07.02.2017 Views

Dan Mayer Essential Evidence-based Medicine

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

The medical literature: an overview 29<br />

Methods<br />

The methods section is the most important part of a research study and should<br />

be the first part of a study that you read. Unfortunately, in practice, it is often<br />

the least frequently read. It includes a detailed description of the research<br />

design, the population sample, the process of the research, and the statistical<br />

methods. There should be enough details to allow anyone reading the study to<br />

replicate the experiment. Careful reading of this section will suggest potential<br />

biases and threats to the validity of the study.<br />

(1) The sample is the population being studied. It should also be the population<br />

to which the study is intended to pertain. The processes of sample selection<br />

and/or assignment must be adequately described. This includes the eligibility<br />

requirements or inclusion criteria (who could be entered into the<br />

experiment) and exclusion criteria (who is not allowed to be in the study<br />

and why). It also includes a description of the setting in which the study<br />

is being done. The site of research such as a community outpatient clinic,<br />

specialty practice, hospital, or others may influence the types of patients<br />

enrolled in the study thus these settings should be stated in the methods<br />

section.<br />

(2) The procedure describes both the experimental processes and the outcome<br />

measures. It includes data acquisition, randomization, and blinding conditions.<br />

Randomization refers to how the research subjects were allocated<br />

to different groups. The blinding information should include whether the<br />

treating professionals, observers, or participants were aware of the nature<br />

of the study and if the study is single-, double-, or triple-blinded. All of<br />

the important outcome measures should be examined and the process by<br />

which they are measured and the quality of these measures should all<br />

be explicitly described. These are known as the instruments and measurements<br />

of a study. In studies that depend on patient record review,<br />

the process by which that review was carried out should be explicitly<br />

described.<br />

(3) The statistical analysis section includes types of data such as nominal,<br />

ordinal, interval, ratio, continuous, or dichotomous data; how the data are<br />

described, including the measures of central tendency and dispersion of<br />

data; and what analytic statistical tests will be used to assess statistical relationships<br />

between two or more variables. It should also note the levels of α<br />

and β error and the power.<br />

Results<br />

The results section should summarize all the data pertinent to the purpose<br />

of the study. It should also give an explanation of the statistical significance<br />

of the data. This part of the article is not a place for commentary or

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!