These articles should report original research studies that can improve decision making in medical practice, policy, education, or research and will be understandable by general medical readers.

All research studies published in the BMJ should be morally acceptable, and must follow the World Medical Association's Declaration of Helsinki. To ensure this, we aim to appraise the ethical aspects of any submitted work that involves human participants, whatever descriptive label is given to that work including research, audit, and sometimes debate. This policy also applies on the very rare occasions that we publish work done with animal participants.

To learn more about the kind of research we give priority to, and what services we offer to authors of research, please read this editorial: Why submit your research to the BMJ? You may also want to know more about the BMJ's continuous online publishing model, and read our FAQs about it - for example on how to cite BMJ articles.

We audit the performance of all BMJ research articles, using a wide range of indicators to assess their impact on readers and their dissemination to the wider world.

Open access

The full text of every research article published in the BMJ is immediately accessible on bmj.com through open access to everyone at no charge. The full text of all research articles is also sent, without further intervention from the author, to PubMed Central, the National Library of Medicine's full text archive, which makes it fully accessible without delay. This means that the BMJ immediately fulfils the requirements of the US National Institutes of Health, the UK Medical Research Council, the Wellcome Trust, and other funding bodies to make publicly funded research freely available to all.

Trial registration

In accordance with the ICMJE uniform requirements, the BMJ will not consider reports of clinical trials unless they were registered prospectively before recruitment of any participants. This applies to trials which commenced after 1 July 2005: for older trials retrospective registration will be acceptable, but only if completed before submission of the manuscript to the journal.

Eligible trials have been defined by ICMJE since 1 July 2008 as "where human participants are prospectively assigned to one or more health-related interventions [including health services and behavioural interventions] to evaluate the effects on health outcomes", and before that were defined more narrowly as trials "where human participants are prospectively assigned to investigate the cause and effect relationship between a medical intervention and health outcome".

This means that:

• trials randomising human participants to investigate the cause and effect relationship between a medical intervention and a health outcome that commenced before 1 July 2005 can be registered retrospectively, but this must be done before submission to the BMJ
• trials randomising human participants to investigate the cause and effect relationship between a medical intervention and a health outcome that commenced after 1 July 2005 must have been registered prospectively, ie before enrolment of any participants
• trials randomising human participants or groups of humans to one or more health-related interventions [including health services and behavioural interventions] to evaluate the effects on health outcomes and that commenced after 1 July 2008 must have been registered prospectively, ie before enrolment of any participants

The BMJ’s criteria for a suitable public trial registry are: free to access, searchable, and identifies trials with a unique number; registration is free or has minimal cost; registered information is validated; registered entry includes details to identify the trial and investigator and includes the status of the trial; and the research question, methodology, intervention, funding, and sponsorship must all be disclosed at registration.

The BMJ does not consider posting of protocols and results in clinical trial registries to be prior publication.

BMJ policy on drugs and devices trials

We welcome submission of any drug or device trial that asks an original research question that will sufficiently aid doctors' decisions. This is most likely to be a trial that compares a new drug or device (or new regimen/indication) head to head against the best current treatment(s) using clinically valid doses/administration of both study and comparator interventions. Placebo controlled trials often have much more limited relevance to practice than head to head trials and may not sufficiently help BMJ readers' decisions, but we welcome emailed presubmission inquiries about these too.

In addition, we will give greater priority to a drug or device trial if it:

• has a main outcome measure that's sufficiently clinically relevant and, if it's a composite outcome, matters enough to patients
• has important results: please note that we welcome "negative" trials as long as their research questions are important, new, and relevant to general readers and their designs are appropriate and robust
• is reported fully in line with the CONSORT statement or the relevant CONSORT extension statement, and has sufficient internal and external validity
• is submitted with the original study protocol, for use in confidence during peer review
• is reported transparently as explained in our detailed advice below on reporting industry-sponsored trials
• is a phase III, IIIb, or IV trial. Trials done for "label extension" may be useful to BMJ readers if they ask research questions that are sufficiently new and relevant to practice.

Industry-sponsored trials

If you are submitting a report of such a clinical trial please follow the guidelines on good publication practice and on properly reporting the role of professional medical writers. Please provide the trial registration details; declare the details of all sources of funding for the study; provide statements of competing interests and contributorship; fully describe the role of the study sponsors; provide a statement on the independence of researchers from funders; and state whether all authors had full access to and can take responsibility for the data and analyses. All of these items are explained in more detail below.

Data sharing

We ask authors to include a data sharing statement at the end of each original research article. The statement should explain which additional data from the study—if any—are available, to whom, and how. Those data could range from additional explanatory material to the complete dataset. People allowed access to the data might range from fellow researchers to everyone. And data might be available only on request, accessible online with a password, or openly accessible to all on the web with a link on bmj.com.

We understand that many authors wish to guard data until they have published all their own papers, and we know that data sharing is hard to do. But we hope that authors will, increasingly, set the data free, perhaps after a set period of personal use.

Abridged research articles

For research articles we routinely post the full, citable, indexed version only on bmj.com, with open access, and prepare an abridged version for publication in the print journal. We believe that a well written short version in the print BMJ can encourage casual readers to read something they might otherwise miss, while the full version on bmj.com allows serious readers the detail they need.

We are currently piloting a new way of abridging research articles for the print BMJ - publishing what is essentially an evidence abstract called BMJ pico - and here are some real examples. We hope that you will want to take part in this pilot if your research article is accepted. This would involve you producing, using a template from us, a succinct evidence abstract of your own article after acceptance. There is no need to prepare a BMJ pico in advance, however - please wait until we have offered to publish your article.

We’re doing this to increase readership of research articles in the print BMJ and to give authors more control over the abridging.  If you prefer not to try BMJ pico we will abridge your research article using our ELPS  - electronic long, paper short  - process.

How to prepare BMJ original research articles (full versions)

Here is all the information you need:

No word limit

We do not set fixed limits for the length of BMJ research articles and can be flexible. Nonetheless, please try to make your article concise and make every word count. Think hard about what really needs to be in the paper to get your message across accurately and what can be left out. You will be prompted to provide the word count for the main text (excluding the abstract, references, tables, boxes, or figures) when you submit your manuscript.

Title page

This should give the title of the article, including the study design. Please give for each author his or her name and initials, full address including postal code and one main work position (job title) at the time of writing the paper. We do not need authors’ qualifications. For the corresponding author please provide an email address and the best contact address: this may differ from his or her work address.

Overall style

Original research articles should follow the IMRaD style (introduction, methods, results and discussion) and should have a structured abstract, a structured discussion, and a succinct introduction that focuses  - in no more than three paragraphs - on the background to the research question.

Structured discussion

Please ensure that the discussion section of your article comprises no more than five paragraphs and follows this overall structure, although you do not need to signpost these elements with subheadings:

• statement of principal findings
• strengths and weaknesses of the study
• strengths and weaknesses in relation to other studies, discussing important differences in results
• meaning of the study: possible explanations and implications for clinicians and policymakers
• unanswered questions and future research

What other information do we need?

Please see our general requirements for all BMJ manuscripts. For original research articles in particular, please note that we need, as appropriate:

In the manuscript:

• a competing interest statement -  either a statement describing the interests of all authors or a declaration "All authors declare that the answer to the questions on your competing interest form are all No and therefore have nothing to declare". We no longer need to see a hard copy of the signed form
• details of contributors - giving their names and specific roles - and the name of the guarantor(s) for the study
• a statement that any identifiable patients have provided their signed consent to publication. Please submit, as a supplemental file, the signed BMJ patient consent form giving consent to publication in the BMJ of any information about identifiable individual patients. Publication of any personal information about a patient in the BMJ, for example in a case report or clinical photograph, will normally require the signed consent of the patient
• a statement that the study obtained ethics approval (or a statement that it was not required), including the name of the ethics committee or institutional review board and a statement that participants gave informed consent before taking part
• a statement giving the details of all sources of funding for the study
• description of the role of the study sponsor(s) or funder(s), if any, in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication
• a statement of the independence of researchers from funders 
• a statement that all authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis
• a data sharing statement such as "Data sharing: technical appendix, statistical code, and dataset available from the corresponding author at <email address or url>". If there are no such further data available, please use this wording: "Data sharing: no additional data available"
• trial registration number and name of register for a clinical trial – in the last line of the structured abstract.
• for a clinical guidelines article explicit statements of the quality of evidence and strength of recommendations, preferably using the GRADE system (we encourage but do not insist on this)
• references to any non-standard statistical packages used to analyse the data. There is no need to cite a package as widely used as STATA, but please say if you have used a package that might be unfamiliar to general readers eg RevMan for a systematic review

 As supplemental files:

• the original protocol for a clinical trial or, if the protocol has been published in an open access online journal, its reference and url
• for a randomised controlled trial, the appropriate completed CONSORT checklist showing on which page of your manuscript each checklist item appears, the CONSORT-style structured abstract, and the CONSORT flowchart  (CONSORT has several extension statements, eg for cluster RCTs). To find research reporting guidelines and statements such as CONSORT you may find it easiest to go to the website of the EQUATOR network, where they are all available in one place. Because we aim to improve BMJ papers’ reporting and increase reviewers’ understanding we ask our research authors to follow such reporting guidelines and to complete the appropriate reporting checklist before submission (or before external peer review if not done sooner). We do not, however, use reporting guidelines as critical appraisal tools to evaluate study quality or filter out articles.
• QUOROM checklist and flowchart for a systematic review
• MOOSE checklist and flowchart for a meta-analysis of observational studies
• STARD checklist and flowchart for a study of diagnostic accuracy
• STROBE checklist for an observational study
• for an economic evaluation (noting that we do not usually consider economic evaluations of clinical or health services research studies unless we have already published or are currently considering those primary studies) the BMJ health economics checklist
• original raw data if you think they will help our reviewers (and maybe readers), or if we specifically request them
• video and audio files that will add educational value to your article, for example by explaining the intervention in a trial
• copies of any non-standard questionnaires and assessment schedules used in the research  
• copies of patient information sheets used to obtain informed consent for the study or to comprise or deliver the intervention in a clinical trial
• copies of closely related articles you have published (this is particularly important when details of the study methods are published elsewhere)
• copies of any previous reviewers' reports on this article. We appreciate that authors may have tried other journals before sending their work to the BMJ, and find it helpful if you let us know how you have responded to previous reviewers' comments

In the cover letter:

• details of previous publications from the same study - including in scientific abstracts or partial reports by the media at scientific meetings and in foreign language journals
• details of any previous publication of the same study in electronic form. For example, the BMJ does not consider posting of protocols and results in clinical trials registries to be prior publication, but we would like to know if results have been posted, and where (please provide urls or trial registration details). And we are pleased to consider articles based on longer systematic reviews and meta-analyses published at the Cochrane Library or HTA database
• names and contact details (including email addresses) of suitable peer reviewers; we often find authors' suggestions helpful, though this is optional
• assurance that a study funded or sponsored by industry follows the guidelines on good publication practice. These guidelines aim to ensure that such studies are published in a responsible and ethical manner. The guidelines cover companies’ responsibility to endeavour to publish results of all studies, companies’ relations with investigators, measures to prevent redundant or premature publication, methods to improve trial identification and the role of professional medical writers
• assurance that any article written by a professional medical writer follows the guidelines by the European Medical Writers' Association on the role of professional medical writers. The guidelines emphasise the importance of respecting widely recognised authorship criteria, and in particular of ensuring that all people listed as named authors have full control of the content of articles. The role of professional medical writers must be transparent. Please name any professional medical writer among the list of contributors to any article for the BMJ (not only original research articles), and specify in the formal funding statement for the article who paid the writer. Writers and authors must have access to relevant data while writing articles. Medical writers have professional responsibilities to ensure that the articles they write are scientifically valid and are written in accordance with generally accepted ethical standards.

Structured abstract

Please ensure that the structured abstract is as complete, accurate, and clear as possible—but not unnecessarily long—and has been approved by all authors. We may screen original research articles by reading only the abstract. For randomised controlled trials please provide all the information required for a CONSORT style abstract. Please note the general rules for abstracts in the BMJ:

• should be 250- 300 words long: you may need up to 400 words, however, for a CONSORT abstract
• use active voice but avoid “we did” or “we found”
• numbers over 10 do not need spelling out at the start of sentences
• sentences starting with a number do not require a capital letter
• p values should always be accompanied by supporting data and denominators should be given for percentages
• abstracts do not need references

The first few items (objective, design, setting) may be note-like and need not form full sentences. The results and conclusions sections should be written properly. Do not mix notes and full sentences in one section.

If the standard headings do not suit the type of study, substitute something sensible such as "population" as a heading instead of "participants" in an economics article. Please do not simply delete the heading.

For standard original research articles please provide the following headings and information (for RCTs please add the trial registration details - but there is no need to provide the additional subheadings which are used in the CONSORT statement on abstracts, as long as you include all the required information):

• objectives - a clear statement of the main aim of the study and the major hypothesis tested or research question posed 
• design - including factors such as prospective, randomisation, blinding, placebo control, case control, crossover, criterion standards for diagnostic tests etc 
• setting - include the level of care eg primary, secondary; number of participating centres. Be general rather than give the name of the specific centre, but give the geographical location if this is important
• participants (instead of patients or subjects) - numbers entering and completing the study, sex, and ethnic group if appropriate. Give clear definitions of how selected, entry and exclusion criteria
• interventions - what, how, when and for how long. This heading can be deleted if there were no interventions but should normally be included for randomised controlled trials, cross over trials, and before and after studies. 
• main outcome measures - those planned in protocol, those finally measured (if different, explain why) 
• results - main results with (for quantitative studies) 95% confidence intervals and, where appropriate, the exact level of statistical significance and the number need to treat/harm. 
• conclusions – primary conclusions and their implications, suggesting areas for further research if appropriate. Do not go beyond the data in the article. Conclusions are important because this is often the only part that readers look at.
• trial registration - registry and number (only for clinical trials)

Please note that confidence intervals should be written in the format (15 to 27) within parentheses, using the word "to" rather than a hyphen.

Abstracts for meta-analyses and systematic reviews should have these headings:

• objective – what the review set out to determine
• design – type of meta-analysis, systematic review
• data sources - where included studies were retrieved from
• review methods - inclusion and exclusion criteria
• results - main findings with 95% confidence intervals
• conclusions -  primary conclusions and their implications

Abstracts for qualitative research articles should follow the standard style but may need fewer headings:

• objective      
• design
• participants
• setting
• results
• conclusions             

Quality improvement reports also have their own style of structured abstract:

• problem
• design
• setting
• key measures for improvement
• strategies for change
• effects of change
• lessons learnt

“What this paper adds” box

Please produce a box offering a thumbnail sketch of what your article adds to the literature, for readers who would like an overview without reading the whole article It should be divided into two short sections, each with 1-3 short sentences.

section 1:  What is already known on this subject
In two or three single sentence bullet points please summarise the state of scientific knowledge on this subject before you did your study and why this study needed to be done. Be clear and specific, not vague.

For example you might say: “Numerous observational studies have suggested that tea drinking may be effective in treating depression, but until now evidence from randomised controlled trials has been lacking/the only randomised controlled trial to date was underpowered/was carried out in an unusual population/did not use internationally accepted outcome measures/used too low a dose of tea.”

or: “Evidence from trials of tea therapy in depression have given conflicting results. Although Sjogren and Smith conducted a systematic review in 1995, a further 15 trials have been carried out since then…”

section 2:  What this study adds
In one or two single sentence bullet points give a simple answer to the question “What do we now know as a result of this study that we did not know before?” Be brief, succinct, specific, and accurate. For example: “Our study suggests that tea drinking has no overall benefit in depression”.

You might use the last sentence to summarise any implications for practice, research, policy, or public health. For example, your study might have: asked and answered a new question (one whose relevance has only recently become clear) contradicted a belief, dogma, or previous evidence provided a new perspective on something that is already known in general provided evidence of higher methodological quality for a message which is already known.

Summary statistics to clarify your message

We do want your piece to be easy to read but also want it to be as scientifically accurate as possible. Please include in the results section of your structured abstract (and in the article's results section) the following terms, as appropriate:

 For a clinical trial:

• Absolute event rates among experimental and control groups
• RRR (relative risk reduction)
• NNT or NNH (number needed to treat or harm) and its 95% confidence interval (or, if the trial is of a public health intervention, number helped per 1000 or 100,000)

 For a cohort study:

• Absolute event rates over time (eg 10 years) among exposed and non-exposed groups
• RRR (relative risk reduction)

 For a case control study:

• OR (odds ratio) for strength of association between exposure and outcome

 For a study of a diagnostic test:

• Sensitivity and specificity
• PPV and NPV (positive and negative predictive values)

The box stating what is known and what this paper adds (see below) should also reflect accurately the above information. Under what this paper adds please give the one most useful summary statistic eg NNT.

Please do not use the term "negative" to describe studies that have not found statistically significant differences, perhaps because they were too small. There will always be some uncertainty, and we hope you will be as explicit as possible in reporting what you have found in your study. Using wording such as "our results are compatible with a decrease of this much or an increase of this much" or “this study found no effect” is more accurate and helpful to readers than “there was no effect/no difference”.  Please use such wording throughout the article, including the structured abstract, and the box stating what the paper adds.

If you are sending us a revised article

Please provide all of the above, as appropriate (if not done earlier), as well as a detailed covering letter explaining how you have responded to editorial and peer review comments and other guidance from the BMJ. All of this should be submitted via your author area at our online editorial office.

Commentaries on research

We often publish original research articles with an accompanying commentary of up to 500 words and five references, commissioned to help readers interpret the research or place it in context. If we commission a commentary on your article we will send you a copy of it before publication.

If we ask you to write a commentary, please provide in the manuscript a title for your piece; a title page giving your name, position, and contact details including email address; and statements of competing interests and – if appropriate -  contributorship and funding. Please say in your covering letter or email which BMJ article you are commenting on and give its BMJ manuscript number.