Data from Teschke et al., Route Infrastructure and the Risk of Injuries to Bicyclists, Table 4, descriptions modified, *P less than .05
The Atlantic's Cities blog has a decent summary of a recent paper aimed at determining the impact of various kinds of cycling infrastructure on injury rates. Among the study's conclusion is that bike lanes significantly reduce the rate at which cyclists are injured.
Terminology & Denver's Paths
At the outset, I should note that study's terminology doesn't match up well with how we in Denver refer to things. According to the study's definitions, the Cherry Creek and Platte Greenways are "multiuse" paths ("meant for nonmotorized use by pedestrians, cyclists, skaters and others, either alongside city streets or away from city streets") not "bike paths."
I'm not sure how the study would classify Denver's designated sidewalk/bike routes. Those monstrosities fit both the definitions of sidewalk ("paved path meant for pedestrian use, either alongside city streets or away from streets") and multiuse path. (As far as I know, we don't have any "bicycle-only paths" in Denver.)
Multi-Use Paths Are (Relatively) Unsafe
The study's results are interesting for several reasons. First, in terms of safety, multi-use paths are less safe than everything except sidewalks and major streets with parked cars and no bike infrastructure.
That probably surprises a lot of planners. Multi-use paths are assumed to be the safest place for cyclists to be.
If the results apply to Denver, Denver has built (and designated) a lot of infrastructure that is less safe to use than simply riding on a major street that doesn't have parking. In particular, where Denver has designated sidewalks as bikeways it has created conditions that are about as unsafe as they can be for cycling.
Parking Poses Problems
Another interesting thing is just how much difference street parking makes. Eliminating parking makes even major streets safer than adding a bike lane or sharrows would. (The risk for major streets with no parking and no infrastructure are not statistically significant.)
If cyclist safety is an issue, planners should consider removing parking before considering adding cycling infrastructure. (Anyone who has worked on these issues in Denver knows that's wishful thinking. A bike planner once told me -- sarcastically -- that City Council, given a choice between a few dead cyclists and the calls triggered by eliminating a few parking places, would choose dead cyclists every time. He might have been sarcastic, but I think he was right.)
What's Hurting Us
Finally, how cyclists were injured surprised me. Only about a third of cyclist injuries were caused by collisions with motor vehicles. About a quarter were caused by collisions with things like train tracks, potholes, or rocks. Collisions with animals, other cyclists, and pedestrians accounted for only about 7% of the injuries. Looks as if keeping streets maintained on bike routes would help a great deal. Drivers would appreciate it, too.
Having said all that, I'm not sure how applicable the study's conclusions are to Denver or much weight to give them.
A Concern & The Author's Response
One concern I had is that the study may have underestimated injury rates on roads. That could be the reason that the study reports that multi-use paths are less safe than roads.
The study only included cyclists who could recall their routes. Some cyclists could not -- for example, the two cyclists who were dead. It seems likely that those who could not recall their routes were more likely struck by cars than injured in some other way. If this is true, the road injury rates might actually be greater than reported.
One of my compatriots asked Dr. Teschke about this. Here's her, convincing, response:
You are the first person to suggest that not including those who couldn't recall their routes might have underestimated the risk of streets (compared to separated routes). Several people (who seem not to like the result for cycle tracks) have suggested the opposite. Yours is a very logical suggestion, and it is possible.
We have done a severity analysis (being written up now) and it shows, as do many other studies, that motor vehicle collisions were more severe (more likely to result in being unable to continue riding, being transported by ambulance, or being admitted to hospital). However, we think that the number of people not included because of severe head injuries is likely small. We are quite sure of the fatalities missed - just 2. We excluded 26 people who reported to us that they could not remember their route. The mystery is what proportion of those we couldn't contact had brain injuries. In other studies that studied cyclists attending a hospital emergency department and that tallied the types of injuries of every case, the proportion of brain injuries has been low. Using data from the biggest of these studies (Thompson DC, Rivara FP, Thompson RS. Effectiveness of bicycle safety helmets in preventing head injuries. JAMA. 1996:276:1968-1973), I estimate that there would have been 43 brain injured people among all 2335 cyclists who attended the emergency departments in our study. Given our eligibility criteria, slightly more than half of these would not have been eligible for other reasons (mainly because they were injured or resided outside the study cities). If this estimate is correct, then the impact on the results should be small.