Is writing really just words? Pearls and pitfalls to getting a manuscript accepted for publication
After enduring potentially rigorous research processes including designing an appropriate study, data collection, and analyzing results, the final step prior to submitting to a research journal is formulating a well-written manuscript. However, this final step can be one of the most challenging, as the way a reviewer interprets the language and findings of a manuscript can have a significant influence on their ultimate decision. The following commentary will provide the pearls and pitfalls to getting a research article accepted for publication by providing information pertaining to writing format, style, and perspective.
The abstract should provide a concise overview of all sections of the manuscript. What is required in the abstract will vary depending on the journal being submitted to; however, the majority of journals require some form of a background/study purpose, methods, results, conclusion, level of evidence, and key words. It is often easiest to write the abstract first, as it is an easy way to create a “template” and theme for how the rest of the manuscript writing will fall into place. Keep in mind that each journal has a specific word limit for the length of the abstract.
The level of evidence should be specific to your study design and decided a priori. If you are unsure of the level of evidence, it can be checked against descriptions of study designs and levels of evidence from either the journal you are submitting to or another validated source.
The key words should capture the main ideas of the study and it is often helpful to think which words you would type into Google or PubMed if you wanted to find your study. Most journals require a minimum of 5 key words.
The introduction should be concise and restricted to three total paragraphs. It is best to avoid wordy and verbose writing in this section, as it may influence the reviewer’s perception of your work prior to reading the subsequent sections. A well-written introduction should be no longer than one to one and one-half pages at the most.
The first paragraph should capture the reviewer’s attention and present what is currently known about the subject – similar to the background in the abstract. It can be useful to give important numbers such as costs or incidence of the event in order to help capture the attention of the reviewer. However, keep in mind that these should always directly relate to your topic, and being non-specific should be avoided.
The second paragraph should provide the rationale for your study based on the previous information – why is your study needed and what will it add to the literature? The end of this paragraph should set a transition to present the purpose and hypothesis of your study.
The final paragraph should very concisely state the purpose of the study (important to make this purpose identical to the purpose written in the abstract), followed by another sentence regarding the a priori hypothesis for the study. Examples include determining a functional outcome after a procedure or investigating the incidence of a disease process after a procedure. The hypothesis should be as specific as possible!
As this section can become lengthy and challenging to read if the reviewer is not entirely familiar with the topic, sub-headings are an excellent way to split up the text and facilitate easier reading. For the purpose of an original article on an intervention (medical or surgical), the methods should follow this order: (1) Study design; (2) Patient population (important to report how many patients considered and how many lost to follow-up); (3) surgical technique (cite previous literature if applicable to shorten this section); (4) postoperative rehabilitation protocol; (5) primary outcome of interest and secondary outcomes; and (6) statistical analysis.
The study design and patient population should include the type of study, the location and setting of where the study was performed, the inclusion and exclusion criteria applied to the patient population, and the indications for treatment. Inclusion and exclusion criteria should not overlap. For example, if inclusion criteria was a minimum of two-year follow-up, it is not necessary to say that exclusion criteria was less than two-year follow-up. For non-randomized studies, it can be helpful to include a STROBE diagram to depict the flow of how patients were recruited and any reasons they were lost to follow-up.
The outcomes of interest should include all variables studied, how the outcomes were determined and by what means collected, and by whom they were collected. All outcomes, treatments, predictors, and potential confounders should be clearly defined. Give diagnostic criteria, if applicable. If the outcome is being used in a non-standardized way, it is important to describe this process in-depth.
Provide detail on follow-up if applicable, including follow-up frequencies/intervals and relevant assessments (e.g. physical examination particulars, laboratory tests, imaging). Describe how relevant measurements were made (e.g. ROM or imaging), with citations of methods. If applicable, describe inter-observer variability for the measures where inter-observer variability would be anticipated and provide values. Describe any efforts to address sources of bias. (i.e. for subjective outcomes – explain here which outcomes, at which time points. For objective outcomes, describe which outcomes were collected at which time points the machine that was used). Attempts to describe rates of missing data can help provide clarity and decrease potential concern for bias.
The statistical analysis should be thorough and reproducible. If a qualitative statistical analysis was performed, describe how data was presented in the results section. This may be in the form of narrative descriptions of the findings or simple descriptive statistics including means, ranges, and frequencies.
If a quantitative statistical analysis was performed, describe how confounding factors were controlled for, how missing data was handled and accounted for, and how subgroups were compared. If performing a comparative data, present the a priori power analysis shows whether or not the study was powered to detect a difference in the primary outcome of interest. Data should also be explored prior to analysis to determine whether parametric or non-parametric tests should be performed. Always indicate which statistical software was used to perform the analyses (including company and location) and what p-value was used to indicate statistical significance (usually alpha = 0.05, unless Bonferroni adjustments are warranted).
The first paragraph should describe the study population including baseline characteristics and demographic data. It is important that the reviewer understands the study population for readers such that they can understand whether they can apply your findings to their patient population. If relevant, include age, gender/sex distribution, BMI, and other factors that describe the population of patients involved in the analysis.
Begin with a paragraph reporting numbers of individuals at each stage of study (numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and analyzed). Provide reasons for non-participation or loss to follow-up (and indicate when the losses occurred, eg, between which months). Consider the use of a flow diagram such as STROBE as mentioned in the methods. For prospective studies of therapy/treatments, indicate whether intention-to-treat or per-protocol analysis was performed. Present the minimum follow-up time, mean or median follow-up time (as applicable), and standard deviation or range of follow-up.
For outcomes of interest, the most important information should be written in the text, while other information should be referred to in the form of figures and tables. Do not repeat information in the tables by describing it in the text. This should be repeated for each outcome (or category of outcomes). Do not interpret the results in this section. All results should be simply stated, while interpretation of results should occur in the discussion section.
The discussion is known as one of the most daunting parts of the manuscript to write. The discussion section should present important results of the study and extrapolate these findings into a clinical context.
The first paragraph of the discussion should always be a concise summary of the most important finding of the study, without any background information. Whether or not the study hypothesis was confirmed or rejected can also be described here.
Each paragraph of the discussion should (1) introduce a main finding, (2) discuss this finding in the context of published literature by comparing and contrasting the results of the study to other studies, (3) provide an explanation for discrepancies or agreement, and (4) discuss why the finding is clinically important. This process should be repeated for each main finding listed in the first paragraph of the discussion and in this order.
The limitations should be presented as the last section of the discussion section prior to the conclusion. Limitations should be specific to your study and consider all types of potential bias pertaining to study design, study population, and analytic methods. Provide arguments as to why these limitations do not jeopardize your conclusions, or how readers should use them when evaluating your conclusions. It is not necessary to write a separate paragraph regarding the strengths of your study.
Take one paragraph to summarize the implications of your findings and their clinical relevance. What are the most important messages you wish readers to remember? Share how your findings will change clinical practice, and offer any treatment recommendations you can support based on your data. Offer directions for future study, if appropriate. This paragraph must be identical to the conclusions in the abstract.
References should be from peer-reviewed literature and formatted for the specific journal being submitted to. References should be contemporary when possible (within last 5-10 years is an appropriate goal). Older references, if important to the study methods or other sections, are appropriate to include.
The Figures and Tables
Figures and tables should complement your results and not repeat them as previously stated. Check the figure and table restrictions for the journal being submitted to. Always place figure legends after the reference sections (unless journal allows figures to be kept in the body of the text). Figure legends should be stand alone, meaning a reader should be able to easily understand the figure by reading the legend alone and without reading the manuscript.
Your manuscript is now ready for submission! Make sure all authors have agreed on the content of the manuscript and contributed to the manuscript. Follow the author instructions for submission on the journal website. After peer review, you will be contacted regarding the decision on your manuscript.