Helmet manufacturers recommend replacement after a time period due to material aging, UV, and experience with their products from field evaluations. Every time your helmet takes a hit - thing like dropping it on a hard surface - you cause compressive damage to the material of the helmet that can cumulatively lessen it's ability to provide protection.

Qualitatively, as a ski patroller, when I look at my helmet after three or so years of use, it's got dings all over it from my handling of it and from getting normal bumps in the head in the course of it's skiing usage. Same goes for my bike helmets. We use them a lot and they are not babied but used in real life. The material in them is designed to crush on impact and it does not spring back, hence the damage is cumulative.

Furthermore, our son had a bad ski crash that resulted in a severe TBI. His helmet was obviously no longer any good based upon the crash that he experienced. However, fairly involved inspection of his helmet revealed no apparent damage other than that caused by first responders (cut chin strap). The only way to tell was to have the helmet examined by the manufactured to make sure that the foam had not been crushed significantly - in other words, the assessment tools and processes are pretty much out of the scope of being done by an end user. There is no way to really tell if your helmet has retained its full protection capability without sending it in the the manufacturer for testing (which they will do). But in the meantime, you have no helmet.

For those reasons, and because the testing is difficult, it's a reasonable thing to replace helmets on a schedule or after they are involved in a crash. In our case, we choose to follow the manufacturer's recommendation because we think it's prudent and because we're seeking the maximum effective protection (why wear one otherwise?). Materials age and we don't necessarily take care of things in an ideal manner. Our helmets are heavily used.

In a crash, there is a non linear response of the brain to injury. As the acceleration and forces on the brain increase, the damage is relatively low as the normal protections of the head mitigate injury. As the accelerations and forces reach the point at which damage starts to accelerate, the damage increases disproportionately to the increase in accelerations. As a result, a little decrease in the accelerations can pay a significant dividend in the reduction of severity of the injury. This suggests that little decreases in the effectiveness of the helmet can be significant.

Then there is the impacts of a TBI - even a mild to moderate one - which can have lasting and permanent impacts on the patient. The cost of the helmet is, by far, the least important piece of that equation. Cognitive issues, motor control, and virtually any operational system of the human body, all can be permanent disabled to a greater or lesser degree. Even a mild impairment is not worth the risk to me/us. The risk/reward equation - after having first hand experience with this - tips heavily for us to agreeing with replacing helmets on a schedule. The probability of a problem is relatively low but the potential for catastrophe is high should the circumstance arise.

This doesn't disprove anything. It simply proves that a collection of helmets of which there is nothing known about their use or maintenance were shown to be able to pass a test that has been increasingly seen to be non representative of actual use in practice.

For example, if I had a bike helmet and I rode my bike once a year and the helmet was then safely stored in a closet the rest of the 5 years minus 5 days it was used, I'm positive it would pass this test. I'm also sure that a huge percentage of the helmets tested were used in a similar manner rather than the level of use by, say, a cycling enthusiast who rides several thousand miles a year like I do. So what does this then mean to someone like me? Little to nothing.

More significantly, they "eliminated any that showed damage." Seriously? Wasn't that part of the point to see how many helmets in use were still effective?
J.