The prevalence and reliability of visibility aid and other risk factor data for uninjured cyclists and pedestrians in Edmonton, Alberta, Canada

This study was conducted to determine the prevalence and reliability of risk factors collected on uninjured cyclists-pedestrians in Edmonton, Alberta, and what characteristics predict cyclist-pedestrian visibility. At randomly selected locations from July 2004 to August 2004, two independent observers recorded cyclist-pedestrian characteristics such as age, sex, clothing color, use of reflectors, flags, helmets, and a subjective impression of visibility. Data were collected on 836 individuals; most were either walking/jogging (approximately 63%) or cycling (approximately 33%). For non-cyclists, the prevalence of bright colored clothing on the trunk ranged from 12.7 to 14.7%. Few people used any kind of reflective strips. Inter-observer agreement (Kappa) ranged from 0.37 (visibility assessment) to 0.99 (sex). For cyclists, 17-19% of headgear was brightly colored, and 13-14% was white. Approximately one-fourth had a front light; half had a rear reflector. Few cyclists used a flag and just over half used spoke reflectors. Kappa ranged from 0.35 (observer assessed speed) to 0.95 (head gear and sex). A major trunk color of orange, red, yellow or white resulted in a higher visibility rating for both cyclists and pedestrians. The results indicate a low prevalence of visibility aid use among cyclists and pedestrians, but there appears to be acceptable inter-observer reliability for most data collected. Further work is required before an overall visibility rating can be used in place of component scores. (c) 2006 Elsevier Ltd. All rights reserved.