Self-Reported Mental Health Measures of Incoming Collegiate Student-Athletes With a History of COVID-19

As part of a larger concussion research project, we enrolled 178 student-athletes of 324 participants (54.9%) who self-selected into this component by completing the COVID-19 question and all associated mental health measures. Included were incoming freshmen or transfer varsity student-athletes or cheerleaders. To achieve ecological validity of student-athletes, we used no exclusion criteria. All participants provided written and oral informed consent, and the study was approved by the University of Delaware Institutional Review Board. A total of 79 athletes with a history of COVID-19 (COVID+; n = 79, 44.4%; age = 18.90 ± 0.16 years) and 99 athletes without a history of COVID-19 (COVID−; n = 99, 55.6%; age = 18.95 ± 0.16 years) were included in our analyses. We observed no differences between groups for a previous diagnosis of psychiatric disorder, anxiety, or depression (P > .05). Complete demographic and health history information are presented in Table 1.

Participants completed 3 commonly used mental health screening questionnaires.

The Satisfaction With Life Scale (SWLS) measures subjective well-being (college student sample: Cronbach α = 0.82).¹⁴  It consists of 5 questions scored on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Scores range from 5 (extremely dissatisfied) to 35 (highly satisfied), with a score <20 considered dissatisfaction with one’s life.¹⁴ 

The Hospital Anxiety and Depression Scale (HADS) is a valid and reliable instrument (adult sample: Cronbach α range = .68–.93) for detecting states of anxiety and depression with subscales measuring the severity of emotional distress and has been used in collegiate student-athletes.¹⁵  The survey consists of 7 questions for anxiety and 7 for depression that are scored separately. Subscores for anxiety and depression are categorized as normal (0 to 7), borderline abnormal (8 to 10), or abnormal (11 to 21).

The State-Trait Anxiety Inventory (STAI) consists of 40 questions separated into 2 forms (state and trait) scored on a 4-point Likert scale, with responses ranging from 1 (not at all) to 4 (very much so) for state anxiety and 1 (almost never) to 4 (almost always) for trait anxiety (range = 20–80). As suggested by previous researchers,¹⁶  we used a cutoff of 40 to evaluate the presence or absence of state and trait anxiety. The STAI has demonstrated good internal consistency (Cronbach α = .90) and test-retest reliability (r = 0.70, P < .001). We analyzed both state and trait anxiety total scores.

All surveys were completed via convenience sampling during the athletes’ preseason concussion baseline testing session. Testing occurred individually in a dedicated laboratory space under quiet conditions, and participants were informed that their results would not be shared with the coaching staff or teammates. The COVID-19 history question was part of a larger set of questionnaires on the overall medical history approved by our institutional review board for a separate branch of the study from the required baseline testing. Therefore, completion of these additional assessments was optional, and athletes were informed that they were not required to take part in the study. The total time to complete the entire battery (Qualtrics) was approximately 30 minutes.

Descriptive statistics were calculated to describe the sample demographic and outcome measures. To test our hypotheses, we compared the groups (COVID+ and COVID−) in a series of unadjusted models (model 1) across all patient-reported outcomes (SWSL, HADS anxiety, HADS depression, and state and trait anxiety) using 1-way analyses of variance. Effect sizes were calculated using the Cohen d and interpreted as small (d = 0.20), moderate (d = 0.50), or large (d = 0.80).¹⁷  To control for potential contributors to model 1, we assessed the groups in model 2 using a 1-way analysis of covariance adjusted for the covariates of sex and race and ethnicity. Both sex and race and ethnicity were independently associated with mental health disorders^18,19 ; thus, their inclusion in our model was essential. Both models are presented because model 1 is more clinically meaningful for athletic trainers. Furthermore, a 2-way analysis of covariance was computed to determine the interaction of COVID-19 group and sex in patient-reported mental health outcomes. Subsequent post hoc Bonferroni testing was performed to identify specific differences between groups. Finally, we used χ² analyses to compare the number of athletes who met the standard clinical cut points for both groups. All analyses were conducted using JMP (version 16.0; JMP Statistical Discovery LLC), and the α level was set a priori at .05.

Means, SDs, and 95% CIs for each mental health patient-reported outcome (ie, SWLS, HADS anxiety, HADS depression, STAI) are provided in Table 2. In the unadjusted model, we observed a difference between groups for HADS depression (F_1,176 = 2.83, P = .047), whereby athletes in the COVID+ group had higher depression ratings (2.86 ± 0.26) than those in the COVID− group (2.04 ± 0.25), with a large effect size (Cohen d = 3.21). No other models were different (Table 2). Individual responses are shown in the Figure.

After including sex and race and ethnicity as covariates, we found a group difference for the SWLS (F_1,176 = 2.27, P = .02, Cohen d = 0.47), HADS anxiety (F_1,173 = 5.88, P = .003, Cohen d = 5.26), and STAI state anxiety (F_1,171 = 2.36, P = .01, Cohen d = 0.21) such that satisfaction with life, HADS anxiety, and state anxiety scores were higher for the COVID+ group. No group differences were found for the HADS depression or STAI trait anxiety scores in model 2.

Model 3 yielded group interaction effects for the HADS anxiety (F_1,173 = 13.76, P = .003) and STAI trait anxiety ratings (F_1,171 = 6.74, P = .03; Table 3). Post hoc testing revealed that female athletes in the COVID+ group had worse HADS anxiety (P = .01) and STAI trait anxiety (P = .002) ratings than all other athletes.

We observed no differences between groups for the percentage of respondents who were classified as abnormal for any of the outcome measures (Table 4).

Given the possible deleterious effect of a positive COVID-19 diagnosis on mental health, we described and compared patient-reported measures among incoming intercollegiate athletes who did and those who did not report having had a COVID-19 diagnosis. We addressed this question using both unadjusted, clinically relevant models and adjusted models. Overall, athletes in the COVID+ group reported higher depression ratings than the athletes in the COVID− group. In addition, sex was a covariate linked to decreased life satisfaction, heightened depression, and increased anxiety in female compared with male student-athletes. Our findings suggest that the COVID-19 pandemic may result in persistent adverse effects on the mental well-being of otherwise healthy young adults. Thus, prioritizing athletes’ mental health support is crucial for schools. High-resource schools likely have access to a team of sports psychologists working closely with the athletic department, whereas schools with limited resources may need to forge relationships with student health services to ensure that athletes receive the necessary mental health care.

Anxiety and depression frequently occur together but reflect different sequelae. Depression is often linked to past events in which something or someone valued is lost and perceived as challenging or impossible to regain.¹²  Furthermore, major life events (eg, a pandemic) and other personal problems such as social isolation are believed to increase depressed feelings.²⁰  Depression was higher in the COVID+ group, which may be a psychological manifestation of not only the isolation and time lost from participation associated with the pandemic but also the systemic inflammation caused by COVID-19.¹³  Conversely, anxiety reflects uncertainty about a potential future crisis.¹⁸  Interestingly, anxiety scores did not differ between the groups. By the summer of 2021, vaccines for COVID-19 had become available, which could have contributed to a more positive outlook moving forward. Student-athletes appeared to remain optimistic about their future, as demonstrated by the low anxiety ratings on the HADS.

Although we demonstrated higher depression scores with a large effect size (Cohen d = 3.21) for the COVID+ group, most participants (94.5%) were still considered normal and nearly 4 points below the mild depression classification criterion.²¹  Therefore, these statistical differences likely have limited clinical meaningfulness; however, athletic trainers should be aware that these individuals could have an increased risk of clinical depression if another event (eg, concussion or anterior cruciate ligament injury) occurs. Nearly all participants in both groups (COVID+ = 96.4%, COVID− = 95.7%) were within the normal anxiety range. As the prevalence of mental health concerns continues to increase globally,²²  the authors of a large study focused on physical activity as a protective mediator of anxiety and depression.²³  Involvement in sports mediates the presentation of psychological disorders and may serve as an effective treatment approach.²⁴  In addition, organized sports engagement (eg, being on a team, including in individual sports such as tennis) was associated with a decreased risk of developing anxiety and depression compared with physical activity alone.¹⁰  Hence, continuing to be physically active and being on a team may explain why our sample, regardless of COVID-19 history, had scores below the clinical thresholds for anxiety and depression. However, because we do not have data specific to summer 2020 or earlier, we do not know if mental health declined when sport participation was temporarily cancelled.

Anxiety is further subdivided into state and trait anxiety, with state anxiety defined as a more transient reaction to an adverse situation and trait anxiety defined as a more stable personality attribute that affects long-term thought processes.²⁵  More than 83% of student-athletes in both groups had state or trait anxiety scores below the clinical cutoffs. This result aligns closely with prepandemic anxiety levels, when 30% of female and 25% of male student-athletes self-reported anxiety.¹⁸  This lack of group differences indicates that COVID-19 may not have resulted in lasting, meaningful changes in either state or trait anxiety. Still, we did observe differences in HADS and trait anxiety scores between male and female student-athletes such that female student-athletes had higher (worse) levels. This finding is similar to most of the literature, with researchers¹⁸  consistently documenting more anxiety in women than men. Therefore, critical demographic characteristics such as sex may affect the mental health response of student-athletes after COVID-19 recovery and should be considered by the evaluating sports medicine professionals.

Satisfaction with life refers to a cognitive judgmental process in which individuals assess their quality of life according to criteria they deem important.¹⁴  It requires people to compare their current circumstances with what they believe is the standard. The COVID-19 pandemic likely influenced this factor across the general population. Our incoming student-athletes had higher SWLS scores (approximately 28.9) than previously reported for typical college students (approximately 23.5¹⁴ ; Tables 2 and 3); however, these scores were lower than the SWLS scores of athletes before COVID-19 (range = 29.2–29.5) in a Concussion Assessment, Research and Education Consortium study.²⁶  We did observe a large effect size that was not different (P = .73; Cohen d = 1.05) but no clinically meaningful difference (difference between scores <1) between groups. Starting their college experience and joining a university sport team is an exciting period for student-athletes. The higher-than-average SWLS scores for college students in our study may reflect this optimism for the future, yet these scores were still lower than those before COVID-19. Involvement in sports and greater team cohesion have been associated with better SWLS scores in athletes.²⁷  Furthermore, postconcussion musculoskeletal injury rates in intercollegiate student-athletes were lower in those with higher SWLS scores.²⁸  Future authors should examine how SWLS evolves over an individual’s collegiate athletics career.

Our work had limitations. All mental health conditions were self-reported by the participants, which is an inherent limitation of questionnaire-based research. We could not account for illness severity (eg, symptom magnitude and days to recover) in athletes who had COVID-19. Furthermore, our investigation was limited to a single mid-Atlantic university, and surveys were administered only to new, incoming student-athletes; however, this represents a large ecologically valid patient sample. We reported low power (range = 0.06–0.39) for our unadjusted model, but the sample size of 178 participants was sufficient. Finally, future researchers should evaluate both pathophysiological and psychosocial contributions to post–COVID-19 mental health challenges in student-athletes and, more importantly, treatment strategies focused on the specific demands of this unique population.

Incoming collegiate student-athletes who reported a previous COVID-19 diagnosis displayed higher depression scores, suggesting that clinicians may need to provide appropriate identification and referral for mental health concerns. Nonetheless, it was encouraging that most participants had mental health scores that did not exceed established threshold values. Thus, although the pandemic resulted in a global emotional upheaval, this change may be transient in healthy young adults. Female athletes may have a higher risk for mental health challenges, and athletic trainers should continue to incorporate mental health specialists in the treatment of student-athletes.