Originally Posted by
Dahon.Steve
If you look at figure 2 in the study, you'll see that quite a few of the original odds ratios, with their 95% confidence intervals, are disjoint. The original studies are not in agreement, so this study author throws them together in a hopper to make one big average.
The other interesting tidbit is in table 3. It shows that one of the studies indicates that for people over 25, helmets INCREASE risk!
Speedo