Start with the question a paper asks, then work outward to what its results can support.
Identify the study model
Look for whether the work used a cell system, an animal model or human participants. Those settings answer different questions.
Check what was measured
Find the endpoint, timing and comparison condition. An observed change in one measurement is not a complete account of an effect.
Read the limitations
Check sample size, study design, potential bias and whether other work has reproduced the finding. Preserve uncertainty when summarizing the paper.
Return to the original source
Keep the citation and consult the methods and results rather than relying on a headline. PubMed is one place to find indexed literature.