The title of the article has to be studied carefully and remembered because the rest of the article i.e. introduction, methods, results, discussion and conclusion should justify the title. At times there may be a need for change in the title.
The abstract of the article should convey the intellectual plot of the article. Check whether there is a hypothesis or thought process behind doing this research. One need to recognize whether it is an original research article which is based on prospective series of cases, case controlled study or randomized controlled study or an epidemiological study or an audit or opinion paper. This distinction will help to scrutinize the article better as the methodology and results should reflect this.
The key words will help one to check similar articles and to compare this study with what exists in the literature. Based on the abstract the article can be assigned into a) Interesting but is not true e.g. pre-eclampsia was due to some worms in the blood); b) True but not interesting (Prostaglandin 2 mg is better than Foley’s catheter for induction of labour) and c) Interesting and may be true (Uterine transplantation and pregnancy). The third variety is the one that journal should publish to attract more readers.
The introduction must state why the study was needed, its importance, whether it is a new concept that is being explored or it has been done in the past but it is repeated. If so there should be statement/clear explanation as to why it is repeated e.g. earlier studies did not have enough numbers or there were inconclusive results, or the technology or medicine to do the study had improved, etc. If the author claims that this is the first time this study has been done or states that it is the largest study on the subject it will be easy to check this by doing an internet search.
Methods should mention ethics committee approval to conduct the study and appropriate consent has been taken where it is needed. There should be detailed description of the methodology used. If it is a lab experimentation, the details of the experiment should be provided or an appropriate reference given where the methodology is described so that another group can repeat the experiments and endorse or disagree with the results. The numbers that need to be studied should be identified based on power calculation which in turn depends on the incidence of the condition studied and the primary and secondary outcome measures. As a reviewer one should be satisfied with the methodology and how the primary and secondary outcome indicators are defined and measured. Methods used should be appropriate and adequate to achieve this outcome. In a RCT it should spell out how the randomization was carried out (e.g. individual case based randomization, cluster randomization, wedge design study, etc.) and whether it was pre-determined that the results would be analyzed using the intention to treat basis.
Results should be given in a clear concise way preferably using tables and figures with the statistical analysis so it is clear and easily understood by the reader. Results need not always show the expected results. The analysis should provide the basic characteristics of the subjects studied to show that these characteristics were unlikely to influence the results. If there is doubt about power calculations or the statistical significance achieved then please ask the journal to send the article to have a review by a statistical referee. Note what they have claimed is what they originally stated that they were trying to achieve. If there is doubt or discrepancy please ask them to send you a copy of the research ethics committee application.
Discussion should provide the argument that their findings have advanced the knowledge to improve basic understanding of the issue or has provided means of improving clinical practice. The authors should demonstrate that they have proven the hypothesis on which the research was performed. The significance of the research and its reach should be clear and will determine whether the article is worthy of publication. References should be appropriate and as indicated in the text.