The assessment identified 136 schools with playground equipment
that represented a severe hazard (i.e., an imminent risk of
serious and permanent injury, usually indicating risk of a fall
from more than 1.5 m or a fall onto unsuitable surfacing), did not
meet CSA standards and was impractical to retrofit to make safer
(Fig. 1). The equipment was removed from all 136 schools in the
summer of 2000. By Dec. 31, 2001, the equipment had been fully
replaced in 86 schools, which constituted the intervention group;
it had not been fully replaced in the remaining 50 schools, and
they were excluded from analysis (Fig. 2). Injury rates before and
after equipment replacement were compared in the intervention
schools. A total of 225 schools had equipment that did not require
replacement, and they constituted the nonintervention
schools; injury rates in this group served as an indicator of stability
of baseline injury rates in this natural experiment. A total of
34 557 students attended the intervention schools, and 88 417 attended
the nonintervention schools. All schools included grades
1 through 6 and were public schools.
A database of incident reports from the Ontario School Board
Insurance Exchange was used to identify all injury events occurring
at TDSB schools between January 1998 and December 2002
inclusive. Information in the database is provided by school staff
whenever an incident occurs in the school during school hours.
The threshold for completing a report is “whenever medical or
dental attention is required,” and this includes injuries attended
to by teachers or school staff, as well as those in which the child
went home or to a health facility. We included all injuries to children
4–11 years of age that occurred within the school playground.
Research assistants, blinded to the study purpose and
date of injury (i.e., before or after equipment removal), coded all
playground incidents on the basis of written descriptions from
the database. Injuries where equipment was explicitly mentioned
in the written description were flagged as “equipment related”
for subanalysis