The Physician and Sportsmedicine 17(11):100-113 (1989)

Injury Rates in a National Sample of College Football Teams: A 2-Year Prospective Study

Eric D. Zemper, PhD

____________________________________________________________________________

In brief: The National Sports Injury Surveillance System* collected data during the 1986 and 1987 playing seasons from a national sample of 6,229 college football players. The overall injury rate for the two seasons was 6.32/1,000 athlete-exposures, or 45.27/100 athletes. Offensive players incurred more injuries than defensive players. The knee and ankle were the most common injury sites, and sprains were the most common injury. Injuries during games occurred most frequently in the third quarter and least often in the first quarter. This finding suggests that players might benefit from warming up and stretching during half time rather than resting and cooling off.

____________________________________________________________________________

Dr Zemper was director of research at the International Institute for Sport and Human Performance, University of Oregon, Eugene, when this paper was written; he is now president and director of research for Exercise Research Associates of Oregon, Eugene. He is a member of the American College of Sports Medicine.

*NOTE: The National Sports Injury Surveillance System was the forerunner of the Athletic Injury System currently being operated by Exercise Research Associates of Oregon.

____________________________________________________________________________

The total number of football injuries incurred by high school and college athletes each year is high because there are many large teams across the country (an average of approximately 100 players per team at the college level). However, the number of injuries per 1,000 athlete-exposures (AEs) for football is actually lower than that of some other sports such as wrestling, men's and women's soccer, and women's gymnastics (1-4).

There have been short-term studies of football injuries (5,6) and long-term studies of football fatalities and catastrophic injuries (7). But the only attempts at long-term, national studies of football injuries have been made by the National Athletic Injury/Illness Reporting System (NAIRS), which operated from 1975 to 1982 (8, 9), and the National Collegiate Athletic Association Injury Surveillance System (NCAA/ISS), which began collecting data in 1982(1). More recently the National High School Injury Registry began collecting football injury data during the 1986 season, under the sponsorship of the National Athletic Trainers Association (10), and the National Sports Injury Surveillance System (NSISS) began collecting data on college football injuries, also during the 1986 season.

The NSISS, housed at the International Institute for Sport and Human Performance at the University of Oregon, is a longitudinal data collection system for sports injuries that provides analyses of national injury rates and patterns. Its purpose is to provide data that will assist in developing ways to prevent or to reduce the severity of sports injuries through rule changes, equipment modifications, or changes in coaching techniques. It also provides a vehicle for investigating special issues. For example, ongoing studies are being conducted of injury rates on artificial vs natural surfaces, brands of helmets worn by players who incur a cerebral concussion (11) and the efficacy of prophylactic knee braces (12).

Initial reports on two special NSISS studies have been published (11, 12). One concerns the relationship between various brands of football helmets and the rate of cerebral concussions (11). Results from the first 2 years of operation of NCAA/ISS showed no difference in concussion rates (1). This pattern continued with the NSISS data during the 1986 season. However, during the 1987 season there was a significant difference between two brands. Further data are needed to determine whether this difference is real and a result of changes in helmets, or if it is a statistical anomaly in 1 year's data. Because of frequent changes in helmet design and material, these data need continual monitoring. The other special NSISS study (12) concerns the efficacy of prophylactic knee braces. Analysis of the first 2 years' data indicates that the incidence of knee injuries is slightly higher among players wearing braces. In addition, braces do not reduce the number of injuries to the medial collateral ligament (MCL), and there is no apparent difference in injury incidence among the various brands of braces. Nor was any difference found in the severity of knee injuries or MCL injuries in braced vs unbraced players when measured in days lost from play or degree of MCL injury. When injuries in games were isolated to control the "intensity" of exposure to the risk of injury, the results were the same as those for games and practices combined.

Until now the NSISS has been operating on a pilot basis, collecting data from a national sample of college football teams, but there are plans to expand its operation to include larger samples and a variety of sports at different levels (eg, collegiate, high school, and youth sports). The basic operational format for the NSISS was first developed and implemented by the author in 1982, while at the NCAA, as the NCAA Injury Surveillance System. The NSISS has since been used to collect injury rate data for a number of sports (1, 4, 13).

This article summarizes two years of football injury rate data collected by the NSISS from a national sample of college football teams.

Methods

Data were collected during the 1986 and 1987 college football seasons. The NSISS meets the major criteria for reliable studies of sports injury rates outlined by the American Orthopaedic Society for Sports Medicine (14). We used a stratified sample of NCAA and National Association of Intercollegiate Athletics (NAIA) members sponsoring intercollegiate football. Stratification was based on four geographic regions and three NCAA/NAIA divisions. The sample was designed to represent approximately 5% of all NCAA and NAIA intercollegiate football teams, including both varsity and subvarsity. Therefore the results can be generalized to the total population of intercollegiate football players. The sample included all football players at the participating institutions. During the 1986 season, data were collected for 32 teams (3,431 players); during the 1987 season, data were gathered for 27 teams (2,798 players). The average team size over the two seasons was 105.6 players.

Only time-loss injuries were recorded--ie, those that kept a player from full participation for one day or more beyond the day of the injury. The injury rate is the number of injuries per 1,000 AEs (an AE occurs each time a player takes part in a practice or a game, thus being exposed to the risk of injury). Therefore, if a football team had 100 players who took part in five practices during a given week, that team had 500 AEs in practices for the week. If 40 players played in the game on Saturday, there were 40 game exposures, for a total of 540 AEs for the week for that team.

Injuries per 1,000 AEs is considerably different from injuries per 100 players per season, which is most commonly used in research. However, the former is a much more precise measure of injury rates. The use of injuries per 100 players per season assumes that all players participate in every practice and game and that all teams have the same number of practices and games. Since this is obviously not the case, injuries per 100 players per season can be quite misleading, especially when trying to compare data from different teams or studies, or trying to compare injury rates across different sports. Use of injuries per 1,000 AEs compensates for the fluctuating number of players who attend practices each week and who play in games. In most cases the number of those who play in the game is much lower than the total number of team members. This method also takes into account the varying number of practices for each team.

Prior to the start of each season the head athletic trainer of each participating team was sent copies of brief, easy-to-use forms for reporting AE and injury data on a weekly basis, along with detailed instructions. From the first preseason practice until the final regular season or post-season game, the trainers returned forms that listed the number of practices and games played during the week and the number of players participating in each. They also returned separate forms that described each football injury.

Hard-shell helmets and shoulder pads are designed to protect the wearer from injury. However, because there is some concern about the role of hard-shell helmets and shoulder pads in causing injuries to other players, the trainers were asked to indicate whether each reported injury was directly caused by impact from another player's helmet or shoulder pad. If a medical staff person directly observed that this happened, or if an injured player recalled it, a positive response was recorded. Otherwise the response was recorded as negative or unknown.

In reporting the number of days lost because of an injury, the trainers were instructed to disregard the end of the season and indicate the number of days before an injured athlete returned to unrestricted activity, as if the season were continuing indefinitely. This provided a consistent and accurate record of time lost for injuries that occurred at the end of the season. An injury that kept a player out for more than 90 days was considered season-ending.

Each returned form was screened for completeness and consistency before the data were entered into a computer for later analysis. Trainers were contacted for clarification of any incomplete or inconsistent information on a form. During the two seasons, 98.8% of the forms were returned.

Results

For the 1986 and 1987 football seasons there was a combined total of 6,229 players in the study, with 445,856 AEs. Of the AEs, 410,165 (92.0%) occurred in practices and 35,691 in games (8.0%). A total of 2,820 injuries were reported, for an overall injury rate of 6.32/1,000 AEs, or 45.27/100 players. The rate in practices was 3.93/1,000 AEs; in games, 33.68/1,000 AEs. Thus the injury rate was 8.6 times higher in games than in practices (in which 57.1% of total injuries occurred). This major difference in injury rate between competition and practice is consistent with most other sports that have been studied (1-3).

Taking into account recurring injuries from the same season and athletes who were injured more than once, 32.1% of the players were injured at least once. The median number of days lost per injury was 4 (1986) and 5 (1987). Over the 2 years, 8.3% of the injuries were season-ending. Surgery was required for 8.5% of the injuries. Approximately half of these were treated arthroscopically, which allowed many players to return to football before the end of the season.

Based on positive responses, the impact from another player's helmet was the direct cause of injury in at least 14.4% of injuries; from a shoulder pad, 7.0%. During the 1986 season, injuries caused by impact from another player's knee brace were also noted and were reported as the cause of only 0.3% of injuries. Thus wearing a knee brace appears to play a negligible role in causing injuries to other players.

The knee and the ankle were injured most frequently (table 1), followed by the shoulder, quadriceps, head, neck, and hamstrings. Sprains were by far the most common type of injury (table 2), followed by muscle strains, contusions, fractures, and cerebral concussions. Being tackled, blocking, tackling and being blocked were the most common mechanisms of injury (table 3), followed by noncontact categories ("sprinting/running" "no evidence of contact").

_______________________________________________________________________________________

Table 1. College Football Injuries by Anatomic Site

_______________________________________________________________________________________

 

Total Injuries

Injuries/1,000 AEs*

Knee

522

1.17

Ankle

446

1.00

Shoulder

297

0.67

Quadriceps

180

0.40

Head

155

0.35

Neck

133

0.30

Hamstring

130

0.29

Lower back

108

0.24

Finger(s)

85

0.19

Pelvis, hip

84

0.19

Lower leg

81

0.18

Hand

58

0.13

Foot

54

0.12

Elbow

53

0.12

Ribs

52

0.12

Groin

51

0.11

Toe(s)

46

0.10

Thumb

38

0.09

Wrist

26

0.06

Clavicle

24

0.05

Other

24

0.05

Heel/Achilles tendon

23

0.05

Forearm

17

0.04

Patella

17

0.04

Eye(s)

13

0.03

Sternum

13

0.03

Chin

12

0.03

Upper back

12

0.03

Nose

11

0.02

Stomach

9

0.02

Upper arm

7

0.02

Spine

7

0.02

Kidney

5

0.01

Jaw

4

0.01

Testicles

4

0.01

Face

3

0.01

Mouth

3

0.01

Teeth

3

0.01

Coccyx

3

0.01

Scapula

2

0.00

Buttocks

2

0.00

Spleen

2

0.00

Ear

1

0.00

Totals

2,280

6.32

_______________________________________________________________________________________

* Athlete-exposure: each time a player takes part in a practice session or game, thus being exposed to the risk of injury.

_______________________________________________________________________________________

 

_______________________________________________________________________________________

Table 2. College Football Injuries by Type

_______________________________________________________________________________________

 

Total Injuries

Injuries/1,000 AEs*

Sprained ligament

899

2.02

Strained muscle

524

1.18

Contusion

438

0.98

Fracture

193

0.43

Cerebral concussion

143

0.32

Complete ligament tear

88

0.20

Nerve injury

87

0.20

Torn cartilage

66

0.15

Other

58

0.13

Subluxation

54

0.12

Dislocation

53

0.12

Laceration

43

0.10

Hyperextension

35

0.08

Separation

31

0.07

Tendinitis

25

0.06

Heat exhaustion

15

0.03

Inflammation

15

0.03

Bursitis

10

0.02

Stress fracture

9

0.02

Torn tendon

9

0.02

Abrasion

8

0.02

Infection

6

0.01

Hemorrhage

3

0.01

Hernia

3

0.01

Internal injury

2

0.00

Blister

1

0.00

Burn

1

0.00

Heat Stroke

1

0.00

Totals

2,820

6.32

_______________________________________________________________________________________

* Athlete-exposure: each time a player takes part in a practice session or game, thus being exposed to the risk of injury.

_______________________________________________________________________________________

 

_______________________________________________________________________________________

Table 3. Mechanism of Injury in College Football

_______________________________________________________________________________________

 

Total Injuries

Injuries/1,000 AEs*

Tackled

512

1.15

Blocking

497

1.11

Tackling

458

1.03

Blocked

397

0.89

Sprinting/running

277

0.62

No evidence of contact

157

0.35

Impact with playing surface

119

0.27

Overuse/gradual onset

119

0.27

Stepped on/kicked

97

0.22

Other

77

0.17

Blocked below waist

72

0.16

Clipped

17

0.04

Lifting weights

16

0.04

Blocking a kick/punt

5

0.01

Totals

2,820

6.32

_______________________________________________________________________________________

* Athlete-exposure: each time a player takes part in a practice session or game, thus being exposed to the risk of injury.

_______________________________________________________________________________________

Offensive players incurred 50.9% of the reported injuries, defensive players 45.2%, and special team players and kickers/punters 3.9% (table 4). Injury rates were adjusted for the number of players in each position on the field at any one time (eg, one quarterback, two offensive tackles). These adjusted rates indicate that the flankers/wide receivers and running backs apparently had the highest rates, followed by offensive ends. Defensive linebackers and halfbacks/cornerbacks had the highest rates among defensive players, which were similar to the injury rate of offensive ends. Defensive safeties and offensive slotbacks/wingbacks had the lowest rates. All other positions were roughly equal.

Injuries occurred most often during the third quarter and least often during the first quarter (table 5).

The overall injury rate was 60% higher on artificial turf than on natural grass (table 6). Of the 445,856 AEs over the two seasons, 386,957 (86.8%) were on grass and 58,899 (13.2%) on artificial turf. In games, 27,978 AEs (6.2%) were on grass and 7,713 (1.7%) on artificial turf; in practices, 358,979 AEs (80.5%) were on grass and 51,186 (11.5%) on artificial turf.

_______________________________________________________________________________________

Table 4. College Football Injuries by Playing Position

_______________________________________________________________________________________

 

No. of Injuries

Injuries/1,000 ___AEs*___

No. of Players at Position

Injuries/1,000 AEs (Adjusted)

Offense

 

 

 

 

Running back

402

0.90

2

0.45

Flanker/wide receiver

237

0.53

1

0.53

Tackle

221

0.50

2

0.25

Guard

201

0.45

2

0.23

End

138

0.31

1

0.31

Quarterback

121

0.27

1

0.27

Center

91

0.20

1

0.20

Slotback/wingback

24

0.05

1

0.05

Totals

1,435

3.21

11

 

Defense

 

 

 

 

Linebacker

444

1.00

3

0.33

Down lineman

427

0.96

4

0.24

Halfback/cornerback

271

0.61

2

0.31

Safety

133

0.30

2

0.15

Totals

1,275

2.87

11

 

Special teams

88

0.20

 

 

Kicker/punter

16

0.04

 

 

Other

6

0.01

 

 

Totals

2,820

6.32

 

 

_______________________________________________________________________________________

* Athlete-exposure: each time a player takes part in a practice session or game, thus being exposed to the risk of injury.

Adjusted for the number of players in each type of position on the field at any one time (eg, one quarterback, two offensive tackles).

_______________________________________________________________________________________

 

_______________________________________________________________________________________

Table 5. Injury Rates in College Football Games by Quarter

_______________________________________________________________________________________

 

Total Injuries

Injuries/1,000 AEs*

1

186

5.21

2

342

9.58

3

382

10.71

4

292

8.18

Totals

1,202

33.68

_______________________________________________________________________________________

* Athlete-exposure: each time a player takes part in a practice session or game, thus being exposed to the risk of injury.

_______________________________________________________________________________________

Table 6. Football Injuries per 1,000 AEs* During Practice Sessions and Games by Type of Surface

_______________________________________________________________________________________

 

Practices

Games

Total

Natural

3.84

31.27

5.82

Artificial

4.20

43.43

9.34

_______________________________________________________________________________________

* Athlete-exposure: each time a player takes part in a practice session or game, thus being exposed to the risk of injury.

_______________________________________________________________________________________

 

Discussion

The overall injury rate found by the NSISS during the 1986 and 1987 seasons (6.32/1,000 AEs) was similar to rates reported in previous years by the NCAA/ISS (1-3), which used a study design similar to that of the NSISS but a different sample. Injury rates during practices (3.93/1,000 AE) and games (33.68/1,000 AE) also were similar to the NCAA reports. The distribution of injuries by site (table l) and type (table 2) was consistent with previous results showing that the knee is the most frequent site of injury and that sprains, strains, and contusions were the most common types of injuries. The combined data indicate that ankle sprains were the most common injury, followed by knee ligament sprains and other knee injuries (eg, complete ligament tear or torn meniscus). Upper leg strains (quadriceps and hamstrings) were the next most frequent injuries.

Although the head is protected by a helmet, cerebral concussion was the fifth most frequent injury (table 2). This is cause for concern, especially when one remembers that every concussion recorded kept the player out for 1 day or more. In other words, these were more serious head injuries than the common "bell ringers" that cause transient disorientation but are followed by immediate return to play. As noted by Mueller and Schindler (7), coaches and referees must do a better job of emphasizing and enforcing the rules against using the head as an initial contact point.

Between 8% and 9% of all injuries were season-ending, and a similar proportion (but not all the same athletes) required surgery. The percentage of season-ending injuries probably would have been higher, if not for the increasingly common use of arthroscopic surgery, which has enabled many players to return to finish the season.

One would expect that most football injuries result from contact, primarily from tackling, being tackled, blocking, and being blocked. Thus it is interesting that despite the heavy contact involved in the sport, approximately 20% of the injuries occurred in the noncontact categories (sprinting/running, no evidence of contact, overuse/gradual onset, and lifting weights) (table 3). Many of these injuries probably could be avoided with the use of proper technique, conditioning, and stretching.

The rate of injuries by position (table 4) seemed reasonably equivalent across most positions after being adjusted for the number of players on the field at each position. It is not surprising that the running backs, who are the primary ball handlers, had a relatively high injury rate, especially since being tackled was the most frequent mechanism of injury. A high injury rate for flankers/wide receivers would be expected because they tend to be involved in high-speed collisions. The low rate for slotbacks/wingbacks was not expected, but all teams may not use the same terms to describe this position. In such cases the slotbacks/wingbacks most likely would be categorized as flankers/wide receivers, which would help explain the high injury rate in that category and the low rate for slotbacks/wingbacks.

The injury rates in games by quarter (table 5) suggest a way to reduce these injuries. The lowest rate was in the first quarter and the highest during the third quarter; this pattern also was observed in the NCAA/ISS data (1). It is usually assumed that injury rates increase during each half as a result of increasing exhaustion as a game progresses. However, if exhaustion were a primary factor, one would expect the injury rate to be higher in the second quarter than in the first, and lower again in the third quarter after the players had rested at halftime, with the rate increasing again in the fourth quarter. The pattern we found suggests that the injury rate was lowest in the first quarter because the players had warmed up and stretched before the game and was highest during the third quarter because the players had rested and cooled down during halftime. This hypothesis relates only to injuries that are affected by warming up and stretching.

Sprains and muscle strains also occurred most frequently during the third quarter and least often during the first quarter. This implies that coaches, trainers, and team physicians could help reduce injuries during the third quarter by ensuring that players have an opportunity to warm up and stretch before the third period starts. However, this is likely to happen only if the NCAA Football Rules Committee mandates an extra amount of time between halves--eg, 5 minutes during which the players must be on the field warming up and stretching.

Exhaustion no doubt plays a role in the incidence of injuries as a game progresses. But it apparently is not the most important factor, at least not in the first quarter, when warming up and stretching seem to reduce injuries, or in the third quarter, when the lack of warm-up and stretching may lead to increased injuries. During the second and fourth periods exhaustion may become a more important factor, with resulting intermediate rates of injury.

Controversy continues as to whether playing on artificial turf causes more injuries than playing on natural grass. Data from the first 2 years of operation of the NCAA/ISS (1) showed a consistent pattern of injury rates that were 50% higher on artificial turf, whether looking at total injuries, major injuries( > 7 days lost from play), total knee injuries, or knee injuries that required surgery. The data gathered by the NSISS from a different sample showed a similar pattern (table 6). This pattern of higher injury rates on artificial surfaces has also been noted in men's and women's lacrosse (Zemper ED: unpublished data). This consistent finding over a 6-year period, by two data collection systems using two different samples, indicates the need for a thorough investigation of this question.

No conclusions have been made regarding the effect of the brand or age of the artificial surface used. New surfaces may have characteristics that make them significantly better than the earlier types. Because the injury rate may be greatly affected by a particular brand and by the age of the surface, the NSISS has been collecting data on these factors. However, not enough data have been gathered yet for reasonable analyses.

Concluding Comments

The results presented here from the first 2 years of operation of the NSISS are consistent with earlier results obtained with a similar data collection system (NCAA/ISS) using a different sample. These results demonstrate the practicality of the system and its ability to provide descriptive summaries of epidemiologic data and special analyses of specific questions. The NSISS data collection forms are brief and easy to use, yet they provide sufficient data to allow analyses of national injury patterns and trends and of special questions such as the studies of knee braces, helmets, and artificial turf. The high return rate of completed forms indicated excellent cooperation by the trainers. As with any data collection system of this type, it is most useful when it is longitudinal. The NSISS is attempting to achieve this goal by continuing to operate and provide reports on injury rates in football and other sports.

Acknowledgment

This study was funded by the international Institute for Sport and Human Performance, University of Oregon, Eugene.

References

1. Zemper ED: NCAA Injury Surveillance System: initial results. Presented at the 1984 Olympic Scientific Congress, July 19, 1984, Eugene, OR

2. NCAA: Injury data released by NCAA. The NCAA News 1988;(0ct 3):1-2

3. NCAA: Study shows injury trends in fall sports for '88 season. NCAA News 1989; Feb 1:2

4. Zemper ED, Pieter W: Injury rates during the 1988 US Olympic team trials for taekwondo. Br J Sports Med 1989, to be published

5. Robey JM, Blyth CS, Mueller FO: Athletic injuries. Application of epidemiologic methods. JAMA 1971; 217(2): 184-189

6. Martin GL: The 1970 Intercollegiate Tackle Football Injury Surveillance Report. Prepared for the Joint Commission on Competitive Safeguards and Medical Aspects of Sports, Lincoln, University of Nebraska, 1970

7. Mueller FO, Schindler RD: Annual Survey of Football Injury Research 1931-1988. Prepared for American Football Coaches Association, National Collegiate Athletic Association, National Federation of State High School Associations. American Football Coaches Association, Orlando, FL, February 1989

8. Alles W, Powell JW, Buckley WE, et al: Three-year summary of NAIRS football data. Athletic Training 1980; 15(2):98-100

9. Buckley WE: Five year overview of sport injures: the NAIRS model. JOPERD 1982;53(5):36-40

10. NATA: National High School Injury Registry: summary reports released by the National Athletic Trainers Association. Feb 3, 1987

11. Zemper ED: Cerebral concussion rates in various brands of football helmets. Athletic Training 1989; 24(2):133-137

12. Zemper ED: A two-year prospective study of prophylactic knee braces in a national sample of college football players. Sports Medicine, Training and Rehabilitation 1989, to be published

13. Caine D, Cochrane B, Caine C, et al: An epidemiological investigation of injuries affecting young competitive female gymnasts. Am J Sports Med 1989, to be published

14. Thompson N, Halpren B, Curl WW, et al: High school football injuries: evaluation. Am J Sports Med 1987; 15(2):117-124 (Published erratum appears in Am J Sports Med 1987;15(3):257)