## Problems with medical numeracy.

One of the big problems in medical and dental research is getting used to making sense of the statistics. Not only at the reader level but more importantly at the level of the researcher.

A couple of years ago I was researching for a presentation on clinical risk and did not have access to an electronic medical library like I do now. To check some of the facts I had to go back to some of my old dental books published in the 1990’s. This is still possible since some fundamentals of medical science have not changed in the past 50 years and what do you think I noticed?

Medical science is now at least 100x more precise than it was then say in 1975 (Hamp, S. Nyman, S. Lindhe 1975) not because the data is better but that the answer is to 4 significant figures. It’s not more accurate, it just appears more precise and this false precision is dangerous. It gives confidence and certainty where none exists.

Much of this problem could be resolved if we understood the meaning of ‘significant figures’ and that the answer for say average survival should not be more precise than the input data. Here is an example from a recent paper I reviewed:

“5.6% (4/72) teeth were effected”

In division, the answer should not have more significant figures than the divisor so instead of 5.6% it should be just 6%, otherwise, the answer is 10x more precise than the unit being measured.

To help me improve my numeracy and general feel for the numbers I just bought a 1963 Faber-Castell slide rule from eBay. The advantages beyond not needing electricity are many, it:

• demonstrates elegant simplicity (few pieces, only two moving parts).
• has total transparency (every register and scale is readily visible).
• works with numbers of any size
• has precision to 3 significant figures with an accuracy of 0.1%
• functions include multiplication, division, chain calculations, reciprocals, squares, square roots, powers and logs plus, plus, plus.
• easy to learn and simple to operate.
• provides automatic parallel computations.

Interested in Slide Rules? Take a look at the  The Oughtred Society.