Lack of data may also be used in the risk assessment matrix effectively.
An absence of no means yes?
For the accident statistics to be valid in a risk assessment, you can only use the accident data where CO relevance was measured in either the gas or the victim. I haven't seen many studies reporting the percentage of accidents where CO was causative, out of accidents where CO was investigated as a potential contributor. I haven't seen any from outside North America.
There is a difference between no data, and data supporting an absence of risk factors. If you try to use a numerically scored risk assessment, you wouldn't be able to assess CO risk based on recreational scuba accident data sets due to the absence of consensus in the reliable data - the score is undefined, not zero. That doesn't mean it's low risk, it means you can't predict the risk based on recreational scuba accident reports alone.
You can look at studies by various compressed gas associations (some referenced in another post) on contaminant levels found in compressed gases and score the hazards on that basis. The studies I've seen indicate a low risk for CO contamination, about 1 to 3 percent, which aligns pretty well with the number of CO contaminated tanks I've encountered here in the middle east.
In the US there's a good chance your filler has a CO monitor on his compressor, which gives you a better comfort factor in your personal risk assessment. I've yet to see a CO monitor on a compressor over here. I have seen an LDS go 3 years without changing the compressor oil (they just kept adding vegetable oil when the level was low). And I have seen several shops use the same molecular sieve well past its' expiration (something's better than nothing theory).
Back to the original hijack; comparatively speaking on a global basis, the risk to an OW recreational scuba diver of a CO hit seems higher than the risk of being tox'd out by a rich nitrox blend inadvertently mixed in a rental tank marked air. While both may be considered low risks, I've never heard of a diver toxing out on a rental tank marked air, but have seen several incidents per year of divers dying from CO contaminated air. If an OW diver were inclined to invest in only one analyzer, the CO analyzer addresses the higher risk amongst the two IMO.