Team Physician’s Corner

Reliability, Validity, and Injury Predictive M

Value of the Functional Movement Screen
A Systematic Review and Meta-analysis
Nicholas A. Bonazza,* MD, Dallas Smuin,y BS, Cayce A. Onks,z DO, MS, ATC,
Matthew L. Silvis,z MD, and Aman Dhawan,*§ MD
Investigation performed at the Milton S. Hershey Medical Center,
Penn State Health, Pennsylvania State University College of Medicine, Hershey, Pennsylvania, USA

Background: The Functional Movement Screen (FMS) is utilized by professional and collegiate sports teams and the military for
the prevention of musculoskeletal injuries.
Hypothesis: The FMS demonstrates good interrater and intrarater reliability and validity and has predictive value for musculoskel-
etal injuries.
Study Design: Systematic review and meta-analysis.
Methods: A systematic review and meta-analysis were conducted using a computerized search of the electronic databases
MEDLINE and ScienceDirect in adherence with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) guidelines. Extracted relevant data from each included study were recorded on a standardized form. The Cochran
Q statistic was utilized to evaluate study heterogeneity. Pooled quantitative synthesis was performed to measure the intraclass
correlation coefficient (ICC) for interrater and intrarater reliability, along with 95% CIs, and odds ratios with 95% CIs for the injury
predictive value for a score of 14.
Results: Eleven studies for reliability, 5 studies for validity, and 9 studies for the injury predictive value were identified that met
inclusion and exclusion criteria; of these, 6 studies for reliability and 9 studies for the injury predictive value were pooled for quan-
titative synthesis. The ICC for intrarater reliability was 0.81 (95% CI, 0.69-0.92) and for interrater reliability was 0.81 (95% CI, 0.70-
0.92). The odds of sustaining an injury were 2.74 times with an FMS score of 14 (95% CI, 1.70-4.43). Studies for validity dem-
onstrated flaws in both internal and external validity of the FMS.
Conclusion: The FMS has excellent interrater and intrarater reliability. Participants with composite scores of 14 had a signifi-
cantly higher likelihood of an injury compared with those with higher scores, demonstrating the injury predictive value of the test.
Significant concerns remain regarding the validity of the FMS.
Keywords: injury prevention; Functional Movement Screen

Professional and National Collegiate Athletic Association

§
Address correspondence to Aman Dhawan, MD, Department of
Orthopaedics and Rehabilitation, Penn State Health, 30 Hope Drive,
Building B, Suite 2400, Hershey, PA 17033-0850, USA (email: adhawan
@hmc.psu.edu).
*Department of Orthopaedics and Rehabilitation, Milton S. Hershey
Medical Center, Penn State Health, Pennsylvania State University College
of Medicine, Hershey, Pennsylvania, USA.
y
Pennsylvania State University College of Medicine, Hershey, Penn-
sylvania, USA.
z
Department of Family and Community Medicine, Milton S. Hershey
Medical Center, Penn State Health, Pennsylvania State University College
of Medicine, Hershey, Pennsylvania, USA.
One or more of the authors has declared the following potential con-
flict of interest or source of funding: A.D. is a consultant for Smith &
Nephew and is on the speakers’ bureau for Smith & Nephew and Biomet.
The American Journal of Sports Medicine, Vol. 45, No. 3
DOI: 10.1177/0363546516641937
Ó 2016 The Author(s)
(NCAA) sports teams and the United States (US) military
rely on physically healthy people to compose their respec-
tive work forces. Musculoskeletal injuries are a major
source of lost participation time, lost income, and medical
resources for the care of these injuries.35 The Functional
Movement Screen (FMS) is a screening test that was devel-
oped with the goal of identifying deficits in movements that
may predispose an otherwise healthy person to injuries
during activity.6-9 Preparticipation examinations have
long been used to assess a person’s ability to safely partic-
ipate in physical activity at the time of examination, but no
existing screening test has been shown to predict a person’s
risk of injury while participating in future activities. Kiesel
et al21 were the first to explore the possible predictive
value of the FMS when they found that lower FMS scores
were predictive of a significantly higher risk of injury in

725
726 Bonazza et al The American Journal of Sports Medicine

Initial Search Terms: MEDLINE Search

Functional Movement Screen Science Direct Search
Functional Movement Screening Total Articles Found Duplicates Removed
111

AND [Predict OR Prediction OR Injury

Predict OR Injury Prediction OR Predictive AND [Valid OR Validity
AND [Reliable Value or Injury Predictive Value] OR Effective or
OR Reliability] Effectiveness]

Articles Screened Articles Screened Articles Screened

30 45 20 33 33 43

Full Text Screened Full Text Screened Full Text Screened

4 15 4 13 5 10

Inclusion/Exclusion Inclusion/Exclusion Inclusion/Exclusion

Qualitative Analysis Qualitative Analysis Qualitative Analysis

5 11 9 5

Quantitative Analysis Quantitative Analysis

6 9

Figure 1. Search methodology.

professional football players in a 2007 study. The value of
such a screening test was quickly realized, and the FMS
was widely adopted in organizations such as the National
Football League (NFL), the National Hockey League
(NHL), and the US military.20,21,28,33,35
Subsequent studies regarding the FMS, however, have
produced varied results in regard to the injury predictive value
as well as the validity of the FMS as a screening
test.4,11,14,15,20,39,40 Analyses of the structure of the FMS have
questioned the ambiguity inherent in its grading structure
and its ability and sensitivity to identify functionally relevant
movement limitations.1,2,5,14,15,40 Interrater and intrarater reli-
ability were also introduced as possible sources for error in the
FMS, although some early studies found high reliability
among examiners with varying levels of experience.27
Given the implementation of the FMS in numerous
organizations and the growing body of literature examin-
ing the FMS, we performed a systematic review of the lit-
erature and meta-analysis to determine whether (1) the
FMS is a reliable screening tool; (2) the FMS is a valid
tool to identify functional asymmetries; and (3) if a lower
score, and what specific score, on the FMS correlates
with a higher risk of musculoskeletal injuries. We hypoth-
esized that the FMS demonstrates both interrater and
intrarater reliability with validity that can be used to iden-
tify people at a higher risk of an injury during activity.
searched for relevant studies with the primary search terms
‘‘Functional Movement Screen’’ OR ‘‘FMS’’ and secondary
search terms ‘‘reliable’’ OR ‘‘reliability,’’ ‘‘valid’’ OR ‘‘validity,’’
‘‘effective’’ OR ‘‘effectiveness,’’ ‘‘predict’’ OR ‘‘prediction,’’
‘‘injury predict’’ OR ‘‘injury prediction,’’ and ‘‘predictive
value’’ OR ‘‘injury predictive value.’’ Inclusion criteria
included (1) English-language studies in peer-reviewed jour-
nals and (2) use of the FMS to assess uninjured people before
participation in their respective activities. Any reviews, case
reports, technique articles, or abstracts were excluded. The
references of articles that met inclusion/exclusion criteria
were also hand reviewed to ensure that any additional rele-
vant studies were not missed.
Duplicates were removed from the results of each of the 3
separate searches. The titles and abstracts for all of the
studies were then screened by the senior author (A.D.) to
remove studies that did not involve the FMS. Each relevant
study was then reviewed by the senior author and assessed
using our inclusion and exclusion criteria for appropriate-
ness for qualitative and quantitative analyses in our study.
Qualitative and quantitative analyses were then performed
by all authors. Quantitative analysis was performed for all
articles in which data were sufficient to be included with
other studies in our meta-analysis as described below. A dia-
gram of our search methodology can be found in Figure 1.

Data Extraction and Statistics

METHODS
A written protocol was developed in adherence to the Pre-
ferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) guidelines to conduct a systematic
review and meta-analysis of the available literature.24 The
MEDLINE and ScienceDirect electronic databases were
Interrater and Intrarater Reliability. Interrater and
intrarater reliability were assessed using the intraclass
correlation coefficient (ICC). Where variance was not
directly reported, the confidence interval was used to
determine the variance using the Fisher method.13,19 An
evaluation for heterogeneity using the I2 statistic
AJSM Vol. 45, No. 3, 2017 Functional Movement Screen Meta-analysis 727

TABLE 1
Reliability Studiesa

Study (Year) Level of Evidence Study Population ICC or k (95% CI)

Elias12 (2013) 3 Physical therapists (N = 20) ICCInter = 0.906b

Gribble et al17 (2013)
Gulgin and
Hoogenboom18 (2014)
Minick et al27 (2010)
Onate et al29 (2012)
Parenteau et al30 (2014)
Schneiders et al34 (2011)
Shultz et al36 (2013)
Smith et al37 (2013)
Teyhen et al38 (2012)
Wright et al41 (2015)
3 Athletic training faculty and students (N = 38) ICCIntra = 0.754 (0.526-0.872)
3 Physical therapy students (n = 3); expert ICCInter = 0.88 (0.767-0.948)
FMS rater (n = 1)
3 Novice FMS raters (n = 2); expert FMS raters (n = 2) kInter = 0.79-1.0
3 Experienced FMS rater (n = 1); rater with no ICCInter = 0.98b
experience with FMS (n = 1) ICCIntra = 0.92b
3 FMS-certified physical therapists (N = 4) ICCInter = 0.96 (0.92-0.98)
ICCIntra = 0.96 (0.92-0.98)
3 Noncertified FMS researchers (N = 2) ICCInter = 0.971b
3 Undergraduate student (n = 1); physical therapist (n = 1); ICCInter = 0.29 (0.03-0.55)
athletic trainers (n = 2); strength and conditioning ICCIntra = 0.6 (0.35-0.77)
coaches (n = 2)
3 Rater with experience with FMS (no certification; n = 1); ICCInter = 0.89 (0.85-0.94)
rater with experience with FMS (certification; n = 1);
athletic training faculty (no certification; n = 1);
physical therapy student (no certification; n = 1)
3 Physical therapy students (N = 8; interrater: n = 4, ICCInter = 0.76 (0.63-0.85)
intrarater: n = 4) ICCIntra = 0.74 (0.60-0.83)
3 Researchers (n = 2) kInter = 0.11-0.83
kIntra = 0.23-0.87

a
FMS, Functional Movement Screen; ICC, intraclass correlation coefficient; Inter, interrater; Intra, intrarater.
b
95% CI not reported.

suggested significant heterogeneity between studies. A
meta-analysis was performed using the DerSimonian and
Laird10 random-effects model. A sensitivity analysis was
performed for the inclusion of the study of Smith et al,37
which reported 4 separate analyses of intrarater reliability
(1 for each subtype of observer). As Smith et al37 included
the individual ICC for each rater, the data were pooled to
reduce the analysis to a single ICC, which was then
included in the meta-analysis of the ICC.
Injury Predictive Value. The pooled effect measure that
was subjected to the meta-analysis was the odds ratio (OR)
for failure of functional movement. The numerical cutoff of
the FMS was evaluated at a value of 14, which was the
only value consistently utilized in all of the studies that
met inclusion/exclusion criteria. Again, there was signifi-
cant heterogeneity between studies based on the Mantel-
Haenszel Q statistic. Therefore, the DerSimonian and
Laird10 random-effects model was used to estimate the
pooled OR.26 All analyses were performed using R (version
3.1.3) statistical software32 and the rmeta package.25
Validity. Because of insufficient reporting and heteroge-
neity of the data, a pooled quantitative analysis of validity
could not be performed. Any reported data are included as
part of the results and discussion of the articles from which
they are extracted as part of our qualitative review.
RESULTS
The initial search using the primary search terms resulted
in 111 articles. Inclusion of the secondary search terms and
removal of duplicates resulted in 45 articles addressing
reliability, 43 articles addressing validity, and 33 address-
ing the injury predictive value, with several articles
addressing more than 1 aspect of the FMS. Inclusion and
exclusion criteria resulted in 11 studies evaluating reliabil-
ity for qualitative analysis, which can be found in Table 1,
and 6 studies for quantitative analysis/data synthesis.
Results that did not include the ICC with associated
95% CIs were unable to be included in the meta-analysis.
Inclusion and exclusion criteria resulted in 9 studies eval-
uating the injury predictive value for qualitative analysis
and 9 for quantitative analysis, which are described in
Table 2. Inclusion and exclusion criteria resulted in 5 stud-
ies evaluating validity for qualitative analysis (Table 3).
Given the variability of these studies for validity, a quanti-
tative analysis could not be performed.
Reliability
Interrater Reliability. Ten studies evaluated interrater
reliability of the FMS. All studies examined the reliability
of scores of more than 1 examiner grading the same partic-
ipants. Significant differences were seen in the character-
istics of the raters included in the studies. Five of the 10
studies included examiners of various levels of experience
with the FMS. Only a few studies included raters specifi-
cally certified in the FMS.
Nine of the 10 studies found acceptable interrater reli-
ability, with ICC values of 0.76 to 0.98. Shultz et al36
was the only study to report poor interrater reliability
with a Krippendorff alpha (a) value of only 0.38, which is
well below the 0.8 cutoff considered acceptable. The Cohen
kappa (k) coefficient was reported as a measure of reliabil-
ity for each individual FMS test component in 6 studies.
These values varied, despite most finding overall scores
had acceptable interrater reliability. Of the individual
728 Bonazza et al The American Journal of Sports Medicine

Study (Year)
3
Butler et al (2013)
Chorba et al4 (2010)
Dossa et al11 (2014)
Garrison et al16 (2015)
Kiesel et al21 (2007)
Kiesel et al22 (2014)
Knapik et al23 (2015)
28
O’Connor et al (2011)
Warren et al39 (2015)
TABLE 2
Injury Predictive Value Studiesa
Level of FMS
Study Design Evidence Study Population Injury Definition Cutoff Value OR (95% CI)
Prospective cohort 3 Firefighter trainees (N = 108) Missed time 14 8.31 (3.2-21.6)
Prospective cohort 3 NCAA Division II athletes Required medical 14 3.85 (0.980-15.130)
(N = 38; all female) attention
Prospective cohort 3 Major junior hockey players Missed time 14 2.33 (0.37-14.61)
(N = 20; all male)
Prospective cohort 3 NCAA Division I athletes Required medical 14 5.61 (2.73-11.51)
(N = 160) attention
Prospective cohort 3 Professional football players Missed time 14 11.67 (2.47-54.52)
(N = 46; all male)
Prospective cohort 3 Professional football players Missed time 14 2.33 (1.14-4.77)
(N = 238; all male)
Prospective cohort 3 Coast Guard cadets Required medical 14 1.42 (1.05-1.93)
(N = 1045; 770 male, 275 female) attention
Prospective cohort 3 Marine officer candidates Required medical 14 2.00 (1.29-3.08)
(N = 874; all male) attention
Prospective cohort 3 NCAA Division I athletes Required medical 14 1.01 (0.53- 1.92)
(N = 167; 89 male, 78 female) attention

a
FMS, Functional Movement Screen; NCAA, National Collegiate Athletic Association; OR, odds ratio.

Study (Year)
1
Beach et al (2014)
Clifton et al5 (2013)
Frost et al14 (2013)
Kazman et al20 (2014)
Whiteside et al40 (2015)
TABLE 3
Validity Studiesa
Level of External/Internal
Study Design Evidence Study Population Validity Significance
Case control 3 Male firefighters External FMS movements are difficult to replicate
and question external validity.
Prospective cohort 2 General population Internal Postural fatigue after exercise did not
(male/female) alter FMS scores.
Prospective cohort 2 Firefighters Internal Participants were able to increase composite
(male/female) scores with knowledge of scoring criteria.
Cohort 3 Marine officers Internal FMS movements are not interrelated as a
(male/female) unitary sum. Individual movement scores
may be more informative.
Cross-sectional 3 NCAA Division I External Manual grading is not sensitive enough
basketball players to detect cues for defects in joint angles
(male/female) and is subjective to raters.

a
FMS, Functional Movement Screen; NCAA, National Collegiate Athletic Association.

test components, the in-line lunge, rotary stability, and the
hurdle step were all implicated as the least reliable compo-
nent by at least 1 study.27,29,30,34,38,41
There were 5 studies that measured interrater reliabil-
ity among raters with varying levels of FMS experience.
Only 1 of those found unacceptable interrater reliability.
Additionally, Gulgin and Hoogenboom18 and Shultz
et al36 found that overall scores did not differ significantly
between raters of different experience levels. Shultz et al36
found overall unacceptable interrater reliability, but fur-
ther analysis of their data did not show that experience
had any affect. Both interrater reliability of raters with
less than 1 year of experience (ICC, 0.44; 95% CI, 0.12 to
0.67) and that of raters with more than 2 years of experi-
ence (ICC, 0.177; 95% CI, –0.15 to 0.46) were unacceptable.
Five studies were finally included in the meta-analysis
for interrater reliability as the remaining 5 studies did not
have sufficient data for inclusion as previously described.
Pooled quantitative analysis demonstrated that the mean
ICC was found to be 0.81 (95% CI, 0.70-0.92) (Figure 2),
indicating acceptable interrater reliability.
Intrarater Reliability. Six studies evaluated intrarater
reliability of the FMS. All studies examined the reliability
of scores of the same examiners grading the same partici-
pants at 2 points separated by times ranging from 48 hours
to 4 weeks. Five of 6 studies included multiple examiners of
various levels of experience. Four of 6 studies had exam-
iners evaluate video-recorded FMS tests, while 3 had
examiners evaluate participants in real time.
All studies found acceptable intrarater reliability, with
ICC values ranging from 0.6 to 0.96. As with interrater
reliability, the level of experience did not consistently
affect intrarater reliability. Gribble et al17 showed that
intrarater reliability increased with experience. However,
Smith et al37 found no difference in their 4 raters and actu-
ally found that the only certified FMS rater among their
group had the lowest intrarater reliability. Individual com-
ponents of the FMS showed significant variability with
regard to intrarater reliability, again similar to the find-
ings for interrater reliability.
A meta-analysis of intrarater reliability was performed,
and a pooled mean ICC of 0.77 (95% CI, 0.58-0.96) (Figure
AJSM Vol. 45, No. 3, 2017 Functional Movement Screen Meta-analysis 729

Study ICC (95% CI) Smith et al37 ICC (95% CI)

Gulgin et al 18 2014 0.88 (0.80-0.96) Rater 1 0.90 (0.81-0.99)
Parenteau et al30 2014 0.96 (0.94-0.98) Rater 2 0.81 (0.66-0.96)
36
Shultz et al 2013 0.29 (0.03-0.55) Rater 3 0.91 (0.83-0.99)
37
Smith et al 2013 0.89 (0.82-0.96)
38 Rater 4 0.88 (0.78-0.98)
Teyhen et al 2012 0.76 (0.69-0.83)
Summary 0.89 (0.84-0.94)
Summary 0.81 (0.70-0.92)
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2
0.2 0.4 0.6 0.8 1 1.2

Figure 2. Analysis of interrater reliability. ICC, intraclass cor- Figure 4. Analysis of intrarater reliability of raters per Smith
relation coefficient. et al.37 ICC, intraclass correlation coefficient.

Study ICC (95% CI)

Study ICC (95% CI) Gribble et al 2013 0.75 (0.45-1.06)
Gribble et al17 2013 0.75 (0.45-1.06) Parenteau et al30 2014 0.96 (0.94-0.98)
30
Parenteau et al 2014 0.96 (0.94-0.98) Shultz et al36 2013 0.60 (0.47-0.73)
36
Shultz et al 2013 0.60 (0.47-0.73) Smith et al37 2013 0.89 (0.84-0.94)
38
Teyhen et al38 2012 0.74 (0.67-0.81) Teyhen et al 2012 0.74 (0.67-0.81)

Summary 0.77 (0.58−0.96) Summary 0.81 (0.69-0.92)

0.2 0.4 0.6 0.8 1 1.2 0.2 0.4 0.6 0.8 1 1.2

Figure 3. Analysis of intrarater reliability excluding Smith et al.37 Figure 5. Analysis of total intrarater reliability. ICC, intraclass
ICC, intraclass correlation coefficient. correlation coefficient.

3) was obtained, which signified acceptable intrarater reli-

ability. A synthesis of the data presented by Smith et al37 Study OR (95% CI)
(Figure 4) was included in the analysis, and an ICC of 0.81 Butler et al3 2013 8.31 (3.20-21.59)
(95% CI, 0.69-0.92) was observed, which also signified Chorba et al4 2010 3.85 (0.98-15.13)
acceptable intrarater reliability (Figure 5). Dossa et al11 2014 2.33 (0.37-14.61)
16
Injury Predictive Value. Nine studies with a total of Garrison et al 2015 5.61 (2.73-11.52)
21
2696 participants were identified that evaluated the injury Kiesel et al 2007 11.67 (2.48-54.81)
predictive value of the FMS. All studies, per the inclusion Kiesel et al22 2014 2.33 (1.14-4.77)
Knapik et al23 2015 1.42 (1.05-1.93)
criteria, evaluated healthy people before participation in
O'Connor et al28 2011 2.00 (1.29-3.08)
their respective activities. As seen in the studies evaluat- 39
Warren et al 2015 1.01 (0.53-1.92)
ing reliability, a significant variation was seen in the
raters and in the populations being tested. Summary 2.74 (1.70-4.43)
Definitions of injury were based on either time lost or
0 2 4 6 8 10
treatment sought. The Kiesel et al,21,22 Butler et al,3 and
Dossa et al11 studies used time lost from activity or sport,
Figure 6. Analysis of injury predictive value. OR, odds ratio.
while Chorba et al,4 Garrison et al,16 Knapik et al,23
O’Connor et al,28 and Warren et al39 based their definitions
of injury on seeking medical care. for women. Additionally, the risk ratio was optimized at a cut-
All 9 studies included in the quantitative synthesis com- off of 12 for men and 15 for women.
pared participants who scored 14 with those who scored Six of the 9 studies found that participants with an FMS
.14. This cutoff was first established by Kiesel et al21 in score of 14 had a statistically significant higher risk of
a 2007 study in which a receiver operating characteristic injury during subsequent activity than those with scores
(ROC) curve of their data showed that a cutoff of 14 maxi- of .14. Studies consisted of mostly male participants,
mized sensitivity and specificity and was affirmed in studies although Garrison et al16 and Knapik et al23 included
by Butler et al3 and Garrison et al.16 Chorba et al,4 Dossa female participants. The ORs ranged from 1.42 to 11.67.
et al,11 and Kiesel et al22 used a cutoff of 14 based on the pre- Three studies with a total of 225 participants did not find
vious studies. O’Connor et al28 and Warren et al39 developed a statistically significant correlation between FMS scores
ROC curves with their data but found no value that opti- and the risk of injuries. Chorba et al,4 the only study
mized sensitivity and specificity and thus used 14 as a cutoff with all female participants, as well as Dossa et al11 and
per prior studies as well. Knapik et al23 reported data using Warren et al,39 found ORs that ranged from 1.01 to 3.85
14 as a cutoff but found that sex affected optimal values. In but were not statistically significant.
their analysis of 1045 Coast Guard cadets, they found that A pooled quantitative synthesis using all 9 studies was
the FMS score cutoff that maximized sensitivity and specific- performed using a score of 14 points as a cutoff cumulative
ity, determined by the Youden index, was 11 for men and 14 score. An OR of 2.74 (95% CI, 1.70-4.43) was found (Figure
730 Bonazza et al
6), indicating that participants who scored 14 on the FMS
had a 2.74 times greater probability of sustaining an injury
during subsequent activity than those who scored .14 on
the FMS.
Validity
Ten studies were identified in the initial search that evalu-
ated the validity of the FMS. We defined validity based on
the described purpose of the screen to identify deficiencies
in movements that may contribute to a higher risk of inju-
ries.6 The application of the screen as an injury prediction
tool is evaluated separately. The full text of all 10 studies
was then reviewed independently by 2 of the authors
(C.A.O., M.L.S.), who agreed independently on the appropri-
ateness of 5 of the 10 studies for inclusion in the qualitative
analysis. One study did not meet inclusion/exclusion criteria
as it was an abstract only and was excluded. One study
administered the FMS differently than it has been described
in the literature and was excluded as well. Two studies were
excluded as they sought to correlate scores with athletic per-
formance, and the last excluded study looked at changes in
FMS scores with training intervention, all outside our focus
on the validity of the screen itself. A pooled quantitative
synthesis could not be performed on these 5 studies because
of the absence of a standard quantitative value, which is
used to assess the validity of the FMS.
Kazman et al20 used the FMS results of 934 Marine offi-
cer candidates to evaluate the factor structure of the FMS
using the Cronbach alpha value and exploratory factor
analysis (EFA). The Cronbach alpha value was found to
be 0.39, less than 0.5, which signifies unacceptable inter-
nal consistency of the test. EFA additionally showed that
the different test components within the FMS differ in
their contributions to the total score, suggesting that the
FMS is not unidimensional and cannot be used in its cur-
rent form as a unitary construct to predict injuries.
Two articles evaluated the grading used in the FMS.
Frost et al14 hypothesized that instruction on grading cri-
teria could improve the performer’s score on the FMS.
Twenty-one firefighters completed the FMS first and
then were educated on the specific grading criteria used
to assess each move. After the instruction, the participants
were tested again, and the score was found to be signifi-
cantly higher for 4 of the 6 moves. Their conclusion was
that additional performance instruction significantly
changed FMS scores, not only functional asymmetries,
speaking to potential flaws in the validity of the FMS.
Whiteside et al40 compared manual real-time testing to
objective kinematic grading utilizing motion capture in
the evaluation of 11 NCAA Division I athletes. They found
significant differences between the 2 types of grading for
all of the FMS exercises tested (only 6 of the 7 moves
were tested), pointing to the ambiguous criteria for grad-
ing. This study also highlighted the difficulty in assessing
multiple aspects of a movement from one vantage point.
Two studies evaluated the FMS against other measures
of functional movement. Clifton et al5 sought to validate
the FMS by attempting to correlate scores with measure-
ments in static balance during a single-leg stance: center
of pressure velocity, center of pressure area, and center
The American Journal of Sports Medicine
of pressure standard deviation in the medial-lateral and
anterior-posterior directions. These measures had previ-
ously been validated to measure fatigue, which is a risk
factor for musculoskeletal injuries.5 They tested active peo-
ple, defined as participants aged 18 to 50 years who exer-
cised 3 times per week for at least 30 minutes per
workout, both before and after exercise. They hypothesized
that fatigue would lead to a decrease in static balance after
exercise and would correlate with changes in FMS scores.
Although the static balance measurements decreased after
exercise as expected, they found that FMS scores did not
change. They concluded that the FMS was not a useful pre-
dictor of who will experience greater balance deficits after
exercise. This finding does not support the idea that the
FMS can be used in all settings as an injury predictor.
Beach et al1 hypothesized that if the FMS identified defi-
ciencies in functional movements, scores may correlate with
more activity-specific parameters. A total of 30 firefighters
were evaluated: 15 who scored .14 on the FMS and 15
who scored 14. The participants were height and weight
matched. They were asked to perform the standardized
task of lifting a box. Lumbar spine loading magnitudes
and lumbar spine angles were measured and compared
with FMS scores. They did not find a statistically significant
difference between the 2 groups, which calls into question
the ability of the FMS to measure core stability.
DISCUSSION
The popularity and utilization of the FMS have grown rap-
idly since its development, bolstered by evidence in the lit-
erature to suggest its injury predictive value. Its adoption
at the highest level of athletics as well as the military and
other public service organizations has further contributed
to its rise in popularity, despite the existence of conflicting
literature evaluating not only the injury predictive value
but also the validity and reliability of the FMS. Given
this, we sought to assess and assimilate the relevant litera-
ture where appropriate to determine whether the FMS is
a reliable, valid screening test with an injury predictive value.
It is essential that limitations in the screen are understood to
eliminate to the extent possible inaccurate evaluations that
can result in significant consequences for both the screened
participant and his or her respective organization.
The reliability of the examiners has been thoroughly
investigated, especially given the theoretical concern that
varying levels of experience as well as the presence or
absence in certification would result in significant differen-
ces in examinations. The variety in examiners evaluating
participants undergoing the FMS and the methodology for
how participants were examined are clearly evident in our
included studies for both reliability and the injury predic-
tive value, introducing numerous biases that could affect
the results of the respective individual studies and, ulti-
mately, the results of the meta-analysis. Overall, there is
significant evidence that the composite FMS test is reliable
and can be replicated by raters with varying degrees of
experience with the FMS. Gribble et al17 had the only study
evaluating interrater or intrarater reliability that found
that increasing experience led to increased reliability. Smith
AJSM Vol. 45, No. 3, 2017
et al37 was the only other study to report the ICC for inter-
rater reliability of multiple raters who were both certified
and not certified with different experience levels, and they
found no difference. None of the 5 studies evaluating inter-
rater reliability with raters of different experience levels
found any effect on reliability. Only Shultz et al36 found
poor interrater reliability, but as stated above, further anal-
ysis found that dividing more and less experienced raters
did not improve their interrater reliability. The results of
our meta-analysis, which showed high interrater and intra-
rater reliability, suggest that level of experience and formal
certification by Functional Movement Screen Inc have little
effect on scoring of the FMS.
Interrater and intrarater reliability for individual tests
of the FMS did vary greatly among all types of raters,
which may indicate a lack of specificity in the grading cri-
teria or simply a level of difficulty in grading certain sub-
tests that could be a source of error in the screen. This
may be an area for further research. While the ICC is
used as a quantitative measure of reliability, individual
test subsection analysis was often reported using the
kappa coefficient. While it would have been ideal to ana-
lyze each individual subtest within the FMS, a degree of
variance was often not reported with the kappa coefficient.
Because of these restraints in statistical methodology, we
were unable to calculate a P value for our calculations
based on insufficient reporting of raw data and/or variance.
Our study demonstrates that the historical pass/fail cutoff
of 14 points is valid in predicting those at a higher risk of
injury. Evidence supporting 14 points as the optimal cutoff,
however, is limited as only 2 studies replicated the finding
of Kiesel et al21 via independent ROC curves. These findings
were also limited to only male participants, which may be
a limitation. Knapik et al23 found that the optimal cutoff
may differ by sex. O’Connor et al28 and Warren et al39 inde-
pendently found no optimal value in their data. Studies
with mixed-sex populations did find statistically significant
results, but the effect of sex and other population characteris-
tics on a cutoff that optimizes the injury predictive value of
the FMS may be an area of further research.
Garrison et al16 found that a history of injuries alone
could identify those at a higher risk of future injuries
and that combining a history of injuries with an FMS score
of 14 suggested a significantly higher risk of future inju-
ries, increasing their ORs from 5.61 (95% CI, 2.73-11.51) to
15.11 (95% CI, 6.60-34.61). This is consistent with the prior
finding that a history of injuries is associated with lower
FMS scores.31 Further research may identify other factors
that, combined with the FMS, significantly affect its injury
predictive value.
The included studies regarding the validity of the FMS
point to several concerns about its structure and its ability
to detect abnormal movement patterns. Kazman et al20
showed that the composite score of the FMS is not valid as
a unitary construct as it is often used. Grading may be flawed
by somewhat ambiguous criteria, and Frost et al14 showed
that educating those being screened on the criteria can signif-
icantly affect scores, suggesting that scores may be reflecting
more than just the physical characteristics that they intend
to assess (ie, learned behavior). Comparisons with other
Functional Movement Screen Meta-analysis 731
measures of movement and balance also did not find a corre-
lation to FMS scores, again questioning its accuracy or at
least sensitivity in detecting physical abnormalities. Because
of the absence of any gold-standard comparison and signifi-
cant heterogeneity of the existing data, it is difficult to derive
any definitive conclusions from the current literature as to
whether the FMS is a valid tool for the measurement of func-
tional limitations and asymmetries.
Limitations of our meta-analysis of both reliability and
the injury predictive value included insufficient reporting
of raw data, P values, and variance. Other observed poten-
tial biases in analysis include the need to determine vari-
ance using the Fisher method in situations where these
data were not directly reported.13 This has been well
described previously in the literature.13
Although all studies included either time missed from
sport or activity because of an injury or that which
required medical attention, the variability in the definition
adds an element of heterogeneity to our analysis. As these
studies include a variety of participants including athletes,
military personnel, and public servants such as fire-
fighters, much of this variability is inherent to their vary-
ing respective activities and demands.
Based on the results of this systematic review and meta-
analysis, the FMS as a composite score has excellent inter-
rater and intrarater reliability and can be effectively admin-
istered by raters of varying levels of experience with the
FMS both with and without formal certification. Partici-
pants who score 14 on the FMS have greater than twice
the odds of sustaining a musculoskeletal injury than those
with scores of .14. However, the FMS lacks validation of
its structure as a composite score of multiple subtest scores
and of its ability to accurately and sensitively measure def-
icits in posture and balance. Despite this demonstrated
injury predictive value of the FMS, the clinical application
of the FMS should be exercised with caution until further
studies can confirm the screen’s validity.
An online CME course associated with this article is avail-
able for 1 AMA PRA Category 1 CreditTM at http://www
.sportsmed.org/aossmimis/Members/Education/AJSM%20
Current%20Concepts%20Store.aspx. In accordance with
the standards of the Accreditation Council for Continuing
Medical Education (ACCME), it is the policy of The Amer-
ican Orthopaedic Society for Sports Medicine that
authors, editors, and planners disclose to the learners all
financial relationships during the past 12 months with
any commercial interest (A ‘commercial interest’ is any
entity producing, marketing, re-selling, or distributing
health care goods or services consumed by, or used on,
patients). Any and all disclosures are provided in the
online journal CME area which is provided to all partici-
pants before they actually take the CME activity. In accor-
dance with AOSSM policy, authors, editors, and planners’
participation in this educational activity will be predicated
upon timely submission and review of AOSSM disclosure.
Noncompliance will result in an author/editor or planner
to be stricken from participating in this CME activity.
732 Bonazza et al
REFERENCES
1. Beach TA, Frost DM, Callaghan JP. FMS scores and low-back load-
ing during lifting: whole-body movement screening as an ergonomic
tool? Appl Ergon. 2014;45(3):482-489.
2. Beach TA, Frost DM, McGill SM, Callaghan JP. Physical fitness
improvements and occupational low-back loading: an exercise inter-
vention study with firefighters. Ergonomics. 2014;57(5):744-763.
3. Butler RJ, Contreras M, Burton LC, Plisky PJ, Goode A, Kiesel K.
Modifiable risk factors predict injuries in firefighters during training
academies. Work. 2013;46(1):11-17.
4. Chorba RS, Chorba DJ, Bouillon LE, Overmyer CA, Landis JA. Use of
a functional movement screening tool to determine injury risk in
female collegiate athletes. N Am J Sports Phys Ther. 2010;5(2):47-54.
5. Clifton DR, Harrison BC, Hertel J, Hart JM. Relationship between
functional assessments and exercise-related changes during static
balance. J Strength Cond Res. 2013;27(4):966-972.
6. Cook G, Burton L, Hoogenboom B. Pre-participation screening: the
use of fundamental movements as an assessment of function, part
1. N Am J Sports Phys Ther. 2006;1(2):62-72.
7. Cook G, Burton L, Hoogenboom B. Pre-participation screening: the
use of fundamental movements as an assessment of function, part
2. N Am J Sports Phys Ther. 2006;1(3):132-139.
8. Cook G, Burton L, Hoogenboom BJ, Voight M. Functional movement
screening: the use of fundamental movements as an assessment of
function, part 1. Int J Sports Phys Ther. 2014;9(3):396-409.
9. Cook G, Burton L, Hoogenboom BJ, Voight M. Functional movement
screening: the use of fundamental movements as an assessment of
function, part 2. Int J Sports Phys Ther. 2014;9(4):549-563.
10. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin
Trials. 1986;7:177-188.
11. Dossa K, Cashman G, Howitt S, West B, Murray N. Can injury in
major junior hockey players be predicted by a pre-season functional
movement screen: a prospective cohort study. J Can Chiropr Assoc.
2014;58(4):421-427.
12. Elias JE. The inter-rater reliability of the functional movement screen
within an athletic population using untrained raters [published online
July 8, 2013]. J Strength Cond Res. doi:10.1519/JSC.0b013e3182a1ff1d.
13. Fisher RA. Statistical Methods for Research Workers. 14th ed. Edin-
burgh: Oliver and Boyd; 1970.
14. Frost DM, Beach TA, Callaghan JP, McGill SM. FMS scores change
with performers’ knowledge of the grading criteria: are general
whole-body movement screens capturing ‘‘dysfunction’’ [published
online November 20, 2013]? J Strength Cond Res. doi:10.1519/
Jsc.0b013e3182a95343.
15. Frost DM, Beach TA, Callaghan JP, McGill SM. Using the functional
movement screen to evaluate the effectiveness of training. J Strength
Cond Res. 2012;26(6):1620-1630.
16. Garrison M, Westrick R, Johnson MR, Benenson J. Association
between the functional movement screen and injury development
in college athletes. Int J Sports Phys Ther. 2015;10(1):21-28.
17. Gribble PA, Brigle J, Pietrosimone BG, Pfile KR, Webster KA. Intra-
rater reliability of the functional movement screen. J Strength Cond
Res. 2013;27(4):978-981.
18. Gulgin H, Hoogenboom B. The functional movement screening
(FMS): an inter-rater reliability study between raters of varied experi-
ence. Int J Sports Phys Ther. 2014;9(1):14-20.
19. Ip EH, Wasserman R, Barkin S. Comparison of intraclass correlation
coefficient estimates and standard errors between using cross-
sectional and repeated measurement data: the safety check cluster
randomized trial. Contemp Clin Trials. 2011;32(2):225-232.
20. Kazman JB, Galecki JM, Lisman P, Deuster PA, O’Connor FG. Factor
structure of the functional movement screen in Marine officer candi-
dates. J Strength Cond Res. 2014;28(3):672-678.
For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.
The American Journal of Sports Medicine
21. Kiesel K, Plisky PJ, Voight ML. Can serious injury in professional foot-
ball be predicted by a preseason functional movement screen? N Am
J Sports Phys Ther. 2007;2(3):147-158.
22. Kiesel KB, Butler RJ, Plisky PJ. Prediction of injury by limited and
asymmetrical fundamental movement patterns in American football
players. J Sport Rehabil. 2014;23(2):88-94.
23. Knapik JJ, Cosio-Lima LM, Reynolds KL, Shumway RS. Efficacy of
functional movement screening for predicting injuries in Coast Guard
cadets. J Strength Cond Res. 2015;29(5):1157-1162.
24. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that eval-
uate healthcare interventions: explanation and elaboration. BMJ.
2009;339:B2700.
25. Lumley T. rmeta: meta-analysis. R package version 2.16. 2012. Avail-
able at: http://CRAN.R-project.org/package=rmeta. Accessed July
16, 2015.
26. Mantel N, Brown C, Byar DP. Tests for homogeneity of effect in an
epidemiologic investigation. Am J Epidemiol. 1977;106(2):125-129.
27. Minick KI, Kiesel KB, Burton L, Taylor A, Plisky P, Butler RJ. Interrater
reliability of the functional movement screen. J Strength Cond Res.
2010;24(2):479-486.
28. O’Connor FG, Deuster PA, Davis J, Pappas CG, Knapik JJ. Func-
tional movement screening: predicting injuries in officer candidates.
Med Sci Sports Exerc. 2011;43(12):2224-2230.
29. Onate JA, Dewey T, Kollock RO, et al. Real-time intersession and
interrater reliability of the functional movement screen. J Strength
Cond Res. 2012;26(2):408-415.
30. Parenteau GE, Gaudreault N, Chambers S, et al. Functional move-
ment screen test: a reliable screening test for young elite ice hockey
players. Phys Ther Sport. 2014;15(3):169-175.
31. Peate WF, Bates G, Lunda K, Francis S, Bellamy K. Core strength:
a new model for injury prediction and prevention. J Occup Med Tox-
icol. 2007;2:3.
32. R Foundation for Statistical Computing. R: A Language and Environ-
ment for Statistical Computing [computer program]. Vienna: R Foun-
dation for Statistical Computing; 2015.
33. Rowan CP, Kuropkat C, Gumieniak RJ, Gledhill N, Jamnik VK. Inte-
gration of the functional movement screen into the National Hockey
League Combine. J Strength Cond Res. 2015;29(5):1163-1171.
34. Schneiders AG, Davidsson A, Horman E, Sullivan SJ. Functional
movement screen normative values in a young, active population.
Int J Sports Phys Ther. 2011;6(2):75-82.
35. Shah VM, Andrews JR, Fleisig GS, Mcmichael CS, Lemak LJ. Return
to play after anterior cruciate ligament reconstruction in National
Football League athletes. Am J Sports Med. 2010;38(11):2233-2239.
36. Shultz R, Anderson SS, Matheson GO, Marcello B, Besier T. Test-
retest and interrater reliability of the functional movement screen.
J Athl Train. 2013;48(3):331-336.
37. Smith CA, Chimera NJ, Wright NJ, Warren M. Interrater and intrarater
reliability of the functional movement screen. J Strength Cond Res.
2013;27(4):982-987.
38. Teyhen DS, Shaffer SW, Lorenson CL, et al. The functional move-
ment screen: a reliability study. J Orthop Sports Phys Ther.
2012;42(6):530-540.
39. Warren M, Smith CA, Chimera NJ. Association of the functional
movement screen with injuries in Division I athletes. J Sport Rehabil.
2015;24(2):163-170.
40. Whiteside D, Deneweth JM, Pohorence MA, et al. Grading the func-
tional movement screen: a comparison of manual (real-time) and
objective methods. J Strength Cond Res. 2015;29(1):254-261.
41. Wright MD, Portas MD, Evans VJ, Weston M. The effectiveness of 4
weeks of fundamental movement training on functional movement
screen and physiological performance in physically active children.
J Strength Cond Res. 2015;29(1):254-261.