A study about an integrative or complementary approach should be evaluated by the same questions as other health research: who was studied, what was compared, what changed, how the results were measured and what remains uncertain. A treatment label or a promising headline does not answer those questions.
This checklist helps prepare a discussion with a qualified clinician, including a consultation at Dr. Fred Bloem’s practice. It does not endorse a particular therapy, certify a practice’s evidence standards or determine whether a treatment is suitable for you.
Find the actual research
Begin with the article rather than a testimonial, advertisement or news summary. Record its title, authors, publication date and link. Check whether you have the full paper or only an abstract; a short summary may omit important methods and limitations.
The National Center for Complementary and Integrative Health provides a guide to reading scientific journal articles. Use it to distinguish the methods, results and authors’ interpretation rather than treating the conclusion as the entire evidence base.
Identify the research question and design
Write the question in one sentence: which intervention, for which people, compared with what, and for which outcome? Check whether the work involved laboratory models, animals, observations in people or a clinical trial. A mechanism observed outside a clinical setting does not by itself establish a treatment benefit in patients.
In a trial, ask how groups were formed and whether there was an appropriate comparison. Random assignment and blinding can reduce some forms of bias, but a study’s label alone does not guarantee good methods. When blinding is difficult, look for an explanation of how the researchers addressed that limitation.
NCCIH’s sections on study design and bias explain why the comparison and the knowledge of treatment assignment matter.
Check who participated and for how long
Note the number of participants, their relevant characteristics, eligibility criteria and follow-up duration. Ask whether the people studied resemble the people described in the treatment claim. A result in one condition or patient group should not automatically be generalized to another.
Look for how many participants completed the study and how missing data were handled. Small samples and short follow-up may leave important uncertainty, especially about uncommon or delayed harms. A larger study is not automatically free of bias; size and design answer different questions.
Read the outcomes, not just the headline
Identify what the researchers actually measured. Was it a symptom, everyday function, quality of life, a laboratory marker or something else? Ask whether the outcome supports the claim being made. A change in a marker is not automatically a demonstrated improvement in how patients feel or function.
Look at the size of the difference between groups and its uncertainty, not only whether a result was called statistically significant. NCCIH’s results guide distinguishes statistical findings from clinical importance. A statistically significant difference may still be too small to matter to patients.
Ask how harms were examined
Check whether adverse events were actively collected, how they were defined and how long participants were monitored. A statement that no serious event was observed in a small study does not prove the intervention is risk-free. Missing safety information should remain a question, not become a reassurance.
Ask about interactions, the practical burden of treatment and what alternatives exist. Do not change prescribed treatment on the basis of one article without discussing the evidence and your circumstances with the responsible clinician.
Put the paper into the wider evidence
Compare its findings with other relevant studies and current professional guidance. A single favorable study rarely settles a clinical question. A systematic review can help organize the evidence, but its usefulness still depends on its methods, included studies and relevance to the proposed use.
Read the authors’ limitations, funding and disclosed conflicts of interest. A conflict does not automatically invalidate a result, and a lack of a disclosed conflict does not prove the study is unbiased. Check whether the conclusions go beyond the data and whether later evidence changes the interpretation.
Bring a short appraisal to the consultation
The exact treatment claim and the paper used to support it.
The population, comparison, outcomes and follow-up.
The size of the reported benefit and important uncertainties.
How harms, missing data and possible bias were addressed.
Whether other studies or guidance agree, and what remains unknown.
Ask the clinician how those findings relate to your options and what would change their recommendation. This turns a broad claim of “evidence-based care” into specific questions without assuming that a study proves personal benefit.
Is every randomized trial conclusive?
No. Methods, size, follow-up, outcome selection and relevance still need appraisal.
Does preliminary evidence prove that a treatment works?
No. It may justify further research while leaving the clinical benefit uncertain.
Can a checklist decide my treatment?
No. It supports an informed discussion alongside clinical assessment, alternatives and your priorities.
