References should help readers check a claim, not just make a page look well researched. This policy explains how authors and editors at FreeBooks4Doctors.com should find, assess and cite the evidence behind medical articles.
Start with the question
Start by deciding what the article needs to answer and who will read it. A parent, a medical student and someone comparing treatments may need quite different explanations. The country, patient population and care setting matter too. Make those limits clear before searching.
Record the date your research search was completed. Also, keep enough detail so an editor could duplicate your process. An article must not be described as a systematic review unless it used and reported an appropriate systematic method.
Choosing evidence
When researching issues related to the delivery of care, authors must first seek out current clinical guidance, and high quality evidence synthesis. Depending on the question being asked, the author may also need to research the original trial(s), observational studies, diagnostic studies or general reference materials.
Your initial searches may start with recognised public health organisations, professional associations and published scientific literature. PubMed is a helpful resource to use when locating biomedical publications. However, simply identifying a paper in PubMed does not eliminate the responsibility to evaluate the content of the paper. NICE and other national clinical guideline groups may be very useful resources when writing about the delivery of healthcare within their jurisdiction.
Ultimately, the best source is the one that supports the specific claim you want to make. Textbooks provide an excellent foundation for understanding fundamental anatomy but would be insufficient when verifying recent warnings regarding medication safety. Similarly, while small case reports may illustrate possibilities they do not demonstrate frequency.
Read beyond the headline
Read the methods as well as the conclusion. Authors and reviewers should consider the design, who took part, the comparison group, the outcomes measured and how long follow-up lasted. Funding, declared interests and the study’s limitations also belong in that assessment.
Findings in animals or laboratory settings cannot be claimed as proven benefits for humans. Associations cannot be implied as causal relationships. Findings from one population may not generalize to others such as children, pregnant women or those with comorbidities.
In most cases it is reasonable to read the entire article rather than rely solely on an abstract. However, if only an abstract is available, do not assume that additional details (not found in the abstract) have been validated.
Make the references usable
Each significant clinical assertion should provide clear access to its supporting source (i.e., link to or clearly identified reference). Each reference should contain sufficient detail to locate the referenced material: typically, this includes title, author or organisation name, publication year, and a DOI or similar stable identifier, where available.
Do not create a reference for something you did not see. Do not imply that a source provides more information than it actually does. Editors should confirm that hyperlinks lead to the correct content and that quotes or paraphrases match the original content. Social media posts, paid advertising summaries and search engine snippets should never be the primary basis for providing clinical advice when more substantial evidence exists.
Readers should not have to buy access to understand our article. When a cited paper is behind a paywall, an accessible abstract, guideline or public explanation may help. We must be clear about which material was actually checked.
Handle uncertainty openly
When credible sources present conflicting opinions the article must discuss the conflict and implications. The article should not selectively highlight findings consistent with the desired conclusion.
Contextualize numbers. When presenting risk data indicate the population and time frame involved. Compare absolute vs. relative differences where applicable (as the two are interpreted differently). Indicate whether preliminary findings and/or preprints represent preliminary data. These types of data should not supplant existing clinical guidelines without a clear explanation and medical review.
Retractions, corrections and updates to clinical guidelines may necessitate updating an article. The “date completed” indicates when the last evidence search occurred; it is not a guarantee that subsequent studies have been evaluated.
Books, recommendations and commercial links
A freely accessible book may still be old, incomplete or intended for a different country. Directory entries should identify the edition or publication date when available. Inclusion in the directory is not a medical endorsement.
A commercial relationship must not determine which clinical conclusion an article reaches. Paid or affiliate links require disclosure under our Conflicts of Interest and Funding Policy. If you find a missing, broken or misleading citation, use Contact Us and include the article address and the reference concerned.