| reply stolen from elsewhere
It is important to differentiate between absolute risk and relative risk. The absolute risk for alzheimer's in this study was 17% for non-diabetics and 28% for diabetics, a difference of 11%. The relative risk, 28% vs 17%, meant that 65% more diabetics than non-diabetics got alzheimers, but still 72% of diabetics didn't get alzheimers. A 65% difference in risk is much more dramatic sounding than an 11% difference in risk, so it is important to aware of which kind of risk is being talked about to begin to understand the significance of the results.
The smaller the absolute risk, the more inflated the relative risk can sound. I don't remember the exact numbers but in the heart protection study looking at the risk of heart attacks in poeple using zocor relative to those not using zocor the relative risk was fairly high, let's say on the order of 50% higher in the non-zocor group, but the absolute risk was low in both groups, maybe 4% in the zocor group and 6% in the non-zocor group. (I am making up these numbers but the principles is true) Then you have to know over how many years the risk was measure, 1 year, 5 years, 10 years, to figure out how many more people actually suffered a heart attack in any given year. if in a group of 100 people the risk of getting a heart attack was 4% in zocor group and 6% in nonzocor group that means that 4 people in the zocor group had a heart attack and 6 people in the non-zocor group had a heart attack. The relative risk was 50 % greater but it still amounted to only a difference of 2 people. If this was measured over 5 years, which I think it was in the zocor study, that would mean a difference of less than a person a year. Also the size of the group studied makes a difference. If you study a very large group then small differences can be statistically significant but actually only represent very small differences in absolute risk. I am afraid this is getting too complicated, but the point is we need to become savvy consumers of research in order to put the results of the studies into perspective. An 11% difference in absolute risk of alzheimer's certainly seems high enough to me to demand further study, but a difference of 11% doesn't sound nearly as bad a difference of 65%. Then you can ask how they defined diabetics, did they look at blood glucose control etc etc. This is preliminary research that raises some important questions, but it answers very few.  | 
