How I Evaluate Research

How Research Quality Is Evaluated

Not all research is equally informative. A single dramatic study should not be held up as proof of harm. On the other hand, a single inconclusive study shouldn’t be grounds for dismissal. A lack of research on a topic doesn’t mean it’s not worth paying attention to.

Interior Medicine evaluates research based on what model was studied, at what dose, under what conditions, what the authors actually concluded, and what the broader body of evidence says.

Four questions are applied to any single study:

• What was studied: Humans, animals, or cells in a dish. Cell and animal studies are useful for screening but do not always translate to humans. Observational studies in humans are stronger but cannot prove causation alone.

• At what dose: Doses 1,000 times higher than typical residential exposure are common in screening studies. An effect at high doses does not automatically scale down to an effect at low doses. Many substances have threshold effects or non-linear dose-response curves.

• Under what conditions: Sealed chamber at elevated temperature is not equivalent to regular use at room temperature. An effect under extreme conditions is not the same finding under normal ones.

• What the authors concluded: Paper titles often imply more certainty than the conclusion section supports.

Beyond individual studies, weight is given to the broader body of evidence.

• Strong evidence: Multiple independent research groups, study types, and regulatory agencies have reached consistent conclusions across countries and years. This is where meaningful health guidance is established.

• Substantial but contested evidence: A real body of research exists and the hazard is real, but mechanism, threshold, or significance at realistic exposure levels is still debated.

• Limited evidence: Some research exists but is preliminary, narrow in scope, or conducted in models distant from human biology.

• Data gap: No meaningful research exists on the specific question. Absence of evidence is treated as a data gap, not as reassurance or cause for dismissal.

An in-depth, example-based version of this approach is outlined in Part 3 of my course, How to Read a Toxicity Study in 4 Steps.