Customize your JAMA Network experience by selecting one or more topics from the list below.
Benson BW, Mohtadi NGH, Rose MS, Meeuwisse WH. Head and Neck Injuries Among Ice Hockey Players Wearing Full Face Shields vs Half Face Shields. JAMA. 1999;282(24):2328–2332. doi:10.1001/jama.282.24.2328
Context Speculation exists that use of a full face shield by ice hockey players
may increase their risk of concussions and neck injuries, offsetting the benefits
of protection from dental, facial, and ocular injuries, but, to our knowledge,
no data exist regarding this possibility.
Objective To determine the risk of sustaining a head or neck injury among intercollegiate
ice hockey players wearing full face shields compared with those wearing half
Design, Setting, and Participants Prospective cohort study conducted during the 1997-1998 Canadian Inter-University
Athletics Union hockey season of 642 male hockey players (mean age, 22 years)
from 22 teams. Athletes from 11 teams wore full face shields and athletes
from 11 teams wore half face shields during play.
Main Outcome Measure Reportable injury, defined as any event requiring assessment or treatment
by a team therapist or physician or any mild traumatic brain injury or brachial
plexus stretch, categorized by time lost from subsequent participation and
compared by type of face shield.
Results Of 319 athletes who wore full face shields, 195 (61.6%) had at least
1 injury during the study season, whereas of 323 who wore half face shields,
204 (63.2%) were injured. The risk of sustaining a facial laceration and dental
injury was 2.31 (95% confidence interval [CI], 1.53-3.48; P<.001) and 9.90 (95% CI, 1.88-52.1; P
= .007) times greater, respectively, for players wearing half vs full face
shields. No statistically significant risk differences were found for neck
injuries, concussion, or other injuries, although time lost from participation
because of concussion was significantly greater in the half shield group (P<.001), than in the group wearing full shields.
Conclusions These data provide evidence that the use of full face shields is associated
with significantly reduced risk of sustaining facial and dental injuries without
an increase in the risk of neck injuries, concussions, or other injuries.
Ice hockey has been characterized as the fastest and most violent team
sport in the world played with clubs (hockey sticks), a bullet (puck), and
knives (skates).1 The high intensity of the
sport results in frequent collisions between players and forceful impacts
with the side boards, goal posts, pucks, and hockey sticks. The head and neck
are particularly vulnerable to hockey-related injury,2- 10
and the search for improved player safety to prevent these injuries has been
an ongoing process for sports governing bodies and researchers.11
Ice hockey associations from Canada and the United States have introduced
head and neck risk management strategies, the most significant being the mandatory
use of full facial protection for athletes across many different age groups
and levels of play. These rules, combined with advancements in face shield
standards, have been effective in reducing the frequency of facial and eye
However, a trend of increasing catastrophic hockey-related injuries to the
cervical spine has led to speculation that use of the full face shield may
increase players' risk of sustaining a neck injury, possibly due to biomechanical
alterations or changes in the style of play.15- 20
To our knowledge, no prospective cohort or experimental research has been
conducted in the sport of ice hockey to address this issue.
The purpose of this study was to determine the risk of sustaining head
or neck injury among intercollegiate ice hockey players wearing full face
shields compared with those wearing half shields.
This prospective cohort study was conducted during the 1997-1998 Canadian
Inter-University Athletics Union (CIAU) hockey season. The CIAU is a national
league consisting of 4 divisions with identical officiating and player eligibility
rules, except that each division may mandate use of protective equipment.
The Ontario Universities Athletic Association (OUAA) requires that all athletes
under its jurisdiction wear full face shields for every practice and game
throughout the varsity season. In contrast, the Canada West Universities Athletic
Association (CWUAA) and Atlantic Universities Athletic Association (AUAA)
require all athletes to wear, at a minimum, half shields (visors) for all
practices and games. A half shield is a clear plastic visor that is attached
to a helmet and extends down to the tip of a player's nose. A full face shield
extends down to the bottom of a player's chin and covers the entire face.
For this study, 2 cohorts were defined by their associations' facial protection
The study population included 642 consenting male ice hockey players
from 11 CIAU university teams competing in the OUAA, 7 teams in the CWUAA,
and 4 teams in the AUAA. Thus, athletes from the 11 OUAA and 11 CWUAA/AUAA
teams formed the cohorts wearing full and half face shields, respectively.
The 11 OUAA teams were selected based on geographic proximity to a preseason
meeting in Toronto, Ontario, between the investigator and team therapists.
The 7 CWUAA team therapists had been collecting data for the Canadian Intercollegiate
Sport Injury Registry (CISIR), Calgary, Alberta, during the previous 3 seasons,
and the 4 AUAA teams were selected based on geographic proximity to a preseason
meeting in Halifax, Nova Scotia. The purpose of the meetings was to discuss
the data collection protocol with team therapists so that they would have
a thorough understanding of what was required of them in terms of data collection.
Approved by the University of Calgary's institutional review board,
this study required team therapists to obtain informed verbal consent from
all players prior to participation. Previous injury status was documented
for each CWUAA player using a standardized medical form during a mandatory
preseason medical examination with the team physician. Because no standardized
form was uniformly used by team physicians of institutions participating in
the OUAA and AUAA, athletes competing in these divisions completed an injury
history questionnaire with their team therapist during a preseason meeting.
The injury information recorded on these forms served as a baseline for the
1997-1998 hockey season. From the first practice of the season, team therapists
used a standardized weekly exposure sheet21
to record the level of individual participation (full, partial, or none) and
the type of face shield worn for every practice and game throughout the season.
If a player sustained an injury that met the reportable injury definition,
team therapists, physicians, or both were required to complete an injury report
form.21 A subjective assessment of whether
an injury was the result of an illegal action during a game and subsequently
whether a penalty was called by a referee was included on the injury report
form so that a descriptive comparison could be made with respect to style
of play between facial protection divisions. The completed forms were sent
to the CISIR, where they were checked, coded, and entered into a database
using a dual entry system to minimize data entry errors.21
Athletic therapists were important members of the investigative team
because of their medical orientation and diagnostic skills; their relationships
with athletes, coaches, and team physicians; and their daily presence at both
practices and games. Canadian Athletic Therapist certification requires: (1)
enrollment in a university undergraduate degree program in kinesiology or
physiotherapy, (2) the completion of 1200 hours of practical experience (600
clinical and 600 field hours), and (3) successful completion of a national
examination consisting of both written and practical components. During training,
student athletic therapists are under the direct supervision of certified
Canadian Athletic Therapists. Each participating coordinating athletic therapist
received a small honorarium for overseeing data collection throughout the
The main outcome (dependent) variable measured in this study was an
injury defined as "any event requiring assessment or treatment by a team therapist
or physician" and "any mild traumatic brain injury or brachial plexus stretch
(ie, burner/stinger)." Reportable injuries were then categorized by the amount
of time lost from subsequent participation. To avoid potential injury reporting
biases between institutions, the following injury definition was used for
data analyses: (1) any injury received during an organized practice or game
during the 1997-1998 CIAU hockey season that required assessment or treatment
by a team therapist or physician and resulted in at least 1 missed participation,
or (2) any facial laceration/fracture, dental injury, eye injury, traumatic
brain injury, or brachial plexus stretch (regardless of playing-time lost).
The independent variable of interest was full face shield vs half facial
shield. Individual athlete-participation (exposure),
defined as 1 player participating in 1 practice or game in which the athlete
was exposed to the possibility of injury, was measured to determine the amount
of time under the different exposure conditions during which each individual
was at risk of injury. Athlete-exposures were calculated by weighting a full
session of participation as "1," a partially missed session as "0.5," and
a completely missed session as "0." Goal tender exposure and injury information
was not included in the analyses so that a true comparison could be made between
players wearing full face shields vs half shields. For example, 20 athletes
fully participating in 10 games would yield 200-game athlete exposures. In
addition, several potential confounding or effect-modifying variables were
assessed, including player position (forward vs defense), injury status (new
vs recurrent), and injury setting (game vs practice).
The sample size chosen for this study was calculated based on injury
data from the CISIR database during the previous CWUAA varsity season, which
showed that neck injuries accounted for approximately 7% of the total reported
player injuries wearing half shields. Using a 2-sided test (α = .05, β
= .20), it was estimated that 300 subjects in each group would be required
to achieve 80% power to detect a relative risk of 2.0 or greater between the
2 study groups. The test-based method of Miettinen22
was used to calculate 95% confidence intervals (CIs) for relative risks based
on incidence density ratios23 and Epi Info 6 (Version 6.04a, Centers for Disease Control and Prevention,
Atlanta, Ga) was used for all P value computations
All 642 varsity athletes from the 22 participating teams consented to
have data collected and sent to the CISIR . In addition, athlete-exposure
information was 100% complete throughout the 1997-1998 CIAU hockey season.
Furthermore, the weekly exposure sheets were verified when they were received
at the CISIR for any indication that a player had missed time due to an injury.
In all cases but 1, an injury report form was sent to the CISIR (99.9% completion
rate). It was not possible to determine the completion rate for injuries that
did not result in time lost from participation.
The median age for athletes in both categories was 22 years (range,
17-29 years, full face shield division; range, 18-29 years, half face shield
division). Of the 319 athletes wearing full face shields, 195 players (61.1%)
were injured during the 1997-1998 season compared with 204 (63.2%) of the
323 athletes wearing half face shields. In addition, the 2 study groups, classified
as rookies, had a similar level of playing experience, with 118 (41.9%) of
the 281 players wearing full face shields, compared with 116 (39.5%) of 294
players wearing half shields. Furthermore, 280 (95.2%) of the 294 athletes
competing in half shield divisions had 1 or more injuries prior to the study
compared with 247 (87.9%) of 281 players competing in the full shield division.
Of note, 38 athletes (11.9%) competing in the full shield and 29 athletes
(9%) competing in half shield divisions did not have medical forms or injury
history questionnaires returned to the CISIR. These individuals participated
for a minimal amount of time as varsity players or joined a team later in
Excluding goal tenders, players wearing full face shields accumulated
24,147.5 athlete-exposures during the 1997-1998 varsity season, while players
wearing half shields amassed 26,823.0 athlete-exposures.
Crude relative risk estimates of head and facial injuries, neck injuries,
concussions, and other types of injuries sustained by players from each cohort
are shown in Table 1. Although
we found a significant difference in rates of head and facial injuries between
the 2 groups (P<.001), there was no significant
difference in risk of sustaining a concussion, neck, or other injury (overall)
for athletes wearing half shields compared with those wearing full face shields.
Determined by multivariate (stratified) analysis, these findings were not
confounded or modified by injury setting (game vs practice), injury status
(new vs recurrent), or position (forward vs defense), which were recorded
on the injury report forms by the team therapists and physicians. The same
differences between face shield types were found across all strata. The incidence
rates of head injuries sustained for each cohort are shown in Figure 1. There were no significant differences in the number of
concussions sustained between the 2 study groups (P
The risk of sustaining a facial laceration was 2.31 times greater for
players wearing half shields compared with those wearing full face shields
(95% CI, 1.53-3.48; P<.001). Based on the absolute
number of facial lacerations sustained during the 1997-1998 season (half shield,
77; full shield, 30), an average of 7 facial lacerations per team are expected
during any given season for athletes wearing half face shields, compared with
2 to 3 injuries per team for athletes wearing full face shields.
The incidence rates for various types of facial lacerations sustained
by athletes in the 2 study groups are presented in Figure 2. The lip and eyebrow are 2 areas of the face that receive
the most lacerations among players wearing half shields. The majority of lacerations
sustained by athletes wearing full face shields were to the chin. Although
there was a higher incidence of chin lacerations among players wearing full
face shields, the incidence was not significantly greater than for players
wearing half shields (P = .25).
Stick contact to the face was the predominant mechanism of lacerations
for athletes wearing half shields, whereas contact with a stick or an opponent
via body check or collision caused the majority of lacerations among athletes
wearing full face shields (Table 2).
Six facial lacerations sustained by athletes wearing half shields were caused
by a skate.
Eleven players wearing half face shields sustained dental injuries (ie,
teeth fractures) during the 1997-1998 varsity season. Nine of the 11 dental
fractures sustained by athletes wearing half shields were caused by contact
with a hockey stick, 1 resulted from contact with a player's own visor, and
the remaining mechanism was unknown. Seven of the 11 players injured in this
cohort were wearing mouthguards at the time of injury. Only 1 athlete, who
was wearing a full face shield but not a mouth guard, sustained a dental injury
when he was punched in the face during an organized practice. The risk of
sustaining a dental injury was 9.90 times greater for players wearing half
shields than for those wearing full facial shields (95% CI, 1.88-52.1; P = .007). One eye injury (traumatic iritis) was sustained
by a player wearing a half shield. This injury occurred when an opposing player's
elbow struck the injured player's eye during an organized game. No eye injuries
were reported for players wearing full face shields.
We found no statistically significant risk differences in the number
of concussions sustained by players in either study group (P = .90). However, the time lost from practices and games associated
with these injuries was significantly greater for players wearing half shields
than for those players wearing full face shields (P<.001).
Forty-one players who sustained concussions while wearing half shields missed
a total of 166.5 sessions, and 38 players who sustained concussions while
wearing full face shields missed 64.5 sessions. In addition, 24 of the 41
players who experienced concussions while wearing half shields missed 1 or
more complete sessions compared with 13 of the 38 players wearing full face
shields. Using a time-lost definition, the injury rate is approximately double
for players wearing half face shields. Concussive head injuries were randomly
distributed between the participating teams; that is, there was no clustering
of concussions on any 1 particular team within either cohort.
Only 5 neck injuries experienced by athletes wearing full face shields
resulted in 1 or more completely missed games or practices, compared with
7 for athletes wearing half shields. The risk of neck injury was not significantly
different between the 2 comparison groups (P = .78).
High sticking, slashing, and checking from behind, respectively, were
the 3 most common types of illegal play causing injury reported by therapists
of athletes competing in half shield divisions. In comparison, slashing, checking
from behind, and cross-checking, respectively, were the top 3 reported types
of illegal play causing injury in full face shield divisions.
Team therapists from the half shield divisions reported that 41.3% of
the injuries sustained during games were caused by an illegal action, compared
with 32.2% of the injuries sustained by players competing in the full face
shield division. In addition, it was reported that a referee called a penalty
after 10% of injuries sustained by players competing in half shield divisions
vs 8.5% of injuries of players in the full face shield division.
An accurate assessment of the penalty information reported by team therapists
was not possible because official game sheets were not collected for verification.
Public and professional concern about the number of head and neck injuries
occurring in ice hockey is increasing. There is speculation that while full
face shield use reduces facial, dental, and ocular injuries,24
it may increase injury rates to other anatomic regions such as the neck, brain,
and spinal cord, possibly due to biomechanical alterations25,26
or changes in the style of play.27 In this
study, the risk of injury was prospectively evaluated in a natural experimental
setting using 2 groups similar in every respect, apart from face shields use.
There were no differences between sites, league divisions, or physicians and
therapists in their threshold to decide to remove a player or limit a player's
participation in a game or practice, and there was no differences in player
skill levels or training. Furthermore, the data were collected using a previously
validated system of injury surveillance,21
and data collection was nearly 100% complete.
For intercollegiate ice hockey players wearing half shields compared
with full face shields, we found that the risk of sustaining a head injury
(excluding concussions), facial laceration, and dental injury was 2.52, 2.31,
and 9.90 times greater, respectively. We found no evidence in this study to
support the speculation that full face shield use increases players' risk
of sustaining a neck injury or concussion. Furthermore, the overall risk of
injury to an anatomic region other than the head or neck was not significantly
different between the 2 comparison groups.
Several other findings warrant comment. First, a significant number
of facial lacerations in the half shield group were to periorbital regions,
which carries a risk of catastrophic eye injury. This concern is compounded
when the evidence shows that stick and skate related facial injuries are far
higher in this group.
Second, the finding that 7 of the 11 athletes who sustained dental injuries
in the half shield group were wearing mouth guards at the time of injury suggests
that the use of such protective equipment in combination with half shields
is not enough to offer protection from these injuries. The use of full face
shields significantly reduces players' risk of sustaining a dental injury.
Third, concussions sustained by players wearing half shields resulted
in significantly greater time lost from competition than for players wearing
full face shields (P<.001). Possible explanations
for this finding may be that many players wearing half shields place their
helmets in a manner such that the visors are raised above the level of their
nose in order to get a clearer view of the ice and surrounding area. Since
the helmet is placed further back on the head, the protective effect of the
padding is minimized on the forehead area. In addition, a loose chin strap
may allow a properly fitted helmet and half shield unit to shift from its
original position during impact with an opposing player, thereby decreasing
its protective effect. Furthermore, a direct blow to the exposed jaw of players
wearing half shields may increase concussion severity compared with the full
facial shield, which because of its chin piece may cushion the jaw. The chin
piece of the full face shield also helps hold the helmet in place during impact,
thereby maintaining maximum player-protection from brain injury. However,
data or detailed evaluations of concussions, such as grade or associated or
persistent neurologic deficits, were not collected, so the relative severity
of injury between the groups could not be determined.
Finally, these findings are based on data from collegiate hockey players
and may not be generalizable to high school and younger, skeletally immature
In this study, use of full face shields was associated with significant
reductions in the risk of facial and dental injuries without increasing the
risk of neck injuries, mild traumatic brain injuries, or other injury rates
(overall). Sports governing bodies at the intercollegiate level of competition
should seriously consider mandating full facial protective equipment for all
participants under their jurisdiction.