Manuscript preparation and publication.

Stephen Jay Gould14 wrote, “I always look to closing paragraphs as indications of a book’s essential character.” Although that may be germane for books, in the case of manuscripts, it may be just the reverse; it may be that Abstracts are the indications of the essential character of as are the indications of the essential character of a manuscript. The structure of the Abstract varies depending on the journal’s requirements. Some journals specify a structured Abstract; in others, the form must be unstructured. Some in others, the form must be unstructured. Some journals require a condensed abstract of only a sentence or 2, sometimes no more than 50 words. Preparation of the Abstract requires a great deal of attention because it is seen first. It may be the most important piece of information in helping the reviewer or prospective reader decide on the merits of the article. The Abstract should set the background stage succinctly in 1 or 2 sentences (why it was done), then identify the study design (how it was done and in whom), and then create a slightly longer section presenting the results before coming to a concise statement about conclusions and implications. Some journals require separate paragraphs for each of these elements. Typically, an abstract must be written in 250 words; the instructions to authors will provide this information. The word limit is very important; few things are more discouraging in a literature search than to come across the phrase “abstract truncated at 250 words” while the reader is still midway through the Results section. Introduction The Introduction should be short and focused. Young investigators should avoid the temptation to include in the Introduction much information that should be in the Discussion. An Introduction has several purposes: 1. It should attract the reader’s attention. This requires that the author address the needs of the intended audience. 2. It should mention the questions or issues that form the background of the study. Information about the relevant literature with key references to highlight the questions or issues to be addressed should be included succinctly and without elaboration. 3. The final part of the Introduction identifies the hypothesis to be treated and the questions addressed in the manuscript. Materials and Methods This section has many required elements. 1. Documentation of Institutional Review Board or Animal Care and Use Committee approval is mandatory. Without it, no article can be published. 2. The population in which the hypothesis is to be tested must be indicated. It may be a patient series, in which case it is optimal to have it be consecutive patients. This portion includes patient selection, the inclusion/exclusion criteria. These criteria can be listed in a table, which is more user friendly. If the subject of the article is an animal experiment, then the specifics are given (eg, species, size, and gender). 3. The details of the study, ie, design and implementation, come next. Details here vary widely depending on the study, whether it was a population-based study, a new device, a new medication or strategy, or an animal experiment. There must be enough detail so that the reader understands the study design and could duplicate it if desired. This may be more crucial for basic science experiments than for clinical registries. 4. The primary end point of the study is stated, which refocuses the hypothesis to be tested. In some studies, the primary end point is a composite (eg, major adverse cardiac events), whereas in other studies it is a single and discrete end point (eg, death). This section also presents secondary end points or end points in prespecified subgroups. This latter point is important because posthoc analyses, although interesting for hypothesis generation, lack the scientific merit of prespecified end-point analyses. 5. Definition of terms is integral so that the reader can put the results into context. Well-accepted definitions should be used; it is important to reference the articles that have defined the terms. Terminology that is unique to the manuscript should be avoided. For example, if the author’s definition of cardiogenic shock is blood pressure 100 mm Hg but the standard definition in the literature is 90 mm Hg, the results may be discounted because the patient population differs from the accepted shock population. 6. The statistical section requires great attention.15 Part of the educational value of manuscript preparation resides in having the writer learn about the underpinnings of statistical calculations. Study design considerations here include issues related to superiority versus noninferiority, techniques for data analysis such as propensity scoring and bootstrapping, statistical power to detect differences between groups in the study, and sample size calculations. Input from a statistician is essential in this section. The statistician should typically be a coauthor. Results The construction of the Results section requires considerable thought. Depending on the study, some results may best be displayed in tabular or graphic form. For example, baseline demographics of patient groups are best placed in a table. The text can then be used to highlight crucial patient characteristics or differences between the groups rather than to completely reduplicate material in the table. The number of tables and figures will vary, depending on the subject, the journal, and the audience. Gopen and Swan16 have written extensively on writing with the “reader in mind.” The Table illustrates their point that there are several ways to present the same data. For data such as these, tabular form is obviously much more understandable. They have also explored such technical writing issues as subject/verb separation, sentence structure, topic position, and logical gaps. The Results section should be written to stand on its own merits. After the population has been detailed, the primary end point is the focus of attention. Use of graphic material can be of great value in highlighting specific points, but duplication of information in text and illustrations should be avoided. Holmes et al Manuscript Preparation and Publication 909 by gest on A ril 2, 2017 http://ciajournals.org/ D ow nladed from After the primary end point, the secondary end points are then detailed and documented. Some authors make the mistake of repeating or including information that should have been in the statistical portion of the Materials and Methods; this should also be avoided. It must be kept in mind, as Gopen and Swan16 write, that “we do not start with the strawberry shortcake and work our way up to the broccoli.” Thus, some of the very important or potentially controversial secondary end points should be placed at the end of the Results section to continue to stir and keep reader interest. Illustrations Graphics are an important part of any manuscript. The number and kind of illustrations vary. Some information is more readily assimilated in tabular form (the Table). Tables should include only relevant material; extraneous data that do not affect the manuscript should not be included. The information presented in graphic or tabular form should not merely repeat what is in the text but should complement and highlight it. Kaplan–Meier statistics are much better presented as graphs. The reader must be kept in mind when constructing legends. All pathological/histological material should be labeled explicitly, which often requires a more detailed legend than usual. Definitions of any acronyms and abbreviations should be included in the legend for a table or figure. It is important to remember in this age of electronic manuscripts that color printers are not universally available. Accordingly, the lines on graphs should allow the reader to understand the data in black and white using, for example, dots and crosses or dashed lines (Figure 1). The scale used in graphs is also important to highlight differences in populations; the vertical scale should be adjusted accordingly. Alternatively, this can be accomplished with an inset in a graph to point out differences in population outcomes (Figure 2). Pathological or histological figures often require color. Given the cost of printing color in most journals, which is usually the responsibility of the author, great care should be taken to optimize the number of examples. Discussion There is significant literature on the components of an optimal Discussion section, including what should be avoided. The list is extensive. Jenicek17 and others18,19 have identified multiple issues. Key components include the following. 1. A recitation of the major findings of the study. This shouldappear in the first paragraph of the Discussion and is oftenin a point format (eg, the major findings of this study areA. . ., B. . ., C. . .). It is neither necessary nor advisable toinclude statements such as “This is the first report of X.”2. Positioning of these findings against the background ofpublished information, highlighting areas of concordanceand discordance. Other studies should not be denigrated.However, the shortcomings of other studies, either gaps orunrecognized biases, should be noted in this section.3. Exploring the basic mechanisms of the findings of thestudy. This may involve the underlying pathophysiologyin an animal model or social or technical issues arising inand from a clinical study.4. Application of the findings to other populations. It isimportant not to overstate the findings of the study ormake unwarranted speculations. However, some applica-tions to broader patient populations are reasonable. It isalso reasonable to call for more information in a largerpatient base to further study the issues.5. Limitations. This is an exceedingly important section. Adiscussion of the shortcomings of either the experimentaldesign or the statistical methods, as well as the generaliz-ability of the findings, is crucial. This is important forreader understanding and the reviewers’ evaluations of themanuscript. For an experimental study, issues that shouldhave been dealt with by an improved study design shouldnot be identified as limitations.6. Conclusions. This requires a succinct re-presentation offindings. It can be accompanied by statements of the needto gather more data either from larger series or from otherpatient groups, although this should be limited. It isoptimal to concentrate on the presentation of the findingsfrom the data set studied. ReferencesReferences should receive considerable attention. In general,published manuscripts are preferred in any reference list. Iffor any reason there is a long delay between planning andcompleting a manuscript, the literature search should beTable. Example of 2 Different Layouts for Presentation of Data Nontabular Presentation of Data t (time) 45 , T (temperature) 36°; t 30 , T 32°; t 25 , T 30°;t 15 , T 29°; t 10 , T 28°; t 0 , T 27° Tabular Presentation of Data Time, minTemperature, °C