The prevention of injuries among youth basketballers according to the “Sequence of Prevention’’: a systematic review

s against the inclusion criteria, 57 potentially relevant studies were included for the full text review.[20-76] After screening the full text, 30 articles were excluded.[20-49] The reasons for exclusion were: the articles did not meet the requirements for a prospective cohort or an RCT (n = 15) [2022,28,30,31,34-36,39-43,48], the articles were not specifically about youth basketball players (n = 13) [21,23-26,29,32,37,38,44-47] ,or there were no outcome measures (n = 2).[27,33] Figure 1 presents the search procedure.

injuries among boys. [13] The extent of the injury problem in youth basketballers calls for preventive action based on the results of epidemiological research. The number of injuries among youth basketball players should be reduced to prevent long-term complaints or complaints into adulthood and to reduce healthcare costs. The 'van Mechelen sequence of prevention model for sports injuries' describes four consecutive steps that lead to efficacious preventive interventions. [14,15] Steps 1 and 2 consist of exploring the incidence and aetiology of musculoskeletal injuries. Steps 3 and 4 consist of developing and evaluating preventive interventions. [14] A systematic review on van Mechelen's quadrants in adult basketball players has already been published, but not yet in youth basketball players. [16] Therefore, we aimed to gather epidemiological information to answer the following questions: (1) What is the incidence of musculoskeletal injuries among youth basketball players?; (2) What are the risk factors of these musculoskeletal injuries among youth basketball players?; (3) What are the interventions available for the prevention of musculoskeletal injuries among youth basketball players?; and (4) How effective are these interventions on the reduction of musculoskeletal injuries among youth basketball players?

Design
A systematic review was conducted on sports injury prevention among youth basketball players. This systematic review has been written in accordance with the PRISMA guidelines. [17]

Data sources and searches
Search strategies (Appendix 1) were composed by using three groups of keywords, namely: 'injury', 'youth basketball', and 'cohort study'. The search strategies were entered into two databases, namely Medline and SPORTDiscus. Medline was searched from October 2, 2018 up to February 7, 2019. SPORTDiscus was searched from October 2, 2018 up to January 8, 2019. Different filters were used: Humans, English, Randomised Controlled Trial, Systematic Review and/or Academic Journal. All search terms were combined with 'AND' and 'OR'.

Eligibility criteria
The inclusion criteria were:  The population consists of youth (boys and/or girls) basketball players (age [6][7][8][9][10][11][12][13][14][15][16][17][18].  The article is written in English.  If the article is about descriptive epidemiology (Step 1 of van Mechelen's model), prospective cohort design is used.  If the article is about descriptive epidemiology, incidence rates or prevalence rates are reported.  If the article is about aetiology (Step 2 of van Mechelen's model), prospective cohort design is used.  If the article is about aetiology, a risk estimate is reported.  If the article is about prevention (Steps 3 and 4 of van Mechelen's model), randomised controlled trial is conducted.  If the article is about prevention, incidence rates and/or effect sizes are reported.

Study selection
Titles and abstracts of the retrieved citations were independently screened by two researchers (DA and VG). When the title and abstract met the inclusion criteria, the article was included for the full text selection. When the title and abstract did not contain sufficient information, it was not included for the full text selection. Then the full text articles were independently assessed by two researchers (DA and JA).
Where doubts arose concerning inclusion or exclusion of an article, a third researcher was consulted (MB).

Data extraction
The data from the included articles were extracted by one researcher (DA) in a standardised table and checked by another researcher (MB). The data extraction focused on: article information (author, year), study population (numbers, age and gender), injury definition and injury incidence. If there was information about the risk factors, preventive measures and the effect of these preventive measures, it was also included in the data collection.

Risk of bias appraisal
To assess the methodological quality of the included articles, two different checklists were used. For the articles related to descriptive epidemiology and aetiology, the Quality in Prognosis Studies (QUIPS) tool was used (Appendix 2). The Cochrane Collaboration's tool was used for the articles related to prevention (Appendix 2). For both the QUIPS and the Cochrane Collaboration's tool, six potential bias domains were assessed with a high, moderate or low risk of bias. For assessments using the QUIPS tool, a study was considered to have a low risk of bias rated as low or moderate in all six domains, with at least four domains being rated as low. [17] If two or more domains were scored as high, the study was rated as having a high risk of bias. [17] Studies that were in between were scored as having a moderate risk of bias. [18] For assessments using the Cochrane Collaboration tool, a study was assessed with a low risk of bias when all items were assessed as low. [18] When at least one item was assessed as moderate, the article received a score with a moderate risk of bias. A high risk of bias was rated when at least one item was assessed as high. [18] The checklists were assessed and crosschecked by two researchers (DA and JA). If a difference of opinion arose concerning the scoring of an item, a consensus was reached.

Data synthesis and analysis
The data were processed according to the four steps of van Mechelen's 'Sequence of Prevention' so that the collected information was presented clearly. The following outcome measures were used for the incidence and the risk factors: exposure, hours of exposure, hours of game exposure, athlete exposure, and percentages. Only the results that were expressed in hours of exposure were included in the results of the overall injuries. For the effectiveness of the preventive measures, it was considered whether a reduction was found on these outcome measures. Risk reduction rates were used.

The incidence of musculoskeletal injuries in youth basketball
Of the 27 included articles for data extraction, 19 articles contained information on the incidence of musculoskeletal injuries in youth basketball. [49][50][51][52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67] Of these 19 studies, 13 studies were about girls and boys, five studies were about girls only and one study was about boys only. The age of all participants in these 19 studies together ranged from 8 to 20 years. Twelve studies were about American basketball players and seven studies were conducted in other countries (five in Europe, one in Japan and one in Nigeria).

Specific injuries
McGuine et al. investigated the incidence of ankle sprains in youth basketballers; the rate of ankle sprain was 1.56 per 1 000 exposure (1.68 for boys and 1.44 for girls per 1 000 exposure). [63] Two studies investigated the incidence of patellofemoral pain (PFP) in female youth basketball players. Herbst et al. found an incidence rate for development of PFP of 0.97 per 1 000 athlete exposures (AE), the study of Myer et al. found an incidence rate of 1.09 per 1 000 AE. [65,66] Symptoms of anterior knee pain were likely to persist to after middle school-aged onset and to reach peak prevalence during the high school years. [68] The shoulder injury rate ranged from 0.045 to 0.061 per 1 000 AE. [60,61]

Risk factors of musculoskeletal injuries in youth basketball
Eleven articles presented information on the risk factors of musculoskeletal injuries in youth basketball. [50,51,58,59,63,65,66,[68][69][70][71] Of these 11 studies, five were about girls and boys, four were about girls only and two studies were about boys only. The age of the participants in these 11 studies together ranged from 9 to 20 years. In seven of the 11 studies, research was done on American youth basketballers. In the other four studies, research was done in different countries (three in Europe and one in Taiwan).

Game vs practice
The injury risk for basketball injuries in youth was higher in games than in practices in all included studies. Most game injuries resulted from body contact, 46% in the study of Kuzuhara et al. and 1.32/1 000 AE for boys and 1.55/1 000 AE for girls in the studies of Clifton et al. [51,58,59] In the study of Pasanen et al. body contact with another player was the most frequent injury situation (25%), followed by stepping or landing on another player's foot (23%) or landing from a jump (16%). Proportions of contact injuries, indirect contact injuries, and non-contact injuries were 49%, 17%, and 34%, respectively. [50] Information on the aetiology of injuries is presented in Table 2, Appendix 3 and Appendix 5.

Risk factors of lower extremity injuries
An anterior right/left reach distance difference measured with the Star Excursion Balance Test (SEBT) is a risk factor for lower extremity injuries in youth basketball players. [69] Logistic regression models indicated that players with an anterior right/left reach distance difference greater than four cm were two and a half times more likely to sustain a lower extremity injury (p<0.05). [69] Girls with a composite reach distance less than 94% of their limb length were six and a half times more likely to have a lower extremity injury (p<0.05). [69] Risk factors of ankle injuries In boys' high school basketball players, high variations of postural sway in one leg standing is a risk factor for developing an ankle injury. [70] [70] Subjects who demonstrated poor balance (high sway scores) had nearly seven times as many ankle sprains as subjects who had good balance (low sway scores) (p<0.001). [63] Risk factors of knee injuries Players with an ankle dorsiflexion range less than 36.5 degrees had a risk of 19% to 29% of developing patellar tendinopathy (PT) within a year, compared with 1.8% to 2.1% for players with an ankle dorsiflexion range greater than 36.5 degrees. [71] The ankle dorsiflexion was measured with the established weight-bearing lunge test. Young female basketball athletes with greater hip abduction strength have an increased risk for the development of PFP-need to write this out in full first.

Preventive interventions and related effectiveness
Four studies regarding injury-preventive interventions were included. [72][73][74][75] [72] The CORE group showed a reduction in injuries for basketball players (p = 0.02). [72] The absolute risk reduction rate per 1 000 AEs was: 2.73 (95% CI  [72] LaBella et al. evaluated the effectiveness of coach-led neuromuscular warm-up on reducing lower extremity injuries in young female soccer and basketball athletes. [73] The warmup was similar to previously studied NMT programmes, combining progressive strengthening, plyometric, balance, and agility exercises. Athletes were instructed to avoid dynamic knee valgus and to land from jumps with flexed hips and knees. Coaches for the control group used their usual warm-up. Compared to controls, athletes in the intervention group had lower incidence rates per 1 000 AEs of gradualonset lower extremity injuries (0.43 vs 1.22, p<0.01), acuteonset non-contact lower extremity injuries (0.71 vs 1.61, p<0.01), non-contact ankle sprains (0.25 vs 0.74, p=0.01) and lower extremity injuries treated surgically (0.00 vs 0.17, p=0.04). [73] Emery et al. studied the effectiveness of a sports-specific balance training programme in reducing injury in adolescent basketball. [74] The training group and the control group were taught a standardised warm-up programme. The training group was also taught an additional warm-up component and a home-based balance training programme using a wobble board. The injury rate in the control group was 33.1 injuries per 100 participants per season (95% CI; 28.64-37.79); in the training group it was 26.3 injuries per 100 participants per season (95% CI; 22.48-30.43). [74] The basketball-specific balance training programme was protective with regard to acute onset injuries in high school basketball (RR = 0.71 95% CI; 0.5-0.99), [71] but not significant, like all the results from this study. Selfreported compliance to the intended home-based training programme was poor (60%). [74] McGuine et al. investigated the effect of lace-up ankle braces on the incidence and severity of acute first-time and recurrent ankle injuries sustained by high school basketball players. [75] Athletes were instructed to wear McDavid Ultralight 195 braces over a single pair of socks on both ankles for each teamorganised conditioning session, practice, or competition throughout the season. The rate of acute ankle injuries was 0.47/1 000 exposures in the braced group and 1.41/1 000 exposures in the control group (Cox hazard ratio [HR] 0.32; 95% CI 0.20-0.52; p=<0.001). [75] For players with a previous ankle injury, the incidence of acute ankle injury was 0.82/1 000 exposures in the braced group and 1.79/1 000 exposures in the control group (Cox HR 0.30; 95% CI 0.17-0.90; p=0.028). [75] Information about the preventive interventions and effectiveness is presented in Table 3, Appendix 4 and Appendix 5.

Discussion
The results showed that the overall injury rate for youth basketball players ranged from 2.64 to 3.83 per 1 000 hours of exposure. [50,53,56] Ankle injuries (22%-37%) and knee injuries (5%-41%) were the most common injuries. [50,51,54,56,57] Several risk factors for developing these injuries were mentioned in the Results section, including an anterior right/left reach distance difference, high variations of postural sway and an ankle dorsiflexion range of less than 36.5 degrees.
The anterior right/left reach distance and the variations of postural sway were both measured with a static balance on one leg. This may indicate that poor static balance on one leg, in particular, is a predictor of developing lower extremity injuries in youth basketball players.
For the preventive measures, results showed that a CORE intervention focused on the trunk and lower extremity led to a reduction in injuries (p=0.02). [72] The basketball-specific balance training programme was protective with regard to acute-onset injuries in high school basketball. [74] Another effective preventive measure was wearing a McDavid Ultralight 195 brace. This brace reduced acute ankle injuries.

Similarities and differences with other studies
In 16 of the 27 included articles, research was done on American youth basketball players. The rules of the basketball federation in America (NBA) and the rules of the basketball federation in Europe (FIBA) are different. The study of Madarame suggested that women's basketball games are played in a different manner in each region of the world. [76] This could mean that the incidence and risk factors of injuries are different in each region of the world for women/girls. For European basketball girls, and probably also boys, more research must be done into incidence, aetiology and preventive measures.
Compared with the overall injury rates in adult basketball players (ranged from 0.05 to 12.92 per 1 000 hours of exposure), [16] the overall injury rate in youth basketball players (2.64 to 3.83 per 1 000 hours of exposure) is less and closer together. However, in the review of Kilic et al. there is only one article that describes the incidence in 1 000 hours of exposure. In accordance with the review of Kilic et al. [16] , the ankle and knee are the most common injuries in youth basketball. For ankle injuries, a high postural sway was a risk factor in both reviews. [16,70] Because ankle and knee injuries are most common among the youth and adults, it seems best to reduce these injuries in the youth. It would therefore be best to start with preventive measures for the youth and to adjust the corresponding youth exercises for adults later.
The injury rate in games was higher than in practice in all included studies. This is similar to other sports. [9] In youth basketball, body contact is the main reason for injuries in games. [50] At training sessions, there are forms of practice with no body contact. The competitive pressure will probably also be a reason for more injuries in games and this should be investigated in future studies.

Methodological aspects
It was hard to compare the findings between studies because the denominator varied from 1 000 person-days to 1 000 exposures and athletes per season. For the results of the overall injury rates, only results expressed in hours of exposure were included. The influence that this can have on the overall injury rates seems to be small because the studies with other outcome measures show roughly the same results. In future, it is advisable to use one outcome measure for all studies reporting epidemiological data. Increasingly, incidence rates in all sports are being expressed as rates per 1 000 hours. This is a good fdfdsfsd Self-reported compliance to the intended homebased training programme was poor (298/494 or 60.3%).
The programme was effective in reducing acute onset injuries in high school basketball. There was also a clinically relevant trend found with respect to the reduction of all lower extremities and ankle sprain injuries. For players with a previous ankle injury, the incidence of acute ankle injury was 0.82/1 000 exposures in the braced group and 1.79/1 000 exposures in the control group ([HR] 0.30; 95% CI 0.17,0.90; P = 0.028).

McGuine
For players who did not report a previous ankle injury, the incidence of acute ankle injury was 0.40 in the braced group and 1.35 in the control group ([HR] 0.30; 95% CI 0.17, 0.52, P <0.001).
The use of a lace-up ankle brace reduced the incidence but not severity of acute ankle injuries in male and female high school basketball athletes by 68% regardless of sex, age, level of competition, or BMI compared with wearing no brace. N, number of participants; G, gender; A, age; C, country where study was conducted; D, design; F, follow-up period; RCT, randomised controlled trial; AE, athlete exposures; HR, cox hazard ratio; CI, confidence interval; RR, relative risk approach and allows some comparison across sports. [77] It is better than per exposure because not every individual takes part in a training session or game for an equal period of time.
Some limitations need to be addressed. The search for articles was done in two databases and only articles written in English were included. In this review, only studies with a prospective design were included to formulate valid answers to the research questions while maintaining the highest scientific quality. The databases and selection criteria used might have led to the exclusion of relevant studies with a different design or in another language. We believe that these limitations did not significantly affect the findings, because the two databases used were the most obvious ones and prospective studies were used to formulate valid answers.
There is a lack of scientific literature on the aetiology of basketball-specific shoulder and lower back injuries among youth basketball players. The search for articles only revealed articles about the aetiology and prevention of ankle and knee injuries. Studies on the aetiology and prevention of other specific regions of the body are lacking. Little is known about the incidence of injuries to other regions of the body and this information is presented in Table 2. Because there are no studies into the aetiology of those injuries, prevention studies cannot be drawn up for these few common injuries.
Several risk factors for injuries in youth basketball players were found in this overview. Conclusions on the risk factors for youth basketball injuries are derived from only one study. Therefore, it is important to be cautious when interpreting these risk factors. More research into risk factors for youth basketball injuries is recommended.
In some studies, the ages of the children used were not specifically described and in other studies an average age was used. In two studies, the maximum age of the included participants was 20 years. [49,71] Because participants older than 18 years were a small group of the participants of all included articles, we believe that our results are still representative of youth basketball players.
All included studies scored a low or moderate risk of bias. This means that we must be careful with some of the conclusions. This systematic review scores a level of evidence 2. A level of evidence 2 applies to the results of the incidence and risk factors. A level of evidence 1 applies to the conclusions from the studies on preventive measures, except for the study by LaBella et al., which scores a level of evidence 2.

Implications for practice
Based on the results from the included studies on risk factors, it seems advisable to do a screening at the start of the season for ankle mobility, the strength of the hip abductors and the reach distance of the lower extremities. If there are risk factors, they will have to be addressed in a preventive program to prevent injuries. [78] Also, it is advisable to perform a neuromuscular warm-up in combination with performing weekly strength and stability exercises for the trunk and lower extremity and to wear a lace-up ankle brace around both ankles in each training session and each game.

Conclusion
The conclusions of this systematic review are predominantly based on American youth basketball players, showing an overall injury rate ranging from 2.64 to 3.83 per 1 000 hours of exposure. Ankle and knee injuries were the most common injures among youth basketball players. The main risk factors for injuries in youth basketball were: playing games, anterior right/left reach distance difference greater than four cm, an ankle dorsiflexion range less than 36.5 degrees and high variations of postural sway in one leg. Physical therapists and coaches can use the SEBT to identify youth basketball players who are at increased risk for a lower extremity injury. The hip strength and the ankle dorsiflexion can also be tested in knee complaints or to prevent PT and PFP. A neuromuscular warmup in combination with performing weekly strength and stability exercises for the trunk and lower extremity currently seems to be the best training method for preventing injuries in youth basketball players. Acute ankle injuries can be reduced by wearing a McDavid Ultralight 195 brace.
For the preparation of specific prevention programmes in youth basketball, further research on the incidence and especially on the aetiology is needed.

Conflict of interest and source of funding:
The authors declare no conflict of interest and no source of funding.

Author contributions:
DA wrote this manuscript as a graduation project for his course (Master of Pediatric Physical Therapy), collected the data and performed the analysis. VG was the project leader for this manuscript, conceived and designed the study, collection of data, interpretation of the results and the writing/adaptation of the text. MB was the project supervisor and was involved in the design, collection of data, interpretation of the results and the writing/adaptation of the text. EK works at the Dutch Consumer Safety Institute and was involved with the writing/adaption of the text and with preparing the manuscript for publication.