Centre For Local Research into Public Space (CELOS)

See also Site Map

Citizen-Z Cavan Young's 2004 film about the zamboni crisis





Manufacturing risk data: Pick a number, assign a cause

From: The Toronto playgrounds story (Jutta Mason, 2007)

If Health Canada’s web site is to be believed, between 1982 and 1999 Canada was the world capital for children being strangled in playgrounds. The website states that during those years 17 children died of strangulation in Canadian playgrounds. The same number is often quoted in other articles, and in risk presentations. When there is a footnote, it says “personal communication” from a Health Canada program consultant, and after that the trail goes cold.

A U.S. article on playground injuries counted 147 playground deaths in the U.S. between 1990 and 2000, of which 82, out of a much larger child population than Canada's, were said to be due to strangulation. Outside of North America, reports of child strangulation in playgrounds were rare. I e-mailed the general information link at Health Canada to ask how they had arrived at their number of 17. To my surprise, I got a letter back by regular mail, from the head of Canada’s Public Health Agency, Dr. David Butler-Jones: “the number of strangulations did not come from…coded data, but was based on cumulative information gleaned from coroners’ reports, incidents reported to the Product Safety Program at Health Canada and the Canadian Standards Association, and reports from other public health bodies. There is no single source that may be used to reference this information.” Dr.Butler-Jones said he had no more details.

I spent an afternoon looking up journal references at the university medical library. All but two, maybe three, of the playground strangulation numbers turned out to be free-floating anecdotes, unconnected to either a specific story or a source in the medical journals that dealt with child injuries and accidents.

As any first-year statistics student can tell you, a number without a verifiable source won’t get you a pass on your term paper. But the story remains influential. A 2007 press release from the Canadian Standards Association (CSA), about its newest version of playground standards, still quoted the strangulation number first in its list of references. Many playground manufacturers and safety officials use the number, and many journalists repeat it, to this day.

How do playground injury numbers come into being? What separates guesses from fact? When little Janie plays tag with her best friend Sam at the park playground, and she runs into a corner post of the monkey bars and knocks herself out, what does a harried emergency room resident write on the data sheet as the cause of injury? When three children tie a fourth playmate to a playground slide with a skipping rope, as part of a make-believe game, and by a horrible fluke the playmate slides just far enough down the slide that she can’t breathe – and the plastic rope cinches around her neck so tight that her friends can’t get her loose – what was the cause of the accident? Skipping ropes made of plastic? Slides that a rope can be tied to? Children interested in acting out pirate stories? Parents in a small town who let their kids go the playground on their own, as the parents themselves did, hundreds of times, when they were kids?

Who knows?

In the face of uncertainty, the simplicity and power of numbers is very attractive. And so the search is on, to discover numerical patterns of mishap that can be controlled.

In April 1990, Health Canada established the Canadian Hospitals Injury Reporting and Prevention Program (CHIRPP for short) at 10 pediatric hospitals. Between 1991 and 1995 CHIRPP added 6 additional (general) hospitals, specifically to monitor emergency room visits of children. The categories of injury were not always clear, the emergency room staff didn’t welcome the extra paperwork, and since there were then over 750 hospitals in Canada, the numbers were hardly complete. But they didn’t have to be complete to be useful. For example: Sally Lockhart, a program consultant at Health Canada, worked for the "Safe and Supportive Environments," Childhood and Youth Division. It was her job to figure out how the federal government could intervene to make all children safe right across Canada. Such an immense project needs a lot of numbers. On the Health Canada web site, Ms.Lockhart praised the new CHIRPP data collection. “By providing the rich detail on circumstances surrounding these injuries, it is now possible to focus interventions more effectively.”

Not everyone would describe the CHIRPP data as “rich in detail.” Little Janie’s crash during the game of tag, for instance, would have been listed as a “head injury” resulting from a “playground climber,” and the E.R. resident may have been too busy to note that this was not a fall but a collision. Janie might have had a goose-egg on her forehead and a black eye, and maybe would have been kept in overnight for a concussion watch. But even if Janie’s mom took her home for breakfast, smiling and hungry, and Janie was out in front of her house by mid-morning, showing off her black eye to her best friend and tag partner Sam – in the government’s data base she would most likely have been put into the serious-injury group as a “head injury” with “hospitalization of at least one night” caused by a “playground climber.” The resulting number would then be available to be marshaled at some point for a safety intervention – replacing playground climbers, for instance, like the one little Janie ran into, with new climbers lower to the ground, to reduce the fall distance that could produce serious head injuries. The new CSA-approved climbers would still have corner posts, tall enough for kids to run into them during an exciting game of tag, and maybe even knock themselves out...

What was missing in detail and in completeness in the new government data after 1990 was soon filled in by extrapolation and guesswork. Sally Lockhart wrote on Health Canada’s web site, “Each year, more than 10,000 Canadian children are injured on playgrounds. Many of these children require hospitalization, and, tragically, some die.” This number was based on yet another, unnamed Health Canada staff person’s “personal communication.” The same number was repeated by insurance companies for a few years, including the Ontario School Boards’ Insurance Exchange (OSBIE), which suggested that possibly all these 10,000 children “required medical treatment.” Then in 1998, OSBIE put out a special bulletin with a new number. The insurance exchange had “received approximately 28,000 playground incident reports in 1997 for Ontario [school playgrounds].” Teachers told us that by then they were required to do a report on any scrape requiring a bandaid. OSBIE guessed that of the 28,000 school reports “an estimated 1,200 would have been serious enough to require medical attention.”

28,000 is a more impressive number than 10,000. In 2000, that Ontario schools incident-reports number seems to have migrated over to a new Sick Children’s Hospital organization called "Safe Kids Canada," and slipped into their booklet of advice for parents. The booklet gave the Canadian playground injury number as “more than 28,500 each year.” It cited Health Canada program consultant Tammy Lipskie as the source, but she couldn’t pin down further details when questioned later by playground advocates. In 2001, a Health Canada staff epidemiologist, Susan Mackenzie, told Today’s Parent magazine that the annual number of playground injuries might be 25,000, “estimating in a really crude way.” And in May 2005, when the Globe and Mail gave the Canada-wide playground injury number as 25,000 and the National Post said 28,500, the newspapers themselves came up with these calculations, multiplying some Ontario hospital data to make a national number.

Making up data by simple multiplication, without telling, is not considered reputable statistical practice. But this numeration was not about careful study of the evidence. It was about getting a number big enough to be a quick-click icon, to be turned into a banner under which playground safety advocates would then march across the land, taking out swings, climbers, slides, and whole playgrounds on their way.

Content last modified on November 11, 2017, at 01:06 PM EST