Lucious Hellcat:
Sorry this was so long in coming, last time the computer bluescreened on me and it's taken a while to get back to it. I'll try to be clear in my response.
You make a good point about torsional injuries that occur in various sports, including equestrian and skiing. However you must consider that the masses involved in an automobile accident, as well as the speeds involved in airbag deployment, bear dramatically different physiological effects.
So, you've been struck in your Honda Element, "T-Boned" perpendicular on the drivers side by a crazy person. Here's what happens.
Your car moves to the right. However, you are not attached to your car, so inertia acts upon you, essentially causing a force from your body, into the door, to the left.
The driver moves from right to left, into the door.
As the driver moves, the SRS seatbelt tensioners engage around the left shoulder and waist, decellerating the driver in their approach into the door. However, the right shoulder is not attached, so the driver's body turns toward the door, and they will be striking the door in a facing/diagonal manner.
The side impact airbags deploy. The only way to countermand a force from right to left, is with a force from left to right. The bags deploy from within the driver's side seat at app. elbow level (see owners manual), and canopy forward and to the right.
The side impact airbags strike the driver with a force from left to right.
The driver's body is....hopefully....stopped from going into the door as their left shoulder comes in contact with the airbag. However, the left-to-right force from the airbag also must stop the encroaching right shoulder, which is twisting from back right (against the seat facing forward) to forward left (towards the and facing the door). In doing this, a strong force vector must be generated opposite that circular vector.
A force is applied against a twisting body.
When this happens, again...inertia takes over. Your head is attached to your body by just about the same relative strength that your body is attached to the car. When we're talking about twisting and the speeds involved, that's just about nil. Your head continues to whip to the left against a...hopefully....now stationary body.
If your head does this, and even maybe moves sideways (left ear to your shoulder) you could make it with just some achyness. If the airbag actually contacts your head, or if it pushes your elbow up so that the outside of your left arm pushes against your head, there will now be a force to the right which exceeds the decellerating force on your body.
Your head moves from left to right while your body moves from right to left, or remains stationary.
This torsional velocity has been present in a 'statistically significant'** number of side-impact crashes where the victems received concurrent nerve damage, and a connected loss of olfaction - sense of smell.
Hope that clears the way for you.
You can see how securing the right shoulder with a Y-strap that clicks in above the right shoulder would help.
**statistically significant - refers to an analysis of variance (ANOVA) probability value (P-Val), in comparisons done for side impact with nervous injury, side impact with no nervous injury, and impact (non-side)with and without nervous injury. A conclusion is significant when a P-Val is gleaned that is .05 or less. That is, if your hypothesis is that there might be a connection between side impacts and nerve damage, a P-Val of .05 states that there is a 95% level of confidence that your hypothesis cannot be rejected. That does not prove the connection, it just means that you cannot reject the hypothesis.