“Linked to” and “causes” are easy to confuse in health news. A study may find that cannabis exposure and a health outcome occur together more often than expected. That association matters, but interpreting it requires more than replacing one phrase with another. The right questions concern timing, comparison groups, measurement and alternative explanations.
Start with how the study was designed
The NIH guide to clinical studies explains that researchers use different designs for different questions. Observational research records what happens without randomly assigning an intervention. Randomised controlled trials assign participants to groups, helping researchers assess whether an intervention produces a difference.
Observational evidence is valuable, especially when a trial would be impractical or unethical. Its limitations should shape the wording of a claim. A snapshot survey, for example, may show that two things coexist without establishing which came first. Following people over time can clarify sequence, although sequence alone does not eliminate other explanations.
Look for confounding and reverse causation
A confounder is a factor related to both the exposure and the outcome that can distort their apparent relationship. Consider a hypothetical survey finding that people reporting cannabis exposure also report poorer sleep. Stress, existing illness, working patterns or other substances could influence both answers. Alternatively, sleep problems might have preceded the exposure.
This example illustrates possibilities, not an explanation for any particular study. Researchers can adjust analyses for measured factors, but “adjusted” does not mean every relevant difference has disappeared. Factors may be omitted, recorded inaccurately, or represented by categories too broad to capture meaningful differences.
Keep the exposure specific
A label such as “cannabis use” may combine different histories, substances and methods of exposure. That makes broad headlines difficult to interpret. The CDC's heart-health overview notes that much research linking cannabis with heart attacks and strokes concerns smoking, and that separating cannabinoid effects from the effects of smoke is difficult. It also identifies the need for further cardiovascular research.
That uncertainty does not establish safety. Equally, an association observed in a particular setting should not be presented as a precise risk estimate for every person or every cannabinoid. Responsible reporting preserves both the warning signal and the boundaries of the evidence.
Read the numbers behind the headline
Suppose a hypothetical outcome occurs in two people per thousand in one group and three per thousand in another. The relative increase is 50%, while the absolute difference is one additional person per thousand. Both descriptions are mathematically valid; neither alone describes the full picture. These invented figures are not cannabis risk estimates.
- Ask for the baseline risk and the period over which outcomes were measured.
- Check whether a reported number describes risk, odds, or another measure.
- Look for uncertainty intervals and the authors' stated limitations.
- Check whether the result was the main planned analysis or one of many exploratory comparisons.
Make a conclusion that fits the evidence
Good reporting asks whether findings are consistent across studies and plausible alongside other evidence. One study rarely settles a complex health question. “Association does not establish causation” should prompt closer reading, rather than automatic dismissal. For readers, the practical goal is a proportionate conclusion: take credible safety signals seriously, avoid exaggerated certainty, and discuss individual health concerns with a healthcare professional.