. 2021 Apr 9;11(4):e044199. doi: 10.1136/bmjopen-2020-044199

Table 2.

Article data extraction

Author (year)	Sample Descriptors Country of origin n= (sex, %) Race (%) Age M±SD; range Recruitment source Sport (%) Level of sport History of sport engagement (frequency/years)	Injury Description Definition of injury (yes/no: definition) Sport injury/severity Time out of sport; M(SD) Rehabilitation protocol and surgery details	Study Design and Objectives Study design Primary objective Secondary objective	Model or Theory Referenced Authors (year) Model name	Outcome Measures AIMS: 7 or 10 items Timeline of administration Group; score (M±SD) Names of additional measure(s) used	Key Findings Pertaining to AI	Study Strengths and Limitations
Padaki et al (2018)⁴⁹	USA n=24 (male, 50) — 14.5±2.7 Tertiary care centre Single sport (29.2); multisport (58.3) — —	Yes: ‘ACL rupture requiring surgery’ ACL tear; 41.7% reporting concomitant meniscal injury — —	Cross-sectional To examine the psychological trauma, including potential PTSD symptomatology, following ACL rupture among young athletes.	— —	10 items Baseline: pre-operation Sex: male=53.4 vs female=56.6 Sport involvement: single sport=57.5 vs multisport=52.8 Age: ≤14 years old=54.5 vs 15–21 years old=54.1 SDs not provided 4. Level of sports specialisation; IES-R	Single sport athletes had significantly higher AIMS scores than multisport athletes No significant difference in AIMS scores by age group (≤14 years old vs 15–21 years old) No significant difference on IES-R between high (AIMS score: >50) and low AI groups (AIMS score: ≤49)	Strengths: Only study to group athletes by sport specialisation (as per the American Orthopaedic Society for Sports Medicine definition; that is, single vs multisport athletes) and compare AIMS scores between groups Only study to examine psychological trauma associated with a sport injury Limitations: Small sample size Unknown how long athletes were removed from sport Figures are provided, but exact values are not referenced Does not appear that tests of statistical significance were conducted to compare high and low AI groups No pre-injury data available Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Hilliard et al (2017)⁵⁵	USA n=79 (male, 64.6) Caucasian (70) 19.96±1.56 Athletic training clinics in colleges or universities in Midwestern USA Football (35); soccer (18); basketball (11); track (10); baseball (6); volleyball (6); gymnastics/dance (6); swimming (4); cross-country (3); field hockey (3); lacrosse (1); wrestling (1); not specified (2) Division I (26%); Division II (15%); Division III (40%) and NAIA (19%) 14.19±9.40 hours spent training/week prior to injury; 10.45±4.46 years involved in sport	Yes: ‘experiencing an MSK injury considered moderate in severity that results in at least 7 days of missed practice or competition and receiving physiotherapy for the injury’ ACL tear (13.9%); sprains (12.6%); fractures (6.3%); undefined injury, only general area reported (eg, right knee, lower back, etc) (67%) As per definition, ‘…at least 7 days of missed practice or competition…’; median of 4 weeks reported since time of injury (range 1–63 weeks) 42% of injuries required surgery, not otherwise specified	Cross-sectional convergent parallel mixed methods To explore what aspects of AI might predict overadherence to rehabilitation. To get a better understanding of participants’ views of their athletic participation and rehabilitation adherence.	Wiese-Bjornstal et al (1998)³⁰ Integrated Model of Response to Sport Injury	7 items Baseline: post-injury 5.78±0.72† ROAQ	Positive moderate and significant association between AIMS score and overadherence to rehabilitation protocols Positive moderate and significant association between AIMS score and attempts to expedite rehabilitation process Positive moderate but non-significant association between AIMS score and willingness to ignore practitioner recommendations pertaining to rehabilitation AIMS negative affectivity subscale independently predicted likelihood that athlete would: (1) ignore practitioner recommendations and (2) attempt to expedite the rehabilitation process	Strengths: Sample is described clearly and thoroughly (eg, clear definition of injury, sport, level of play, frequency of sport involvement, type of sport injury, time removed from sport) Range of sports and levels of play captured increase the generalisability of findings Study design used does not prioritise one aspect of the research over the other (ie, quantitative vs qualitative) Regression models have sufficient power Captured a range of MSK injuries Clear operational definition of injuries eligible for inclusion Limitations: Only one additional outcome measure administrated ROAQ assesses athlete beliefs, not actual behaviours Sample is predominantly male Statistical tests comparing AIMS scores with subscale scores increases likelihood of multicollinearity Large variation in ‘time since injury’: 1 week (acute) vs 63 weeks (chronic) No pre-injury data available
O’Rourke et al (2017)⁵⁴	USA n=51 (male, 52.9) — 14.53±1.85 Athletes presenting to a local hospital or university-affiliated outpatient concussion clinic Soccer (24); lacrosse (10); football (8); other (58; skiing, volleyball, hockey, swimming, ultimate frisbee, cheerleading and wrestling) — —	Yes: suffered a concussion in the past 14 days; unknown diagnostic criteria Concussion — —	Prospective longitudinal To assess the role of psychological factors on self-reported post-concussion recovery in youth athletes within an existing theoretical and empirically supported framework. To assess non-psychosocial variables previously shown to influence concussion symptomatology (eg, age, gender, number of days post-concussion and number of previous concussions).	Wiese-Bjornstal et al (1998)³⁰ Integrated Model of Response to Sport Injury	7 items Time 2: ~14–21 days post-concussion 38.25±6.23 SCAT-2; AGS-YS; MCS-YS; PIMCQ-2; SMS; SAS-2; SNS	Moderate positive and significant association with AIMS score: mastery-orientation, ego-orientation, parent ego climate, intrinsic and extrinsic motivation, social network size, post-concussion symptoms at time 2 and 3 Small negative and significant association between AIMS score and social network satisfaction Stronger AI significantly predicted more severe post-concussion symptoms at time 3 (~21–28 days post-concussion)	Strengths: Only study to capture and compare AI with presence of post-concussion symptoms at multiple time points in the acute recovery phase Similar number of male and female athletes captured in sample Thorough evaluation of athlete motivation captured via measures administered Limitations: Poorly described sample with respect to level of and frequency of sport involvement Use of a hospital-based clinic as a recruitment source may have biased the study sample (ie, captured athletes with more severe concussion symptoms) Follow-up measures administered in close proximity (time 1: ~1–14 days post-concussion; time 2: ~14–21 days post-concussion; time 3: ~21–28 days post-concussion) Diagnostic criteria for concussion not stated No pre-injury data available
Baranoff et al (2015)⁵²	Australia Time 1: n=44 (male, 61.4); Time 2: n=26 (male, 46.1) 27±9.4 Physiotherapy clinics Australian rules football (29.5); netball (18.2); basketball (13.6) — —	Yes: ACL tear ACL tear Mean time between injury and surgery: 7 weeks, 6 days (SD=9 weeks, 4 days) ACLR rehabilitation protocol; ACL allograft reconstruction (11.4%); ACL autograft reconstruction (89%)	Prospective longitudinal To assess the roles of catastrophising and acceptance in relation to depression, pain intensity and substance use to cope with an injury 2 weeks post-ACL reconstructive surgery (time 1) and 5 months of ACLR rehabilitation (time 2).	— —	7 items Baseline: 0–2 weeks post-operation 31.0±9.0 AAQ; PCS; DASS 21	Strong positive and significant association between AIMS score and depressive symptom severity	Strengths: Equal representation of males and females in sample T-tests conducted to determine if there was a significant difference between athletes who submitted questionnaires at both time points versus at time 1 only; no significant difference between groups on measures of depression Measure mean/SD provided for both groups (ie, athletes who completed questionnaires at both time points vs time 1 only) Limitations: Small sample size Only three sports captured Frequency and years of sport involvement not provided for sample ~8 weeks between occurrence of injury and questionnaire completion No pre-injury data on AI Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Samuel et al (2015)⁴⁵	Israel n=6 (unknown) — 21.83±2.93 Sports medicine centres Basketball (33.3); judo (33.3); track and field (16.7); gymnastics (16.7) Internationally ranked (83.3%); nationally ranked (16.7%) 11.17±3.41 years involved in sport	Yes: ACL tear ACL tear Range: 7–12 months —	Prospective longitudinal To examine competitive athletes’ experience of severe injuries.	Samuel et al (2011)⁷⁵ Scheme of Change for Sport Psychology Practice	7 items Multiple: time 1: 2.25 months from date of initial injury; time 2: 6.58 months from date of initial injury; time 3: 10.08 months from date of initial injury Time 1=45.17±1.83 Time 2=43.33±3.83 Time 3=44.55±3.50 4. CEI; BCope	No significant difference between AIMS scores as assessed at different time points	Strengths: Years of sport involvement provided AI was assessed at multiple time points, with sufficient time between follow-ups Limitations: Small sample size Participant raw data provided; means/SDs not calculated Sex distribution of sample not provided Recruitment source not provided Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Kroshus et al (2015)⁴²	USA n=146 (baseline); n=116 (post-season) (male, 100) — — Collegiate teams Ice hockey Division I (NCAA) —	Yes: NCAA definition of concussion Concussion — —	Prospective cohort To assess the association between pre-season individual characteristics and post-season recall of within-season concussion symptom-reporting behaviours.	Cialdini and Trost (1998)⁷⁶ Social Influence: Social Norms, Conformity and Compliance	7 items Baseline: pre-season, pre-injury 39.79±4.73 Concussion history; CKI; CAI; HIQ	Significant interaction identified between perceived concussion reporting norms and AIMS score with respect to predicting non-reporting behaviours; stronger AI was associated with non-report AIMS score alone did not significantly predict non-reporting behaviours	Strengths: Only study to exclusively examine concussion reporting behaviours Homogeneous sport sample captured; all participants were NCAA Division I ice hockey players Large sample size Limitations: All male sample; not generalisable to females Reporting behaviours subject to recall bias; follow-up questionnaires were administered at the end of hockey season Reporting behaviours based on presence of post-impact concussion symptoms rather than incidence of unreported suspected concussions
Madrigal and Gill (2014)⁴⁴	USA n=4 (female, 100) — Only range was provided: 20–21 years old NCAA Division I school teams; by referral via team athletic trainer Softball; women’s soccer NCAA Division I —	Yes: ‘sport injury that is expected to prevent/limit his/her sport participation for at least 4 days’ Meniscus tear, leg injury (not otherwise specified), broken bone in hand, labrum tear in shoulder Range: 5 weeks–8 months 50% required surgery	Prospective longitudinal To examine an athlete’s psychological strengths (ie, mental toughness, hardiness and optimism) and emotional response to sport injury and rehabilitation and coping resources. To examine individual differences and changes over time from injury to being cleared to play.	1a. Wiese-Bjornstal et al (1998)³⁰ 2a. Integrated Model of Response to Sport Injury 1b. Lazarus and Folkman (1984) ³³ 2b. Stress Appraisal and Coping	10 items Multiple: time 1: pre-season; time 4: cleared to play Time 1=54.25±7.80 Time 4=53.67±8.74 4. MTS; PPI-A; LOT-R; BCope; PRSII; RAQ; DRS	No significant difference identified between AIMS score as measured at pre-season and return to play following injury	Strengths: equal representation ofmales and females in sample Assessed AI prior to injury Captured a range of MSK injuries Limitations: Frequency and years of sport involvement not provided Small sample size Measure means/SDs not calculated for sample; participant raw data provided Results were presented for each athlete, rather than summary for the entire sample Narrow age range captured (20–21 years old)
Masten et al (2014)⁵⁸	Slovenia n=68 (male, 69.1) — M=23.4; range: 16–40 years old Orthopaedic clinic in Ljubljana, Slovenia Handball (20.6); football (20.6); basketball (19.1); volleyball (6); alpine skiing (<3); ice hockey (<3); judo (<3); snowboarding (<3); tennis (<3); running (<3); gymnastics (<3); rugby (<3); standing/acrobatic skiing (<3) World-class and internationally ranking (41.2%); national ranking or uncategorised (58.8%) —	Yes: according to a previously proposed injury rating scale; individuals categorised to be in group 4 (ie, rehabilitation time expected to be up to 1 month) or group 5 (ie, rehabilitation time expected to be over 1 month and up to 6 months) Meniscus tear; ACL/PCL; patella injury; unreported (% not reported); group 4 (8.8%), group 5 (76.5%) As per inclusion criteria, removed from sport for at least 1 month Standard rehabilitation protocol, not otherwise specified; ‘knee surgery’, not otherwise specified	Cross-sectional To examine if athletes differ from each other in depression, general irritability and inhibition of behaviour regarding injury severity. To examine the psychological response to injury on the basis of specific dispositional characteristics to identify those personality and dispositional traits that make athletes more prone to injury.	— —	7 items Baseline: pre-operation — FPI; STAI-X1; SIP 15; SIRBS; 6-item author-developed scale assessing social support provided by family, coach and sport colleagues, and athlete’s motivation for rehabilitation	AIMS scores independently predicted an athlete’s motivation to engage in rehabilitation as well as their subjective value of rehabilitation; athletes with stronger AI were significantly more likely to have greater motivation and positive views towards rehabilitation	Strengths: Only study to exclusively capture high-ranking athletes (eg, world class, international and national) Compared athletes by injury severity (more severely injured (expected rehab time >1 month but ≤6 months) vs less severely injured (expected rehab time ≤1 month)) Diverse group of athletes captured Wide age range captured (16–40 years old) Limitations: AIMS mean/SD not provided or compared between more severely injured versus less severely injured athletes Level of sport involvement was not provided for majority of sample Questionnaires only administered at one time point; unable to make any conclusions about changes to AI as a result of sport injury
Petrie et al (2014)⁴¹	USA n=26 (male, 100) Black (52.2) 20.08±1.46 Football teams from the Southwestern USA Football NCAA Division I —	Yes: ‘(an injury) defined as having occurred as a result of participation in an organised intercollegiate practice or game, requiring medical attention by a team athletic trainer or physician, and having resulted in the inability to participate for one or more days beyond the day of injury’ Lower extremity not otherwise specified (69%); upper extremity (31%) 11.88 days±27.71 —	Prospective longitudinal To determine the direct effects of life stress, different sources of social support, AI and mental toughness on athletic injury over the course of a competitive season. To examine the potential moderating effects of social support, AI and mental toughness on the life stress–injury relationship.	Andersen and Williams (1988)⁷⁷ A Model of Stress and Athletic Injury	6 items; 1 item removed due to lack of variability Baseline: pre-season (ie, pre-injury) 32.23±5.71 LESCA; MSPSS; SMTQ	No significant associations between AIMS score: (1) life stress, (2) injury outcome, (3) social support or (4) mental toughness were identified AIMS score was not a significant predictor of ‘time lost’ (ie, number of days removed from sport due to injury); AIMS score interaction terms with (1) positive and (2) negative life stress were also non-significant	Strengths: Homogeneous sport sample captured; all participants were NCAA Division I football players Sample was racially diverse Assessed AI prior to sport injury Clear operational definition of injuries eligible for inclusion Limitations: Small sample size Frequency and years of sport involvement not provided Findings not generalisable to females No post-injury assessment of AI No comparison between injured and uninjured athletes with respect to AIMS baseline scores
Brewer et al (2013)⁵⁰	USA n=91 (male, 63.7) Caucasian (92) 29.73±10.24; range 14–54 years old Physical therapy clinics — Competitive (43%); recreational (54%) —	Yes: ACL tear ACL tear At least 6 weeks Accelerated ACL rehabilitation protocol as developed by Shelbourne et al; emphasis placed on early attainment of ROM, quadriceps strength and normal gait. Exercises tailored to and considered safe for patients’ stage of recovery, patients may be encouraged to exceed the prescribed number of sets to hasten their recovery	Prospective longitudinal To identify predictors of adherence to a postoperative ACL home rehabilitation programme.	1a. Lazarus and Folkman (1984) ³³ 2a. Stress, Appraisal and Coping 1b. Wiese-Bjornstal et al (1998)³⁰ 2b. Integrated Model of Response to Sport Injury	7 items Once: pre-operation 30.07±9.73 NEO-FFI- Neuroticism; LOT-R; POMS-B; Subjective Pain Rating; Subjective Daily Stress Rating	AIMS score did not significantly predict home exercise completion ratio (ie, number of sets of home exercises completed compared with what was prescribed) Significant interaction identified between AIMS score and daily stress as predictors of home exercise completion ratio; when daily stress was high, individuals with stronger AIs were more likely to complete their prescribed exercises	Strengths: Similar distribution of competitive versus recreational athletes One of three studies that assessed actual rehabilitation behaviours (eg, home exercise completion, cryotherapy) Limitations: Sample was predominantly Caucasian; findings may not be generalisable to other racial groups Sample was predominantly male Sample was poorly described; frequency and years of sport involvement and sports captured were not provided Exclusively captured ACL injuries; findings may not be generalisable to other injuries
McKay et al (2013)⁴³	Canada n=316 (male, 100) — Median=15; range 13–17 years old Elite ice hockey teams in Calgary, Alberta Ice hockey AAA, AA, A Bantam age group: mean of 8.06 years of organised hockey; midget age group: mean of 9.57 years of organised hockey	Yes: ‘any injury that required medical attention, resulted in the inability to complete the current session of activity, and/or required the cessation of sporting activity for at least 24 hours; Subsequent injury: ‘any injury that occurred during the season, after the first reported injury, regardless of anatomical position or injury type’ n=143 injuries reported: concussion (22.4%); muscle strain (14.7%); joint/ligament sprain (14.7%) As per definition —	Prospective cohort To determine the risk of injury associated with AI, attitudes towards body checking, competitive state anxiety and re-injury fear in elite youth ice hockey players. To determine if there is an elevated risk of subsequent injury associated with return to play before medical clearance.	— —	10 items Baseline: within 3 weeks of hockey season start, pre-injury 55.72±7.54 CSAI-2R; BCQ; FRQ; MPQ-SF	Athletes with AIMS score below the 25th percentile were at greater risk for incurring an injury; this finding was significant * Findings omitted due to publishing authors’ error; discrepancy between findings communicated in text of the Results section and tables	Strengths: Large sample size Athletes grouped by age for analysis Only study to examine AI in relation to injury risk Injuries were reported by an external source Homogeneous sport sample captured; all participants were elite male ice hockey players Only study to capture concussion and MSK injuries Clear operational definition of injuries eligible for inclusion Limitations: Reporting discrepancy in findings pertaining to AI; authors were contacted for clarification but no response was provided No post-injury assessment of AI Findings not generalisable to females Narrow age range captured (13–17 years old)
Podlog et al (2013)⁵⁶	Study 1: USA n=118 (male, 51.7) — 15.97±1.41 Teams in Texas Football (36); basketball (24); soccer (11); volleyball (8); track and field (5); baseball (4); softball (4); cheerleading (3); tennis (1.7); dance (0.8); swimming (0.8) School teams, local clubs or community leagues 14.18±8.93 hours per week spent training prior to injury; 6.69±2.80 years involved in current sport (range: 1–14 years) Study 2: USA n=105 (male, 59) — — NCAA teams across the USA Football (21); basketball (15); soccer (11); volleyball (9); track and field (4); baseball (16); softball (3); cheerleading/gymnastics (9); tennis (5); golf (0.9); rugby (0.9); swimming (2); lacrosse (2); snowboarding (2); missing (0.9) NCAA Division I, II, III 14.06±6.14 hours per week spent training prior to injury; 9.74±4.60 involved in current sport (range: 1–20 years)	Study 1: Yes: ‘were currently experiencing an injury requiring a minimum 2-week absence from sport training and competition, and currently receiving physiotherapy for their injury’ ACL tear (34.7%); medial malleolus/fibula/distal tibia fracture (22.9%); shoulder dislocation (7.6%); carpal tunnel syndrome (<1%) M=2.7 months (SD=2.01); range: 0.5–7 months 57.6% required surgery, not otherwise specified Study 2: Same as above ACL (17.1%); fractured humerus/femur/clavicle (14.3%); shoulder dislocation (8.6%); sprain (7.6%) M=2.49 months (SD=2.10); range: 0.5–7 months 50.5% required surgery, not otherwise specified	Study 1: Cross-sectional To provide initial validation of a novel injury-rehabilitation overadherence measure. Study 2: Cross-sectional To examine correlates of overadherence and premature return to sport.	Study 1 and Study 2: Wiese-Bjornstal et al (1998)³⁰ Integrated Model of Response to Sport Injury	Study 1: 7 items Baseline: post-injury 5.67±0.90† SPSQ†; ROAQ†; I-PRRS† Study 2: 7 items Baseline: post-injury 5.63±0.96† SPSQ†; ROAQ†; I-PRRS†	Study 1 only: AIMS scores significantly predicted attempts to expedite the rehabilitation process; athletes with a stronger AI were significantly more likely to think and behave in a way that would expedite rehabilitation Study 1 and study 2: Small positive and significant association between AIMS score and tendency to ignore practitioner rehabilitation recommendations AIMS scores significantly predicted rehabilitation tendencies; athletes with a stronger AI were significantly more likely to ignore practitioner recommendations	Strengths (Study 1 and Study 2): Samples captured were thoroughly described Wide range of sports and levels of involvement captured Large sample size Similar number of males and females captured Captured a range of MSK injuries Clear operational definition of injuries eligible for inclusion Limitations: No post-injury assessment of AI (both studies) Large variation in time lost (ie, number of days removed from sport) due to sport injury (both studies) Sample age (mean/SD) not provided in study 2
Weinberg et al (2013)⁵⁷	USA n=130 (52.3) — 20.03±1.60; range: 18–24 years old Intramural teams at a midsized university in the Midwestern USA Basketball (100) Recreational 6.64±3.98 years involved in sport	Yes: ‘playing through injury was defined in the current study as participating while still feeling pain so that (a) the pain/injury needs some sort of mental attention during participation, (b) involves some sort of loss of or change in function that would directly affect performance capabilities, therefore indicating a threat to well-being, and (c) a decision process was necessary as to whether participation should and/or would be initiated and continued during the experience of pain/injury‘ — — —	Cross-sectional To determine whether athletes’ attitudes and behavioural intentions regarding playing through pain and injury differ as a function of their level of AI and their gender.	— —	10 items Baseline: post-injury 4.15±1.21† RPIQ†; PIB†	Men scored significantly higher on each AIMS subscale compared with women AI significantly predicted athlete attitudes towards sport risk, pain and playing through pain; athletes scoring ≥75th percentile on the AIMS were more likely to have positive attitudes and behavioural tendencies to play through pain and injury compared with the moderate (between 25th and 75th percentile) and low AI groups (≤25th percentile) AIMS exclusivity and negative affect subscales significantly predicted RPIQ toughness (in regards to risk, pain and injury in sport), social role choice (willingness to accept risk, pain and injury in sport), and ‘pressed’ (perceptions of pressure exerted by others to play with pain and injury) subscale scores; athletes scoring higher on the exclusivity and negative affect AIMS subscales were more likely to endorse toughness (ie, risk, pain and injury) AIMS negative affect subscale scores significantly predicted athlete behavioural intentions to play through an injury; athletes with stronger AIs were more likely to play through an injury	Strengths: Large sample size Equal representation of men and women Limitations: Homogeneous sample of intramural basketball players; findings not generalisable to other sports Few details provided about injury Reporting behaviours subject to recall bias; questionnaires administered at an unknown time point following injury Did not assess actual behaviours following injury; operational definition (‘playing through injury as defined…’) applied as an inclusion criteria only Narrow age range captured (18–24 years old)
Brewer et al (2010)⁴⁸	USA n=108 (men, 66.7) Caucasian (90) 29.38±9.93; range: 14–54 years old Physical therapy clinics — Competitive 47%; recreational 49%; non-athletes 4% —	Yes: ACL tear ACL tear — —	1. Prospective longitudinal 2/3. To test the following predictions in a sample of physically active people who tore their ACL and underwent reconstructive surgery and rehabilitation: (a) decreasing one’s AI after ACL surgery could help to preserve self-esteem in the face of formidable threat to short-term and potentially long-term sport participation, and (b) greater decrements in AI are expected for those individuals who are experiencing slow post-operative recovery.	— —	7 items Multiple: time 1: pre-operation; time 2: 6 months post-operation; time 3: 12 months post-operation; time 4: 24 months post-operation Time 1=32.14±8.83 Time 2=31.62±8.23 Time 3=29.07±8.47 Time 4=28.45±8.09 4. Subjective rating of rehabilitation progress (%)	Time 1 and time 2, time 3 and time 4 AIMS scores were not significantly different; all other time point comparisons were significantly different and adjusted for age and gender Subjective ratings of rehabilitation progress significantly predicted AIMS score differences between time 2 and 3 after adjusting for time 1 AIMS score, gender and age; athletes who experienced a slower recovery were more likely to experience greater decreases to their AI	Strengths: Sufficient time between follow-up points Long-term follow-up; only study to gather information 2 years post-injury Bonferroni correction applied to tests of multiple comparisons Equal distribution of competitive and recreational level athletes Wide age range captured (14–54 years old) Limitations: Details about sports captured not provided Frequency and years of sport involvement not provided Details about sport injury not provided Men and Caucasians were over-represented in the sample; findings not generalisable to women and other races Small number of cases included in the data set for analysis (53.7% of total sample); no indication if tests of significance were conducted between included/excluded cases Limited number of covariates included in regression models Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Brewer et al (2007)⁴⁷	USA n=91 (male, 63.7) 29.73±10.24; range: 14–54 years old Physical therapy clinics — Competitive 43%; recreational 54%	Yes: ACL tear ACL — ACLR rehabilitation	Prospective longitudinal To examine predictors of daily pain and negative mood over the first 6 weeks of rehabilitation following ACL reconstruction.	Wiese-Bjornstal et al (1998)³⁰ Integrated Model of Response to Sport Injury	7 items Baseline: preoperation 30.36±9.71 NEO-FFI-Neuroticism Subscale; LOT-R; PDS; number of physical therapy appointments per/day; HOMEX (frequency of exercise completion with and without videocassette use); HOMEXRAT (division of HOMEX by number of sets of home rehabilitation exercises prescribed for a given day); EXERCISE (number of minutes spent ‘on vigorous physical activity other than their rehabilitation exercises’); NRS; POMS-B	AIMS score did not significantly and independently predict average daily pain AIMS score did not significantly and independently predict negative mood Significant interaction between AIMS score and number of days since surgery with respect to predicting negative mood; athletes with stronger AIs experienced greater decreases in negative mood as number of days since surgery increased	Strengths: Similar representation of recreational and competitive level athletes One of three studies that assessed actual rehabilitation behaviours (eg, home exercise completion, cryotherapy) Wide age range captured (14–54 years old) Limitations: Details about sports captured not provided Frequency and years of sport involvement not provided Details about sport injury not provided Males and Caucasians were over-represented in sample; findings not generalisable to females and other races Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Brewer et al (2003)⁴⁶	USA n=61 Caucasian (92) 26.03±7.99; range: 14–47 years old Physical therapy clinic — Competitive 57%; recreational 41% —	Yes: ACL tear ACL — ACL reconstruction; accelerated rehabilitation protocol	Prospective longitudinal To investigate whether prospective associations among psychological factors and rehabilitation adherence differ as a function of age through re-analysis of data from a previously published report.	1a. Wiese-Bjornstal et al (1998)³⁰ 2a. Integrated Model of Response to Sport Injury 1b. Brewer (1994)³¹ 2b. Cognitive Appraisal Models of Adjustment	10 items Baseline: ~10 days pre-operation 44.16±9.98 SMI; SSI; BSI; SIRAS*; ratio of appointments attended to those scheduled: home rehabilitation adherence–exercise completion; home rehabilitation adherence–cryotherapy	Significant interaction between age and AIMS score with respect to predicting: (1) home exercise adherence and (2) cryotherapy use; younger athletes with stronger AIs were more likely to complete at-home exercises and to use cryotherapy	Strengths: One of three studies that assessed actual rehabilitation behaviours (eg, home exercise completion, cryotherapy) Wide age range captured (14–47 years old) Limitations: Competitive athletes were over-represented in sample Males and Caucasians were over-represented in sample; findings not generalisable to females and other races Details about sports captured not provided Frequency and years of sport involvement not provided AIMS only assessed at one time point Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Manuel et al (2002)⁵³	USA Time 1 (baseline): n=48 (female, 58.3); time 2 (3 weeks) n=44; time 3 (6 weeks) n=40; time 4 (12 weeks) n=34 Caucasian (85) Range: 15–18 years old MSK Outpatient Physical Therapy Department at Wakeforest University Males: football (56); baseball (11); wrestling (11); females: soccer (25); basketball (21); track (14); volleyball (7) — —	Yes: ‘athletes who would be out of sports for at least 3 weeks’ Most common injury was ACL (no % provided); Injury Severity Scale as completed by the attending orthopaedic surgeon. Scores range from 1 to 4, with a lower score indicating a less severe injury; M=2.50 (SD=1.26) As per definition, out of sport for at least 3 weeks —	Prospective longitudinal To explore patterns of psychological distress in adolescents experiencing sport injuries.	— —	10 items Baseline: post-injury 47.20±9.78 ISS*; APES; PRQ-R-S; ACS; BDI	AIMS score significantly predicted depression scores; athletes with stronger AIs were more likely to experience more severe depressive symptoms	Strengths: Range of sports captured One of two studies to assess injury severity (based on physician rating) Limitations: Frequency and years of sport involvement not provided Few details provided with respect to injuries captured Small sample size Caucasians were over-represented in sample; findings may not be generalisable to other races AIMS only assessed at one time point Narrow age range captured (15–18 years old)
Green and Weinberg (2001)³⁵	USA n=30 (male, 60) Caucasian (93.3) M=30.8 (SD=missing); range: 19–70 years old Sport medicine clinics, physical therapy clinics and orthopaedic centres — — Minimum of 30 min of sport or physical activity/week	Yes: ‘discontinuance of regular physical activity/sport that was operationally defined as 30 min of physical activity a week, for a period of at least 6 weeks’ 50% knee injury; 26.7% other (three foot injuries, one broken tibia/fibula, one herniated disc, one broken arm); 10% shoulder injury; 6.7% hip injury; 3% ankle injury As per definition ‘at least 6 weeks’, no additional data provided —	Cross-sectional To examine coping skills and social support to better understand those individuals who are most vulnerable to injury.	1a. Kubler-Ross et al (1969)⁷⁸ 2a. Stage Models of Grief 1b. Brewer (1994)³¹ 2b. Cognitive Appraisal Models of Adjustment 1c. Lazarus and Folkman (1984) ³³ 2c. Cognitive Appraisal Models of Adjustment 1d. Andersen and Williams (1988)⁷⁷ 2d. A Model of Stress and Athletic Injury 1e. Wiese-Bjornstal et al (1998)³⁰ 2e. Integrated Model of Response to Sport Injury	10 items (note: 5-point Likert response scale used) Baseline: post-injury 43.10±11.51 ACSI; POMS; PSPP; SSQ	Negative but non-significant association between AIMS score and depressive mood Moderate positive and significant association between AIMS score and physical conditioning AIMS score did not significantly predict depressive symptom severity	Strengths: Captured a range of MSK injuries Wide age range captured (19–70 years old) Clear operational definition of injuries eligible for inclusion Limitations: Information about sports and levels of athlete sport involvement not provided Caucasians were over-represented in sample; findings may not be generalisable to other races Small sample size AIMS only assessed at one time point
Brewer et al (2000)⁵¹	USA n=95 (male, 70.5) Caucasian (88) 26.92±8.23 Physical therapy clinic — Competitive (52%); recreational (43%); non-athletes (3%); missing (2%) —	Yes: ACL tear ACL tear — Accelerated ACL rehabilitation protocol as developed by Shelbourne et al; emphasis on early attainment of ROM, quadriceps strength and normal gait	Prospective longitudinal To examine the relationships among psychological factors, rehabilitation adherence and rehabilitation outcomes after ACL reconstruction.	1a. Brewer (1994)³¹ 2a. Cognitive Appraisal Models of Adjustment 1b. Wiese-Bjornstal et al (1998)³⁰ 2b. Integrated Model of Response to Sport Injury 1c. Self-developed by authors 2c. Adapted model based on above referenced models (see article)	10 items Baseline: ~10 days preoperation 41.65±12.16 SMI; SSI; BSI; SIRAS*; ratio of appointments attended to scheduled: home rehabilitation adherence–exercise completion; home rehabilitation adherence–cryotherapy; KT 1000 (joint laxity); one leg hop distance; LKSS	Small positive and significant association between AIMS score and motivation Moderate positive and significant association between AIMS score and joint laxity as measured 6 months following ACL reconstructive surgery Small positive and significant association between AIMS score and (1) one leg hop distance and (2) knee function as measured 6 months following ACL reconstructive surgery AIMS score significantly predicted joint laxity as measured 6 months following ACL reconstructive surgery; athletes with stronger AIs were more likely to have similar knee joint stability between the affected and unaffected leg	Strengths: Large sample size Only study to measure functional injury outcomes (eg, joint laxity, one leg hop distance, pain) using objective measures Limitations: Exclusively captured ACL injuries; findings may not be generalisable to other injuries Frequency and years of sport involvement not provided Males and Caucasians were over-represented in sample; findings may not be generalisable to females and other races AIMS only assessed at one time point Exclusively captured ACL injuries; findings may not be generalisable to other injuries
Brewer (1993)²⁶	Study 3 USA n=121 (male, 66.9) — — Sport medicine clinics in Phoenix, Arizona — — — Study 4 USA n=90 (injured: 16.7%); (male, 100) — — University of California Varsity Football Team — — —	Study 3 No Physician-rated injury severity on a 3-point scale (1=mild, 2=moderate, 3=severe); M=2.10 Injury status at time of enrolment on a 7-point scale (1=acutely injured, 7=completely recovered) M=3.53 — Study 4 No — — —	For both studies Cross-section observational To test the prediction that individuals who maintain strong, exclusive identification with the athlete role are more likely to become depressed following an athletic injury than individuals without such an identification. Study 3 3. To assess the extent to which AI was related to depressed mood in a sample of athletes who were already injured. Study 4 3. To investigate the relationship between AI and depressed mood in a sample of both injured and uninjured athletes.	1a. Abramson et al (1989)⁷⁹; Alloy et al (1988)⁸⁰; Beck (1967⁶⁴, 1970⁶⁸); Dance and Kuiper (1987)⁶⁶; Linville (1987)⁶⁷; Robins and Block (1988)⁸¹ 2a. Cognitive Diathesis–Stress Models of Depression 1b. Oatley and Bolton (1985)⁸² 2b. Social Cognitive Theory of Reactive Depression	Study 3 10 items Baseline: ~2 weeks following injury 47.93±9.98 PSPP-G; SARRS; POMS-D; BDI Study 4 10 items Baseline: pre-season Injured=48.47±9.09 Non-injured=51.60±9.09 4. PSPP-G; SARRS; POMS-D; BDI	Study 3: AIMS score was not significantly associated with depressive symptom severity AIMS score was a significant independent predictor of depressive symptom severity; athletes with stronger AIs were more likely to experience more severe symptoms of depression Small positive and significant association between AIMS scores and physician-rated injury severity Study 4: Significant interaction between AIMS score and physician-rated injury severity in regard to predicting depressive symptom severity; athletes with a stronger AI and more severe injury were more likely to experience depressive symptoms of a greater severity No significant difference in AIMS score between injured and uninjured groups	Strengths (study 3): One of two studies to assess injury severity (based on physician rating) Large sample size Limitations (study 3): Males were over-represented in sample; findings may not be generalisable to females Strengths (study 4): Only study to compare AIMS scores between injured and uninjured group of athletes Limitations (study 4): Exclusively captured male football players; findings may not be generalisable to females and other sports Very small proportion of injured athletes captured (20% of total sample) Strengths (both studies): Cross-validated depressive symptom severity using two measures of depression Limitations (both studies): Details pertaining to sport injury not provided Frequency and years of sport involvement not provided AIMS only assessed at one time point No operational definition of sport injury provided

Each number with — means there is missing data point.

*Clinician-reported data.

†Item mean score.

AAQ, Acceptance and Action Questionnaire; ACL, anterior cruciate ligament; ACLR, anterior cruciate ligament reconstruction; ACS, Adolescent Cope Scale; ACSI, Adolescent Coping Skills Inventory; AGS-YS, Achievement Goal Scale for Youth Sports; AI, athletic identity; AIMS, Athletic Identity Measurement Scale; APES, Adolescent Perceived Events Scale; BCope, Brief COPE; BCQ, Body Checking Questionnaire; BDI, Beck Depression Inventory; BSI, Brief Symptom Inventory; CAI, Rosenbaum and Arnett’s Concussion Attitudes Index; CEI, Change-Event Inventory; CKI, Concussion Knowledge Index; CSAI-2R, Competitive State Anxiety Inventory 2-R; DASS 21, Depression, Anxiety and Stress Scale; DRS, Dispositional Resiliency Scale; FPI, Freiburger Persönlichnkeitsinventar-Personality; FRQ, Fear of Reinjury Questionnaire; HIQ, Head Injury Questionnaire; HOMEX, home exercise completion with and without videocassette; IES-R, Horowitz Impact of Event Scale-Revised; I-PRRS, Modified Injury Psychological Readiness to Return to Sport Scale; ISS, Injury Severity Scale; LESCA, Life Events Survey for Collegiate Athletes; LKSS, Lysholm Knee Scoring Scale; LOT-R, Life Orientation Test-Revised; MCS-YS, Motivational Climate Scale for Youth Sports; MPQ-SF, McGill Pain Questionnaire-Short Form; MSK, musculoskeletal; MSPSS, Multidimensional Scale of Perceived Social Support; MTS, Mental Toughness Scale; NAIA, National Association of Intercollegiate Athletics; NCAA, National Collegiate Athletics Association; NEO-FFI-Neuroticism, NEO Five Factor Inventory-Neuroticism Subscale; NRS, Numerical Rating Scale; PCL, posterior cruciate ligament; PCS, Pain Catastrophizing Scale; PDS, Perceived Daily Stress; PIB, Perceived Injury Behaviour; PIMCQ-2, Parent-Initiated Motivational Climate Questionnaire-2; POMS, Profile of Mood States; POMS-B, Profile of Mood States-B; POMS-D, Profile of Mood States-Depression; PPI-A, Psychological Performance Inventory-A; PRQ-R-S, Personal Resource Questionnaire-Revised-Social Support; PRSII, Psychological Response to Sport Injury Inventory; PSPP, Physical Self-Perception Profile; PSPP-G, Physical Self-Perception Profile-Global Physical Self-Worth Subscale; PTSD, post-traumatic stress disorder; RAQ, Rehabilitation Adherence Questionnaire; ROAQ, Rehabilitation Over Adherence Questionnaire; ROM, range of motion; RPIQ, Risk of Pain and Injury Questionnaire; SARRS, Social and Athletic Readjustment Scale; SAS-2, Sport-Anxiety Scale-2; SCAT-2, Sport Concussion Assessment Tool; SIP 15, Sports Inventory for Pain; SIRAS, Sport Injury Rehabilitation Adherence Scale; SIRBS, Sport Injury Rehabilitation Belief Scale; SMI, Self-Motivation Inventory; SMS, Sport Motivation Scale; SMTQ, Sports Mental Toughness Questionnaire; SNS, Social Network Scale; SPSQ, Self Presentation in Sport Questionnaire; SSI, Social Support Inventory; SSQ, Social Support Questionnaire; STAI-X1, State Anxiety.