For example, if the question is whether increased heat shock protein (hsp) 70 results in improved myocardial recovery in a dog model of stunning, then search the keywords “hsp70” and “stunning.” If instead the question is whether hsp70 is associated with improved vascular reactivity in the dog coronary artery, then search the keywords, “hsp70” and “coronary bioassay.” If seeking out papers on the prognosis of a condition, searching for the MeSH term “prospective studies” may be better. In the case of a new diagnostic test, searching for the relevant MeSH terms “sensitivity and specificity” may help gain the needed information.

To gain new leads for references, look at the associated keywords and assigned MeSH terms, as well as explore reference lists at the end of retrieved articles that are particularly relevant. Learn what the key journals are in your chosen field. Journals generally have a systematic peer review process, but it is important to remember that the skills and biases of individual reviewers and editors may create some limitations. It is important that the literature search is thorough; realize that no one source will likely be complete for all needed references. For example, while PubMed is an excellent source of information, not all journals are indexed there. Or there may be similar key terms that you have not searched, resulting in a significant portion of the literature being left out.

When reading a journal article containing primary research, many aspects must be considered. The content of the literature and its conclusions cannot simply be accepted at face value. What is the goal of the study? What is the hypothesis is being tested? Did the methodology actually test the hypothesis? Is the methodology chosen flawed? Are there clear inclusion and exclusion criteria when appropriate? How is the data presented? What statistical tests and what level of significance were chosen? Is the analysis of the results appropriate? Are the discussion and conclusions supported by the data? While these are a lot of factors to consider, these considerations are important in deciding whether the conclusions of the different studies are valid.

Recognize what information to expect when reading an original research article [1–4]. The Introduction of the article describes the broad problem to be addressed and provides the background for the basis of the study. This section should end with a purpose and Hypothesis, which provides a framework for the scientific study.

The next section is Materials and Methods. While it is tempting to give this section a cursory look, reading this section carefully is important. What techniques were chosen to test the hypothesis and how the data will be interpreted need to be clearly outlined? The methodology should be completely provided through direct description or appropriate referencing. If significant gaps exist, then the validity of the paper comes into question. Additionally, if the experiments performed do not address the hypothesis, nor use appropriate methodology or analysis, the study results and conclusions will likely not be valid. Deciding if the methodology is “appropriate” may not be easy this early in your research career and may warrant input from your mentor or others with more experience.

The Results section should provide a logical layout of the data, follow the experimental design outlined in the methods, and should either support or refute the hypothesis. Enough raw data should be available to allow independent assessment by the reader. Not all data may be present in the body of the written text; supplemental data may be referenced and available through the Internet. The appropriate analyses of results are critical as different analyses can yield different conclusions. The analysis choices may be difficult for you to independently assess, but accepting the authors’ results and analyses at face value may lead to inappropriate conclusions. Common problems include statistical method and sample size determined after the study is complete, lack of difference found when there is one (Type II error) because of inadequate sample size, lack of power calculations to determine appropriate sample size, and multiple outcome measures or comparisons in the absence of appropriate multivariate procedures.

Finally, the Discussion and Conclusion section is reached. Again, do not simply rely on the interpretation and opinions of the authors; make your own assessment of how the methodology and results tested the hypothesis. However, this section is where the author provides their perspective of the problem studied and interpretation of the data and puts the study findings in context with the existing literature. In the discussion section, be aware of (1) the potential for author bias to influence the interpretation of the results, (2) findings and analyses unrelated to the study hypothesis or stated study objective (such findings should be a basis for further studies and should not support a firm conclusion), and (3) conclusions beyond the data presented or applying the study findings to other models of study. If the conclusions are not supported by the data, then the reference may not be valid.

Another source of information can be by way of conference proceedings. Conference proceedings can contain abridged manuscripts and abstracts [1]. Sometimes they are available electronically as part of meeting programs at society websites, are published in a major journal, or are only located in the program book of a society scientific meeting. Although the information is not in depth for a detailed assessment and the review process is less rigorous than for a journal (and sometimes chosen for the novelty of the idea to generate discussion), the information is often more timely than books and journal articles, and knowing the latest work in your specific area of study is useful.

After a detailed literature search, summarize your knowledge of the subject. You must know the topic well to be able to ask the right questions. What is the most relevant literature? What is known about what you are studying? Where are the gaps in knowledge? What questions need to be answered? Understanding what is not known and how that knowledge could potentially affect outcome allows you to refine the questions you want to answer and the methods for addressing them.