Principles of Scientific Writing in Biomedical Sciences
The structure of the abstract should be such that it enables the reader to assess the study hypothesis and methods quickly and easily.
- The context for research question and the hypothesis (objective) should be clearly stated (For example, To determine whether Enalapril reduces left ventricular mass…)
- Methods: Mention clearly the study design (randomized control trial, crossover trial, cohort study, etc.), the population (diabetic patients, epileptic patients, asthma patients, etc….), and setting from which the sample was drawn. A basic explanation of statistical analyses (For example, The screening test was validated using a bootstrap procedure and performance tested using an ROC curve)
- Mention the main outcome (prognosis) of the study
- Results: Some explanation of the effect size, if appropriate, with point estimates and confidence intervals to describe the results
- Conclusion: There should be no over interpretation of the data
- A concise review of the relevant literature to provide the context for the study question and a rationale for the choice of a particular method
- The study hypothesis must be clearly stated in the last sentence before the methods section
- No results and conclusions must be presented in this section
This section should enough information to enable a knowledgeable reader to reproduce the study and verify the results with the reported data. Components should include as many of the following as are possible parameters
- Study Design (Refer Types of Study Design)
- Year(s) and (month if appropriate) in which the study was conducted
- Disease or condition to be studied–how was it defined?
- Setting in which subjects were studied (community based, referral population, primary care clinic, volunteers)
- Subjects studied–who was eligible; inclusion and exclusion criteria. If all the subjects were not included in the analysis, reasons for exclusion; information consent and institutional review board approval when appropriate. If results for any of the subjects have been previously described, provide citations for all reports or ensure that different reports of the same study can readily be identified (e.g., by using a unique study name)
- Intervention, including the length of intervention and enough information to allow a knowledgeable reader to reproduce the intervention, or define the exposure adequately to allow comparison of different studies.
- Outcomes and how they were measured, including reliability of measures and whether investigators determining outcomes were blinded to which groups received the intervention or underwent the exposure.
- Independent variables and how they were measured–for example, demographic variables and risk factors for the disease.
- Preliminary analysis: if the current study is a preliminary analysis of an ongoing study, then the reason for publishing data before the end of study should be clearly stated along with information about when the study is to be completed.
- Source to obtain original or additional data if somewhere other than from the authors. For example, data tapes are often used from the US government; the source should be stated. The National Auxiliary Publications Service (NAPS) and the World Wide Web can be used to store or display data or information that could not be included in the manuscript. The source may also listed in the acknowledgment.
- Statistical methods, including which procedures were used for which analyses, what level was considered acceptable, power of the study (if calculated before the study was conducted), assumptions made, any data transformation, multiple comparison procedures performed, steps used for developing a model in multivariate analysis, and pertinent references for statistical tests and types of software used. Results should be presented in terms of confidence intervals wherever possible.
- If the study has been registered in a central trial registry, the name of the registry and the trial number should be provided.
Power and Sample Size
Power: the ability to detect a significant difference with the use of a given sample size and variance; determined by frequency of the condition under study, magnitude of the effect, study design, and sample size. Power should be calculated before a study is begun. If the sample is too small to have a reasonable chance (usually 80% or 90%) of rejecting the null hypothesis for a true difference, then a negative result may indicate a type II error rather than a true acceptance of the null hypothesis.
Power calculations are important to perform when designing a study; a statement providing the power of the study should be included in the methods section of all randomized controlled trials and it is appropriate for many types of studies.
A power statement is especially important if the study results are negative to demonstrate that a type II error was not the reason for the negative result. Performing a post hoc power analysis is controversial, especially if it is based on the study results, but if included, it should be placed in the discussion section and the fact that it was performed post hoc must be stated clearly.
Example: We determined that a sample size of 800 patients would have 80% power to detect the clinically important difference of 10% at = 0.05
Multiple comparison procedures: any of several tests used to determine which groups differ significantly after another. A number of general tests has identified that a significant difference exists but not between which groups. These tests are tended to avoid the problem of type I error caused by sequentially applying tests as the t test not intended for repeated use.
Some test result in more conservative estimates (less likely to be significant) than others. More conservative tests include the Tukey test and the Bonferroni adjustment; the Duncan-multiple range test is less conservative. Other tests include the Scheffe test, the Newman-Keuls test, and the Gabriel test.
Type 1 error: data demonstrating a statistically significant result, although no true association or difference exists in the population. The level is the size of a type I error that will be permitted, usually. .05
A frequent cause of a type I error is performing multiple comparisons, which increase the likelihood that a significant result will be found by chance. To avoid a type I error, one of several multiple comparisons procedures can be used.
Bonferroni adjustment: statistical adjustment applied when multiple comparisons are made. The level (usually 0.05) is divided by the number of comparisons to determine the level that will be considered statistically significant. Thus, if 10 comparisons are made, and of 0.05 would become = 0.005 for the study. Alternatively, the P value may be multiplied by the number of comparisons, while retaining the of 0.05. For example, a P value of 0.02 obtained for 1 to 10 comparisons would be multiplied by 10 to get the final results of P = 0.20, a nonsignficant result.
The Bonferroni test is a conservative adjustment for large numbers of comparisons (i.e. less likely than other methods to give a significant result) but is simple and used frequently.
Duncan-multiple range test: modified form of the Neuman-Keuls test for multiple comparison
Newman-Keuls test: type of multiple comparisons procedure used to compare more than 2 groups; the first thing is to compare the 2 groups that have the highest and lowest means. Then, we sequentially compare the next most extreme groups, and stop when a comparison is not significant.
Multivariate analysis: Any statistical test that deals with 1 dependent variable and at least 2 independent variables. It may include nominal or continuous variables, but ordinal data must be converted to a nominal scale for analysis.
Compared to bivariate analysis, the multivariate analysis has three advantages:
- It allows for investigation of the relationship between independent and dependent variables while controlling for the effects of other independent variables.
- It allows several comparisons to be made statistically without increasing the likelihood of type I error
- It can be used to compare how well several independent variables individually can estimate the values of the dependent variable.
Some examples of multivariate analysis are as follows: Analysis of variance, multiple (logistic or linear) regression, analysis of covariance, etc.
This section must include the following
- Number of subjects in the study at its inception
- Statistics describing the study population
- The Number of subjects excluded, dropped out, lost to follow-up
- Discussion of prognosis (primary and secondary outcomes)
- Discussion of post hoc analyses, but the content should clearly mention this analyses: it is used to generate hypothesis, NOT for testing hypothesis.
- If one statistical test is used through the study, then mention it in the Methods Section. If more than one statistical test has been used, then they must be discussed in the Methods Section and the specific tests used must be reported along with their results in the Results Section
This section would elaborate the following
- Whether the hypothesis was supported or refuted by the results must be elaborated
- Study results must be interpreted in the context of published literature
- Discuss the limitations of the study, including possible sources of bias that affect the generalization of results; they would create issues with conclusions
- Evidence to support or refute the problems introduced by limitations
- Implications of clinical practice
- Specific directions for future studies
- Conclusion should not be beyond data; it must be based on the study results and population.