Abstract
You have just completed months of research; your data have been reviewed and checked for accuracy; your results are well documented and data are carefully stored; your statistics have been checked; and your data are novel and of high significance. You are ready to sit down and write-up your manuscript. How do you start? First, consider writing an outline. It can often serve as a useful guide. Scientific manuscripts that describe original data follow a rather standard format in most medical journals: title page, abstract, key words, introduction, methods, results, discussion, conclusion, references, tables and legends, and figure legends and figures.
A brief outline for each of the major components of the manuscript listed above will help organize your thoughts. Outlines are especially useful if you are writing a review article or editorial that does not follow the same basic format as a manuscript describing original data.
Title Page
Provide the full title of the manuscript, all authors, affiliations of the authors, and contact information for the corresponding author. Some journals also request a brief running title and key words.
Abstract
Think of the abstract as a “mini” manuscript with an introduction or background, purpose, methods, results, and conclusion sections. Include the key data only; primary outcome data, some of the secondary data, and safety data, if crucial to the study. In most cases, the abstract should be ≤250 words.
Introduction
The introduction should provide key background information including the rationale for the study. What are the key concepts that are known about the scientific problem that you are addressing and what is unknown? What are the knowledge gaps? Why are you doing this study? In the introduction, you should specifically state the purpose of your study. State to the reader the scientific question that you are asking in your study. The introduction should only be about 2 to 3 paragraphs; it should not be an extensive review of the literature (which is a common mistake that is made by new authors).
Methods
The methods should be written very clearly so that a reasonable scientist in your field could repeat your studies. Describe the basic study design. Provide n values for the total number of subjects (people, animals, or if an in vitro study—number of cells, organs, biopsies, etc) that initially entered the trial or study. Provide both inclusion and exclusion criteria. Where possible, describe whether randomization and/or blinding procedures were involved, and if so the details of these procedures. If a pharmacologic therapy is given, be specific—provide dose, route of administration, timing of administration in relationship to the protocol, and duration of therapy. Describe exactly what the control group(s) received—placebo, saline, vehicle, and so forth. How was the control administered? If a patient or experimental subject is also receiving other treatments, these should be described. The primary, secondary, and other outcomes to be measured should be specifically stated. Details of how measurements were made should be provided. In some cases, where these methods are standard and details of them were published previously, you can refer to the previous articles that provided the methods. The statistical tests used should be included.
Results
List the most important results (such as the primary end point) first, followed by secondary findings, other measurements, safety end points, and then unexpected findings. Use of tables and figures are important for keeping the text of the results section as concise as possible. I prefer simple 1 or 2 panel figures that are not overly busy—if your figures are too busy, the main message is sometimes lost.
P values for significance should be provided where appropriate. Give n values of final number of subjects in each group that successfully completed the trial, and describe any eliminations of subjects and why (a common error is for the authors to leave out the n values). Provide baseline demographics or characteristics of various groups, so the reader can determine whether the groups were comparable.
Discussion
The first paragraph of this section starts with a brief overview of your main findings. Then, you should relate your data to what is already in the literature. How does the new data in your study add to the existing knowledge base of the subject? The discussion is not meant to be a full review of the literature on the topic but meant to show how your study relates to what is already published, adds to it, agrees with it, or in some cases disagrees with it. If your study disagrees with the literature already published—what are some of the possible reasons for the discrepancy? Describe potential limitations of your study. How generalizable is your study? Perhaps, your study may be relevant to one population; but perhaps, it is less relevant to another population. What is the research or clinical implications of your findings? I usually end the discussion with a very brief concluding paragraph, summarizing the key take home message or messages. Try to limit the key messages to about 3 or less.
Formats for tables, figures, and references must follow the published instructions for authors in the specific journal. Tables and figures should be cited in the text of the result section in the order in which they appear. Be sure that figures and tables are clear, easy to read, and illustrate the key points that you are trying to make. Write clear figure legends. Define all symbols in a figure or table legend.
Proofreading
Proofreading is crucial prior to submission of your manuscript. It is often useful to have several of the coauthors independently proofread a manuscript prior to submission. The more independent “eyes” that look at the manuscript, the better. Strive for internal consistency among the various presentations of quantitative data. Thus, if the abstract provides a number that is, for example, a mean ± standard error of the mean value of a blood pressure measurement, be sure that the number, standard error, and units are identical in the abstract, text of the results, and if relevant, in the table or figure in the results section. Round off to the same decimal in all of these locations within the manuscript. Check to be sure that these numbers are correct when compared to the primary source data. Be sure that n values are correctly reflected in the results section and in the tables and figures. If figures include histograms, it is useful to count the points against the final n value in each group for consistency (of course, sometimes points on a graph overlap, so it is not always possible to find each separate value of a subject on a histogram). Recheck all P values against source documents. Be sure P values are consistent among those provided in the abstract, text of results, figures, and tables. Carefully proofread your reference section. Be sure that all authors’ names are correctly spelled. Be sure that reference numbers given in the body of the text correspond to the correct references in the reference section.
Dr William C. Roberts, editor of the American Journal of Cardiology, often describes the phenomenon of “red-eye”. If you have been working on and reworking a manuscript, without a break, your eyes and brain may tire and it becomes easy to miss obvious mistakes. One suggestion is to put the manuscript away for a while (∼1 week) and then come back to it later. With fresh eyes and a rested brain, it is often amazing how many mistakes you will pick up. Do not be afraid to revise your manuscript multiple times. Our research group typically revises manuscripts that we submit for submission at least 5 to 7 times or more.
Revising Your Paper
If your manuscript is not accepted, do not discount the comments of the reviewers. Where possible, take their comments to heart and revise your manuscript as best you can and submit to another journal. If the reviewers have accepted your manuscript, pending major and/or minor revisions, again try to revise your manuscript based on as many of their comments as possible. Where revision is not possible, explain to the reviewers why it is not possible and supply a gentle rebuttal that outlines your stance. Avoid simply ignoring a reviewer’s comments. When you write your response to a reviewer’s comments, always repeat their comment and then provide a response. Show the reviewer exactly where you have made the suggested revisions in the manuscript.
Help With the Writing
There are many fine articles, books, and manuals about writing style and improving English grammar and punctuation. For authors whose native language is not English, consider working with a professional translating service. Most charge a fee, but it may be worth it. See http://www.sagepub.com/journalgateway/engLang.htm. I recommend that all potential authors read a few key articles regarding how to write a scientific medical manuscript. A few of my favorites are Quinn CT, Rush AJ. Writing and publishing your research findings. J Investig Med. 2009;57:634-639 and Roberts WC. Formulating an answerable question, displaying data, illustrating, writing, reviewing, and editing manuscripts for publication in medical journals. Am J Cardiol. 2012;110(2):290-306.
I hope that these comments help you as you prepare your manuscripts for submission to JCPT or other journals.