Study Method and Findings

In the eight months prior to our study the Fire Department had rejected 28 applicants for failing the vision requirements. Standards were not tied to the job via job analysis and existing standards were considered far too strict by firefighters.

The need for a standard for vision in the absence of corrective lenses (an uncorrected standard) is determined by the need for the firefighter to do the job safely on those occasions when corrective lenses are missing. Setting a fair standard requires balancing two demands: 1)the need for safety and protection of firefighters and the public, and 2) the avoidance of unfair discrimination against firefighter applicants who can perform the job adequately, even though they lack perfect vision.

Our review of existing vision research turned up no studies which would have allowed transportability of findings. Several research hers have provided empirically valid police vision standards. In contrast, it appears that vision standards for firefighters have not been based on formal analyses of the visual demands in firefighting. Further, few tasks performed by police are also performed by firefighters, except driving and driving-related tasks, therefore, transporting police vision standards transporting police vision standards to firefighters is inappropriate.

Our research program followed several phases. We performed a survey of firefighters to determine breakage of glasses and the consequences thereof. All career and volunteer firefighters were surveyed. Of the 814 career and approximately 800 volunteer firefighters 525 surveys were returned. Response rate for career firefighters was 80 percent.

Later, critical firefighter vision tasks were empirically determined by a survey of 29 glasses-wearing experienced firefighters. These raters were asked to rate each task on its consequences if not performed. The tasks were than rank ordered by relative importance. simulations of the most important tasks were created with the assistance of fire service training experts. Those most important vision tasks included spotting a victim of the roof, spotting a human victim behind the window on the fourth floor, reading codes on a tanker, and identifying dangerous elements in an artificially lit room from a distance of 20 feet during overhaul.

Subjects were six career firefighters who had 20/20 vision. All had between 5 and 13 years of first-line firefighter experience. Age ranged from 21 to 31. The repeated measure experimental designs provided statistical power (the ability to detect a true effect) in excess of .80. Firefighter subjects were clinically decorrected to four levels of acuity by standards opthalmic technique. The levels spanned the gradations between the then current uncorrected standard for the better eye (20/40) and the lowest level judged feasible (20/200).

On the day of the field testing, we first field-tested firefighter subject's uncorrected vision and their vision with each pair of glasses at 20 feet to verify the decorrection levels. Each simulation took place on a separate side of the training building at the fire academy, so subject firefighters could not see the next simulation. No firefighters were present during the testing of any other firefighters.

Scoring criteria for each simulation were established in advance; subsequent comparison of independently assigned scores revealed an interrater agreement of 100 percent on each simulation. We assessed each firefighters' performance on each of the simulated critical tasks at each of the five levels of acuity: 20/20; 20/40; 20/70; 20/100; and 20/200. Trials began at 1:00 PM and ended at 7 PM on July 7, 1988.

An expert panel, composed of fire chiefs, fire training officers, and one psychologist, had determined acceptable cutoff scores for adequate performance for each simulation. Firefighters performed at levels judged satisfactory at binocular acuity levels (both eyes open) of 20/100 in two simulations and 20/200 in the other two simulations. Statistical analysis (single factor univariate anovas) revealed large differences in performance due to acuity, with F values ranging from 6 to 62, but little difference due to intra subject variability.

The results of the field experiments (the four simulations) clearly indicated the appropriateness of radical changes in existing selection standards. Based on the findings of this research program, we recommended setting the uncorrected static distance vision standard at 20/100.

This recommendation was reviewed by the National Fire Protection Association Medical Subcommittee and adopted in 1992. The far visual acuity standard in NFPA 1582 is based on this MED-TOX study.