CLINICAL SCIENCES
Predictive Value of Orthopedic Evaluation
and Injury History at the NFL Combine
ROBERT H. BROPHY1, ERIC L. CHEHAB1, RONNIE P. BARNES2, STEPHEN LYMAN1,
SCOTT A. RODEO1, and RUSSELL F. WARREN1
1
Shoulder and Sports Medicine, Hospital for Special Surgery; and 2New York Giants, New York, NY
ABSTRACT
BROPHY, R. H., E. L. CHEHAB, R. P. BARNES, S. LYMAN, S. A. RODEO, and R. F. WARREN. Predictive Value of Orthopedic
Evaluation and Injury History at the NFL Combine. Med. Sci. Sports Exerc., Vol. 40, No. 8, pp. 1368–1372, 2008. Purpose: The
National Football League (NFL) holds an annual combine to evaluate college football athletes likely to be drafted for physical skills, to
review their medical history, and to perform a physical examination. The athletes receive an orthopedic grade on their ability to
participate in the NFL. The purpose of this study was to test the hypothesis that this orthopedic rating at the combine predicts the
percent of athletes who play in the NFL and the length of their careers. Methods: A database for all athletes reviewed at the combine by
the medical staff of one team from 1987 to 2000 was created and linked to a data set containing the number of seasons and the games
played in the NFL for each athlete. Players were grouped by orthopedic grade: high, low, and orthopedic failure. The percent of players
who played in the NFL and the mean length of their careers was calculated and compared for these groups. Results: The orthopedic
grade assigned at the NFL combine correlated with the probability of playing in the league. Whereas 58% of athletes with a high grade
and 55% of athletes with a low grade played at least one game, only 36% of athletes given a failing grade did so (P G 0.001). Players
with a high grade had a mean career of 41.5 games compared with 34.2 games for players with a low grade and 19.0 games for
orthopedic failures. Conclusion: This is the first study to report on the predictive value of a grading system for college athletes before
participation in professional sports. Other professional sports may benefit from using a similar grading system for the evaluation of
potential players. Key Words: AMERICAN FOOTBALL, ACL, SHOULDER, COLLEGE FOOTBALL
T
skeletal disorders in this population and found that the
percentage of players rejected for orthopedic reasons
decreased over the study period (1). The most common
major procedures in this population were anterior cruciate
ligament (ACL) reconstruction and shoulder stabilization
surgery, and both procedures became more common over
the study period. Of note, players with a history of ACL
reconstruction were less likely to be rejected for orthopedic
reasons during the study period.
Ultimately, the goal of the combine is to predict the
probability of success for each athlete in the NFL. Ideally,
the orthopedic rating should predict both the likelihood of
playing in the league and the duration of an athlete’s career
in the league. Enhancing the combine database to include
games played, games started, and seasons played in the
NFL for all combine participants who participated in at least
one game in the NFL regular season would enable an
analysis of how well the orthopedic rating predicts success
in the league. Adding these data would also allow analysis
of how specific diagnoses and surgical procedures, for example, ACL reconstruction, may predict the likelihood of
playing in the NFL.
The purpose of this study was to assess the predictive
value of the combine participants’ orthopedic grades and
specific diagnoses and procedures concerning the length of
career for players who entered the league in 1987 to 2001.
We hypothesized that the orthopedic grade at the combine predicts the percent of athletes who play in the NFL
he National Football League (NFL) holds an annual
combine where individual teams evaluate college
football players likely to be drafted for their physical skills. As part of the process, the teams review the
players’ medical history and imaging studies as well as a
detailed orthopedic examination. Each team separately
assigns an orthopedic grade based on a subjective interpretation of the objective history, exam, and imaging studies.
This grade attempts to predict the athlete’s ability to participate in the NFL.
A database for all players reviewed at the annual combine
by the medical staff of one NFL team from 1987 to 2000
was created. The database includes each player’s position,
collegiate division, and orthopedic rating assigned by one
team at the NFL combine as well as their diagnoses and
surgical procedures. A previous study based on this database reported the prevalence and the trends of musculo-
Address for correspondence: Robert H. Brophy, MD, 14532 South Outer
Forty Drive, Chesterfield, MO 63017; E-mail: brophyr@wudosis.
wustl.edu.
Submitted for publication August 2007.
Accepted for publication February 2008.
0195-9131/08/4008-1368/0
MEDICINE & SCIENCE IN SPORTS & EXERCISEÒ
Copyright Ó 2008 by the American College of Sports Medicine
DOI: 10.1249/MSS.0b013e31816f1c28
1368
Copyright @ 2008 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
TABLE 1. Summary of orthopedic ratings for athletes at the NFL combine 1987–2000.
Rating
High pass
3536
Low pass
1227
Failure
284
and the length of their careers. In addition, the study also
investigated whether specific diagnoses (e.g., ACL tear or
shoulder instability) and procedures (e.g., ACL reconstruction or shoulder stabilization) predicted the likelihood of
playing in the NFL.
METHODS
A database for all players reviewed at the annual combine
by the medical staff of one NFL team from 1987 to 2000
was created, including each player’s position, collegiate
division, and orthopedic rating as well as their diagnoses
and surgical procedures. As described in our previous study
(1), the orthopedic rating ranged from the highest grade for
players with no history of injury to an orthopedic failure
for players with a history of severe or recurrent injury which
was likely to limit their ability to perform in the NFL. The
senior investigator (RFW) directed the assigning of orthopedic grades over the entire study period. We then linked
this database to a data set from the Elias Sports Bureau
(New York, NY) containing the round drafted, the number
of seasons, and the number of games played in the NFL
for each athlete who first played in the league from 1987 to
2001. Players were grouped by orthopedic grade as follows:
high (passing) grade, low (passing) grade, and failing grade
(Table 1). Athletes received a high passing grade if they
had no history of surgery and only minor injuries, which
had not limited their playing time while in college and
would not limit their ability to compete in the NFL. Athletes
were assigned a low passing grade if they were 1) currently
healthy but had a history of injury or surgery, which had
limited their playing time in college or might limit their
playing time as a professional, or 2) had a recent injury from
which they would be expected to recover, that is, ankle
sprain at the end of their senior season. A failing grade was
assigned to athletes who had a history of significant injury
or surgery, which was an active problem or likely to limit
their playing time in the NFL. The percent of players who
played in the NFL and the mean length of their careers were
calculated and compared for these groups.
The percent of athletes who played in the NFL was also
calculated for specific diagnoses and procedures. We compared the percentage of all athletes with a given diagnosis
or procedure, including those with a history of other diagnoses and procedures, playing in the league to the percentage of all athletes without that specific diagnosis or
procedure playing in the league. Based on our previous
PREDICTIVE VALUE OF NFL COMBINE
study, the percent of athletes with a history of ACL surgery
or shoulder stabilization that played in the NFL was
evaluated for change over the study period. We also tested
whether a history of multiple diagnoses or procedures
would affect a player’s prospect of making it in the NFL.
The players with a failing medical grade were not
selected by the team affiliated with this study, but they
may have still been selected by other teams. This group of
athletes was reviewed in isolation as to the impact of
specific diagnoses and procedures on their ability to play
in the NFL. In addition, the talent rating from the combine was obtained for these athletes to assess how talent
influenced the likelihood of playing in the league. This
talent rating was the subjective assessment of the team
scouts based on the athlete’s performance in college and at
the combine.
Chi-square tests were calculated to test the hypotheses
that the percentage of athletes who played in the league
varied by orthopedic grade, diagnosis, and procedure. Length
of career in years and games by orthopedic grade was assessed using ANOVA and post hoc t-tests with a Bonferroni
correction for multiple comparisons. A P value of less than
0.05 was considered statistically significant. All analyses
were performed using SAS for Windows (Cary, NC).
RESULTS
Overall, 56% of the players at the combine played at
least one game in the NFL. The orthopedic grade assigned
at the NFL combine was correlated with the probability of
playing in the NFL. Whereas 58% of players with a high
grade and 55% of players with a low grade played at least
one game in the NFL, only 36% of players given a failing
grade played at least one game in the NFL (P G 0.0001).
Similarly, the length of playing career correlated with orthopedic grade. Among players who played at least one
game in the NFL, players with a failing grade played
significantly fewer years in the league (5.0 T 3.0 yr) than
players with a high grade (6.2 T 3.6 yr; P G 0.001) or a low
grade (5.9 T 3.4 yr; P = 0.049). However, the difference in
terms of games played was slightly more substantial.
Players with a high grade played a mean of 71.0 games,
TABLE 2. Impact of specific diagnoses on probability of playing in the NFL.
Number of Players
At Combine Play Q1 Game Percentage
Diagnosis
More likely to play
Humerus fracture
Spondylolisthesis
Foot injury (not turf toe or fracture)
Fracture MC/phalanges
Nonspecified knee injury
Burner
Hamstring strain
Ankle sprain
Wrist or hand injury (not fracture)
Less likely to play
Shoulder instability
Meniscal injury
ACL
19
51
161
434
517
765
581
1298
770
15
36
108
277
321
467
354
784
463
78.9
70.6
67.1
63.8
62.1
61.0
60.9
60.4
60.1
447
570
371
229
292
176
51.2
51.2
47.4
Medicine & Science in Sports & Exercised
1369
Copyright @ 2008 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
CLINICAL SCIENCES
Qualitative Description
No injury
Minor injury (no surgery, successful recovery)
Successful surgical interventions or
multiple problems
High risk player (incomplete recovery
and/or injury likely to recur)
Reject
Number of Athletes
with this Rating
CLINICAL SCIENCES
TABLE 3. Effect of multiple diagnoses or procedures on likelihood of playing in the NFL.
Percentage of
Athletes Playing
Q1 Game
Diagnoses
Procedures
Less than 4
4 or more
0
1
2 or more
55.7
58.5
57.9*
54.6*
52.2*
Length of Career (in Games)
Among Athletes Playing
Q1 Game (Mean T SD)
70.0
63.3
73.1
60.3
58.3
T
T
T
T
T
54.7*
48.9*
54.9*
50.0*
47.9*
*P G 0.05.
which was significantly higher than players with a low
grade (62.2 games; P G 0.001) and players with a failing
grade (52.3 games; P G 0.001). Combining the likelihood
to play in the league with the length of career for athletes
who actually played in the league, the expected length
of career for any athlete at the combine with a high orthopedic grade was 41.5 games compared with 34.2 games
for players with a low grade and 19.0 games for players
with a failing grade.
Players with nine diagnoses were found to still be very
likely to play in the league (Table 2). The diagnosis with
the greatest percentage of athletes playing in the NFL was
humerus fractures, with 78.9% of players going on to play
in the league. Players with spondylolisthesis were also
very likely to play in the league, with 70.6% of them playing at least one game. Athletes with shoulder instability,
meniscal injury, and ACL injury were less likely to play
in the league than athletes without these diagnoses.
No procedures were associated with an increased likelihood of playing in the league. ACL reconstruction was
associated with a decreased probability of playing in the
league as only 45.9% (130/283) of those athletes played at
least one game in the NFL.
TABLE 4. Impact of specific diagnoses on the probability of playing in the NFL among
athletes who received a failing orthopedic grade at the combine.
Number of Players
Diagnosis
More likely to play
Tarsal fracture
Turf toe
Low back pain
Less likely to play
Spine (noncervical,
nonspondylolysis or listhesis)
At Combine
Play Q1 Game
Percentage
3
13
34
3
9
18
100.0
69.2
52.9
17
1
5.9
Multiple diagnoses and/or procedures were associated
with a shorter playing career (Table 3). Athletes with four
or more diagnoses did not have a decreased probability of
playing in the league, but they played significantly fewer
games over the course of their career compared with athletes with less than four diagnoses. Athletes with more
procedures were both less likely to play in the league and
had shorter careers in terms of games played.
Over the study period, players with a previous ACL reconstruction or shoulder stabilization became more likely
to play at least one game in the NFL (Fig. 1). For example,
in the first 2 yr of the study, 42% of the players with a
previous ACL injury played at least one game in the NFL
compared with 58% in 1999 to 2000. In the first 2 yr of the
study, only 30% of the players with previous shoulder stabilization played at least one game in the NFL compared
with 53% in the last 2 yr of the study period.
Athletes received a failing grade at the combine for a
variety of reasons. The athletes with failing grades who did
go on to play in the league were compared with the athletes
with a failing grade who did not play in the league to search
for any injury patterns that might predict the probability
FIGURE 1—Probability of playing in the NFL after ACL reconstruction and shoulder stabilization: trends from 1987 to 2000.
1370
Official Journal of the American College of Sports Medicine
http://www.acsm-msse.org
Copyright @ 2008 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
TABLE 5. Effect of talent rating on success in the NFL among athletes who received a
failing orthopedic grade at the combine
A
B
C
D
F
Qualitative
Description
Total
Percentage
Played in NFL
Mean Length
of Career
Quality starter
Solid backup
Average backup
Below average backup
Reject
15
31
52
82
66
93
74
54
29
8
69
49
24
12
3
of making it in the NFL. No significant differences were
found. Overall, the mean number of diagnoses (play 3.48,
no play 3.33) and procedures (play 0.87, no play 1.20) was
similar between the two groups. The types of diagnoses and
procedures were also similar between the two groups, with
ACL reconstruction and knee meniscus surgery the most
common procedures in both groups. Among failures only,
athletes with three diagnoses were still very likely to play in
the league whereas athletes with noncervical, nonspondylolysis or listhesis spine injury were very unlikely to play
(Table 4).
Among athletes who were assigned a failing orthopedic
grade, talent was significantly correlated with the likelihood
of playing in the league (Table 5). More talented athletes
were both more likely to play in the league and more likely
to last in the league. Among the failures who went on to
play in the NFL, the average talent rating was a C compared
with an average talent rating midway between D and F for
those who did not play in the league.
DISCUSSION
This is the first study to report on whether evaluating
elite college athletes before participation in professional
sports correlates to their professional career. The orthopedic
rating system used by one team at the NFL combine from
1987 to 2000 was correlated with the likelihood of playing
in at least one game in the NFL and the length of playing
career. A history of specific diagnoses and procedures are
also correlated with the probability of playing in the league.
College football players should be advised of the potential impact of specific injuries on their aspirations for a
career in professional football although the data should be
interpreted with some caution. Obviously, diagnoses with a
significantly higher percentage of athletes playing in the
league do not enhance or preclude ones ability to make a
team. For example, the data do support the conclusion that
a humerus fracture does not preclude a career in the NFL.
Many of the other injuries that were associated with a
higher probability of playing in the league, such as ankle
sprains, hamstring strains, and burners, are very common
and may reflect more exposure, that is, greater playing time,
in college.
Spondylolisthesis does not reduce the probability of
playing the league. This finding is supported by Shaffer
et al. (8) who found that the prevalence of spondylolisthesis
among college football players was 0.9% compared with
1.5% among NFL players. The trends are similar in our
PREDICTIVE VALUE OF NFL COMBINE
Medicine & Science in Sports & Exercised
1371
Copyright @ 2008 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
CLINICAL SCIENCES
Talent Rating
data. In our initial study (1), we reported a prevalence of
1.1% among all athletes at the combine. In the current
study, the prevalence of spondylolisthesis among athletes
who played at least one game in the league is 1.3%. So
although spondylolisthesis may be a problematic diagnosis
for some individual athletes, overall it would appear that
it is more of an incidental finding among football players
rather than a cause for clinical concern.
In stark contrast to spondylolisthesis, injury to the ACL is
particularly devastating with less than half of players with
an ACL injury or reconstruction going on to play professionally. A recent study that examined the impact of ACL
injury on running backs and wide receivers already playing
in the NFL showed that less than 80% of them returned
to play, and the performance of those who did play was
reduced by one third (2). Knee meniscal injury and shoulder
instability also decrease the likelihood of a professional
football career. However, the likelihood of playing after
surgical treatment for ACL injury and shoulder instability
increased over the study period, suggesting an improvement
in outcome after these procedures, possibly due to improved
surgical techniques, rehabilitation protocols, or both.
Shoulder instability is particularly important in light of a
recent study showing that one third of elite college athletes
have a history of shoulder surgery, with anterior instability
reconstruction making up 48% of those surgeries and posterior instability surgery making up another 10% (4). It is
currently unknown how the role of open versus arthroscopic techniques influenced the positive trend in playing
after shoulder stabilization. Numerous studies have shown a
greater rate of failure after arthroscopic stabilization in
American football players (3,5,7,9). In 2002, Pagnani and
Dome (6) reported good results using open stabilization in
football players after switching from arthroscopic stabilization in these athletes in 1993. The trends we observed may
have resulted from better arthroscopic techniques or more
frequent use of open techniques during the study period.
Although this study uses retrospectively collected data
from the combine, given the high stakes at risk, the
assessment process at the combine is likely to be both
thorough and accurate. Career data are an objective and a
quantitative measure of a player’s health, although it is
obviously confounded by the level of player talent. One
measure that might control for talent is percent of games
started, and there was no significant difference between the
orthopedic rating groups in terms of percent of games started.
We also examined the influence of talent on the success
of athletes who were rejected for medical reasons. Clearly,
talent had a significant influence on whether these athletes
went on to play in the league. One confounding factor is
that some of the talent grades may have been reduced due to
awareness about their medical condition. Unfortunately,
talent ratings were not available for the entire cohort, which
might have allowed a more rigorous analysis of the relative influence of talent versus medical rating on ultimate
success in the NFL.
CLINICAL SCIENCES
The utility of this study is limited by the lack of strict
criteria for failing an athlete at the combine. The final rating
reflects a consensus agreement based on history, physical
examination, and imaging studies as interpreted by the
orthopedic staff from one NFL team. As such, it cannot be
reduced to a specific formula or even a qualitative summary description. This study can only affirm that the concept of using a rating system is valid, not the rating system
itself. Looking at the effect of specific diagnoses and
procedures provides some objective criteria upon which to
base decisions about ratings but does not overcome this inherent limitation.
Another limitation of this study is the unknown effect of
league expansion upon our findings. In 1995, the league
expanded from 28 teams to 30 teams. Four years later, an
additional team was added. With more teams, the league
obviously absorbs a greater number of players from the
combine. However, changes in team roster size also oc-
curred over the study period. Although the active roster size
remained steady at 45 throughout the study period, the
inactive limit decreased from 5 in 1987 to 2 in 1988. This
limit increased to 8 in 1993 and remained at that level
throughout the study period. The percentage of failures who
played in the league did not increase over the study period
nor did it jump in response to expansion. Overall, it would
appear reasonable to assume the net effect of changes over
the study period was negligible.
Despite the limitations, this study suggests that the orthopedic evaluation performed as part of the NFL combine
is a useful tool for team physicians, trainers, coaches, and
managers. The orthopedic rating identifies athletes who are
less likely to play in the league (i.e., failures) and predicts
longevity in the league. Other professional sports might
benefit from using a similar grading system for the evaluation of potential players, particularly if specific grading
criteria tailored to each individual sport were developed.
REFERENCES
1. Brophy RH, Barnes R, Rodeo SA, Warren RF. Prevalence of
musculoskeletal disorders at the NFL combine—trends from 1987
to 2000. Med Sci Sports Exerc. 2007;39:22–7.
2. Carey JL, Huffman GR, Parekh SG, Sennett BJ. Outcomes of
anterior cruciate ligament injuries to running backs and wide receivers in the National Football League. Am J Sports Med. 2006;
34:1911–7.
3. Cole BJ, L’Insalata J, Irrgang J, Warner JJ. Comparison of
arthroscopic and open anterior shoulder stabilization. A two to
six-year follow-up study. J Bone Joint Surg Am. 2000;82:1108–14.
4. Kaplan LD, Flanigan DC, Norwig J, Jost P, Bradley J. Prevalence
and variance of shoulder injuries in elite collegiate football players.
Am J Sports Med. 2005;33:1142–6.
5. O’Neill DB. Arthroscopic Bankart repair of anterior detachments of
1372
Official Journal of the American College of Sports Medicine
6.
7.
8.
9.
the glenoid labrum. A prospective study. J Bone Joint Surg Am.
1999;81:1357–66.
Pagnani MJ, Dome DC. Surgical treatment of traumatic anterior
shoulder instability in American football players. J Bone Joint Surg
Am. 2002;84:711–5.
Pagnani MJ, Warren RF, Altchek DW, Wickiewicz TL, Anderson
AF. Arthroscopic shoulder stabilization using transglenoid sutures. A
four-year minimum followup. Am J Sports Med. 1996;24:459–67.
Shaffer B, Wiesel S, Lauerman W. Spondylolisthesis in the elite
football player: an epidemiologic study in the NCAA and NFL.
J Spinal Disord. 1997;10:365–70.
Speer KP, Warren RF, Pagnani M, Warner JJ. An arthroscopic
technique for anterior stabilization of the shoulder with a bioabsorbable tack. J Bone Joint Surg Am. 1996;78:1801–7.
http://www.acsm-msse.org
Copyright @ 2008 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.