Football for individuals with cerebral palsy (CP football) is played by para-athletes with a minimal impairment criterion of ataxia, hypertonia, or athetosis (i.e. three impairment types that are most commonly associated with individuals having neurological impairment). Being eligible to compete in this para-sport means presenting motor control impairment of a cerebral nature, causing a permanent and verifiable activity limitation [1
]. This brain injury may be due to cerebral palsy (CP), stroke, or traumatic brain injury, compromising the function of the lower limbs and constraining the performance of specific skills, such us jumping [2
], sprinting [3
], changing directions [4
], or dribbling skills [3
]. Therefore, athletes competing in CP football have an impairment that leads to a competitive disadvantage, and, consequently, requires a system that minimizes the impact of impairments on sport performance, ensuring that the success of an athlete is determined by his/her skills, fitness, power, endurance, tactical ability, and/or mental focus [5
]. This system is called classification.
Classification is a defining feature of para-sport. It is defined as grouping athletes into sport classes according to how much their impairment affects fundamental activities in each specific sport or discipline, and that those classification systems should be evidence-based and sports-specific [6
]. In CP football, during the last three decades (i.e. since its introduction at the 1984 Paralympic Games), a functional system was applied in this para-sport, developed by the Cerebral Palsy International Sports and Recreation Association [7
]. The system comprises eight functional classes: the first four groups (classes 1–4) correspond to athletes who need a wheelchair to perform any sport, while the last four groups (classes 5–8) host ambulant athletes, that is, those eligible for CP football. Specifically, those with moderate spastic diplegia where the function of lower limbs is limited by bilateral spasticity are grouped in the sport class FT5; those with a moderate ataxic or athetoid profile involving the four limbs and trunk in the sport class FT6; those with moderate hemiplegia where one side of the body (right/left arm and leg) is affected by spasticity belong to the sport class FT7; and those with mild involvement of diplegia, ataxia/athetosis or hemiplegia, also called minimal impairment criteria to be eligible in this para-sport, are allocated in the sport class FT8 [8
Athletes who practice football should have basic movement patterns such as running, jumping, sprinting, crouching, changing direction and rhythm [9
], requiring players perform several actions that utilize strength, speed, power, agility, stability, flexibility, and endurance [10
]. This intrinsic complexity and the multiple skills required in a team sport like this is a challenge when the sport is played by people with brain injury. More specifically, players with hypertonia have an abnormal increase in muscle tension and a reduced ability of the muscle to stretch, those with ataxia exhibit impaired coordination of muscle movements and those with athetosis usually exhibit unbalanced and involuntary movements due to constant changes in muscle tone and difficulty maintaining a symmetrical posture. These impairments can hamper player performance on the field, reinforcing the desirable link between impairment and activity limitation for classification purposes. In addition, football is a sport modality characterized by intermittent motor actions of short duration and high intensity, as well as moments of longer duration and lower intensity [11
] characterized by the requirement of short races, rapid acceleration, and decelerations, turns, jumps, and changes of direction [13
]. During a football match, players make changes of direction every 2 to 4 s [14
] or perform between 1200 and 1400 times during a match [15
]. Change of direction (COD) ability can be described as the ability to change direction while sprinting over a pre-planned course [16
], so linear and change-of-direction speeds are essential qualities for athletes who play field sports, such as football [17
]. In our para-sport, a recent study by Yanci et al. [19
] demonstrated that football players with CP covered less distance at high-intensity running and sprinting, performing a smaller number of moderate- and high-intensity accelerations and decelerations, had a lower player load and performed fewer CODs in official matches as compared with conventional football players as reported in other studies. In addition, it has been demonstrated that those players with lower impairment (FT8) covered more distance at high-intensity running (13.0–18.0 km·h−1) and sprinting (>18.0 km·h−1), and they performed more accelerations, decelerations, and CODs at a high intensity in matches than did other players (i.e. FT5/6 and FT7 groups) [20
Although noticeable steps have been made to describe the relationships between impairment and activity limitation in this para-sport [21
], relevant questions remain to be solved for the development of a sports-specific, reliable, and evidence-based classification system in CP football. First, recent and relevant research on this topic has considered para-athletes with the brain impairments of ataxia, athetosis, or hypertonia as a unique group, so the relationships between impairment and limitation of activity would be biased. Second, the identification of the relationships between impairment and sport performance presents another particular challenge in team para-sports [22
], because there are several performance factors occurring at the same time [21
]. Furthermore, the inclusion of the ball when the relationships between impairment and limitation of activity are studied may also influence testing reproducibility [3
Because football players with CP lack a repertoire of motor skills to perform the varied skills that football requires, and CODs have a prominent role during a football match, there is a need to develop a valid and reliable test to evaluate this. This study aims to explore the validity and reliability of a test battery [23
] to assess CODs, together with ball dribbling in para-footballers with CP. More specifically, the study aims to validate the content validity through evaluation by a panel of experts in the field of CP football to analyze the reliability of the test battery in relation to its internal consistency (i.e., construct validity), intraclass correlation index, temporal stability, and association between the tests. In addition, the performance outcomes by players with different CP profiles (i.e., spastic diplegia, athetosis/ataxia and spastic hemiplegia) and levels of impairment (i.e., ambulant individuals with moderate vs mild CP) are also discussed.
The aim of this study was to evaluate the content and construct validity, and the reliability of four agility tests requiring ball dribbling in football players with CP. The previous research in this population demanding sport-specific skills is scarce [3
], so this study provides new evidence about the feasibility of its implementation for classification or training purposes. Apart from this, this is the first time that the test battery [23
] has been applied to para-footballers with CP after an expert consultation verifying that the test battery was feasible and pertinent for this population. Therefore, a new group of tests to assess CODs in para-footballers with CP is now available with good reliability (α = 0.86–0.97), increasing the recent evidence for the Illinois Agility Test (IAT) or the Stop and Go (S & G) test [3
], unique evidence on this topic to the best of the authors knowledge. The study by Reina et al. [3
] investigated the effect of the ball dribbling when performing the abovementioned agility test and a 40 m straight sprint, with and without ball dribbling. This cross-sectional study involved 82 international para-footballers with CP, but the measurements were done in a single session. As a result, the between-session reliability was not explored (as our study did). The IAT with ball dribbling had an ICC value of 0.84 and an SEM of 6.3%, while the S&G test had an ICC of 0.48 and an SEM of 9.8%. The three tests of this study requiring CODs (#2, #3 and #4) had better ICC and SEM values, except the SEM of test #2 (SEM = 10.8%).
Although test #1 does not really demand a COD due to it consisting of a straight-line sprint with ball dribbling, this study reports its reliability when ball dribbling is required in a shorter run as compared with the study by Reina et al. [3
], where para-athletes had to run 40 m with partial measurements at 10 m and 25 m. Our study exhibits better ICC values (0.95 vs 0.84 at 10 m, 0.76 at 25 m and 0.73 at 40 m) and similar SEM values (6.3% vs 4.5% at 10 m, 6.2% at 25 m, and 6.5% at 40 m). The 20 m sprint has been recently used to explore the relationships between performance and match load using GPS technology [21
], but without involving ball skills during the test. Therefore, this study provides new evidence about the feasibility and reliability for implementing a sport-specific test that better replicates the demands of the real game [19
Test reproducibility deserves a special mention, as significant differences were found between test and retest sessions. This is the first study reporting between-sessions reliability in para-athletes with CP, suggesting the potential bias of a learning effect. The IPC Position Stand on Classification in Paralympic Sport [5
] identifies and describes several criteria for a valid measure of impairment: measures should be objective, reliable, precise, specific to the impairment of interest, parsimonious (i.e. account for the greatest possible variance in sports performance) and, as much as possible, be resistant to the effects of training (i.e. when athletes undertake rigorous, sport-specific training, pre-training measures of impairment should not be significantly different from post-training measures of impairment) [33
]. Although our tests assess performance when sprint/CODs with ball dribbling are required and not the impairment, the improvement in performance after 72 h from the first session is a remarkable finding for classification purposes. Current methods of classification include both novel motor tasks (i.e. tasks that are unlikely to have been practiced by the athlete in the usual course of training for his or her sport) and sport-specific activities (i.e., tasks that are likely to have been frequently practiced by athletes training for the sport), so a familiarization with the classification protocols would be recommended to avoid bias in the decision-making of classifiers, that is, the para-athlete exhibits more limitation of activity than really he/she has due to unfamiliarity with the tests.
The construct validity of the test battery has also been demonstrated, with very large to almost perfect correlations found between tests and testing sessions (r
= 0.72–0.97), so the tests measure similar components of agility. These correlations are higher than those obtained in other studies with this population, correlating the IAT with the S&G test (r
= 0.35, p
< 0.01) and the straight sprint at 10 m (r
= 0.35, p
< 0.01), 25 m (r
= 0.50, p
< 0.01), and 40 m (r
= 0.34, p
< 0.01) [3
]. The correlations in this study are also higher than those obtained in a study with 99 international para-footballers with CP [4
], where the relationship between the IAT (without ball dribbling) and the Modified Agility Test was explored. Whereas the study by Reina et al. [3
] demonstrates a higher variability in sports performance when agility tests performed while ball dribbling is required, our study increases the options for testing agility in para-footballers with CP with shorter courses and time requirements for its completion, especially in tests #2 and #4.
Regarding the between-group comparisons, our results should be interpreted with caution due to the number of players per sport class. It should be mentioned that the sport class distributions obey the para-sport rules, where two players of the sport classes FT5/FT6 must be on the field of play and no more than one FT8 player would be included in the line-up. Nevertheless, the large effect sizes when comparing FT8 para-footballers with the other three sport classes demonstrated that those with a lower level of impairment exhibit better sports performance [2
]. In addition, it was found that FT6 players performed the worst in all the tests, reinforcing the idea that those with dyskinesia or ataxia would be those para-athletes in which brain injury may affect, to a higher extent, sports performance (i.e., limitation of activity) when multiple skills are required. Para-athletes who have ataxia (involuntary movements and impaired coordination), athetosis (involuntary contractions of the muscles) or dystonia (repetitive torsions and movements or abnormal postures) are individuals that have general problems with balance and performing departures, stops and turns during a race [35
]. Accordingly, a recent study conducting a biomechanical comparison of the initial sprint acceleration performance and technique in an elite athlete with CP of this profile and able-bodied sprinters demonstrated the impact of this type of CP on performance in a competition-specific acceleration movement [36
]. Therefore, when the ball is included during testing, driving it makes it more difficult to accomplish the task, constraining their motor control, and they take longer than the other sport classes.
FT5 players have been considered the other “low sport class” sub-group (i.e., more impaired for the real game) [21
], and they were the second sub-group with lower performance (i.e., after the FT6 sport class), except for tests #1 and #2, with scores similar to FT7 para-footballers. Athletes of class FT5 are individuals with spastic diplegia, that is, they present with spasticity in their legs, hips, and pelvis. This high abnormal muscle tone causes impairment in the voluntary and passive movements of the legs, rotational ankle abnormalities followed by poor pelvic alignment [37
]. These compromises cause difficulty and reduced steps for athletes to move around, consequently making them take longer to make a COD. On the other hand, those in class FT7 are para-athletes with spastic hemiplegia, so it is common to observe limitations in gait, increased limb asymmetry or reduced stride length [38
]. Therefore, knee and hip control are also affected by spasticity and possible loss of range of motion due to muscle contraction. However, the similar results in tests #1 and #2 (i.e. the two tests with no or shorter CODs) would be explained by the compensatory strategies that para-footballers may use. That is, those with spastic diplegia have shorter strides but may use both feet for better ball control, while those with spastic hemiplegia usually use the dominant or unimpaired foot for ball control. The differences between these two sub-groups slightly increased in tests #3 and #4, where CODs are more remarkable, so those with spastic hemiplegia can use the unimpaired leg for better stabilization when pivoting to perform the required COD.
Some limitations should be mentioned. First, the number of participants per sport class constrains the between-group comparisons. However, the study has been applied with participants from three different countries, which have national teams participating at an international level. Second, not all the participants had international experience, or they could not be considered well-trained para-athletes, being a potential factor that explains the differences in performance between testing sessions. However, classification is a process that is applied at local, regional, and international stages, so the level of training or the familiarization with the tests must be considered in further research.