Apr 102016
 

A couple of people have suggested I have a look at the ICD-10-AM codes, as this is how a lot of stats are actually recorded.

The trouble was getting hold of a copy of ICD-10-AM to look at in order to determine what is of interest. It’s a very heavily guarded secret apparently. They’re derived from the ICD-10 standard codes.

As it turns out, there’s a site that provides the ICD-10-CM codes .

The codes appear to be similar for things like head injury, so perhaps this will be “close enough”? For privacy reasons, I might not get deep enough for the differences to be significant, but this is a starting point and will at least get us most of the way there.

The ones that appear to be of interest seem to be the following:

Code Description Reasoning
S12 Neck Fractures (vertebrae) Too-heavy protective gear may be a factor in whiplash injury, need to ensure we don’t make other problems worse.
S14 Nerve injuries in Neck Again, related to whiplash and other neck-related injuries.
S16 Tendons/muscles in neck Once again, whiplash and similar conditions
S02 Skull fractures This is the area helmets are supposed to protect!
S06 Brain injuries Again, what helmets are supposed to protect!

I haven’t gone to the deeper levels for two reasons:

  • There may be privacy issues going too much deeper
  • The stuff I’ll be looking at will be according to the ICD-10-AM system, which may have slightly different designators at the lower levels.
Apr 022016
 

Well, after my last couple of emails, I’ve been pointed to some rather interesting reports that peek into the area of interest.

Robin Guda at the Australian Institute of Health and Welfare was able to point me to ” Trends in serious injury due to road vehicle traffic crashes, Australia: 2001 to 2010 ” which sheds a light on serious injuries (requiring hospitalisation) from road accidents, where the patient survived. She was also able to provide a link to where I can request further information . There’s a fee involved in the latter case, but I do not mind paying a smallish amount of money for such information, alternatively, crowd-funding might be an option.

Both of these are a big help, and that article is raising some interesting questions.

One thing that stuck out was this passage (page 8):

Rates of life-threatening cases involving motorcycle riders and pedal cycle riders rose significantly over this period, with average annual increases of 5.2% and 7.5% respectively. Rates of cases involving passengers of motor vehicles and pedestrians fell, with average annual decreases of 1.2% and 1.0% respectively.

Injuries per registered motorcycle did not change much from 2001 to 2010, suggesting that the rise in population-based rates is largely due to growth in the number of motorcycles in use.

Now, due to the fact we do not have bicycle registration here in Australia, we cannot come up with a hard equivalent statistic for bicycles. Two groups I might be able to approach to get some sort of approximate figure though:

  • Peak bicycle bodies, e.g. Bicycle Queensland — who may be able to give me rough figures on the trends in membership over that period.
  • Australian Bureau of Statistics, who do the national census. One of the questions they ask is “how did you get to work that day”, so there’d be a figure there for the number of people cycling to work.

Another prospect might be to see if I can get similar stats out of China. They have a big population, many of whom ride bicycles/motorcycles, and even their bicycles are registered.

Meanwhile, Angela Watson was in touch from CARRS-Q, and was able to shed some light also. One article she posted through was ” Fatal Road Traffic Accidents, Study of Distribution, Nature and Type of Injury “.

This, in contrast to the above AIHW report, looks at accidents where the victim died within 21 days of the accident, for accidents that took place in India between December 2003 and November 2004. India have somewhat different road and safety standards to Australia, and a much bigger population. Table 6 on page 3 gives the stats for avulsion injury (that is, things getting torn off) to contre-coup injuries, cyclists and motorcyclists made up about 7% of cases there in total, with pedestrians ruling the roost.

They note there in the introduction that a big factor here is poor pedestrian infrastructure in India. If we take pedestrians out, the number rises to 7 out of 24 cases, or 29%. Probably a big portion of those were not wearing helmets, but that is pure speculation on my part.

Two other articles Angela sent through, which I’m yet to study in depth are ” Mechanisms of Head and Neck Injuries Sustainedby Helmeted Motorcyclists in Fatal Real-World Crashes:Analysis of 47 In-Depth Cases ” and ” Bicycle helmets are highly effective at preventing head injury during head impact: Head-form accelerations and injury criteria for helmeted and unhelmeted impacts “.

I will certainly have a look at these.

A very valid point was raised though regarding the availability of raw data. Much of the medical type data is heavily guarded for privacy and ethical reasons, and so getting near it may prove to be difficult. In addition, there appears to be no linkage between the stats on road accidents, and the stats on injuries.

Traffic bodies like the Queensland Department of Main Roads are mostly not concerned with details like injuries sustained, and groups like Queensland Health are mostly not concerned about whether the person was wearing a particular model helmet or whether they hit a car bonnet or a bridge railing.

A sub-area of research for me will be to get to understand the classifications for these injury types, which are covered by the ICD-10-AM standard. This is no freebie however, unlike the Australian Standards, which SAI Global sell for about $50 a piece (and I got for free when I was studying at QUT), I haven’t been able to get pricing on the ICD-10-AM standard. I guess we’ll create an account there, and ask the silly question.

The ideal situation for me would be to borrow a copy for a short period and make a note of the relevant codes. I don’t need a permanent copy. Russel suggested a couple of them in a comment a few days ago, so that’s a starting point.

In short, the road ahead is becoming a little clearer, and it’s going to be a long one. Good thing I’m not gunning for the 2016 Hackaday prize, as I’ll almost certainly not have it done by the due date.