What Is the Relationship of Fear Avoidance to Physical Function and Pain Intensity in Injured Athletes?

Abstract

Background

Fear avoidance can play a prominent role in maladaptive responses to an injury. In injured athletes, such pain-related fear or fear avoidance behavior may have a substantial influence on the recovery process. Specifically, it may explain why some are able to reach their preinjury abilities, whereas others are unable to return to sport.

Questions/purposes

(1) Is fear avoidance in athletes associated with decreased physical function after injury? (2) To what degree is fear avoidance associated with athletes’ pain intensity?

Methods

In a cross-sectional study, we recruited injured athletes—defined as patients with sports-related injury, weekly engagement in sport activities, participation in competitive events as part of a team or club, self-identification as an athlete, and a desire to return to sport after recovery—from an orthopaedic sports medicine center at a major urban university hospital. Of 130 approached patients, 102 (84% men; mean ± SD age 25 ± 8.5 years) met the inclusion criteria. Participants completed a demographic questionnaire, the Athlete Fear Avoidance Questionnaire, which assesses injury-related fear and avoidance behavior specifically in an athletic population, the Pain Catastrophizing Scale, the Hospital Anxiety and Depression Scale, and two Patient-Reported Outcomes Measurement Information System measures: Physical Function Computerized Adaptive Testing (CAT) and Pain Intensity CAT.

Results

After controlling for age, injury region (upper versus lower extremity), catastrophic thinking, and emotional distress, we found that an increase in athletes’ fear avoidance was associated with a decrease in physical function (b = -0.32; p = 0.002). The model explained 30% of the variation in physical function with 7.3% explained uniquely by fear avoidance. After controlling for initial appointment/followup, surgery for the current condition, multiple pain conditions, history of prior sport-related injury/surgery, pain medication prescription, catastrophic thinking, and emotional distress, athletes' fear avoidance was not associated with pain (b = -0.14; p = 0.249). The model explained 40% of the variation in pain intensity and pain catastrophizing (b = 0.30; p = 0.001) uniquely explained 7.1% of this variation.

Conclusions

In injured athletes, fear avoidance is independently associated with decreased physical function, whereas pain catastrophizing is associated with high pain intensity. Both level of an athlete’s fear avoidance and catastrophic thinking about pain should be accounted for in clinical interventions aimed at helping athletes improve recovery and return to sport.

Level of Evidence

Level II, prognostic study.

Introduction

Psychosocial factors play an important role in the recovery process of injured athletes [42, 46, 48] and explain why some athletes are able to reach their preinjury level of function, whereas others are unable to return to sport. Among these factors, emotional distress (such as symptoms of depression and anxiety) [2, 26, 47, 61] and maladaptive coping styles (like catastrophic thinking about pain) [51] have been consistently depicted as pivotal in their association with decreased physical function and increased pain in this population. These factors are also associated with prolonged recovery and higher risk of reinjury [10, 53]. Although athletes typically are considered to better endure pain during competition, when it comes to injury, they tend to experience higher levels of emotional distress and catastrophic thinking than the general population as a result of perceived external (coaches, peers, family, and media) or internal (athletic identity, financial pressures, guilt) pressures [6, 9, 31, 48].

In addition to emotional distress and coping, athletes’ fear avoidance—a set of pain-related fear or fear avoidance behaviors in response to a sport injury—may be an important predictor of recovery [15, 30, 43, 53]. The Fear Avoidance Model, originally developed by Lethem et al. [33] in 1983 and expanded by Vlaeyen et al. [57] in 1995, is a cognitive-behavioral model developed to understand why some individuals who experience pain recover, whereas others develop chronic pain and disability. The model depicts fear avoidance as a critical factor in whether patients adjust to their injury and recover or engage in a pattern of avoidance behaviors that impede recovery [19, 32, 58]. The original fear avoidance questionnaire [59] has been developed and validated in the working population with lower back pain to detect fear avoidance beliefs and predict return to work. Its wording has been adapted in some studies for applications in other body regions and injured patients [40, 43, 56], but the questionnaire does not account for circumstances that are specific to athletes. For instance, athletes may not only have concerns about pain and time to return to sport, but also about the risk of reinjury, expectations of peers and trainers, and losing their athletic identity if the injury jeopardizes their future career. The Athletes Fear Avoidance Questionnaire [20] was developed to assess these specific fears associated with athletes' injuries and return to sports.

To our knowledge, no prior studies have investigated if athletes' specific fear avoidance is associated with pain intensity and physical function during the recovery process of injured athletes. The purpose of this study was to determine to what extent athletes' fear avoidance is associated with physical function and pain intensity in athletes presenting to an orthopaedic sports medicine clinic. In conducting this study, we intend to extend prior research on psychosocial factors in injured athletes [3, 5, 15, 16, 30-32, 37, 41, 42, 46, 47, 53].

Therefore, we asked: (1) Is fear avoidance in athletes associated with decreased physical function after injury? (2) To what degree is fear avoidance associated with athletes’ pain intensity?

Patients and Methods

This was a descriptive, cross-sectional study, which took place in a sports medicine center within the Department of Orthopedic Surgery at a major medical center in Boston, MA, USA, between January and August 2016. A research assistant rounded in the sports clinic 2 days a week and approached prospective patients for participation. Enrollment, inclusion and exclusion criteria, and questionnaire completion occurred in the examination room before the meeting with the physician. Participants completed questionnaires on an encrypted iPad^® (Apple Inc, Cupertino, CA, USA) using Assessment Center [25]. Our institutional review board had approved the study and its procedures before enrollment commenced.

The study inclusion criteria were: (1) age 18 years or older; (2) sports-related injury; (3) English fluency and literacy; (4) ability to provide informed consent; (5) being an athlete, which we defined as an individual with participation in sports activities on a regular basis before injury (minimum once per week), participation in competitive team/group events or individual sport competition as a member of a team/club (such as running, or triathlons, cycling team), and self-identification as an athlete; and (6) desire to return to sport. Exclusion criteria per our institutional review board included (1) pregnant women; or (2) self-reported severe and untreated mental health conditions such as bipolar disorder, schizophrenia, or psychotic symptoms.

Demographic information included age, gender, race, years of education, marital status, sports-related information such as level of participation in sports, number of competitive sports events during the past year, current injury specifics (traumatic versus nontraumatic, upper extremity injury [eg, shoulder, elbow] versus lower extremity injury [eg, knee]), sports-related medical history (prior sports-related injuries, prior sports-related surgeries), surgery status (having had prior surgery for the current injury), time since injury/surgery, and prescription of pain medication.

We used the Tegner Activity Scale [52] to assess self-reported preinjury level of activity. The scale provides a measure of working and sporting activity levels, ranging from 0 “sick leave/disability” to 10 “participation in national and international elite competitive sports.”

We used the Patient Reported Outcomes Measurement Information System (PROMIS) Item Bank—Physical Function Version 1.2 with Computer Adaptive Testing (CAT) to assess patients’ self-reported level of physical function. The PROMIS Item Banks have been developed for a standardized assessment of health outcomes from the patient’s perspective [22]. Response items are related to participants' perception of their ability to perform physical activities (eg, “Are you able to exercise for an hour?” or “Does your health now limit you in doing vigorous activities such as running, lifting heavy objects, or participating in strenuous sports?”). Response options on abilities and limitations range on a 5-point scale from 1 being "unable to do/cannot do" to 5 being “[able to do] without any difficulty/not at all [limited].” Using CAT, item selection from the full item bank (121 items) followed a dynamic algorithm in response to the participant's answers [12, 14, 23]. PROMIS scores are rescaled as T-scores representing a standardized mean of 50–equaling the US general population–and a SD of 10 [12, 34]. Therefore, a person who has a T-score of 40 has a physical function 1 SD below the US general population mean. PROMIS Physical Function CAT is a reliable and valid instrument with low floor or ceiling effects [22, 27].

PROMIS Pain Intensity scale 3a (Version 1.0) was used to measure the perceived intensity of pain. It consists of three questions asking “In the past 7 days, how intense was your pain at its worst?”; “In the past 7 days, how intense was your average pain?”; and “What is your level of pain right now?” Each question has five response options ranging from “no pain” to “very severe.” The final score is represented as a T-score, which is weighted based on a mean of 50 and a SD of 10 for all samples derived from the US general population. Therefore, a score of 60 indicates a level of pain intensity that is 1 SD higher than the average of the US general population.

The Athlete Fear Avoidance Questionnaire (AFAQ) was used to measure sports injury-related fear avoidance in athletes. The AFAQ consists of 10 items assessing an athlete’s perceptions regarding an injury (eg, “I will never be able to play as I did before my injury” or “I worry if I go back to play too soon I will make my injury worse”) [20]. Answer options range from 1 being “not at all” to 5 being “completely agree” with the total score ranging from 10 to 50. Higher scores represent a greater degree of sport-specific fear and avoidance. The scale has been validated by correlation analysis with the Pain Catastrophizing Scale (PCS) [50] and the Fear-Avoidance Beliefs Questionnaire (FABQ) [59] in a set of athletes and showed good psychometric properties in measuring athletes' fear avoidance [20].

The PCS is a self-assessment scale to measure catastrophic thinking about pain [50]. Participants have the option to choose on a range between 0 being “not at all” and 4 being “all the time” on 13 items (eg, “I can't seem to keep it out of my mind”; “I wonder whether something serious may happen”; “It's terrible and I think it's never going to get any better”). The PCS has demonstrated good reliability and construct validity in the general as well as athletic population [50]. Scores below 13 translate into low, from 13 to 30 into moderate, and above 30 into high clinically relevant levels of catastrophizing [50].

The Hospital Anxiety and Depression Scale (HADS) was used to measure the symptoms of depression and anxiety [62]. The scale consists of 14 items with responses being scored on a scale from 0 to 3 with higher scores indicating a higher level of symptom frequency [60]. The questionnaire comprises the two subscales of anxiety (eg, “I get a sort of frightened feeling like ‘butterflies’ in the stomach”) and depression (eg, “I have lost interest in my appearance”). The combined scale (emotional distress) ranges from 0 to 42 with scores of 15 or higher indicating a clinically significant mood disorder [17, 28, 62]. The HADS has shown good psychometric properties [38].

Statistical Analysis

We present patients' characteristics by measures of central tendencies (eg, proportion, mean, median) as appropriate. Bivariate relations among continuous variables (eg, AFAQ, HADS, PCS, and PROMIS scores) were tested using Spearman's rs and Pearson’s r, between continuous and dichotomous variables (eg, gender, surgery status, upper/lower injury region, narcotics) using independent t-test, and between continuous and categorical variables (eg, marital status, work status) using one-way independent analysis of variance. Small, medium, and large effect sizes were distinguished by Pearson r values of ± 0.10, ± 0.30, and ± 0.50, respectively [18]. Visualization tools were used to objectify the relationship among multiple variables (Note: to ease the readability of Figure 1, HADS and PCS scores are displayed after transformation into clinically relevant categories.). To estimate whether fear avoidance has an association with physical function above and beyond relevant demographic and clinical variables, we conducted a hierarchical linear regression. In step 1, we included relevant demographic variables with a strong association to physical function in bivariate analysis, intensity in bivariate analysis, in step 2 relevant modifiable psychologic variables, and in step 3 fear avoidance. We conducted a similar analysis for pain intensity. We report significance using an α level of 0.05. We report the percent of variance explained by the entire model, the percent of variance explained in each step as well as the percent of variance explained by each individual variable. All analyses were performed using Stata/MP 13.0 (Stata Corp, College Station, TX, USA).

Fig. 1.

The sum of graphs visualizes the relationships of athlete fear avoidance and psychosocial variables with physical function (left column) and pain intensity (right column).

Ross reported that fear avoidance explained 12% of the variance in physical function after anterior cruciate ligament reconstruction [43]. We a priori calculated that 93 patients are needed to gain 0.80 power in detecting a R² increase of 8% in adding a single factor (athlete fear avoidance) to a hierarchical linear multiple regression model with a total of 10 predictors to explain physical function on the basis of a two-tailed type I error probability of 0.05.

Of the 130 patients approached, one patient refused participation and 129 (99%) consented to participate. Of these, 27 (21%) were not included as a result of one of the following reasons: (1) no sports-related injury; (2) not involved in competition; (3) no regular participation in sports; or (4) not part of a team. One hundred two patients (78%) were fully enrolled in the study.

The majority of enrolled patients (N = 102) were in their early 20s (mean ± SD age, 25 ± 8.5 years), men (n = 86 [84%]), white (n = 99 [97%]), and single (n = 84 [82%]). Approximately two-thirds of the sample reported a history of involvement in national elite (n = 15 [15%]) or lower division (n = 52 [51%]) competitive sports before the injury with an overall median of 50 competitive games/events per year. The majority of injuries was trauma-related (n = 64 [63%]), involved the upper extremity (n = 83 [81%]), and occurred approximately 4 months before enrollment. There was an equal split between those who have been treated with surgery (n = 51 [50%]) and those who were treated nonoperatively or did not receive treatment yet. Seventy-two percent (n = 73) of patients reported that this was not their first sport-related injury, and 33% (n = 34) had already undergone a prior sport-related surgery in their career. Twenty-two percent (n = 22) of patients reported additional painful comorbidities to their primary injury at the time of enrollment, and approximately 58% (n = 59) were prescribed a nonnarcotic, narcotic, or a combination of these medications (Table 1).

Table 1.

Characteristics of the study population (N = 102)

Descriptive Analyses

In our sample of 102 athletes, the average level of physical function (mean = 49) was slightly below the average score of the general US population. Ten patients (10%) scored 1 or more SDs below and eight (7.8%) 1 or more SDs above the general US population. The mean patient-reported pain intensity was distinctively lower than in the general US population (mean = 47) with 19 (19%) patients scoring 1 or more SDs below and three (2.9%) above the general US population (Table 2).

Table 2.

Descriptive analysis of measurements

AFAQ scores ranged from 12 to 46 with a mean of 26 (Table 2). Compared with a previously reported mean by Dover and Amar [20], the AFAQ mean in our sample was higher than in an uninjured/injured mixed athlete population (mean difference = 1.8, p = 0.023). Sixty-eight patients (67%) scored a low, 26 (25%) a moderate, and eight (7.8%) a high level of catastrophizing about pain. A clinically meaningful increased level of emotional distress (HADS ≥ 15) was measured in 17 of 102 patients (17%).

Results

Fear Avoidance and Decreased Physical Function

After controlling for potential confounding variables like age, injury region, catastrophic thinking, and emotional distress, an increase in athletes' fear avoidance was associated with a decrease in physical function (b = -0.32; p = 0.002).

Older age was associated with lower physical function (rs = -0.20; p = 0.044), and patients with an injury of the upper extremity reported higher physical function than patients with a lower extremity injury (mean difference = 6.5; p < 0.001). High fear avoidance was associated with lower physical function. As fear avoidance decreased, physical function increased; this association was strong (r = -0.39; p < 0.01; Table 3) and the effect size seemed to further increase when clinically relevant emotional distress (HADS ≥ 15) was present (Fig. 1). Pain catastrophizing (rs = -0.26; p < 0.05) and emotional distress (rs = -0.33; p < 0.05) were also associated with lower physical function. After accounting for all potential confounders, athletes' fear avoidance and injury region remained the two variables associated with physical function (Table 4). With each unit increase in AFAQ, PROMIS physical function scores dropped by 0.32 points (p = 0.002) and having an upper rather than a lower extremity injury was associated with a 6.2-point increase in PROMIS physical function (p < 0.001). The entire model accounted for 30% in the variation of physical function. Of these, athletes' fear avoidance uniquely explained 7.3% in physical function.

Table 3.

Bivariate correlation coefficients between measurement variables

Table 4.

Hierarchical regression analysis for variables predicting physical function

Fear Avoidance and Pain Intensity

Only having had surgery for the current condition (b = -5.3; p = 0.010) and low pain catastrophizing (b = 0.30; p = 0.001) were associated with lower pain intensity after controlling for a potential confounder; however, fear avoidance was not associated with pain intensity (b = -0.14; p = 0.249).

Followup patients reported lower pain levels than first visit patients (mean difference = -4.5; p = 0.002) and patients who have had surgery for their chief complaint reported lower pain levels (mean difference = -7.6; p < 0.001). Patients receiving narcotics or a combination of narcotic and nonnarcotic pain medication reported lower pain levels than their nonmedicated counterparts (mean difference = -3.9; p = 0.019). Patients with additional pain conditions reported higher pain levels (mean difference = 3.4; p = 0.026), and patients with prior sport-related injury (mean difference = 6.0; p = 0.002) or prior sport-related surgery (mean difference = 5.0; p = 0.001) also reported higher pain levels. An increase in pain catastrophizing (rs = 0.44; p < 0.001), emotional distress (rs = 0.29; p = 0.004), and athlete fear avoidance (r = 0.27; p = 0.006) was also found to be associated with higher pain (Table 3). However, the association of athlete fear avoidance with pain was weak (Fig. 1).

After controlling for potential confounders such as initial appointment/followup, surgery for the current condition, multiple pain conditions, history of prior sport-related injury/surgery, pain medication prescription, catastrophic thinking, and emotional distress, we found that athletes' fear avoidance was not associated with level of pain (b = -0.14; p = 0.249) (Table 5). Prior surgery for the current condition (b = -5.3; p = 0.010) and pain catastrophizing (b = 0.30; p = 0.001) were associated with higher pain. The whole model explained 40% of the variation of pain intensity with surgery uniquely accounting for 4.5% pain catastrophizing for 7.1% of the total variance in pain.

Table 5.

Hierarchical regression analysis for variables predicting pain intensity

Discussion

In injured athletes, psychosocial factors, eg, emotional distress, anxiety, catastrophic thinking, and fear, are frequently associated with prolonged recovery and lower return to sport rates. The AFAQ was developed to identify athletes' specific fearful cognition and avoidance behavior as negative responses to an injury. We conducted this study to assess whether athletes' fear avoidance is associated with reduced physical function and higher pain during the recovery process.

Results of this study need to be interpreted in light of several limitations. First, data were collected in a single urban academic hospital in the United States, which may limit generalizability. Second, the cross-sectional design of the study precludes us from drawing absolute conclusions about the causal directionality of our findings. Thus, we cannot conclude whether athletes' fear avoidance was the reason or the result of reduced physical function. Third, physical function, as assessed by patient-reported outcomes, represents a patient’s perception of what they can or cannot do (eg, subjective rating) rather than actual impairment (eg, objective rating). This subjective report is influenced by level of impairment and psychosocial factors, in this case, athletes’ fear avoidance. In this study, demographic and psychosocial variables were associated with approximately 30% variance in physical function, suggesting that other factors (psychologic, physical, etc) remain unexplained within the current model. Fourth, although the AFAQ [20] and the PCS [51] have been specifically validated for athletes, others have not. However, good applicability of PROMIS and HADS in an athletic patient population has already been reported [11, 35]. Fifth, because the definition of an athlete varies, we set a list of external (eg, regular participation in competition) as well as internal factors (self-identification as an athlete) as inclusion criteria. Because the AFAQ includes team-relevant questions (eg, “I am worried about my role with the team changing”), we excluded athletes who do not practice or compete in a team. This, however, further limits the generalizability of our results to patients who fit this definition of an “athlete.” Sixth, our sample size limited our ability to test whether fear avoidance, physical function, or pain varies with different injuries or surgery types. Larger studies and subgroup analysis would be required to address injury-specific investigations. Seventh, our study population was predominantly male and white with upper extremity injuries. This may have affected power to find a significant difference in lower extremity injuries or in other populations. Future longitudinal studies with larger samples are warranted to bypass these limitations.

We found that demographic, clinical, and psychosocial factors explain a large proportion of variance in physical function. Our findings add to the current understanding by showing that athlete-specific fear avoidance is the most important factor, explaining 7.3% of the unique variation in physical function. This reasonably large single-factor association occurred negatively proportional: the higher athletes scored in fear avoidance, the lower their physical function. This is of particular importance, because factors previously reported as associated with physical function (older age [1, 21], higher emotional distress, and higher catastrophic thinking about pain [37, 45, 49]) were outreached by athletes' fear avoidance in our sample. The central role of fear avoidance in reports of physical function is consistent with prior research that assessed general fear avoidance or aspects of fear avoidance in athletes [16, 43, 53, 57]. Longitudinal studies showed that fear of reinjury predicts return to a previous level of sport activities [53], whereas decreased fear of movement/reinjury contributes to improved knee function [16]. Our results are also in alignment with previous studies that investigated the effect of fear-related thoughts on functional outcomes with different questionnaires. In athletes with anterior cruciate ligament (ACL) reconstruction, concerns about painful sensations resulting from physical activity--assessed by the physical activity subscale of the FABQ, modified for the knee--contributed to explaining residual physical impairment [43]. Similarly, fear of movement/reinjury--measured by The Tampa Scale of Kinesiophobia--was found to be a major contributor to not returning to preinjury sport activity levels [30, 53]. In a retrospective interview, 53% of high school players and 50% of collegiate players who did not return to play after ACL reconstruction identified fear as the major contributing factor to their decision [36]. A systematic review revealed that fear of reinjury was the most common reason cited for a postoperative reduction in sports participation [4].

Catastrophic thinking was independently associated with higher levels of pain intensity, which is consistent with prior research in athletes [51, 55]. Substantial emotional distress (depression and anxiety) [37, 54] and fear [15, 16] in athletes after injury or injury-related surgery have been described; however, our findings suggest that athletes' fear avoidance, emotional distress, and even the prescription of narcotic pain medication did not show such a strong relationship to pain perception as catastrophic thinking.

Both athletes' fear avoidance and catastrophic thinking play a pivotal role in injured athletes’ recovery. We found that athletes’ fear avoidance is the most important factor associated with decreased physical function and catastrophic thinking about pain the most important factor associated with increased pain intensity in injured athletes. These results have several important implications for clinical practice. First, both fear avoidance and catastrophic thinking are malleable and can be changed through skills-based interventions such as cognitive restructuring, exposure to response prevention, and mindfulness meditation [29]. These strategies have strong research support in improving emotional and physical functioning in the general population and athletes [7, 8, 24]. Second, assessing fear avoidance and catastrophic thinking can provide an opportunity for surgeons to detect potential psychosocial barriers to successful rehabilitation. By challenging distorted fear- and pain-related thoughts, we may be able to optimize the recovery of injured athletes and foster their rates to preinjury sport levels. Third, understanding that reports of decreased physical function and increased pain intensity are associated with ineffective coping styles in injured athletes and that prescribing pain medications is not independently associated with higher physical function or decreased pain provides an opportunity for more nonpharmacologic treatment including assessing and addressing modifiable factors depicted. This finding may be particularly relevant as we observe a mounting epidemic of drug prescriptions with devastating consequences [13, 39, 44].