My colleague Gilles Frydman wrote a great entry today over at e-Patients.net about a new report released this month by the Association for Psychological Science. Don’t let the “psychology” fool you, as this report, entitled “Helping Doctors and Patients Make Sense of Health Statistics” (PDF), is about statistics in medicine and the shocking lack of knowledge about basic statistics amongst many physicians and health reporters.
The problem? Too many media outlets report on relative risks and single-event probabilities rather than absolute risks and frequency statements. For instance, reporters will often note that a new treatment helps twice as many people than an existing treatment. But if you look at the study, it may be that it went from 1 in 10,000 people helped, to 2 in 10,000 people helped. That’s a doubling of the treatment’s help, but as you can see, the vast majority still won’t see any beneficial effects from the treatment (including, likely, you).
What can be done? Gilles notes the report’s recommendations:
The authors advocate for teaching, starting in primary and secondary education and continuing in medical school, the methods of statistical thinking and transparent representations. This early teaching first requires familiarizing children with the concept of probability. It also requires transforming the meaning of statistical literacy as the art of solving real-world problems. The authors also note that a major precondition for statistical literacy is transparent risk communication. To change the current situation the authors recommend using:
- frequency statements instead of single-event probabilities,
- absolute risks instead of relative risks,
- mortality rates instead of survival rates, and
- natural frequencies instead of conditional probabilities.
Reporters should be responsible in help putting data into proper context and be sensitive to whether a finding is really going to impact anyone reading the article (or is simply a statistical artifact of intellectual curiosity).
While nobody expects everyone to become a statistical expert, one might expect treatment providers, such as doctors, and those whose job and responsibility it is to accurately report new research results (reporters) to have more than a passing background and grounding in basic statistics.
This report should be a must read for any health reporter, as well as for any treatment provider who relies on research to help inform their treatments.
Read Gilles Frydman’s entry: Lies, Damn Lies And Statistics: Collective Statistical Illiteracy
Read e-Patient Dave’s take on reading the report, too.
(7 votes, average: 5 out of 5)
Loading ...You can leave a response, or trackback from your own site.
One Comment to
“Lies, Damn Lies and Statistics”
Frequencies are relatively easy enough to understand when being presented with the probability something, along with others, will happen. Single-event probabilities are like one piece of the puzzle. Right now I am trying to do my best in Stats 101 so I can be of better service in graduate school (MA in Counseling Psychology). I really want to be able to read a legitimately published medical article and understand the formulas in them, why they were chosen, and if they were done well.
Join the Conversation! Post a Comment:
Last reviewed: By John M. Grohol, Psy.D. on 18 Nov 2008
