|Year : 2021 | Volume
| Issue : 6 | Page : 599-601
How to read, write, and review?
Department of Community Medicine, Dr. D. Y. Patil Medical College Hospital and Research Centre, Dr. D. Y. Patil Vidyapeeth, Pune, Maharashtra, India
|Date of Submission||19-Aug-2021|
|Date of Decision||25-Aug-2021|
|Date of Acceptance||31-Aug-2021|
|Date of Web Publication||2-Nov-2021|
Department of Community Medicine, Dr. D. Y. Patil Medical College Hospital and Research Centre, Dr. D. Y. Patil Vidyapeeth, Pune, Maharashtra
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Banerjee A. How to read, write, and review?. Med J DY Patil Vidyapeeth 2021;14:599-601
Commenting on appreciation of art, Marcel Proust had written, “The only true voyage of discovery would not be to visit strange lands but to possess other eyes. ”Research writing is a combination of science and art. Science is the observation or experiment which generates data, to improve clinical or public health practice and guide further research. Art is the way the science is presented. A manuscript should include robust science and good art. The following paragraphs outline the key points for reading, writing, and reviewing a paper.
Scientific paper writing originated in 1665, evolving over the years to the present day format of IMRAD, i.e. Introduction, Material and Methods, Results, and Discussion. To this, one may add the title and abstract, the label, and the lid, respectively, of the contents.
The label or title should convey the research question which the paper addresses. Additional features of the paper such as the study design if incorporated enable readers to rank the level of evidence in the hierarchy of evidence-based medicine (EBM).
The abstract or the lid is like the package insert which gives an outline of the contents. It should be concise including the salient points covered in the paper. The purpose is to enable the reader to decide if they should read the paper in detail. Most journals insist on a structured abstract for an original article. The subheadings are similar to those that appear in the main text. The abstract usually has a word limit of 250 words.
Following the abstract is the “introduction.” This should have four themes, which can be summarized in separate paragraphs.
The first paragraph should establish the requirement of the study. This can be done by highlighting the public health problem or clinical challenges, if any, which the condition poses supported by relevant references.
The second paragraph should briefly summarize the work already done on the research question which the present study is addressing. This review should be recent and synthetic rather than pedantic and old. The reader should be able to get a quick insight into the state of the art.
The next paragraph should summarize the limitations or the controversies identified in the previous studies. The limitations brought out should be specific. For instance, earlier studies may have used a different study design lower in the hierarchy of EBM, smaller samples, and less precise instruments for the measurements, and so on. Often, there may be no limitations or controversies in earlier studies, but they may have been done in different populations or settings. This may be mentioned.
The last paragraph is most important in deciding the novelty of the present paper. It should bring out in what way it has overcome the limitations of earlier studies or resolved any controversies from the previous research. Often, the author may not find any limitations or controversies in previous studies. In such cases, the setting and different patient sample may provide justification for the study or at the very least an attempt at replication. The novelty factor is missing in such papers. This can be compensated by a rigorous study design and methodology which serves the purpose of a replication study.
The next and perhaps the most important section is the “Material and Method.” It should begin with the details of ethical approval of the study including the letter number with date. All clinical trials have to be registered, and the trial registration details should be included in this study.
This section should restate the research question in specific terms and describe the predictor, outcome, and confounding variables, the potential confounders, if it is a cause-effect study. The methodology should describe in details what efforts were made to minimize the three types of biases which creep into medical research, i.e., selection bias, measurement bias, and confounding bias.
Selection bias is due to improper sample selection. The more the sampling method deviates from the simple random sample, the gold standard of random sampling, and the greater the selection bias. The simple random sample is a misnomer as selecting this poses the greatest logistics problems. The sampling methods with its limitations should be mentioned under material and methods. Besides the selection bias, which can be minimized by a random sample, the random error which depends on chance can be minimized and quantified by selecting a sample of adequate size. The inputs for sample size calculation such as the Type 1 and 2 errors, the effect size, standard deviation, prevalence, and width of confidence interval should be mentioned.
Measurement biases may be due to observer error, instrument error, participant's recall bias, the placebo effect in the participant, and so on. Quality control measures to eliminate these measurement biases should be described in detail. All that come in the results should have been described in the material and methods section. For instance, if body mass index (BMI) is mentioned in the results or tables under results, how BMI was calculated should have been described in the materials and methods. This should include how weight and height were measured ensuring the elimination of instrument and observer errors. And vice versa. If any methods of measurement have been described, then the outcomes of that measurement should be mentioned under the results section. Each section should seamlessly merge and synchronize with the next.
If any laboratory kits and reagents were used, description of these along with details of manufacturers should be included in this study.
Confounding bias, the most challenging, may be reduced by identifying the potential confounders and dealing with them appropriately. These should be measured as meticulously as the predictor and outcome variables. In observations studies, potential confounders such as age, gender, socioeconomic status, smoking status, alcohol use, and others depending on the research question are controlled by three methods, i.e., matching, stratified analysis, or multivariate analysis. In the planning stage, matching may be resorted to in which both the exposed/cases and control groups should have a similar proportion of participants with the likely confounder. During analysis, either stratified or multivariate analysis, may be undertaken to control for the effect of these confounders. The problem is with unknown confounders, the Achilles heel of observational studies. The study design which can address the bias due to unknown confounders is the randomized controlled trial which places it at the highest level of evidence in EBM.
Finally, the material and methods should include the details of statistical analysis and the software with version and manufacturer details.
Next is the results section which is the major scientific contribution. It should bring out the findings related to the research question, as stated in the introduction and how to address it as described under material and methods. The main results should be easy to locate complemented by tables and figures. Results may be descriptive, quantified by summary statistics or analytic conveyed by P values and confidence intervals. Giving only P values with significance levels is not sufficient. The effect size, i.e., the mean difference between the groups or the magnitude of the relative risk or odds ration along with their 95% confidence intervals should be mentioned.,
For comparing two groups, the main emphasis should be on the effect size which quantifies the difference between the groups. For each effect size predictor and outcome, variables should be stated. Measurement scales, i.e. the types of data decide the summary statistics and statistical tests of significance. In observational studies, the results of adjusting for confounders should be mentioned.
The text should complement tables and figures highlighting the points of interest instead of describing all information given in them. Any result not proposed in the original research question, however interesting, should be put in proper context. Data dredging can generate occasional and meaningless significant results due to chance.
The discussion should bring out the implications of the results and the way forward for public health or clinical practice. The common pitfalls revolve round repetition of the results, boring summaries of previous studies, cursory mention of few limitations, and insipid conclusions. All these add up to a dull discussion.
Discussion should stress on the following. What exactly did the study show? What is the relevance of the results? Could the results be interpreted in a different way? Have other studies come to the same conclusions or there is disagreement? If so, what can be the likely reasons? What are the strengths and limitations of the study? How well could it eliminate or control for selection, measurement, and confounding biases? In what direction the limitations of the study are likely to skew the results? A closing paragraph should bring out the conclusions and recommendations based on the study results.
| References|| |
Sollaci LB, Pereira MG. The introduction, methods, results, and discussion (IMRAD) structure: A fifty-year survey. J Med Libr Assoc 2004;92:364-7.
Bowers D, House A, Owens D. Understanding Clinical Papers. 2nd
ed. West Sussex: John Wiley and Sons, Ltd; 2006.
Browner WS. Publishing and Presenting Clinical Research. 3rd
ed. Philadelphia: Lippincott, Williams and Wilkins, A Wolters Kluwer Business; 2012.