European Respiratory Journal

European Respiratory Journal Instructions for Authors

The European Respiratory Journal (ERJ), the official journal of the European Respiratory Society (ERS), publishes 12 issues per year in English.

The ERJ publishes clinical and experimental work dealing with the whole field of respiratory medicine, including cell biology, epidemiology, immunology, pathophysiology, thoracic imaging, paediatric pneumology, occupational medicine, intensive care, sleep medicine, thoracic surgery and thoracic oncology. In addition to such original material, the ERJ prints Editorials, Reviews, Technical Notes, Correspondence and Letters to the Editor.

The ERJ takes part in the HINARI Programme. The HINARI Programme, set up by the World Health Organization (WHO) together with major publishers, enables developing countries to gain access to collections of biomedical and health literature. Over 6,400 journal titles are now available to health institutions in 108 countries, areas and territories benefiting many thousands of health workers and researchers, and in turn, contributing to improved world health. For more information about HINARI, go to www.who.int/hinari/en/

Overlapping publications
Authorship
Collaborators
Copyright
Peer review
Statement of interest
Registering clinical trials
Guidelines on reporting research findings
Permissions and the reproduction of figures
Use and reproduction of questionnaires
Manuscript preparation
Online depository
Letters
Correspondence
CME
ERJ In press
Proofs and reprints
Statistical notes
Sources of information for statistical analyses, study design and data presentation

Overlapping publications

Authors submitting a paper do so on the understanding that neither the work nor any part of its essential substance, tables or figures have been or will be published or submitted to another scientific journal or is being considered for publication elsewhere. This must be stated in the covering letter. This restriction does not apply to abstracts, but includes work published in any other language.

It is the authors' responsibility to ensure that submitted manuscripts are not duplicate publications as specified by the von Elm criteria (JAMA 2004; 291: 974-980). Therefore, authors must declare any simultaneous submissions of similar or related manuscripts and include electronic copies with the submission. If there are any existing concerns following submission, the Editors reserve the right to take appropriate action.

Authorship

There is no maximum for the number of authors of a manuscript, although all work must have been approved by all co-authors.

Authorship credit should be based on the following criteria: 1) substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data; 2) drafting the article or revising it critically for intellectual content; and 3) final approval of the version to be published.

Writing assistance from medical writers or agencies must be included in the Acknowledgements section, or should be given appropriate credit as authorship.

Multicentre groups

When a multicentre group has conducted the study, all individuals who accept direct responsibility for the manuscript should be identified. When submitting a group author manuscript, the corresponding author should clearly identify all individual authors, as well as the group name.

Acquisition of funding, collection of data, or general supervision of the research group does not justify authorship. All persons designated as authors should qualify for authorship, and all those who qualify should be listed. Each author should have participated sufficiently to be able to take public responsibility for appropriate portions of the content.

Collaborators

Contributors who do not meet the criteria for authorship stated above should be listed in an acknowledgements section, by stating their names and affiliation details. Examples of those who might be acknowledged include: groups of persons/collaborating bodies who contributed materially to the paper but whose contributions do not justify authorship; a person who provided purely technical help or writing assistance; a department leader who provided only general support.

If the study was performed in collaboration with or on behalf of a particular body (for example, ISAAC), the names and affiliation details of members of that body or group should be listed in an acknowledgements section.

Copyright

Submission of the manuscript implies that if and when it is accepted for publication, the authors automatically agree to transfer the copyright to the publisher. The copyright protection implies that the publisher holds the exclusive right to reproduction in any form (including publication in another language) and distribution of any of the articles in the journal. Material published in the ERJ may be stored on microfilm or in any electronic format, or be reproduced photographically, only with prior written permission of the publisher.

One copyright form should be completed by the corresponding author on submission of any manuscript to the ERJ, and by all authors when a paper is accepted.

For manuscripts with NIH funding, the Journal acknowledges that the author retains the right to provide a copy of the final, peer-reviewed author-supplied manuscript (before copy-editing and publication) to the NIH upon acceptance for Journal publication, for public archiving in PubMed Central 12 months from publication in a print issue.

If any of the authors of a manuscript have been supported by the Wellcome Trust or other member of the United Kingdom PMC Funders Group, the Journal acknowledges that the author retains the right to provide a copy of the final, peer-reviewed author-supplied manuscript for public archiving on PubMed Central in compliance with requirements.

Peer review

Each manuscript will be reviewed by an Associate Editor and Reviewers. You will be notified by e-mail on receipt of the submitted online draft and provided with a manuscript number for future reference.

You may be requested to make changes to your manuscript arising from reviewer comments. When submitting your revised manuscript, all text added as a result of reviewer comments should be highlighted in red.

Statement of interest

Any conflict of interest for a given manuscript and for all authors of a manuscript must be dealt with according to the statement of the International Committee of Medical Journal Editors (the 'Vancouver Group') as published in Lancet 1993; 341: 742. The Editors and Reviewers of the ERJ must disclose to the Chief Editors any personal or financial relationship that could bias their opinion and decision in the peer-review process.

In the interests of transparency, all authors must complete the Statement of interest form and these must be uploaded during electronic manuscript submission. The authors have to acknowledge all financial support for the work and other financial or personal conflicts of interest that are both connected and unconnected with the work. It is each author's responsibility to fill in this form correctly. Each author must complete their own statement of interest form. Corresponding authors should not complete statement of interest forms on behalf of other authors; however, corresponding authors can upload statement of interest forms for other authors once they have been completed. The ERJ and the Editors reserve the right to take appropriate action if these forms have not been filled in correctly. A statement will then be added to the manuscript at publication.

The ERJ and its Chief Editors adhere to the European Respiratory Society's policy on tobacco industry funding (www.ersnet.org/images/stories/pdf/ERS_by-laws.pdf). Specifically, this means that any author who has current funding from the tobacco industry, or has any perceived links with the tobacco industry (such as holding shares, or speaking at or attending meetings organised by the tobacco industry) will not be allowed to submit and publish in the ERJ. In addition, studies funded in part or wholly by tobacco industry funding will not be accepted for publication.

If an author declares a current (i.e. within the past year) tobacco industry conflict of interest, there will be a 12-month ban before he/she is allowed to submit and publish in the ERJ. In the event that an undeclared conflict is discovered, there will be a 5-year ban before the author is allowed to submit and publish in the ERJ.

In the interests of transparency, it is advised that authors also declare any previous tobacco industry funding on their Statement of Interest form, specifying the duration of funding.

Registering clinical trials

All clinical trials started after January 1, 2006 must be registered. The ICMJE defines a clinical trial as any research project that prospectively assigns human subjects to intervention or concurrent comparison or control groups to study the cause-and-effect relationship between a medical intervention and a health outcome. Medical interventions include drugs, surgical procedures, devices, behavioural treatments, process-of-care changes and the like.

In addition to accepting registration in any of the five existing registries (for example, ClinicalTrials.gov), the ICMJE will now also accept registration of clinical trials in any of the primary registers that participate in the World Health Organization's (WHO) International Clinical Trial Registry Platform (ICTRP). It should be noted that registration in a partner register only will not suffice.

The ICMJE is expanding the definition of the types of trials that must be registered and will begin to implement the WHO definition of clinical trials for all trials that begin enrolment on or after July 1, 2008. The WHO's definition of clinical trials is: "any research study that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcomes".

The ICMJE does not consider results posted in the same clinical trials registry in which the primary registration resides to be previous publication if the results are presented in the form of a brief (≤500 words) structured abstract or table. When submitting to the ERJ, authors should specify where the clinical trial is registered and disclose all posting in registries of results of the same or closely related work.

For further details on current ICMJE policy, see ICMJE or read N Engl J Med 2007; 356: 2734-2736.

Guidelines on reporting research findings

Randomised controlled trials must conform to the CONSORT statement, which provides a set of recommendations comprising a list of items to report and a patient flow diagram.

For other study designs, authors are strongly recommended to consult the following reporting guidelines: studies of diagnostic accuracy (STARD); preferred reporting items for systematic reviews and meta-analyses (PRISMA); observational studies in epidemiology (STROBE); genetic association studies (STREGA); and meta-analyses of observational studies in epidemiology (MOOSE).

The SPIRIT (Standard Protocol Items for Randomised Trials) are currently under development, and guidelines for reporting biomedical images are now available. For more information, go to: http://www.equator-network.org/resource-centre/library-of-health-research-reporting/reporting-guidelines-under-development/

Permissions and reproduction of figures

The ERJ discourages the use of previously published figures and tables, or other material previously published elsewhere, unless absolutely essential. If such figures, tables, etc. must be used, you must obtain permission from the copyright owner, which is usually the publisher and not the original author. Use of the material in question may involve a fee payable to the original publisher. Please note that some publishers will not provide permission for publication, which precludes the material from being displayed in both the printed and the online version of the ERJ.

Once your manuscript has been accepted for publication, the ERJ publications team will contact you requesting that all written confirmation details are forwarded to them for their records.

Use and reproduction of questionnaires

Authors should be aware that such externally produced questionnaires are often copyrighted. For example, the Asthma Control Questionnaire (ACQ) is copyrighted, meaning its use, reproduction and modification is restricted. Before using, modifying or reproducing questionnaires such as the ACQ, the author should contact the copyright holder (the individual/group credited with creation of the questionnaire) to seek their support and gain written permission for reproduction and/or modification of the questionnaire as necessary. When the study is accepted for publication in the ERJ, the publications team will contact the author requesting that all written confirmation and permission details are forwarded to the publications office for their records.

Where authors are unsure about usage, modification and reproduction of questionnaires, they should contact the copyright holder/creator of the questionnaire in the first instance. Should concerns arise regarding the use, reproduction and modification of questionnaires, the ERJ reserves the right to take appropriate action.

Manuscript preparation

Presentation of manuscripts should conform to the updated Uniform Requirements for Manuscripts Submitted to Biomedical Journals (see icmje).

All manuscripts must be submitted electronically using the online submission at: manuscriptcentral, from now on referred to as ERJ Manuscript Central. Detailed instructions of how to submit are available on the website itself and the process is self explanatory. However, if you do experience problems, please contact the Submission Helpline direct on +44 114 2672864 or contact Gill.Archer{at}ersj.org.uk, the manuscript central coordinator.

Before entering the ERJ Manuscript Central online submission area, please read and carefully follow the instructions below.

For further guidance on how to write papers, please see: Sterk PJ, Rabe KF. The joy of writing a paper.Breathe 2008; 4: 224-232, and guidelines for authors on how to write scientific articles to be published in English at http://www.ease.org.uk/pdfguidelines/AuthorGuidelinesHighRes.pdf

General

• The manuscript file you submit must be saved as .rtf (rich text format) or .doc (MS Word document).
• Original articles should not exceed 3,000 words, not including abstract, references, tables and legends. However, if manuscripts do exceed this limit, please state the final word count and explicit reasons for exceeding the limit in your covering letter.
• Figures and/or tables should be limited to no more than eight altogether. Large figures with a high number of parts should be avoided owing to space limitations. Large figures can be presented in the online depository.
• Abbreviations or unusual terms should be described at the first time of use.
• Symbols as defined by the ad hoc working group of the Commission of the European Communities (see Eur Respir J 1993; 6: Suppl. 16) are recommended.
• Système International (SI) units are recommended.
• Equations should be created as normal text.

Title page

• Please provide a concise and informative title, limited to 90 characters, including spaces between words.
• Include a list of all contributing authors and all of their affiliations, with a clear indication of who is associated with each institution.
• Supply the full correspondence details for the corresponding author, including e-mail address. Please note that only one corresponding author per manuscript should be provided.

Abstract

• Please provide an abstract of 200 words or fewer, which is easily understood without reference to the text (see Ann Intern Med 1987; 106: 598-604).
• The abstract must have four separate paragraphs, which correspond to the question of the study, materials/patients and methods, results, and the answer to the question. One or two sentences of background information can be included before the question if necessary. The question and answer should be the same as those in the text.
• Include only a few important values.
• Avoid using abbreviations.

Keywords

• Please provide a list of 6 keywords or fewer.
• The keywords should be listed alphabetically.
• The keywords should be listed in full without abbreviations.

Introduction

• State the question you asked (or hypothesis to be tested) and your considerations leading to the formulation of the question.
• Give only pertinent references.

Material and methods

Study subjects or animals

• Clearly describe how the subjects or experimental animals were identified, including the control subjects when used. For animals, see Laboratory Animals 1985; 19: 106-108.
• Clearly state the eligibility criteria for cases and controls in observational studies, or for subjects in clinical trials.
• All work involving studies on human subjects is expected to have received approval from local Ethical Committees and the regulatory authority (when appropriate, for example, drug trials).
• Animal experimentation must be performed according to the Helsinki convention for the use and care of animals.
• Provide details of the species and/or strain and number of animals involved in the study.
• The Editors reserve the right to refuse work which does not conform to acceptable ethical criteria.

Study design

• Clearly state the main study objective(s).
• Provide an overview of the main tests or experiments.
• Consider sample size and whether you have enough subjects to reliably address the research question.
• Papers on clinical trials should include details of the sample size calculation (i.e. the expected effect size, power, level of statistical significance and one- or two-sided test). The sample size should be reproduced independently.

Methods

• Describe the methods and apparatus in sufficient detail to allow other workers to evaluate or reproduce the tests/experiments.
• For methods that have been published before, provide only a reference or a reference and a brief description.
• Identify drugs and chemicals, including generic name, dosage and route of administration.
• Please provide manufacturer and manufacturer's address for equipment, drugs and chemicals, as necessary, but not in a separate section.
• For systematic reviews, make sure that the keywords used to search electronic medical databases cover different terminology (for example, tumour or cancer) and spelling (for example, randomised or randomized).

Analysis

• Clearly state and define the main outcome measure(s).
• Briefly state the statistical methods used during the analysis if they are standard. New methods should be described with justification.
• In the case of single- or multicentre trials with blinded intervention, the code must have been broken at the end of the study in the presence of the responsible investigator of each centre. The code and the data will then be available to each participating centre. The first author should make provisions so that if needed, the data are available to the ERJ for independent statistical analysis.
• See the Statistical Notes below for further information on analysis, presentation and interpretation

Results

• Keep the Results section brief.
• Describe the baseline characteristics or condition of patients or animals.
• Focus on the important results, i.e. those that help address the research question.
• Present most data in figures or tables, not in the text. In the text, emphasise or summarise the most important observations.

Discussion

• At the beginning of the Discussion, summarise the main results, and show how they have addressed the research question.
• Make sure that the conclusions are consistent with the results and are pertinent to the research question.
• Describe the limitations of the study and/or analysis, and discuss the possible implications on the conclusions.
• Emphasise the new and important aspects of the study.
• Try to explain contradictory or unexpected results, or discrepancies with previous findings.

Acknowledgements

• All acknowledgements should be grouped into one paragraph placed after the Discussion.
• Only acknowledge persons who have made substantial contributions to the study.
• Provide the names and affiliation details of members of collaborating bodies

References

• Number references consecutively in the order in which they first appear in the text, using full size Arabic numerals in square brackets.
• All authors must be included in the reference list.
• For original articles, the number of references should be limited to 30.
• References should conform to the style used in Index Medicus (Vancouver Style) as shown in the following examples:

1. Bannerjee D, Khair OA, Honeybourne D. Impact of sputum bacteria on airway inflammation and health status in clinical stable COPD. Eur Respir J 2004; 23: 685-692.

2. Bourbon J, Henrion-Caude A, Gaultier C. Molecular basis of lung development. In: Gibson GJ, Geddes DM, Costable U, Sterk PJ, Corrin B, eds. Respiratory Medicine. 3rd Edn. Elsevier Science, Edinburgh/Philadelphia, 2002; pp. 64-81.

• Websites should be listed in the reference list, not in the text, and only used when an original citation is unavailable; citations should be listed as follows:

WHO. Severe Acute Respiratory Syndrome (SARS). www.who.int/csr/sars/en/index.html. Date last updated: June 1 2004. Date last accessed: June 1 2004.

• Work which has not yet been accepted for publication and personal communications should not appear in the reference list.
• One copy of papers cited as 'in press' should be uploaded onto manuscript central as supplementary material.

Tables

• Tables should be created and inserted into the text document using the 'Table', 'Insert Table'; function in your word processing package; DO NOT supply tables in a separate file.
• Tables should be numbered consecutively with Arabic numerals.
• Limit decimals to a sensible number.
• Large tables should be avoided due to space restrictions, and if used may be split.
• Please provide a clear footnote for all tables, making sure ALL abbreviations and symbols used are defined.
• Any references in tables should run on in numerical order from the text where the table is cited.
• The ERJ discourages the use of previously published tables unless absolutely essential. If such tables must be used, you must obtain permission from the copyright owner, which is usually the publisher and not the original author. Use of the table may involve a fee payable to the original publisher. Please note that some publishers will not provide permission for publication, which precludes these tables from being displayed in both the printed and the online version of the ERJ. Once your manuscript has been accepted for publication, the ERJ publications team will contact you requesting that all written confirmation details are forwarded to them for their records.

Figures

• Figures constitute a key element of manuscripts submitted to the ERJ. However, figures should be limited to those required to show the essential features described in the manuscript. Large figures with a high number of parts should also be avoided. Redundant or excessive figures will not be published due to space restrictions. Figures such as these can be presented in the online depository.
• All submitted figures must be clearly named and numbered, giving the figure number both as part of the filename and on the figure itself. If labelling is necessary, see a recent issue of the ERJ for the degree of labelling required.
• There is a charge for colour figures; however, a discounted rate is available to Members. Please contact the Publications Office for further details.
• Please attempt to keep each figure under 1 megabyte (MB). If it is over 1 MB, the upload time will be considerably increased; likewise, the download time could be significantly increased.
• Whether for images, drawings, or graphs, we recommend that no more than four components be used for a given figure (e.g. a, b, c, d).
• Images embedded in the .doc or .rtf file WILL NOT be accepted.
• The ERJ discourages the use of previously published figures unless absolutely essential. If such figures must be used, you must obtain permission from the copyright owner, which is usually the publisher and not the original author. Use of the figure may involve a fee payable to the original publisher. Please note that some publishers will not provide permission for publication, which precludes these figures from being displayed in both the printed and the online version of the ERJ. Once your manuscript has been accepted for publication, the ERJ publications team will contact you requesting that all written confirmation details are forwarded to them for their records.

Photographic images

• All photographic images should be provided with a minimum of 300 dots per inch (dpi).
• Figures should preferably be supplied in .jpeg, .tif or .eps format.
• The image mode for colour photographic images should be cyan-magenta-yellow-black (CMYK). For black-and-white photographic images, it should be grey scale.
• An image-editing program such as Adobe Photoshop is recommended for saving images.
• Due to space restrictions, it is essential that each submitted figure show only the areas of interest with enough surrounding area for orientation purposes.
• Radiographic images should be of high quality and combined into one array, such as posteroanterior and lateral views. They should also be sized the same to facilitate reproduction.
• Image size should be submitted as close as possible to print size; one column width is 88 mm, two-column width is 182 mm.
• When several images of a given type are being shown, please reproduce each specific type at the same magnification.
• Photomicrographs must have internal scale markers (linear scale), since the size and magnification may be altered by the publisher.
• Images should correspond in appearance to the tonal relations of the original radiograph (i.e. showing the bones white on a dark background, with the patient's right to the observer's left. CT scans and MR images should employ the internationally-accepted 'view from below'.
• Please label your images such that all important details are clearly visible.
• The ERJ discourages the use of previously published photographic images unless absolutely essential. If such photographic images must be used, you must obtain permission from the copyright owner, which is usually the publisher and not the original author. Use of the photographic image may involve a fee payable to the original publisher. Please note that some publishers will not provide permission for publication, which precludes these photographic images from being displayed in both the printed and the online version of the ERJ. Once your manuscript has been accepted for publication, the ERJ publications team will contact you requesting that all written confirmation details are forwarded to them for their records.

Guidelines for handling image data

• If an image has been enhanced electronically, please explain the alterations that have been made and send in the original image along with the enhanced one. Moreover, keep an electronic set of original images, since our Reviewers might ask you to modify their content and the display modus.
• The Council of Science Editors (CSE) has established four basic guidelines for handling image data, which authors submitting to the ERJ are urged to comply with. 1) No specific feature within an image may be enhanced, obscured, removed or introduced. 2) Adjustments of brightness, contrast or colour balance are acceptable if they are applied to the whole image and as long as they do not obscure, eliminate or misrepresent any information present in the original. 3) The grouping of images from different parts of the same gel, or from different gels, fields or exposures must be made explicit by the arrangement of the figure (e.g. dividing lines) and in the text of the figure legend. 4) If the original data cannot be produced by an author when asked to provide it, the acceptance of the manuscript may be revoked.

Legends

• Please provide a clear legend for each figure.
• Legends should be brief and nonrepetitive of information given in the text.
• They should include the imaging technique used, the body part imaged and any noteworthy details.
• ALL abbreviations should be expanded.
• Use of internal scales should always be mentioned in the legend.

Online depository

The ERJ has an online depository, which can be used to provide more detailed methodology (which does not need to be included in the Materials and Methods section of the paper), supplementary data or figures, and accompanying videos. This is an optional function and can be used at the discretion of the author and/or Editor. Documents should be uploaded as "Supplementary material".

Videos to be presented as part of the online depository should be supplied in one of the following formats: Quicktime; MPEG; Microsoft AVI; Windows media video; Shockwave Flash.

Authors should note that their supplementary material will not be edited by the Editorial Office, and will be put online as it has been supplied.

Letters

Letters are case study articles, preliminary studies or short reports presented in the format of a Letter to the Editors. Please note that Letters differ from Correspondence articles. More on Correspondence articles can be found below in the appropriate section below.

Letters should provide new knowledge on aetiology, mechanism, diagnosis or treatment of a disease.

Letters should not exceed 1,500 words and should have no more than 10 references. One figure or one table can also be included in the Letter.

It is permissible to include more figures and tables, but the word limit of the Letter must be reduced by 300 words for each additional figure or table included. For example, a Letter with two figures (or two tables, or one figure and one table) should be no more than 1,200 words; a Letter with three figures (or three tables, or one figure and two tables, or two figures and one table) should be no more than 900 words.

Correspondence

The Chief Editors welcome the submission of Correspondence articles, which refer to/discuss previously published ERJ articles. Correspondence articles should be limited to 800 words and should only include 5 references.

ERJ In press

Many of the original articles that have been accepted are published online before they have been selected to appear in a printed issue of the ERJ. None of the additional editorial preparation, which includes copy-editing, typesetting and proofreading, has been performed at this stage. ERJ In press provides authors and readers with immediate, subscription-based access to the newest research.

CME

Accepted manuscripts deemed to be of educational value are submitted to EBAP (the European Board for Accreditation in Pneumology) for accreditation as Continuing Medical Education (CME) articles. ERJ readers then have the chance to earn CME credits by reading such articles and correctly answering a selection of multiple choice questions about the manuscript in question.

If your article is deemed to be of educational value, upon acceptance for publication, you may be asked to provide 4 to 6 multiple-choice questions to be published with the article. Prompt composition of such questions will enable the timely publication of the article and the ERJ.

From January 1, 2010, CME credits will be free of charge. For more information, visit dev.ersnet.org/900-cme-credits.htm.

Proofs and reprints

A proof will be sent by e-mail (pdf file) or fax to the corresponding author.

It should be corrected and returned to the ERJ Publications office within 48 hours by fax or email. Late return will delay publication. Modification to proofs should be limited to typographical errors only.

Statistical notes

Seek help or advice from a statistician on the appropriate methods of analyses and whether the results have been interpreted correctly.

Study design

• By clearly describing how the subjects were selected you could see whether there are any factors that could distort the results.
• Where appropriate (particularly observational studies), give reasons why many eligible patients were not included in the study.
• Meta-analyses of only a few small randomised trials, for example, 3 or 4, could be affected by publication bias, and the statistical power to detect heterogeneity is likely to be low. It might therefore be worthwhile checking whether there are large observational studies that could provide additional information.

Displaying data

• Many studies are based on relatively few subjects, for example, less than 50. A scatter plot is a good way of showing the data.
• When examining associations between two continuous variables, present scatter plots, possibly showing a regression line or curve.
• Bar charts are useful when displaying comparisons of percentages (or proportions) between several groups.

Analysis

• Check whether the variables are normally distributed (e.g. using a probability plot). If they are, the statistical methods are usually much easier to perform.
• When analysing data, be clear about what the outcome measure involves:
  • Counting people or objects (binary or categorical data).
  • Taking measurements on people or objects (continuous data).
  • Time-to-event data (for example, survival data).
• When making comparisons between groups, provide estimates of effect sizes and the corresponding confidence interval and p-value (table 1).
  
  Table 1. Effect sizes and common statistical tests that produce p-values
  

  	Counting people or objects	Taking measurements on people or objects	Time-to-event data
  Effect size	Relative risk Risk difference Percentage change in risk	Difference between 2 means Difference between 2 medians	Hazard ratio Risk difference at a specific time point
  Two separate groups, 1 measurement (unpaired data)	Chi-square test	Unpaired or two-sample t-test if the difference between the means is normally distributed#	Log rank test
  		Mann-Whitney test if distribution of the difference is skewed¶
  One group, 2 repeated measurements (paired data)	McNemar's test	Paired t-test if the difference is normally distributed	Not applicable
  		Wilcoxon matched pairs test if the distribution of the difference is skewed
  Allow for other factors	Multivariate logistic regression	Multivariate linear regression+	Cox regression

  ^#: with more than two arms in the trial, the test is ANOVA; ^¶: with more than two arms in the trial the test is Kruskal-Wallis ANOVA; ⁺: outcome measure is approximately normally distributed.

• All main effect sizes should be reported whether statistically significant or not, including confidence intervals and p-values.
• Know which statistical method to use and when.
• When many variables (or associations) are examined in the same data set (multiple testing), researchers are more likely to find a spurious effect or association. The authors should consider adjusting p-values for multiple comparisons or provide 99% confidence intervals.
• When reporting statistical significance, do not present "NS" (i.e. not statistically significant), "p>0.05" or "p≤0.05". Show p-values as <0.001 or to 2 or 3 decimal places otherwise.
• Be clear about repeated measures data. A study of 10 repeated measurements of a variable on 10 subjects is not the same as 1 measurement on 100 subjects; there are still only 10 subjects. Repeated measurements are likely to be correlated with each other, and this needs to be taken into account in the analysis.
• Subgroup analyses tend not to produce reliable results if there are (a) too few patients or (b) too few events.
• For screening or diagnostic studies use "detection rate" (sensitivity) and "false-positive rate" (1 minus specificity). The word sensitivity can sometimes be confused with the same term in laboratory studies. Using specificity to compare two tests can mask an important difference. For example, specificities of 96% versus 98% both look high. However, the corresponding false-positive rates are 4% versus 2% (one is twice as large as the other).
• For screening or diagnostic tests, if the same data set is used both to derive the model and to test the model (i.e. estimate performance), the performance of the test is often over-estimated. Authors should acknowledge this in the Discussion, and recommend independent studies to confirm the findings.

Interpretation

• If a p-value is <0.05, it does NOT mean that there is no effect or association. Furthermore, a p-value of 0.06 can have a different conclusion than a p-value of 0.75. If a result is "not statistically significant", according to the conventional cut-off of 0.05, the possible reasons are:
  • There really is no effect.
  • There is a real effect, but by chance the sample of patients did not show this.
  • There is a real effect, but the study was too small to detect it.
  
  If a clinically important effect is observed but the p-value is, say, 0.06, it is likely that the study was not large enough. Such data are suggestive of an effect.
• Where studies are hypothesis-generating or based on a small sample size, authors should make it clear that any conclusions should be tested in an independent study. The conclusions should not be strong.

Clinical trials and observational studies

• When designing and reporting trials, consider how to write them up.
• See the section Guidelines on reporting research findings for links to published guidelines associated with clinical studies.
• These guidelines provide an idea of how to structure your report and what to include.

Sources of information for statistical analyses, study design and data presentation

The Statistical Advisors of the ERJ recommend the following series of papers on the use of statistics, in particular those articles on repeatability, reproducibility, regression and correlation analyses and confidence intervals, and comparisons of means and proportions.

1. Bland JM, Altman DG. Correlation, regression and repeated data. BMJ 1994; 308: 896.
2. Bland JM, Altman DG. Regression towards the mean. BMJ 1994; 308: 1499.
3. Altman DG, Bland JM. Diagnostic tests 1: sensitivity and specificity. BMJ 1994; 308: 1552.
4. Altman DG, Bland JM. Diagnostic tests 2: predictive values. BMJ 1994; 309: 102.
5. Altman DG, Bland JM. Diagnostic tests 3: receiver operating characteristic plots. BMJ 1994; 309: 188.
6. Bland JM, Altman DG. One- and two-sided tests of significance. BMJ 1994; 309: 248.
7. Bland JM, Altman DG. Some examples of regression towards the mean. BMJ 1994; 309: 780.
8. Altman DG, Bland JM. Quartiles, quintliles, centiles, and other quantiles. BMJ 1994; 309: 996.
9. Bland JM, Altman DG. Matching. BMJ 1994; 309: 1128.
10. Altman DG, Bland JM. The normal distribution. BMJ 1995; 310: 298.
11. Bland JM, Altman DG. Calculating correlation coefficients with repeated observations: Part 1, correlation within subjects. BMJ 1995; 310: 446.
12. Bland JM, Altman DG. Calculating correlation coefficients with repeated observations: Part 2, correlation between subjects. BMJ 1995; 310: 633.
13. Altman DG, Bland JM. Absence of evidence is not evidence of absence. BMJ 1995; 311: 485.
14. Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. BMJ 1995; 310: 170.
15. Altman DG, Bland JM. Presentation of numerical data. BMJ 1996; 312: 572.
16. Bland JM, Altman DG. Logarithms. BMJ 1996; 312: 700.
17. Bland JM, Altman DG. Transforming data. BMJ 1996; 312: 770.
18. Bland JM, Altman DG. Transformations, means and confidence intervals. BMJ 1996; 312: 1079.
19. Bland JM, Altman DG. The use of transformations when comparing two means. BMJ 1996; 312: 1153.
20. Altman DG, Bland JM. Comparing several groups using analysis of variance. BMJ 1996; 312: 1472-1473.
21. Bland JM, Altman DG. Measurement error. BMJ 1996; 313: 744
22. Bland JM, Altman DG. Measurement error and correlation coefficients. BMJ 1996; 313: 41-42.
23. Bland JM, Altman DG. Measurement error proportional to the mean. BMJ 1996; 313: 106.
24. Altman DG, Matthews JNS. Interaction 1: Heterogeneity of effects. BMJ 1996; 313: 486.
25. Matthews JNS, Altman DG. Interaction 2: compare effect sizes not P values. BMJ 1996; 313: 808.
26. Matthews JNS, Altman DG. Interaction 3: How to examine heterogeneity. BMJ 1996; 313: 862.
27. Altman DG, Bland JM. Detecting skewness from summary information. BMJ 1996; 313: 1200.
28. Bland JM, Altman DG. Cronbach's alpha. BMJ 1997; 314: 572.
29. Altman DG, Bland JM. Units of analysis. BMJ 1997; 314: 1874.
30. Bland JM, Kerry SM. Trials randomised in clusters. BMJ 1997; 315: 600.
31. Kerry SM, Bland JM. Analysis of a trial randomised in clusters. BMJ 1998; 316: 54.
32. Bland JM, Kerry SM. Weighted comparison of means. BMJ 1998; 316: 129.
33. Kerry SM, Bland JM. Sample size in cluster randomisation. BMJ 1998; 316: 549.
34. Kerry SM, Bland JM. The intra-cluster correlation coefficient in cluster randomisation. BMJ 1998; 316: 1455.
35. Altman DG, Bland JM. Generalisation and extrapolation. BMJ 1998; 317: 409-410.
36. Altman DG, Bland JM. Time to event (survival) data. BMJ 1998; 317: 468-469.
37. Bland JM, Altman DG. Bayesians and frequentists. BMJ 1998; 317: 1151.
38. Bland JM, Altman DG. Survival probabilities (the Kaplan-Meier method). BMJ 1998; 317: 1572.
39. Altman DG, Bland JM. Treatment allocation in controlled trials: why randomise?BMJ 1999; 318: 1209.
40. Altman DG, Bland JM. Variables and parameters. BMJ 1999; 318: 1667.
41. Altman DG, Bland JM. How to randomise. BMJ 1999; 319: 703±704.
42. Bland JM, Altman DG. The odds ratio. BMJ 2000; 320: 1468.
43. Day SJ, Altman DG. Blinding in clinical trials and other studies. BMJ 2000; 321: 504.
44. Altman DG, Schulz KF. Concealing treatment allocation in randomised trials. BMJ 2001; 323: 446-447.
45. Vickers AJ, Altman DG. Analysing controlled trials with baseline and follow up measurements. BMJ 2001; 323: 1123-1124.
46. Bland JM, Altman DG. Validating scales and indexes. BMJ 2002; 324: 606-607.
47. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ 2003; 326: 219.
48. Bland JM, Altman DG. The logrank test. BMJ 2004; 328: 1073.
49. Deeks JJ, Altman DG. Diagnostic tests 4: likelihood ratios. BMJ 2004; 329: 168-169.
50. Altman DG, Bland JM. Treatment allocation by minimisation. BMJ 2005; 330: 843.
51. Altman DG, Bland JM. Standard deviations and standard errors. BMJ 2005; 331: 903.
52. Altman DG, Royston P. The cost of dichotomising continuous variables. BMJ 2006; 332: 1080.
53. Altman DG, Bland JM. Missing data. BMJ 2007; 334: 424.
54. Petrie A, Sabin C. Medical Statistics at a Glance. 2nd Edn. Blackwell Publishing, 2006.
55. Kirkwood B, Sterne JAC. Essential Medical Statistics. 2nd Edn. Blackwell Science, 2003.
56. Bland M. Introduction to Medical Statistics. 3rd Edn. Oxford University Press, 2000.
57. Barker DJP, Rose G. Epidemiology in Medical Practice. 5th Edn. Elsevier, 1997.
58. Rothman KJ. Epidemiology: an Introduction. Oxford University Press, 2002.
59. Silman AJ, Macfarlane GJ. Epidemiological Studies: a Practical Guide. 2nd Edn. Cambridge University Press, 2002.

Links 1-53 provided with kind permission from the BMJ.