Ten Commandments For Planning a Research Study or Thesis.

Planning a study requires attention to detail in the following aspects:

1. Introduce and Familiarize yourself with the steps in performing quality research
2. Review literature using PubMed, read Journals regularly and work on Hypothesis building.
3. Interact with a statistician before starting your study/thesis
4. Data collection and maintenance is Critical. Use tips for data entry in Excel sheet.
5. Give clear cut guidelines to the statistician for data analysis and interpretation
6. Use a methodical approach with CHECKLISTS for Writing up the manuscript.
7. Choose journal of choice carefully keeping in mind its indexing and your target audience
8. Use reference managers to efficiently maintain references.
9. Be aware of the Common causes of rejection of a manuscript and avoid these mistakes
10. Be positive even with rejection, be persistent and vow to come back stronger

Introduction:

Young ophthalmologists undergoing residency or into early stages of clinical practice are exposed to a wealth of clinical data. However, in India, we still prefer to learn from textbooks and from what is taught to us in medical school, the quality of which varies widely. However, the best way to learn is from the clinical material one is being offered on a daily basis in your clinics (i.e. from patients).

To collate data from a large number of patients with a similar condition, perform meaningful statistics on the collected data and draw inferences from these results is the best way one can learn. This is nothing but performing research, an extension of clinical curiosity that urges you to go the extra mile, and discover what has not been studied till date, thus fill lacunae in existing literature and finally improve patient care. There are many advantages of doing research, even for a busy clinician. These advantages can be seen here.

In this write up, I will introduce few basic concepts in doing a clinical study/thesis, that will help you in creating robust data and publish your results in peer-reviewed journals.

Review of literature and Hypothesis formulation:

The first step in planning a study is the “Study Idea”. The important questions to be answered are

1. “Is this idea novel enough and how much work has already been done on this in the past?”
2. Is it feasible for me to conduct this study within the limitations of my setting?

Answers to these questions go a long way in determining the publication potential of your research. So think about these from the beginning. A thorough review of literature using PubMed (www.pubmed.com) is the best way to discover whether others have already published in your area of interest. Using the Advanced Search builder option in PubMed always yields better results.

Choosing the topic of study, after a thorough literature review, not only tells you that it is worth doing, but also alerts you to the parameters that need to be recorded from each case during the study.

Reading top journals (such as Ophthalmology, AJO, Eye, BJO etc.) regularly is very beneficial. This not only helps you keep track of the recent advances but also gives you an opportunity to decide on topics that you may want to pursue as research topics. Subscribing to the “e – table of contents: (e-TOC) is a very easy way of getting new journal issues in your email box.

Interact with bio-statistician before starting your study/thesis

Getting inputs from the bio – statistician even before starting the study is a very important step. This helps you understand the nuances of sample size and study design, both of which have a tremendous bearing on your results and conclusions. Mistakes in sample size and design cannot be undone once the study has commenced or is completed.

Details needed for sample size calculation are as follows:

1. Previous studies conducted on similar topics
2. Clarity on the Primary Outcome measure of your study.
3. In comparative studies, please provide a Mean and standard deviation of the primary outcome measure for each of the groups to be studied, from previous literature. Alternately, provide proportion of cases (i.e. percentage) satisfying the primary outcome measure in each group.
4. Think about what difference will be clinically relevant between two groups and provide this information to the statistician.

In addition to the sample size, discuss the study design such as prospective vs. retrospective, need for a control group and need for randomization &/or masking with your bio – statistician. These intricate aspects as well as inclusion and exclusion criteria should be discussed with the entire study team before commencement of the study. You can learn more about interacting with a biostatistician here.

Data collection and maintenance – Excel sheet tips.

Gathering data and maintaining it well is an art that needs practice to master. If the study design is prospective in nature, make a clinical record form i.e. proforma of the parameters to be recorded from each patient at every visit. Also, aspects of the study protocol such as masking and performing crucial tests should be done with utmost care to avoid bias. In a retrospective study, data needs to be entered directly into the datasheet from the case files.

While entering data into an excel sheet, keep the following tips in mind

1. Use the first row to label the variables in short.
2. Do not enter alphabets anywhere in the excel sheet except the first row.
3. For categorical variable (such as Yes/No), label the categories with numerical (Yes=1, No=0). Categories – choose the comparison group as “0”
4. Use the “Insert Comment” option (by right clicking a box in Excel) to enter comments where needed and maintain a codebook showing how variables have been labeled.
5. De – identify patient data that can directly identify a patient.
6. Maintain consistency in units for the continuous variables (e.g. all follow up in months)
7. All Vision parameters always in logMAR
8. Right eye and Left eye as separate rows most of the times
9. Maintain ONLY one Excel sheet for the entire study
10. There should be ONLY One entry in one cell. Never as 1,2,3
11. No special features like hiding a column etc.
12. What to do with missing data? Enter a Dot (.) or 999

What to do with data – analysis and interpretation

Once data is tabulated into the Excel, make sure it is free of errors. “Please remember that most statisticians are not clinicians and hence may not be able to understand your project intuitively like your colleagues”. Hence, merely sending the Excel sheet to the statistician will not yield desired results. You should send the following to the statistician:

1. Excel sheet along with the study protocol and a WORD document of what is expected from the statistician.
2. Please make sure that you send a list of what each column in the Excel represents and if you have labeled a variable using numerical, send a list of how each variable has been labeled i.e. codebook. It is best to label each variable using the Edit Comment in Excel
3. Take care to explain the primary outcome measures and give clear-cut guidelines to the statistician on the kind of results you would like to see.

It is a good practice to send sample tables to the statistician so that he/she can send you results of your study using similar tables. Also, request for relevant graphs and figures that will help present your results effectively. For example, you can ask for a scatter plot with a LOWESS curve when two continuous variables are being represented; use box and whisker plots to show variance in the distribution of an important continuous variable. Interpretation of results can be subjective and hence it is a good idea to ask the statistician to guide you through the results in your initial research projects.


Writing up the manuscript – use checklists

Research that remains unpublished in essentially “undone”. If the manuscript of the study is not written well, there is very little chance that it will be published in the best journals. Hence, writing a manuscript is the most critical and rather, the toughest part of the entire process of research. The usual format that a manuscript contains is called IMRAD i.e. Introduction, Methods, Results And Discussion. Each of these sections have their own demands and can be written using checklists. The following link will take you to an article in the Indian journal of Ophthalmology that elucidates the nuances of manuscript writing.

http://www.ijo.in/article.asp?issn=...ssue=11;spage=1089;epage=1093;aulast=Sengupta

The Equator network (www.equator-network.org) is an extremely useful website that gives a comprehensive overview of how to write different kinds of manuscripts. Global leaders in research have constituted the Equator Network and there are different checklists for observational (STROBE http://www.equator-network.org/reporting-guidelines/strobe/), retrospective (RECORD http://www.equator-network.org/reporting-guidelines/record/), RCTs (CONSORT http://www.equator-network.org/reporting-guidelines/consort/), case reports (http://www.equator-network.org/reporting-guidelines/care/) etc.

These checklists are free to download and make manuscript writing very methodical and prevent authors from missing key elements from their manuscripts.

Writing the abstract of the article is one of the, if not the most important aspect of manuscript writing. Abstract writing should be done in the end, after the entire paper is written and each and every aspect of the study has been ironed out. The abstract should acts as the précis and enumerate all the essential methodology, results and conclusions at a glance, yet adhering to the strict word count limit of the journal.

Choosing the title of the study is the next most critical part because this will be used for archiving and indexing of your paper and will determine how many times the study will show up on search engines like PubMed and Google. The title should be descriptive yet succinct so that it will pique the readers’ interest but will not be very long. All coauthors should participate in deciding the title of the study because this is the one single aspect that will influence how many times your article is read and then cited.


Using reference managers

For the novice as well as experienced researcher, an achilles tendon is arranging references in the end. Circulating a manuscript amongst co-authors is a very good idea to improve its quality, however, this messes with the order of references, making it a herculean task to rearrange, especially while writing a thesis involving more than 50 – 75 references. There are many freely available reference managers such as Mendeley (www.mendeley.com), Zotero (www.zotero.org), Qiqqa (http://www.qiqqa.com) etc., which blend seamlessly with the Microsoft WORD Processor and help you organize references.


Choosing the journal of choice

Once the manuscript is ready, it has to be modified a bit to suit the requirements of the journal of your choice. This choice depends upon the

1. Nature of your target audience.
2. Aims and scope of the journal, its impact factor, indexing standards, acceptance rates and rapidity of review process.
3. The novelty of the topic.
4. The way the data is analyzed and manuscript written up.

It is often a good practice to see the journal that makes the most number of appearances in your own references and submit to that journal as your first choice. Discussing the merits and demerits of the study with your coauthors can help in narrowing down the journal of choice. Generally, it is good practice to submit the manuscript to the top ranked journal and make a list of journals to be submitted next in descending order of preference. This ensures that the paper gets a good initial review and even it is rejected, using the reviewer comments helps in improving the paper a great deal.

Common causes of rejection of a manuscript

1. Manuscript sent to the wrong journal, does not fit into scope of journal, fails to engage the issues addressed by the journal
2. Too long (exceeding word limit) or too short,
3. Writing style inappropriate.
4. Poor regard for the conventions of the journal.
5. Nothing substantial in the content of the paper.
6. Not properly contextualized – concentrates on trivial issues and ignores needs of an international readership
7. Poor theoretical framework.
8. Scrappily presented – not proofread,
9. Unduly assertive, unethical and Rude
10. A combination of the above

Reading “Instruction to Authors” on the journal website and meticulously conforming your paper to it is an important step. Contrary to belief, wrong journal and inappropriate writing style are more common causes of rejection than the scientific content of the paper. Its important to know and avoid these mistakes.

Conclusion

This blog introduces you to the basic requirements and care required to conceptualize a study, nurture it, generate robust data and write it to meet international standards. It is important that this approach be adopted from the very beginning of residency, but it’s never too late to start. The thrill of publishing your first paper is immense and one that inspires you to publish again and again. The first step is always the hardest but can be the start of a trail blazing career, leaving behind a legacy and inspiring those who follow. I urge you to follow the above steps and indulge into a joyful research career in addition to clinical and surgical one.

Source