Forward
This is a guide to provide a basic structure to communicate research findings from Epidemiologic research, primarily through a journal article.
This started as notes that I have tried to operationalize, which come via piecemeal assembly from books on academic writing, NIH grant writing, and from assessing journal articles that seem to be well written by my eye. I wrote the original form of this in 2023 as notes from my dissertation process, which I have subsequently revised after finding new perspective in light of more experience afterwards.
I never took a scientific writing class (but plenty of humanities writing classes - I inadvertantly achieved a minor in History at uni), so this guide is a result of me trying to learn scientific writing despite being shown many alternate ways of writing, and to share information with others. Originally, this quick guide was based on observations from reading literature as well as Writing Your Journal Article in 12 Weeks;1 however, afterwards I added more from The Craft of Research (5th Edition).2 While worth reading in their own right, neither of these books focus specifically on epidemiological or scientific writing and are more general. So, I have tried to extract their ideas as they apply to our field and compile them. And, I always have to mention a paper that is extreme, but sometimes interesting for discussion and maybe reading on your own: “A full systematic review was completed in 2 weeks using automation tools: a case study.”3
I am not sure there are any universal formats for scientific products – the presented flowcharts and ideas concerning article structure are suggestions based on general conventions and observations from quality papers in epidemiologic journals, but each publisher may have differing requirements and collaborators may feel strongly about some specific details. One should do what makes sense in their case for each scientific product. However, having a starting point and breaking down the components of scientific products can be extremely helpful, especially for students and early stage researchers.
Some of these pointers may seem overly simple; however, some of the most important things I discovered after my PhD were extremely simple! I hope you find this helpful in any case, and if you already knew all of these pointers, you might consider yourself lucky.
There are some checklists in the Appendices that may be helpful, and of course the list of Works Cited will also provide even more info.
Section 1 - Scientific Projects: Ideas & Organization
What you do first (how you manage your project) is determined by what you have available to you.
If you have the data available - do the analysis first.
If you do not have data (e.g. for a grant or research proposal) - you will have to conceptualize everything beforehand, and make sure you have the correct data needed (see 2.3 Supporting Information).
My alma mater emphasized grant writing to students, and so I had a tendency to think about this in reverse - to write everything out and do the background literature search first, then do the analysis when you have a more thoroughly grounded project. If you do not have any information for your background this might make sense (or simply conduct a systematic or scoping review); however, I would suggest to start with the analysis and then assemble evidence that supports your claim and fills your introduction. It is usually a more efficient use of your time, and often your project will diverge from your initial plans, no matter how considerate you are in when writing them out. You also may simply not know what sources you need beforehand - something I found frustrating and problematic when trying to generate products as a student. This also makes it more difficult to write abstracts before articles to submit to conferences (because you will have no results to present), and grant institutions likely expect preliminary results as well. I have previously had professors suggest I produce abstracts before analyses were begun, and it was a bit confusing for me.
Additionally, working with data really seems to generate questions and support genuine curiosity. You can explore the data, look at different associations, and test ideas, then go to literature for further evaluation afterwards and connect your ideas to others.
You do need to do some preliminary searching to make sure others haven’t done your exact idea, and some reading to stay up on methods and ideas, but I do not believe it is productive to primarily focus on that, unless that is the focus of the product (e.g. reviews). Ideally, researchers should be consistently reading and taking notes on literature anyway.
When getting started, I think it is also important to consider the project’s scope. I initially had a difficult time with this, but the best way I have found to think about this is the way it is presented in The Craft of Research.2 We all seem to start with a topic or an interest. I am interested in Parkinson’s disease. However, I cannot write a succinct or focused article on Parkinson’s Disease (PD). That would be a book. That is an interesting topic though, so let’s start with that.
From a topic, we need to whittle it down with critical thinking - why am I interested in PD, and what do I think is interesting about the research ongoing?
From there, it is a simple matter of asking a good question. You know you have a good question if it answers the “So what?” check, and you think it might lead others who study the topic to change or update their understanding of the subject in some way. I found this helpful when contrasted with previous advice that I had been given to “just come up with something.”
1.1 - Literature and Notes
Taking notes is an essential part of writing; do so consistently and efficiently. Every researcher has had a moment where they remember a paper that would fit for some situation, only to find they can’t remember the paper’s title or author… it can be lost forever, and be genuinely frustrating because you can really only blame yourself. Additionally, personal notes are the most natural place for the start of a draft article or poster.
If you read the books I mention in the forward or have a humanities background, remember that when writing an article for a humanities journal, citing sources exactly is extremely important for constructing an argument, but for our purposes in Epidemiology it’s not an efficient use of one’s time. I do suggesting reading the section in The Craft of Research that discusses note-taking, but avoid wasting time with direct quotes unless you plan on inserting a quote into a project.
What you do need, is an organized system that you like, to efficiently recall papers’ core results and conclusions, and the article’s information for looking it back up or citing it. So, a sentence or short paragraph, maybe some annotations, and the citation should be all you need. There have been so many different ways to do this in academia, now you can simply choose a system:
Zettelkästen - makes use of index cards, the original way of keeping track of info and sources at a large scale. Get a box of index cards, write the reference, core info, and an ID on the index card so you have them indexed in your box. This allows you to always find papers unattached to projects, but the downsides compared to digital methods are obvious. This simple method is covered in The Craft of Research (5th ed., p. 85). However, this can be expanded to be more useful method, as developed by Niklas Luhmann. The extensive Zettelkästen method indexes not really sources, but your ideas and notes with which you will actually write your articles. The best source I have found is the book How to Take Smart Notes by Sönke Ahrens.4 There are a number of people claiming to be gurus online, but Dr. Ahrens presents it in an honest and authoritative form, whereas all the other sources I have seen are more complicated and miss the forest for the trees by focusing on details that aren’t actually necessary. While this was originally analogue-based, there are numerous digital Zettelkästen services like Obsidian or Zettlr, or you can even just create basic text files in an ever-expanding digital directory.
Equivalence of Primary Sources in NotesI should also like to suggest an adaptation of the Luhmann method; treating data as he treated articles. Luhmann’s primary sources were books, articles, theories, i.e., humanities-type sources. In a statistical or more scientific field, we have to realize that we have articles as primary sources, but also datasets. I believe that Zettel should include notes on our ideas from interrogation of datasets and analyses, as well as ideas from articles. The two are really the same thing - especially if articles in our field are to be considered driven by data first. In our writing process, data are where many of our ideas come from, where we apply critical thinking, and articles’ discussion are usually focused on data and analysis, so I believe we should directly connect ideas about datasets and statistics to discussions in articles as part of the same trail of thoughts and dialogue.
Annotated bibliographies - not exactly an easy method since most editors are not geared for it and may need to by typed out completely instead of inserting citations from a citation manager. But, it does help keep track of literature you mean to use for each project, and keeps everything in one place.
Digital reference annotation - Some reference managers allow you to read and annotate articles, retaining the notes with the reference (e.g. Mendeley). Usually the process is more complicated if you want to print an annotated bibliography, but if you just want the notes there on the article, in a digital format, this might work for you. The downsides are the usual digital issues; if you find yourself without your charger or your account with the reference manager is somehow lost (which does happen), you can lose all of your work instantly. It may also just be easier for recall to write things down.
You might be able to find more ways, or come up with your own - and that’s the key. Find a system that works for you so that you can consistently remember sources and ideas, and then use them to easily develop projects.
1.2 - Audiences
When considering the audience, usually the instructions are to consider whether or not the intended audience is a group of professionals, or a layman. I’m here to tell you that that is not good enough and you must be more specific if you want to be published. Each day research becomes more specialized, with not only scientific jargon, but specialist jargon, and layers of layers of conventions and expectations. If you want to publish in an epidemiologic journal, you must know not only what Epidemiologists or Clinicians expect, but also what the editors and readers themselves expect.
If you want to publish in Neurology, you should probably go to a neurology conference and take notes on what the researchers there say and do and how they present their ideas. There is also obviously some acceptance and work with editors, but usually your chances of being accepted should go up with less friction between editors or reviewers and your product.
I think it is just easy to think of the audience in general terms, and that is exactly the issue. Medical doctor so-and-so will read your research, what would they like to see? Then doctor of philosophy in epidemiology so-and-so will read it and what will they think? Once I began to think beyond the idea of my ideas just being thrown to a random mass of people, that helped my writing. It allows you to be more specific and imagine an more productive dialogue.
1.3 - Light Suggestions & Things to Consider
The more I have done past my dissertation, the more I need help remembering what is going on with what project. There is just more to balance. So, I often take notes twice by hand to help with memory; once for Zettelkästen , and once for each project’s file (yes I have bound hard copies and physical files for each project - keeps me from loosing things and I’m not reliant on my PC either). I also like being able to find articles and sources just in their own right in the index cards, and to also keep project-specific notes in the project folder. I’ve found it is slower than I’d like, but it allows your background to come together very quickly, and is faster overall because you are better able to remember what the sources mean and how you want to use them, and you are definitely not disorganized.
I have ADHD and have challenges staying organized naturally; however, I think everyone benefits to some degree from having what I call a “Road Map.” This is just you sitting down first to think of how you will do your analysis. What tables do you want, models will you use, etc., and then what steps exactly in R (or other software) will you need to do? What will the data need to look like/be formatted like and what sort of functions do you need? If you quickly scribble all this out beforehand, it might save considerable time. I’ve also always found that projects become unexpectedly extended… so when I would think that this was not necessary for a short little project, that project can change into a year long project with more moving parts and things can get confusing. This isn’t an analysis plan - you can do that too if you like - this is literally what steps you will do to complete the project. This can be also very helpful for convincing people to join in your project and to manage cooperative work. It doesn’t have to be detailed, just enough to get you from point A to finished.
1.4 - Section 1 Summary
The main theme of this section is organization, and maybe also using critical thinking. To start your project, critically engage with your topic, question it until you have a question for which the answer is interesting, and from that point be surgical in how you approach the projects you will create.
After this point, this guide is more focused on assembling the parts of an article, so it is mostly assumed that you have an idea and have conducted the analysis, or are able to do the analysis as you write.
Section 2 - Product Foundational Writing
2.1 - Formally Outline the Research Question
Assuming you already have an idea and a question that you think will be interesting for your audience, and worth investing your time in, there are some rules-of-thumb you can use to formalize your question so that it is publication-ready. One form that is popular is the PICOT format, which involves the following (in a table so you can copy/paste):
| Population: | |
| Intervention/Exposure: | |
| Comparison: | |
| Outcome: | |
| Time: |
Where population is the sample, the intervention is the exposure or treatment in the exposure group, comparison is the comparison or reference group, the outcome is the metric or measure you will use to examine the effectiveness of the intervention/exposure, and time is the time periods that should be considered (the duration and/or intervals).
Then combine the table components into a scientific question, and use this to guide filling out the table in the next steps.
There are other forms of research question also, such as PCC (population, concept, context), PEO (Population, Exposure, Outcome), SPIDER, SPICE, ECLIPSE… etc. For more frameworks, see Booth et al.5 These forms are helpful as a checklist, and at this point I assume you have a quality question to be put into one of these formats based on 1.1 - Literature and Notes.
2.2 - Defining the Gap
In general, understanding the “gap in the literature” that is being addressed is a good idea for even informal proposals and protocols, and expected by most researchers in published works. It is important to understand what is a compelling argument for your research project. This is no longer explicitly required by NIH grant proposals – but the same information is required in some form. For a simple gap statement, consider statements that generally answer the following questions:
What has been done currently?
What has not been done currently?
Why should it be done? (What is the payoff – OR – what is currently the issue?)
It may make more sense to revisit your research question after considering the gap, or to define the gap before you conceptualize the research question. Their order of development is mostly arbitrary, and most of these foundational pieces can probably be done at the same time.
2.3 Supporting Information
The following supporting information will also be expected, and probably can be extrapolated from the previous two sections fairly easily.
2.3.A - Specific Aim(s)
For a single article, it is likely you will only have a primary aim, potentially with a secondary aim. Specific aims are always independent, and work towards testing the same hypothesis. In general, they are simply statements of explicitly what you are setting out to achieve in a practical and measurable way.
For a single journal article, usually this the operationalization of your research question. For example:
(Our aim was) To examine the association between exposure(s) and outcome(s) in population after accounting for relevant variables and issues (biases,) in order to accomplish a better understanding of the phenomenon in the hypothesis.
Grants usually have explicit guidelines on how Aims pages should look, so be aware of that for grants.
2.3.B - Method(s)
At this stage, a concise yet thoughtful listing of the following is suggested:
The Null hypothesis
The alternate hypothesis
The statistical method you will use, and how it will test your hypotheses to answer your scientific question
2.3.C - Data Needed
List as many ofthe required variables for your analysis as you can - exposures, and outcomes, and any other explanatory variables. This is essentially to help you plan and request data if that is part of the project.
Appendix A provides a table that you can fill out to make sure this is all covered.
2.4 - Selecting a Target Journal
If you need to select a target journal, use the evaluation form found here, and week/chapter 4 in Writing Your Article in 12 Weeks. Below is an example table that can help with deciding.
| Journal Name | Belcher Score | Priority | Proportion Accepted | Publishing Fee |
| Sci. of the Tot. Environ. (STOTEN) | 20 (good) | III | 24% | $3790 |
| Infectious Disease Modelling | 23 (great) | I | 75% | $995 |
| Jour. Of Pub. Health Man. & Pract. (JPHMP) | 21 | II | 40% | $0 |
2.7 - Working with Collaborators
At this point, you can probably both recruit potential collaborators, discuss authorship, and talk about roles and responsibilities.
The criteria for authorship (everyone uses the International Committee of Medical Journal Editors (ICMJE) definition) are as follows:
Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND
Drafting the work or reviewing it critically for important intellectual content; AND
Final approval of the version to be published; AND
Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Defining an Author – beyond the ICMJE definition, there is the 14 suggested contributor roles from CRediT (https://credit.niso.org/)
Conceptualization
Methodology
Software
Validation
Formal analysis
Data curation
Writing – original draft
Writing – review and editing
Visualization
Supervision
Project administration
Funding acquisition
The example CRediT statement from their website is:
Zhang San: Conceptualization, Methodology, Software Priya Singh.: Data curation, Writing- Original draft preparation. Wang Wu: Visualization, Investigation. Jan Jansen: Supervision.: Ajay Kumar: Software, Validation.: Sun Qi: Writing- Reviewing and Editing, …
It is usually helpful to discuss these roles and the expected time spent on the project up front.
Section 3 - Article Composition
This section is the bullet notes of usual and suggested components of an article, along with typical flowcharts of article sections. Much of this comes from Writing Your Journal Article in 12 Weeks,1 however some of this is my own convention - the mesostructures and flowcharts - which are empirical observations from reading journal articles in general.
3.1 Title
Be specific.
Describe what is going on in your research. Avoid attempting jokes, and do not place a question mark or question in the title.
Should be 5-21 words - not too long.
Include keywords if possible (consider MeSH terms)
3.2 The Abstract
There is an Abstract structure table in Appendix C that one can fill out to get started.
Is it self-contained? – Try to write the abstract so that someone might be able to cite the paper by only reading the abstract.
Belcher suggests the following rough guidelines:1
Methods in only one or two sentences
Findings/results only one to two sentences
Recommendations are about one sentence
In reality, these are general starting points; the correct answer is to adequately describe the study and use the full amount of characters or words allowed
Provide key findings plainly, up front
Use strong verbs; e.g. exploring/examining -> argues, demonstrates
- Basically, try to avoid being vague
Include relevant keywords (use MeSH on demand to process abstract – it will list MeSH keywords for you)
Use Past-tense, to tell what you did
Always use active voice (e.g., say “we analyzed the data,” NOT “the data were analyzed.”)
3.3 Section Mesostructures
The following sections cover the sections in a research article, and the components usually found in those sections. There are bullet points, and a flowchart, for each section suggesting a general structure and flow.
3.3.1 Introduction
Start with a “gripping” first sentence
- Can be a statement of issue significance, interesting fact, historical opening, or perhaps a general subject statement or critical opening
Give basic information on your subject (literature review)
There is a table for literature review in Appendix B.
Go from more general focus to more specific
Locate your research within the body of research found in literature - how does it relate to, continue or change current literature suggestions?
Provide relevant theories or information that your audience may be missing
Provide indication of significance
Provide gap
Provide aim(s)
Significance and Gap are often two paired together and are logically intertwined
Structure Flowchart:
3.3.2 Methods
Sample and sampling procedure
State measurement instruments and methods
Population and location
Describe variables
No statistics description/tutorial if it is an established method
No results in the Methods section
Structure Flowchart:
3.3.3 Results
Standardize tables and graphs; unless otherwise instructed:
table names go above tables
Figure names go below figures
Provide interpretation of results – what do you see in your results?
1-2 paragraphs per figure or table (each figure or table should be interpreted in writing, and if it does not add to the paper reconsider its inclusion)
A summary of the overall interpretation of evidence
Do not repeat table information in text – just interpret table
Organize results around variables
No methods or discussion in results
Structure Flowchart:
3.3.4 Discussion & Conclusion
Discuss
Inference – the meaning of the results section
Check for common biases: check Delgado-Rodriguez & Llorca for a list.6
- Which way would the bias pull the results? Is it negligible or large? Is there a way to check (i.e., performing a correction and checking before and after significance)?
Ethical considerations?
Limitations
Discuss each result in the results section (check with good example papers)
Conclude
Did you confirm your hypothesis/achieve your aims?
State significance
Relate your findings to the current understanding in literature
Suggest further research
These are also a “pyramid,” but goes in reverse order to intro/background – start with specifics and then go back to general
Structure Flowchart:
Section 4 - Editing
4.1 - Macro-editing
The article should follow the following general structure. This is probably already well-known by anyone reading this, but here it is just for reference:
Introduction
Introduction to general subject of investigation
Review of literature
Statement of hypothesis
Methods
Specific Methods
Procedures
Materials and instruments
Experiment
Context and setting
Population or sample
Results
- Report results
Discussion
Validity of methods and findings
Context within current scholarly literature
4.2 - Micro-editing
Micro-editing is the checking of grammar and focusing on individual sentences. This is the last phase of the paper writing process, so naturally the macrostructure and mesostructures should be fairly well established. Conversely, any ability to incorporate micro-editing in earlier stages is fine, but it is really not required and might slow you down.
Check for active voice
Check for past tense
Check for strong verbs where appropriate
Check for conciseness (and remove doublings?)
Check for logical flow
Remove pronouns when you can
For more info on correct grammar and editing advice, the classic source to turn to is The Bedford Handbook, which is a generally well-known resource for English grammar, style, and writing conventions.7
Section 5 - Writing Checklist
Foundation
| ☐ | Read up-to-date literature |
| ☐ | Operationalize a tentative research question to guide review |
| ☐ | Have the library conduct a search if available, or conduct thorough search using pearl growing, upstream & downstream searching. |
| ☐ | Construct bibliographic review table and identify gaps while reading |
| ☐ | Identify specific aims |
| ☐ | Identify needed data |
| ☐ | Identify hypothesis and null hypothesis |
| ☐ | Identify appropriate statistical methods |
Title
| ☐ | Approximately 15 – 21 words |
| ☐ | Conciseness (limited description of the article) |
| ☐ | Does Not include question marks; however, may contain semicolons or colons |
| ☐ | Include key words (MeSH terms if possible) |
Abstract
| ☐ | Self-contained/self-explanatory |
| ☐ | Relevant keywords |
| ☐ | Strong Verbs |
| ☐ | uses the maximum allotted characters or values |
| ☐ | most focus on results |
| ☐ | no p-values in abstract |
Introduction
| ☐ | Strong first sentence (Interesting fact, critical statement, historic point, or significance statement) |
| ☐ | Adequate basic information and literature review |
| ☐ | “Inverted Pyramid” structure – start broad, move to specific |
| ☐ | Locate your research within literature |
| ☐ | Significance statement |
| ☐ | Gap in literature you are filling |
| ☐ | Aim(s) explicitly stated |
| ☐ | All the components of a formal scientific research question (e.g. PICOT) are present. |
Methods
| ☐ | Study Design |
| ☐ | Sampling and/or data acquisition |
| ☐ | Location of study |
| ☐ | Description of Population |
| ☐ | Timeframe |
| ☐ | Measures and Instruments |
| ☐ | Methods used |
Results
| ☐ | Standardize tables and graphs |
| ☐ | 2 paragraphs interpretation per figure or table |
| ☐ | Summary of results |
Discussion and Conclusion
| ☐ | Statement of study significance |
| ☐ | Study strengths |
| ☐ | Relate findings to current literature |
| ☐ | “Pyramid Structure” – move from specific to broad |
| ☐ | Discuss each result of main analysis |
| ☐ | Discuss why this topic is important to research (this is your logical backing) |
| ☐ | Limitations and biases (what do these do to your study?) |
| ☐ | Suggest further research |
| ☐ | Strong summary/conclusion for last paragraph |
Macro-editing
| ☐ | Check overall structure and flow from paragraph to paragraph |
| ☐ | Check flow chart for Introduction section |
| ☐ | Check flow chart for Methods section |
| ☐ | Check flow chart for Results sectio |
| ☐ | Check flow chart for Discussion section |
Micro-editing
| ☐ | Active voice |
| ☐ | Past tense |
| ☐ | Reduce pronouns |
| ☐ | Accurate terminology |
| ☐ | Strengthen verbs where applicable |
| ☐ | Conciseness and removal of doublings |
| ☐ | Logical flow between sentences |
Works Cited
Appendices
A - Foundational Information Worksheet
| Article Foundation | |||
| Scientific Question: | |||
| Gap in Literature: | |||
| Specific Aims: | 1° | ||
| 2° | |||
| Alternate Hypothesis: | |||
| Null Hypothesis: | |||
| Statistical Methods: | |||
| Data Needed: | Outcome Variable(s) | Exposures/Predictors | Other Explanators |
a) Y1 b) Y2 c) … |
a) X1 b) X2 c) … |
a) Z1 b) Z2 c) … |
B - Abstract Skeleton/Worksheet
| Title: | ||
| Authors: | ||
| Keywords: | ||
| Background: | ||
| Hook/Orientation | ||
| Importance of Research (gap or in general) | ||
| Aims | ||
| Methods: | ||
| Data | ||
| Timeframe, location, pop. (selection) | ||
| Main analysis method (information) | ||
| Results: | ||
| Key Results | ||
| Conclusion: | ||
| What the study did | ||
| Key Result Interpretation (what the study means in context |
C - Journal Comparison Table
| Journal Name | URL | Priority | Belcher Score* | Proportion Accepted | Publishing Fee |
| Jour. Of Pub. Health Man. & Pract. (JPHMP) | https://journals.lww.com/jphmp/pages/default.aspx | 21 (good) | 40% | $0 | |
| Online Jour. of Pub. Health Informatics (OJPHI) | https://ojphi.jmir.org/ | No Go | 08 (bad) | ? | $1500 |
| Pub. Health in Practice (PHIP) | https://www.sciencedirect.com/journal/public-health-in-practice | $1700 | |||
| Journal of the American Medical Informatics Association (JAMIA) | https://academic.oup.com/jamia?login=false | 24(great) | 20% | $4,126 | |
| American Journal of Epidemiology (AJE) | https://academic.oup.com/aje | ||||
| American Journal of Public Health (AJPH) | https://ajph.aphapublications.org/ | 25 (great) | 17% | None for regular, $2500 for open access | |
| Journal of Public Health | https://www.springer.com/journal/10389 | ||||
| International Journal of Epidemiology (IJE) | https://academic.oup.com/ije | ||||
| Sci. of the Tot. Environ. (STOTEN) | 20 (good) | 24% | $3790 |