his or her selective
citation of the literature, it should be made clear that a
case is being made for a theory and that a balanced assessment
of the state of the field is not being attempted.
Finally, remember you are writing a scholarly
paper and not running a campaign! Do not enter into politics
and polemics. For example, if your main finding is that a
widely used intervention is less favourable than another which
lacks some sort of official approval for general use in your
country, make the case for their relative merits and if appropriate
argue for a policy change, but don't abuse the politicians
and don't keep repeating the argument in more and more florid
and emphatic language. Political battles are not won in the
pages of academic journals.
THE EDITOR’S DECISION
The much-anticipated and sometimes feared letter
from the editor finally arrives on the corresponding author's
desk, together with the statements from the referees.
The editor will often need to reach a decision
based upon the balance between innovation and quality of work.
The perfect manuscript would have important new ideas with
far-reaching importance backed up by sound data obtained by
means of thoroughly validated methods. In reality, such manuscripts
are seen only rarely, if ever, and the editor and the referees
have to make judgements. If the approach to a problem is highly
novel or the study is a potential stimulus for further valuable
work, a manuscript may be accepted with data that are less
than wholly convincing. On the other hand, if there is not
very much that is really new, but the study is the first one
to address a particular methodological weakness of previous
work, then clear data of high quality will probably be essential.
The referees' reports and recommendations inevitably
and appropriately influence the editor's decision but they
are not the sole determining factors. Editors may study a
manuscript in varying amounts of detail and may have concerns
that are not reflected in referees' reports. These concerns
may relate to any of the whole range of issues that the referees
also address, but especially to the appropriateness of the
subject matter for the journal; whether there are any ethical
problems; and whether the importance of work is sufficient
to justify publication in his or her journal rather than in
a publication of lesser status that may be struggling to fill
its pages. Studies may be technically competent and presented
well, but may be unimportant because they merely confirm well-known
facts or because they focus on apparently trivial issues.
When the reports of referees are in agreement
with each other, the editor will most frequently accept the
recommendations made. It is a brave editor indeed who would
overturn the opinions of two independent experts, not least
because referees will soon stop assisting an editor who consistently
ignores the advice given. When referees disagree, the editor
may seek to sort out the matter by studying the manuscript
and coming down in support of one or other referee; this is
the ideal method if the editor can reach a clear view because
a decision is obtained quickly and the valuable time of another
expert is not required. However, sometimes referees reach
opposing conclusions on the basis of equally well-argued cases
and then the editor may feel it is essential to obtain advice
from a third person. This is especially likely to occur if
the work is outside the editor's main area of expertise.
When a third referee reaches a definite view
supporting one or the other of the earlier referees, then
the way forward is clear, but this does not always happen.
If the first referee supports publication strongly and the
second referee recommends rejection, the third referee quite
often says the manuscript is weak but may reach publication
standard after major revision; in such cases the contribution
of the third referee may swing the decision one way or the
other depending upon the journal's needs at the time. If the
journal is trying to raise the standard of published items,
such marginal manuscripts will probably be rejected, whereas
if the study is in a field that is under-represented in the
journal, the editor may wish to include it. An editor may
also seek to publish the article but in a shorter form that
reflects its lesser merits.
The abusive referee is a particular annoyance
to editors. The most commonly identified, although happily
quite rare, form of abuse occurs when a referee attempts a
review of the author instead of the manuscript. It is one
thing, and perfectly acceptable, to state that an argument
is constructed poorly and is unconvincing, that it is presented
badly, or does not take account of previous knowledge; it
is quite another thing to assert that the author is stupid,
careless or ignorant. Editors have a duty to alter or remove
such inappropriate remarks from a referee's report, so that
unnecessary distress is not caused and the author will be
encouraged to improve the manuscript. If the referee is young
and inexperienced, the editor may also explain the problem
with the report, whereas a senior person will more likely
not be invited to review again.
A particularly difficult situation arises if
the review process generates suspicion that the author has
engaged in scientific misconduct and unethical behaviour.
Such misconduct may be either minor or major in nature and
the editor typically has available a range of sanctions to
apply. These may include refusing to consider further work
from the author, reporting the matter to the author's institution
or employer, and publishing a statement in the journal to
alert the scientific community to the issue. The availability
of a code of practice by which editors can abide in such circumstances
is very helpful (see Appendix B for the ethical practice guidelines,
developed by the International Society of Addiction Journal
Editors). Editors are also wary of trying to resolve contentious
ethical issues because they do not have the resources to conduct
a full investigation; equally important, they cannot simply
brush the matter aside by refusing to publish suspect material
but must take reasonable steps to ensure that appropriate
action is taken. These and other related issues are discussed
in greater detail in Chapters 7 and 8, which deal with ethical
considerations in scientific publishing.
RESPONDING TO REFEREES’ REPORTS: GENERAL
RULES OF CONDUCT
Authors who achieve immediate acceptance of
a manuscript as submitted are rare indeed. Revisions are almost
always required prior to acceptance, and in many cases a final
decision cannot be reached until the revised version has been
assessed. Therefore, the way in which authors respond to the
reports of referees and to the editor can have a major influence
upon the outcome. If editors invite resubmission, it means
they are expecting to receive the manuscript back again.
An invitation to resubmit is not a half-hearted
and cowardly way of saying the work is unpublishable, but
rather an implicit suggestion that the editor remains interested
in the paper and that it is likely to be accepted if the author
is responsive to the questions and recommendations of the
reviewers. In such cases, it is nearly always worth resubmitting
unless there is some clear and unavoidable requirement with
which you cannot possibly comply. Some guidance for authors
on how to navigate through this maze successfully is now given.
The overriding aim of the response is to engender
trust among editors and referees. Authors should never claim
to have made changes that in fact they have not done. If the
covering letter says all requested changes have been made
and an editor or referee checks two or three points at random
and finds nothing much has changed, they may reject the manuscript
without looking carefully at the rest of it. If you have made
major changes by rewriting whole sections of the manuscript,
state that is the case and identify the sections. Alternatively,
if just a few words have needed to be inserted or deleted,
make clear which words were changed so that referees can see
what has been done. If you were asked to shorten something,
you should almost always do so and perhaps state by how much
(i.e. by how many words or pages). Don't try to fool the editor
by printing the new version in smaller type or by other stylistic
changes.
Be polite, even if you feel that the referees
have not understood your intentions. When you have been through
all the points of criticism, you should have an idea of the
changes you think are appropriate. Will they be enough? If
after reading the reports, you have concluded that none of
the recommendations is worth accepting and you don't want
to make any changes, it is commonsense to take a break from
the job and look at it again on another day! It is simply
not realistic to expect editors and referees to accept that
noneof the changes they request and the criticisms they make
are well founded. Referees spend anything from an hour to
a full day preparing their reports.
If you dismiss all this effort out of hand you
will get nothing published. You must therefore aim to make
changes to deal with as many as possible of the points raised,
and preferably with a clear majority of them.
THE COVERING LETTER: MAKE LIFE EASIER FOR
THE EDITOR
After completing the changes to the manuscript,
write a detailed reply to the referees. It is worth spending
a significant amount of time getting your reply to referees
as near perfect as you can. Sometimes constructing the letter
may take nearly as long as doing the revisions, but it will
not take as long as botching the job and then being obliged
to reformat the manuscript for another journal in order to
start the whole process over again. Nevertheless, it is best
to keep this reply as short as possible. Typical successful
replies will be in the range of 1-3 single-spaced pages.
If the referee makes a point in just three lines
and you need a page to rebut it, it is likely you have not
got straight to the heart of the matter and your reply will
probably not be convincing (it is best to write the minimum
needed to refute the criticism). If the referee cannot understand
a point that you made in the manuscript, it may be because
he or she is lacking in intellectual capacity (as we often
think when we encounter such comments on our own work). However,
if one person does not follow what you have written, the same
may apply to others. Referees are all published research workers
and are often the very people whom you might hope would read
your paper; if a referee cannot understand your point, try
to analyse your text to see how the misunderstanding may have
arisen. Then make changes to ensure it will not happen again.
Do remember that if a referee asks a question,
other readers may want to know the answer to it too. The answer
should therefore usually be contained in the revised manuscript
and not in the covering letter. The latter need only refer
to the sections of the manuscript where the question is addressed.
Save the referees' time and they will love you; don't get
them to read the answer to a question in the letter and then
refer them to the manuscript where they have to read the answer
over again to check if it is really there!
If possible, reply in numbered sections that
correspond with numbered points of referees. Explain the revisions
you made to deal with most of the criticisms and also explain
why you did not deal with the rest. Describe briefly each
change you made, referring to the relevant page or paragraph
in the revised manuscript. Try not to respond in a combative,
overly-assertive style. If there are major and important changes
recommended that you are sure are wrong, then present a concise,
logically-argued rebuttal. If there are minor changes requested
that you feel don't really improve matters, do them anyway
because it helps a lot if you can truthfully claim to have
dealt with the majority of points. At all stages, remember
that although referees and editors may appear to be distant,
selfopinionated and arrogant, they are also human beings with
their own feelings, emotions and problems. If you want acceptances,
make life easy for them by writing clearly, and don't antagonise
them by criticisms or by gratuitous insults however unwise
and misguided you think they may be.
It is sensible to maximise and stress agreements
with what referees write and to acknowledge their contribution
when they have made suggestions that improve the manuscript.
Don't build minor disagreements into major issues. You probably
only need to make minor changes to accommodate them, and then
mention the changes in the cover letter, instead of wasting
time arguing and putting the referee's back up in the process.
However, it is not necessary or appropriate to minimise disagreements
to the point of dishonesty; they should be dealt with by logical
rebuttal in the cover letter and, sometimes, by acknowledging
and discussing the point in the manuscript.
Perhaps the most difficult case occurs when
you feel that a referee shows a bias towards a theoretical
approach that differs from yours, and therefore undervalues
the work. Here you can explain in the cover letter that there
are different approaches to the problem (state what these
are), that yours is equally valid, that there is a genuine
difference of opinion and that you have a different but scientifically
legitimate point of view. However, this strategy is probably
unwise unless you have a strong case and there is no other
way to deal with the issue. In the end the editor will have
to decide and what one person perceives as objective and unbiased
looks very different from another viewpoint.
At the end of the day, the editor wants to have
papers to publish. The number of acceptances rather than of
rejections is therefore the mark of success and of an editor's
job well done. Authors, editors, referees and publishers must
all work together to ensure the production of a journal of
high quality that achieves its intended objectives.
ACKNOWLEDGEMENTS
I would like to thank Klaus Mäkelä
and Kerstin Stenius for assistance with drafting the material
about qualitative research.
APPENDIX 6.1 GENERAL PUBLICATIONS ON SCIENTIFIC
AND MEDICAL PUBLISHING
This is a very short selection from the huge
number of publications. Many additional works may be found
by searching biomedical databases such as PubMed or on-line
booksellers.
Albert T, 2000. The A-Z of Medical Writing.
BMJ Books. ISBN 0727914871.
American Medical Association Manual of Style,
1998. 9th ed. edited by C. Iverson. Williams & Wilkins.
ISBN 0683402064. American Psychological Association, Publication
Manual, 2001. 5th ed. American Psychological Association.
ISBN 1557987904.
British Medical Association, http://www.bma.org.uk/ap.nsf/Content/LIBReferenceStyles.
This is a useful website established by the
British Medical Association, giving general guidance on resources
for people publishing in the biomedical field (accessed 3
October 2003).
Huth EJ, 1990. How to Write and Publish Papers
in the Medical Sciences. 2nd ed. Williams & Wilkins. ISBN
0683042718.
McInerney DM, 2001. Publishing your psychology
research. Sage. ISBN 0761973370.
Peat J, 2002. Scientific writing. BMJ Books.
ISBN 0727916254.
Richardson P, 2002. A guide to medical publishing
and writing. Quay Books. ISBN 1856422224.
APPENDIX 6.2 CHECKLIST OF ITEMS TO INCLUDE
WHEN REPORTING A RANDOMISED TRIAL
This section consists of a slightly shortened
version of the checklist of Moher et al. (2001).
TITLE AND ABSTRACT
How participants were allocated to interventions
(e.g., ‘random allocation’, ‘randomised’,
or ‘randomly assigned’).
INTRODUCTION
Scientific background and explanation of rationale.
METHODS
Participants Eligibility criteria for participants
and the settings and locations where the data were collected.
Interventions Precise details of the interventions intended
for each group and how and when they were actually administered.
Objectives Subjective objectives and hypotheses. Outcomes
Clearly defined primary and secondary outcome measures and
when applicable, any methods used to enhance the quality of
measurements (e.g., multiple observations, training of assessors).
Sample size How sample size was determined and, when applicable,
explanation of any interim analyses and stopping rules. Randomisation
Method used to generate the random allocation sequence, including
details of any restriction (e.g., blocking, stratification).
Allocation concealment Method used to implement the random
allocation sequence (e.g., numbered containers or central
telephone), clarifying whether the sequence was concealed
until interventions were assigned. Implementation Who generated
the allocation sequence, who enrolled participants, and who
assigned participants to their groups. Blinding Whether or
not participants, those administering the interventions, and
those assessing the outcomes were aware of group assignment.
Statistical analysis Statistical methods used to compare groups
for primary outcome; methods for additional analyses, such
as subgroup analyses and adjusted analyses.
RESULTS
Participant flow Flow of participants through
each stage (a diagram is strongly recommended). Specifically,
for each group, report the numbers of participants randomly
assigned, receiving intended treatment, completing the study
protocol, and analyzed for the primary outcome. Describe protocol
deviations from study as planned, together with reasons. Recruitment
Dates defining the periods of recruitment and follow-up. Baseline
data Baseline demographic and clinical characteristics of
each group. Numbers analysed Number of participants (denominator)
in each group included in each analysis and whether the analysis
was by "intention to treat". State the results in
absolute numbers when feasible (e.g., 10/20, not 50%). Outcomes
and estimation For each primary and secondary outcome, a summary
of results for each group, and the estimated effect size and
its precision (e.g. 95% CI). Ancillary analyses Address multiplicity
by reporting any other analyses performed, including subgroup
analyses and adjusted analyses, indicating those prespecified
and those exploratory. Adverse events All important adverse
events or side effects in each intervention group.
DISCUSSION
Interpretation Interpretation of the results,
taking into account study hypotheses, sources of potential
bias or imprecision and the dangers associated with multiplicity
of analyses and outcomes. Generalisability External validity
of the trial findings. Overall evidence General interpretation
of the results in the context of current evidence. Reference:
Moher, D., Schulz, K.F., Altman, D.G., for the CONSORT Group.
The CONSORT statement: revised recommendations for improving
the quality of reports of parallel-group randomised trials.
Lancet 2001; 357; 1191-94.
Return to Contents
Page here
Chapter Seven
->
parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint parint v
|
