M. Kary’s second comment letter about Harris et al., unedited

Note: this is the second of two letters constituting a review of the Harris et al. study of Vancouver and Toronto bicycle facilities.

You may read an explanation of how these letters came to be posted here. It includes a link to the other letter and to the edited version which is posted with the original article.

Vulnerabilities of the case-crossover method as applied, and unsuitability of the epidemiological approach, to transportation injuries and traffic engineering problems – Part II

M Kary, Mathematician

Montreal, Canada

Re: Comparing the effects of infrastructure on bicycling injury at intersections and non-intersections using a case–crossover design. Harris, et al. 19:5 303-310 doi:10.1136/injuryprev-2012-040561

Several results from this study [1-3] are potentially influential, yet anomalous or internally inconsistent. They exemplify various weaknesses of the case-crossover method as applied, and the general unsuitability of the epidemiological approach, to transportation injuries and traffic engineering problems.

1. The authors note that contrary to previous studies, they found surplus injuries at intersections with the greatest bicycle traffic. They suggest their finding may not be generalisable. But they do not explain why their method should have led to a non-generalisable result.

The case-crossover method as used by the authors does not generally track the effect of an independent variable– in this case, cyclist traffic– as it varies at a fixed location. Instead it typically ranges over entirely different locations, which coincidentally may have different values of the independent variable. Different urban locations in traffic engineering problems are analogous to different individuals in medical problems: they differ in countless unmeasured ways. The authors may believe that since control locations are chosen at random (but see [4]) along each cyclist’s route, these unmeasured covariates get balanced. But cyclists do not choose their routes at random, and many routes may share intersections and links.

In other words, one would expect there to be inherently hazardous locations that attract bicycle traffic because they are in some way inevitable, such as for being the only way to access a bridge. Danger at such intersections might be mitigated by high cyclist traffic, yet the case-crossover method as applied by the authors– being as much or more blind than other epidemiological methods to the difference between causation and association– would score this against it.

Nor is bicycle infrastructure installed at random. Instead, the locations for it are typically chosen either to take advantage of already safe circumstances, or to address special hazards, via multiple special measures. These biases bear on the next set of anomalous or internally inconsistent results:

2. The authors find bicycle-only paths in parks to be 17.6 times as dangerous as bicycle-only paths in (or adjacent to) streets. They find multi-use paths in parks to be 22.8 times as dangerous as bicycle-only paths in streets; and riding in the street to be 20 times as dangerous as riding in a bicycle-only path in the street.

To comprehend these results requires some introduction to the cycle track controversy. The authors explain that previous research, although favourable to cycle tracks, has suffered from defects. There is more to it than that: highly cited favourable research has been thoroughly refuted (e.g. [5, 6]), while other research has come to the opposite conclusion (e.g. [79]).

The problem is not that bicycle paths in (or adjacent to) streets (a.k.a. cycle tracks) are necessarily unsafe. It is that cycle tracks are inherently hazardous [1014]. Inherently hazardous situations can be made safe, but typically at great cost or sacrifice, and often not in a durable manner.

The fundamental (not the only) hazard of cycle tracks is that they force cyclists to be in the path of turning and crossing vehicles at junctions. The protection they can offer is only between junctions, where the absolute risks are lower, while they force cyclists into danger at junctions, where the absolute risks are higher [11]. Engineers attempt to mitigate the hazards at junctions by e.g. imposing multiple dedicated signal phases, or prohibiting turning and crossing movements outright. The inconvenience and frustration for all users generated by these measures means their benefits may not be durable, as was the case for the Burrard Bridge in Vancouver subsequent to the authors’ short study period [15].

The authors’ work estimates only relative risks of cycle tracks, and only between intersections. By missing both absolute risks and the action at intersections, it does not do anything to address the cycle track controversy, and it is wrong to use its results to promote cycle tracks.

The only novel cycle track result from this study is the anomalously large relative benefit it ascribes to cycle tracks between intersections. This limited result suffers from the following weaknesses:

(1) As found by the authors and others, the majority of cyclist injury events, including hospitalisations, result from bicycle-only crashes [2, 16, 17]. As noted by others, if cycle tracks work by protecting cyclists from motor vehicles, how can they reduce injuries by 95%, if the majority of such injuries have nothing to do with motor vehicles?

(2) If cycle tracks work by protecting cyclists, then the comparisons that introduced this section indicate cyclists need protection most of all not from motor vehicles, but from pedestrians (not the reverse? [6, 18]) and squirrels [19].

Rather than a protective effect, a small selection of more reasonable explanations for the authors’ results are as follows:

(1) Hardly any cycle tracks were studied: none at all in Toronto, and at most only four in Vancouver– Burrard, Beach, Stanley Park Drive, and the Carrall Street Greenway. As indicated by their names, two of these were put in place because they were already special, particularly safe and pleasant locations. Likewise, Vancouver’s “Greenways” are special routes that are safe for reasons far beyond anything in the authors’ definition of cycle tracks [20-22].

(2) The Burrard Bridge tracks, newly installed in the latter part of the study period, were instead placed because of a dangerous circumstance. Because of its length and the bridge’s importance in cyclist traffic patterns, its tracks may have been chosen for many control sites. However, unlike typical cycle tracks, they are long, direct, junction-free rights of way, divided from the roadway by contiguous Jersey barriers. Correspondingly and moreover, cyclist transit times on them may be short, despite their lengths. In such circumstances, discounting time of exposure at the expense of distance of exposure biases estimates of risk [23].

(3) Fresh pavement of the newly installed facilities.

(4) Biases in the selection of control sites as previously described [4].

A selection of remaining problems, and the general unsuitability of the epidemiological approach, will be covered in an eventual final instalment.


1. Harris A, Reynolds CCO, Winters M, Chipman M, Cripton PA, Cusimano MD, Teschke K. The Bicyclists’ Injuries and the Cycling Environment study: a protocol to tackle methodological issues facing studies of bicycling safety. Inj Prev 2011;17:e6. doi:10.1136/injuryprev-2011-040071.

2. Teschke K, Harris MA, Reynolds CCO, Winters M, Babul S, Chipman M, et al. Route Infrastructure and the risk of injuries to bicyclists: a case-crossover study. Am J Pub Health 2012;Oct 18:e1-e8. doi:10.2105/AJPH.2012.300762.

3. Harris MA, Reynolds CCO, Winters M, Cripton PA, Shen H, Chipman ML, et al. Comparing the effects of infrastructure on bicycling injury at intersections and non-intersections using a case-crossover design. Inj Prev 2013;0:1?8. doi:10.1136/injuryprev-2012-040561.

4. Kary M. Vulnerabilities of the case-crossover method as applied, and unsuitability of the epidemiological approach, to transportation injuries and traffic engineering problems – Part I. http://injuryprevention.bmj.com/content/early/2013/02/13/injuryprev-2012-040 561/reply#injuryprev_el_9940 (accessed June 16 2013). (Now on this Web site, as the original letter has been abridged.)

5. Kary M. Still more errors and omissions; and Compendium of errors and omissions, or: What is not in this article. http://injuryprevention.bmj.com/content/17/2/131/reply (accessed November 20 2013)

6. Kary M. Compendium of errors and omissions in Risk Of Injury For Bicycling On Cycle Tracks Versus In The Street, or: What is not in this article. 2011. http://john-s-allen.com/reports/montreal-kary.htm (accessed November 20 2013).

7. Wachtel A, Lewiston D. Risk factors for bicycle-motor vehicle collisions at intersections. ITE Journal 1994;September:30-35. http://www.bicyclinglife.com/Library/riskfactors.htm (accessed November 20 2013).

8. Jensen SU. Bicycle tracks and lanes: a before-after study. 2007. http://trafitec.dk/sites/default/files/publications/bicycle%20tracks%20and%20lanes.pdf (accessed November 20 2013)

9. Agerholm N, Caspersen S, Lahrmann H. Traffic safety on bicycle paths: results from a new large scale Danish study. 2008. Paper presented at International Cooperation on Theories and Concepts in Traffic Safety, Melbourne, Australia. http://vbn.aau.dk/files/14344951/agerholm_et_al._bicycle_paths.pdf (accessed November 20 2013).

10. AASHTO Task Force on Geometric Design (1999). Guide for the development of bicycle facilities. Washington, DC: American Association of State Highway and Transportation Officials.

11. Bicycle infrastructure and safety. http://publications.gc.ca/collections/collection_2012/tc/T41-1-90-eng.pdf (accessed November 20 2013)

12. Forester J. Bicycle Transportation: A Handbook for Cycling Transportation Engineers, second edition. Cambridge, Mass.: MIT Press, 1994. http://mitpress.mit.edu/books/bicycle-transportation (accessed November 20, 2013)

13. Forester J. The bicycle transportation controversy. Transportation Quarterly 2001;55(2):7-17. http://www.johnforester.com/Articles/Facilities/TransQuart01.htm (accessed November 20 2013).

14. Forester J. Traffic engineering to accommodate bicycle traffic: how to accept incompetent operation and its consequences. 2009. http://www.johnforester.com/Articles/Facilities/Theory%20for%20Bikeways.pdf (accessed November 20 2013).

15. Downtown Separated Bicycle Lanes Status Report, Spring 2012. City of Vancouver. June 5 2012. http://bikeroute.files.wordpress.com/2012/06/downtown_lanes_report.pdf (accessed November 20 2013).

16. Aultman-Hall L, Kaltenecker MG. Toronto bicycle commuter safety rates. Accident Analysis & Prevention 1999;31(6):675-86.

17. Boufous S, de Rome L, Senserrick T, Ivers RQ. Single- versus multi-vehicle bicycle road crashes in Victoria, Australia. Inj Prev doi:10.1136/injuryprev-2012-040630.

18. Revay T. Bicyclists have right of way on the path? 2001. http://www.bikexprt.com/bikepol/facil/paths/pathatt.htm (accessed November 20 2013).

19. Anon. The Night of Claire. http://www.bicyclinglife.com/NewsAndViews/BikePoetry.htm (accessed November 20 2013).

20. Greenways: Making Vancouver a more walkable, bikeable city. http://vancouver.ca/streets-transportation/greenways-for-walking-and-cycling.aspx# (accessed November 20 2013).

21. Chow L, Kudzius B, Scott D. Carrall Street Greenway: Public Realm Improvements & Community Development. Transportation Division, City of Vancouver, 2009. http://www.tac-atc.ca/english/resourcecentre/readingroom/conference/conf2009/pdf/Chow.pdf (accessed November 20 2013).

22. Kudzius B, Chen-Adams J. Carrall Street Greenway: Detailed design and implementation. City of Vancouver, 2006. http://former.vancouver.ca/ctyclerk/cclerk//20060713/documents/pe3.pdf (accessed November 20 2013).

23. Chipman ML, MacGregor CG, Smiley AM, Lee-Gosselin M. Time vs. distance as measures of exposure in driving surveys. Accident Analysis & Prevention 1992;24:679-684.

Conflict of Interest:

None declared

Leave a Reply

Your email address will not be published. Required fields are marked *

Please answer this to show that you are a human!... *