script src="" type="text/javascript">
STATS at George Mason University    

Body vesalius


How to evaluate health risks


Every day, new health studies fill the media, many of them often contradicting each other and earlier research.

For years, women were told that hormone replacement therapy could reduce their risk of heart disease, stroke and Alzheimer’s; then new research found that it may actually increase the risk for these disorders; then closer analysis of the new data showed that increase only applied to older women — and that HRT may protect against heart disease for younger women.

Fish is a wonderful source of protein and Omega-3 fatty acids, which may reduce the risk of heart disease, and mitigate the effects of alzheimers. But what about studies claiming that many fish are contaminated with high levels of mercury? Chocolate: artery-clogging bad guy or artery-clearing hero?

The way health stories are usually covered by the media, it’s often impossible to tell. The latest news is often rushed into print, over-simplified, or misunderstood

But the following guidelines can help you make sense of the latest risk-touting study, and help you to decide when you should start — or stop — taking a drug, pursuing treatment or changing your diet.



1. Cause or effect?
Only “experimental” or “controlled” studies can prove whether a particular drug or activity is the cause or the
cure of a particular illness.

2. Has the research been replicated?
Even the best-designed study in the world can occasionally produce impressive looking results by chance

3. What was the dosage?
The ancients knew that what distinguishes medicine from
poison is often in the measure.

4. Is that me?

Who participated in the study?

5. Is the risk relative or absolute?

Be skeptical of shocking numbers.

6. Anecdote is not evidence

Clinical observations, like other anecdotes about health,
are subject to all manner of biases.

  7. Follow the money
The source of a researcher’s funding can affect the
outcome of the research.

8. Where was it published?
Prestige and peer-review count.

9. Is there publication bias?

Journals are three times more likely to publish studies
of treatments showing positive results than those
which show no effect.

10. Are the pleasure police on patrol?
If it feels or tastes good, it’s probably bad for you — or is it?

11. Is the risk worth the benefit?
The opportunity costs may not be worth the risk.

12. Science v journalism
Journalists and scientists tend to have
very different perspectives on research.
email this page to a friend

STATS home