Comparing Postconcussion Sway Balance and Cognitive Test Performance With Preinjury Baseline Identifies a Higher Percentage of Impaired Concussed Athletes Than Comparison With Normative Values

Abstract

Background:

Baseline and postinjury neurocognitive and balance testing assist in the assessment and management of sports-related concussion. We evaluated the utility of comparing postinjury performance on the Sway Medical System Sports+ test battery with individualized baseline data and normative reference data.

Hypothesis:

Comparisons with baseline data would identify a higher proportion of concussed athletes with impaired scores, as opposed to comparing postinjury performance with normative data.

Study Design:

Cohort study (Diagnosis); Level of evidence, 3.

Methods:

Baseline assessments were conducted 1 year apart using the Sway Sports+ battery. Postinjury assessments were conducted within 3 days of concussion and compared with both baseline and normative data using thresholds for 70%, 80%, 90% and 95% CIs. Outcomes were evaluated based on the number of modules within the test battery that surpassed confidence interval thresholds.

Results:

Both comparison approaches identified a high proportion of concussed athletes as impaired when considering impairment on ≥1 measures (eg, 95% using baseline data, 89% using normative data [70% CI]; 89% baseline, 80% normative [80% CI]; 82% baseline, 70% normative [90% CI]). However, impairment on a single module was common and occurred in both the concussed and the nonconcussed samples. When impairment on ≥2 modules was considered, the baseline comparison method correctly identified a higher proportion of concussed athletes as impaired than the normative comparison method (eg, 82% vs 68% [70% CI]; 67% vs 58% [80% CI]; 52% vs 47% [90% CI]). Impairment on ≥2 modules was uncommon in the control sample but occurred frequently in the concussed sample.

Conclusion:

Baseline comparisons consistently classified a greater percentage of athletes as concussed than normative comparisons, across all confidence intervals.

Keywords

concussion balance neuropsychological testing baseline evaluation

Sport-related concussion is associated with a varied acute symptom presentation (eg, physical, cognitive, and emotional), along with a decline in cognitive functioning and a disturbance in static balance and postural stability.^10,27,36,52 Consensus experts recognize that postconcussion symptoms may evolve over the first few hours/days following injury and often resolve within days.⁵⁸ However, postconcussive symptoms may persist for weeks or months,⁵⁸ and a significant proportion of children and adolescents experience concussion-related symptoms beyond 1 month.²⁰ Sideline screening measures (eg, the Sport Concussion Assessment Tool [SCAT] and the Sideline Assessment of Concussion [SAC]) were introduced as tools for the on-field assessment of “red flags” and observable signs of symptoms and cognitive and balance impairments following a concussion,⁶³ including “Maddocks” questions to assess orientation.⁵⁰ With the introduction of the “Sports as a Laboratory Assessment Model” in the 1980s,⁷ pencil-and-paper neuropsychological tests started being used to identify cognitive deficits in concussed athletes.^48,51 Within this paradigm, sports medicine professionals utilize preseason baseline performance as a comparator for an athlete's postconcussion test performance. Multimodal screening measures are indicated for the assessment of postconcussion symptoms, balance, and cognition, especially within the acute (eg, within the first 72 hours) phase following injury.³¹ More comprehensive neuropsychological assessments may be required for individuals with postconcussive symptoms that extend to the subacute phase (eg, days to weeks postinjury).⁵⁸ Numerous factors have been associated with prolonged recovery following concussion, including continuing to play while symptomatic,³⁰ female sex, concussion history,⁴⁴ and preexisting conditions such as history of migraines,⁷⁴ anxiety, and depression.⁶ Subsequently, a variety of neuropsychological tests have been implemented to assess and manage concussions in collegiate^48,51 and professional^49,71 athletes, and these tests have moved from paper-based to computer-based measures over the past few decades.^60,61 Concussion surveillance and management programs, using the above-mentioned neuropsychological test instruments, have been implemented in nearly all professional contact and collision sport leagues, including the National Hockey League (NHL), National Football League (NFL), Major League Soccer, NASCAR, Major League Baseball, Boxing, Australian Rules Football, Equestrian Sports, and Rugby.²⁹ These programs utilize tests and screening measures developed to assess an athlete's symptoms, cognition, and balance, both on the day of injury and during the first few days following injury.⁷⁸ Widely used measures include the SCAT,^25,27 the Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT),^1,2 ImPACT Pediatric,⁴¹ ImPACT Quick Test,⁷⁷ CNS Vital Signs,^62,66 and the Sway Medical System.⁷³ While these tools are widely used in youth, high school, collegiate, and professional sports leagues/organizations, ongoing research investigating the reliability, validity, and clinical utility of these measures⁸ is crucial, as the accurate evaluation of cognitive and balance performance is important for the diagnosis and management of concussion as well as reduction of the risk of prolonged symptoms and/or recovery.

The National Collegiate Athletic Association (NCAA) mandates that every athlete receive a preparticipation baseline concussion assessment, which includes a symptom evaluation, cognitive assessment, and balance evaluation.⁵⁵ The NFL requires that every player complete baseline neurocognitive testing, which includes computerized neurocognitive tests, traditional paper-and-pencil tests, or a combination of the 2 (ie, hybrid testing).⁵⁶ As well, the NHL requires its players to complete preseason neurocognitive testing, including the SCAT, ImPACT, and a custom pencil-and-paper test battery.⁵⁷ Annual/updated baseline assessments are recommended for youth athletes, given ongoing developmental changes throughout adolescence,³⁷ as well as college-age athletes,⁵⁵ to ensure the most accurate data are available for postinjury comparisons. However, while professional sports leagues have the resources to implement widespread baseline testing, many collegiate, high school, and youth sports leagues lack the same resources and personnel.¹⁸ Given that limited resources and personnel in smaller or lower socioeconomic settings may restrict access to, feasibility, and consistent implementation of these assessments,^16,17 establishing the utility and necessity of resource-heavy baseline testing programs is necessary. Despite debate over the utility of baseline neurocognitive testing, consensus experts recognize that neurocognitive testing adds value to the assessment of sports-related concussion and its sequelae.⁵⁸

Alongside challenges related to implementation in less-resourced settings, researchers have also debated the utility of annual baseline assessments. Researchers have compared postconcussion performance with normative reference data as a valid alternative when baseline test data are unavailable.²⁶ Others have shown that the use of normative reference data is likely to misclassify a significant percentage of concussed athletes compared with baseline test performance.⁶⁸ While some researchers have questioned the utility of baseline testing in general,⁶⁴ others have identified strategies to improve clinical utility, such as the use of aggregate baselines.^11,12

Despite debate over the utility of baseline neurocognitive testing, consensus experts recognize that neurocognitive testing adds value to the assessment of sports-related concussion and its sequelae.⁵⁸ In the past decade, only 2 of the aforementioned assessment measures, ImPACT (2016) and Sway Medical System (2025), have received US Food and Drug Administration (FDA) clearance as class II medical devices as a computerized cognitive assessment aid for concussion. Given that the preponderance of research has focused on the utility of ImPACT, and Sway has only recently received FDA clearance, there is a need for research documenting the psychometric properties and clinical utility of the Sway product. As such, the purpose of this study was to compare athletes’ postinjury performance on the Sway Medical System using both baseline-based reliable change methods and normative reference data.

Methods

Participants

Institutional review board approval was obtained for retrospective analysis of deidentified data. Participant data were extracted from 2 separate databases:

Baseline (normative) sample (N = 10,920): preseason baseline data were extracted from a larger, deidentified database of collegiate student-athletes who completed 2 annual baseline assessments over the span of 1 year. Specific sports are not identified in this database. This sample composed the normative test-retest sample. Data were extracted from the database using inclusion criteria requiring that participants

a. were college athletes between the ages of 18 and 22 at the time of testing;

b. spoke English as their primary language and completed both assessments in English;

c. completed annual baseline assessments in the years 2023 and 2024, within the last 2 weeks of August through the first 2 weeks of September (to restrict data/participants to only those completing annual preseason evaluations for fall sports);

d. completed symptom surveys as part of the baseline testing; and

e. were assessed within an organization that agreed to allow their deidentified data to be used for research purposes.

2. Postconcussion sample (n = 231): postconcussion data were exported from a larger, deidentified database of athletes who sustained a concussion and subsequently completed a Sway assessment between August 2021 and May 2025. Specific sports are not identified in this database. Data were extracted from the database using inclusion criteria requiring that athletes

a. were college athletes between the ages of 18 and 22 at the time of testing;

b. spoke English as their primary language and completed both assessments in English;

c. sustained a clinically diagnosed concussion, as documented by a qualified examiner;

d. completed a postconcussion assessment within the “acute phase” of 3 days since sustaining a concussion;

e. were symptomatic at the time of assessment, as demonstrated by a change of ≥10 on Post-Concussion Symptom Scale scores from baseline³⁹; and

f. had completed a baseline assessment within the previous calendar year.

As all participants in this study were athletes from NCAA programs within the United States, the clinical reference standard for diagnosing concussion was that of the NCAA Concussion Safety Protocol guidelines,⁵⁵ which stipulate that a “qualified medical professional” must conduct the evaluation and clinical diagnosis using Concussion in Sport Group criteria.⁵⁸

Testing

All participants completed the Sway Medical System Sports+ test battery (heretofore referred to as “Sway Sports+”) using a smartphone or tablet; previous research has found no difference in test performance across device types.⁷⁶ The Sway Sports+ battery includes the 22-item SCAT symptom scale,²⁸ a balance test, and 4 cognitive function tests.^5,54 The balance test is a modified Balance Error Scoring System (mBESS) that includes 5 stances: double-leg stance (feet together), single-leg stance (right or left), and tandem stance (right or left foot in front). The cognitive tests include the following modules: reaction time (a measure of simple visual motor reaction time), impulse control (a measure requiring choice reaction time within a go/no-go test), inspection time (a measure requiring simple visual inspection speed), and memory (a measure of visual working memory and delayed recall). More detailed descriptions of these tests are available in the literature.^15,72,75 Scores from the reaction time, impulse control, and inspection time modules are presented in milliseconds, where higher scores indicate poorer performance. Scores from the balance and memory modules are summarized with an integer-level score, where higher scores indicate better performance.

The psychometric properties of the Sway Sports+ test battery have been reported previously. The test-retest reliability of the mBESS Balance module ranges from 0.66 to 0.88,^{3,15,24,54,59,75} with moderate-to-high correlations with force plate and traditional balance measures.^13,21,43 This balance module also shows good concurrent validity with the BESS.^4,59 The cognitive tests show test-retest reliability coefficients ranging from 0.60 to 0.83.^14,75 The cognitive tests demonstrate strong concurrent validity with the ImPACT and ImPACT Quick Test^70,76 and moderate to moderately strong correlations with established neuropsychological test measures, including the Wechsler Adult Memory Scale, the Delis-Kaplan Executive Function System, and the California Verbal Learning Test–3rd edition.⁷⁰

Statistical Analysis

Analyses were conducted in 4 steps.

Normative Reference and Test-Retest Reliability. Using the baseline sample, descriptive statistics were calculated for each test battery module. Two-way analyses of variance (ANOVAs) were conducted using age (independent groups; ages 18, 19, 20, 21, and 22) and sex (female, male) as independent variables and each of the 5 modules as dependent variables. Test-retest reliability was assessed using intraclass correlation coefficients (ICCs) using an average-rating, absolute-agreement, 2-way mixed-effects model.^33,45,53 ICC values represent the proportion of variance in test scores attributable to true-score variance. ICC values <0.50 generally reflect poor reliability, values between 0.50 and 0.75 reflect moderate reliability, values between 0.75 and 0.90 reflect good reliability, and values >0.90 indicate excellent reliability.⁴⁵ ANOVA and reliability analyses were conducted using Statistical Package for the Social Sciences (Version 31; IBM Corp).

2. Reliable Change Estimates. Reliable change methodology^38,42 was used to estimate measurement error (standard error of measurement [SEM]) in test-retest difference scores using the baseline sample. One-tailed 70%, 80%, 90%, and 95% CIs¹⁹ were applied to determine cutoff values for reliable change. The formulas for calculating reliable change are provided below and were calculated using Microsoft Excel (Table 1). Of note, “time 1” refers to the first baseline assessment and “time 2” refers to the second baseline assessment.

a. SEM₁ = $SD \sqrt{1 - r_{12}}$ = SD from time 1 multiplied by the square root of 1 minus the test-retest coefficient (ICC)

b. SEM₂ = $SD \sqrt{1 - r_{12}}$ = SD from time 2 multiplied by the square root of 1 minus the test-retest coefficient (ICC)

c. SE_diff = $\sqrt{{SEM}_{1}^{2} + {SEM}_{2}^{2}}$ = square root of the sum of the squared SEMs for each testing occasion

d. Reliable change confidence intervals = SE_diff is multiplied by the following z scores: ±0.52 (70% CI), ±0.84 (80% CI), ±1.28 (90% CI) and ±1.64 (95% CI)

3. Multivariate Base Rates. Within the baseline sample, the percentage of athletes who obtained 0, 1, or ≥2 scores exceeding the reliable change cutoff values across modules was calculated to establish multivariate base rates of change over a 1-year interval.

4. Postconcussion Comparisons. Postconcussion performance was evaluated using 2 approaches.

a. Baseline comparison: postinjury scores were compared with each athlete's own baseline using reliable change values.

b. Normative comparison: postinjury scores were converted to Z-scores using sex-specific means and standard deviations from the baseline normative sample using the following formula: Z = ([postinjury score–baseline normative mean]/baseline normative SD).

Table 1

Descriptive Statistics and Test-Retest Correlations for Normative Sample (N = 10,920) ^a

	Time 1			Time 2							Effect Size
	Mean	SD	SEM₁	Mean	SD	SEM₂	ICC	95% CI	Mean_diff	SE_diff	d	95% CI
Balance	86.81	10.45	5.82	86.45	11.11	6.19	0.69	0.52 to 0.69	0.36	8.49	0.64	0.50 to 0.78
Reaction time	230.07	34.87	21.50	351.09	36.52	22.51	0.62	0.61 to 0.63	2.92	31.13	–0.61	–0.75 to −0.47
Inspection time	46.45	26.15	16.12	43.98	26.47	16.32	0.62	0.61 to 0.64	–2.47	22.94	–0.50	–0.63 to −0.35
Impulse control	350.55	62.02	39.22	351.09	79.24	50.12	0.60	0.58 to 0.61	0.54	36.64	–0.50	–0.62 to −0.35
Memory	72.95	6.79	5.26	73.05	7.55	5.85	0.43	0.41 to 0.45	–0.10	7.87	0.24	0.11 to 0.37

ICC, intraclass correlation coefficient; Mean_diff, mean difference score; SE_diff, standard error of difference; SEM, standard error of measurement.

Results

Baseline Normative/Test-Retest Sample

The baseline sample was composed of 10,920 student-athletes, ages 18 to 22 years (mean ± SD, 18.75 ± 1.1 years) with 6315 male athletes (57.8%) and 4605 female athletes (42.2%). Participants completed 2 baseline assessments across a 1-year interval (mean ± SD, 365.0 ± 8.8 days; range, 335-395 days).

There were no significant differences between age groups (P > .05 for all test battery modules), but there were significant differences by sex on all 5 measures (all P < .001). Therefore, sex-specific normative values were used for subsequent analyses. ICC values indicated moderate test-retest reliability across measures (eg, 0.60 to 0.69), except for memory (0.43) (Table 1), which demonstrated poor reliability.

Cutoffs for reliable change for 70%, 80%, 90% and 95% CIs reflect the threshold to identify cases that fall outside the respective confidence intervals, and scores surpassing these values are considered clinically significant (Table 2).

Table 2

Reliable Change Confidence Intervals Across a 1-Year Interval ^a

	70% CI	80% CI	90% CI	95% CI
Balance	4.42	7.13	10.87	13.93
Reaction time	16.19	26.15	39.84	51.05
Inspection time	33.09	53.46	81.46	104.37
Impulse control	11.93	19.27	29.36	37.62
Memory	4.09	6.61	10.07	12.90

Confidence interval for the mean test-retest difference score for each measure (mean ± SD test-retest interval: 365.0 ± 8.8 days; range, 335-395 days), without adjustment for test-retest practice effects. If a retest score changes by that amount or more (positive or negative) performance reflects more change than would be anticipated with normal variation in test performance and measurement error.

Multivariate Base Rates of Change (Baseline Sample)

Within the baseline sample, the univariate base rate matched expectations; that is, for individual battery modules, the percentage of athletes with a score within the reliable change cutoff approximated the respective confidence interval (see Table 3). When performance was interpreted within the context of a multivariate battery, it was common for athletes to show ≥1 score that exceeded the reliable change cutoffs, at all confidence intervals (70% CI, 76.3%; 80% CI, 58.4%; 90% CI, 47.7%; 95% CI, 25.1%). However, it was less common for athletes to obtain ≥2 scores that exceeded the reliable change cutoffs, again at all confidence intervals (70% CI, 39.2%; 80% CI, 21.3%; 90% CI, 8.8%; 95% CI, 4.4%) (Table 3).

Table 3

Percentage of Normative Sample Cases Falling Beyond Reliable Change Cutoffs ^a

	70% CI	80% CI	90% CI	95% CI
Balance	27.2	18.9	11.9	7.9
Reaction time	30.9	20.6	11.2	6.7
Impulse control	24.5	15.0	7.5	4.4
Inspection time	24.2	16.2	9.6	6.2
Memory	26.7	15.6	8.4	5.2
No reliable change	23.7	41.6	62.3	74.5
Change on ≥1	76.3	58.4	47.7	25.1
Change on ≥2	39.2	21.3	8.8	4.4

Percentage of cases exceeding cutoffs on retest (mean ± SD test-retest interval: 365.0 ± 8.8 days; range, 335-395 days). Percentages for individual measures should approximate the confidence interval (eg, 30% falling outside a 70% CI, 20% falling outside an 80% CI, etc). When considering multiple measures, the percentages on the “change on ≥2” line best approximate the respective confidence intervals.

Postconcussion Sample

The postconcussion sample comprised 231 collegiate athletes, aged 18 to 22 years (mean ± SD, 19.23 ± 1.19 years) who completed a postinjury Sway Sports+ assessment within 3 days of sustaining a concussion (injury-testing interval: 0.84 ± 0.89 days). The sample consisted of 126 female (54.5%) and 105 male (45.5%) patients.

Reliable Change Comparisons

A large proportion of postinjury baseline change scores in the concussed athlete sample fell outside of the reliable change cutoff values on ≥1 module of the test battery, at all confidence intervals (70% CI, 95.2%; 80% CI, 89.2%; 90% CI, 82.3%; 95% CI, 69.3%). It was less common but still frequent for postinjury-baseline change scores in the concussed athlete sample to fall outside the reliable change cutoff values for ≥2 modules of the test battery, at all confidence intervals (70% CI, 81.8%; 80% CI, 66.5%; 90% CI, 51.9%; 95% CI, 40.3%) (Table 4).

Table 4

Percentage of Concussed Cases Falling Beyond Reliable Change Cutoffs Compared With Baseline ^a

	70% CI	80% CI	90% CI	95% CI
Balance	59.3	47.2	36.8	27.3
Reaction time	69.7	61.5	53.2	45.0
Impulse control	49.8	39.0	30.7	25.5
Inspection time	53.7	472	36.8	29.4
Memory	41.1	28.6	22.1	15.2
No reliable change	4.8	10.8	17.7	30.7
Change on ≥1	95.2	89.2	82.3	69.3
Change on ≥2	81.8	66.5	51.9	40.3

Percentage of cases exceeding cutoffs following a concussion.

Normative Comparisons

When postconcussion scores were compared with sex-specific normative reference values, the proportion of athletes who exceeded the cutoff values was slightly lower than the reliable change comparisons described above. It was common for athletes to score outside Z-score normative cutoffs on ≥1 test module at all confidence intervals (70% CI, 89.2%; 80% CI, 79.7%; 90% CI, 69.7%; 95% CI, 64.1%) and less common for athletes to score outside Z-score normative cutoffs on ≥2 test modules at all confidence intervals (70% CI, 67.5%; 80% CI, 58.2%; 90% CI, 47.2%; 95% CI, 35.9%) (Table 5).

Table 5

Percentage of Concussed Cases Falling Beyond Reliable Change Cutoffs Compared With Normative Data ^a

	70% CI	80% CI	90% CI	95% CI
Balance	40.7	32.6	23.8	19.9
Reaction time	64.5	56.7	48.5	39.8
Impulse control	45.5	38.5	29.4	24.2
Inspection time	48.5	39.0	32.0	30.7
Memory	34.6	31.6	17.3	10.4
No reliable change	10.8	20.3	30.3	35.9
Change on ≥1	89.2	79.7	69.7	64.1
Change on ≥2	67.5	58.2	47.2	35.9

Percentage of cases exceeding Z-score cutoffs when comparing postconcussion performance with normative reference data.

Discussion

The purpose of this study was to compare 2 approaches for interpreting postconcussion performance on the Sway Sports+ test battery. A normative sample was utilized to derive cutoff values for both a reliable change method and a normative Z-score comparison method. The results indicate that while both approaches identify a high proportion of concussed athletes with impaired scores, the use of reliable change values (ie, postinjury-to-baseline score change method), consistently identified a greater percentage of concussed athletes as impaired than normative comparisons.

Reliable Change Versus Normative Comparisons

The interpretation of postinjury neurocognitive and balance scores depends on the methodology used. Reliable change methods assess whether an individual's postinjury score reflects a statistically significant difference from their own baseline score and accounts for individual variability in preinjury testing. In contrast, normative comparisons evaluate postinjury scores alone (out of context with baseline) in relation to a healthy reference sample. Normative comparison methods eliminate the need for baseline testing but sacrifice the potential for individualized change metrics.

In this study, comparisons between individual postinjury and baseline scores identified a greater proportion of concussed athletes as falling beyond reliable change cutoffs (ie, were more sensitive) than the normative Z-score method, a pattern observed across all tested confidence intervals. This pattern is consistent with previous research that has analyzed both methods in the same concussed sample across a variety of test measures.^{15,19,35,47,68} At the same time, the normative comparison method in this study did identify a high proportion of concussed athletes as falling beyond reliable change cutoffs, aligning with previous research that suggests postinjury testing alone is sufficient²⁶ (eg, when baseline test data are not available for comparison). Combined, these findings suggest that individualized baseline comparisons can identify individuals whom normative comparisons may miss, although normative comparisons are clinically useful in the absence of baseline scores.

Multivariate Interpretation

In the baseline sample, it was common for nonconcussed control athletes to demonstrate reliable change on 1 of the 5 modules of the Sports+ battery over a 1-year interval (eg, 58.4% at the 80% CI). However, while it was relatively uncommon for nonconcussed controls to demonstrate a reliable change in ≥2 of the 5 test modules (eg, 21.3% at the 80% CI), concussed athletes were ≥3 times as likely (66.5% at the 80% CI using the reliable change method) as the control athletes to show impairments using the ≥2 modules requirement. These patterns reinforce the commonly accepted sports medicine philosophy that encourages the use of multiple measures when assessing athletes suspected of having a concussion.^15,58,65 The current results support the use of multivariate, multimodal test batteries in this context. Classification based on multiple impaired modules reduces false positives and helps better differentiate concussion-related changes from normal variability.⁴⁰

Comparison With Other Concussion Assessment Tools

The classification patterns seen in the current study are comparable with or exceed those reported for other commonly used concussion assessment tools. However, the sensitivity and specificity values reported in the literature vary widely depending on the selected tool and the tested sample. Sensitivity values for the SCAT (across versions) range from 70% to 95%, and specificity values range from 60% to 100%.^22,28,32,63 Reported sensitivity values for ImPACT range from 52% to 95%, and specificity values range from 52% to 97%.^{9,19,46,67,69} However, prior research has demonstrated that the symptom component of these assessments is a primary contributor to the higher sensitivity and specificity values (eg, 65% sensitivity and 80% specificity in the absence of neurocognitive test data⁴⁶), and the removal of symptom score substantially reduces the classification performance of these tools.^23,34,65

Although symptom scales were completed at both baseline and postinjury time points in the current study, these data were intentionally excluded from all analyses. The percentage of concussed athletes falling beyond reliable change cutoffs (eg, sensitivity) ranged from 69% to 95%, when compared with baseline levels of performance, and from 64% to 89% when compared with normative data (depending on the confidence interval used). In addition, the percentage of nonconcussed athletes falling beyond reliable change cutoffs on baseline retest performance (eg, specificity) ranged from 25% to 76% (again, depending on the confidence interval used). However, it is important to note that the values reported in this study are based solely on the balance and 4 cognitive measures of the Sway Sports+ battery. The classification performance reported here reflects objective test components and does not rely on self-reported symptoms. In addition, the ranges listed above reflect values for change on “≥1” Sway Sports+ score. Use of “≥2” scores decreases the percentage of concussed athletes falling beyond reliable change cutoffs and increases the percentage of nonconcussed athletes falling beyond reliable change cutoffs on baseline test-retest, as described above in the Multivariate Interpretation section.

Limitations and Terminology and Classification Considerations

This study is not without its limitations. While the samples are relatively large, data were exported retrospectively and lack the precise controls that often are present in a prospective study. Given that the only available information about the participants was captured within the Sway Sports+ user application, we did not have access to detailed information about how the person was injured, what sport the person participated in, or the severity of the injury. As such, it is possible that some of the individuals were injured in recreational activities or falls taking place outside of sports, but they were evaluated by their school athletic trainer. In this context, we did not have information about who made the clinical diagnosis of concussion, the person's professional background, training, or expertise, or whether the diagnosis was accurate or verified through any other source. In addition, we attempted to strengthen the validity of the study by requiring concussed individuals to demonstrate an increase in symptoms from baseline. However, given these limitations, the current results may only be generalizable to youth athletes ages 18 to 22 who sustained a sport-related concussion while participating in institutional athletics.

It is important to note that none of the tests examined and referenced in this study are diagnostic tools, as there is currently no definitive standalone diagnostic test for concussion. Although the terms sensitivity and specificity have been used historically in the concussion literature, they are used as classification criteria based on test outputs rather than a true diagnostic gold standard. The use of these terms in the current study is intended only to facilitate comparison across tools and methodologies.

Conclusion

This study provides evidence of the clinical utility of the Sway Sports+ test battery for assessing suspected concussions. The data indicate that baseline-referenced reliable change methods enhance detection of within-person declines in neurocognitive and balance performance, that normative comparisons remain clinically useful but may underidentify postconcussion impairments, and that multivariate assessments and interpretations are essential. Future research should examine longitudinal changes in scores postinjury using the Sway Sports+ test battery and evaluate clinical outcomes associated with both reliable change and normative comparison methodologies.

Footnotes

Final revision submitted June 4, 2026; accepted June 9, 2026.

One or more of the authors has declared the following potential conflict of interest or source of funding: P.S. is a scientific advisor to Healthy Roster. J.C. is the director of research for Healthy Roster.

Ethical approval for this study was obtained from the Saint Joseph's University Institutional Review Board for the Protection of Human Subjects in Research (protocol No. 2416747-1).

ORCID iD

Philip Schatz

References

Alsalaheen

Stockdale

Pechumer

Broglio

SP.

Measurement error in the Immediate Postconcussion Assessment and Cognitive Testing (ImPACT): systematic review. J Head Trauma Rehabil. 2016;31(4):242-251.

Alsalaheen

Stockdale

Pechumer

Broglio

SP.

Validity of the Immediate Post Concussion Assessment and Cognitive Testing (ImPACT). Sports Med. 2016;46(10):1487-1501.

Amick

Chaparro

Patterson

Jorgensen

Test-retest reliability of the SWAY Balance Mobile Application. J Mob Technol Med. 2015;4(2):40-47.

Amick

Jansen

Chaparro

Hakansson

Patterson

Jorgensen

MJ.

Comparison of the SWAY Balance Mobile Application to the Abbreviated Balance Error Scoring System. Athl Train Sports Health Care. 2015;7(3):89-96.

Anderson

Gatens

Glatts

Russo

SA.

Normative data set of SWAY Balance Mobile Assessment in pediatric athletes. Clin J Sport Med. 2019;29(5):413-420.

Anish

Thomas

Zynda

, et al. Effect of co-morbid anxiety and depression history on multidomain concussion clinical outcomes and recovery time. Neurol Sci. 2026;47(4):321.

Barth

Alves

Ryan

Macciocchi

Rimel

Nelson

. Mild head injury in sports: neuropsychological sequelae and recovery of function. In: Levin

Eisenberg

Benton

, eds. Mild Head Injury. Oxford University Press; 1989:257-275.

Bauer

Iverson

Cernich

Binder

Ruff

Naugle

RI.

Computerized neuropsychological assessment devices: joint position paper of the American Academy of Clinical Neuropsychology and the National Academy of Neuropsychology. Arch Clin Neuropsychol. 2012;27(3):362-373.

Broglio

Macciocchi

Ferrara

MS.

Sensitivity of the concussion assessment battery. Neurosurgery. 2007;60(6):1050-1057; discussion 1057-1058.

10.

Broglio

Puetz

TW.

The effect of sport concussion on neurocognitive function, self-report symptoms and postural control : a meta-analysis. Sports Med. 2008;38(1):53-67.

11.

Bruce

Echemendia

Tangeman

, et al. Two baselines are better than one: improving the reliability of computerized testing in sports neuropsychology. Appl Neuropsychol Adult. 2016;23(5):336-342.

12.

Bruce

Echemendia

Meeuwisse

Hutchison

Aubry

Comper

Measuring cognitive change with ImPACT: the aggregate baseline approach. Clin Neuropsychol. 2017;31(8):1329-1340.

13.

Burghart

Craig

Radel

Huisinga

Reliability and validity of a mobile device application for use in sports-related concussion balance assessment. Curr Res Concussion. 2017;4:e1-e6.

14.

Burghart

Craig

Radel

Huisinga

Reliability and validity of a motion-based reaction time assessment using a mobile device. Appl Neuropsychol Adult. 2019;26(6):558-563.

15.

Caccese

Teel

Van Patten

Muzeau

Iverson

VanRavenhorst-Bell

HA.

Test-retest reliability and preliminary reliable change estimates for Sway Balance tests administered remotely in community-dwelling adults. Front Digit Health. 2022;4:999250.

16.

Campbell

Howland

Hess

, et al. Disparities in baseline neurocognitive testing for student concussion management in Massachusetts high schools. BMJ Open Sport Exerc Med. 2020;6(1):e000752.

17.

Covassin

Elbin

3rd Stiller-Ostrowski

JL.

Current sport-related concussion teaching and clinical practices of sports medicine professionals. J Athl Train. 2009;44(4):400-404.

18.

Covassin

Elbin

Stiller-Ostrowski

Kontos

AP.

Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT) practices of sports medicine professionals. J Athl Train. 2009;44(6):639-644.

19.

Czerniak

Liebel

Garcia

G-GP

, et al; CARE Consortium Investigators. Sensitivity and specificity of computer-based neurocognitive tests in sport-related concussion: findings from the NCAA-DoD CARE Consortium. Sports Med. 2021;51(2):2.

20.

Davis

Anderson

Babl

, et al. What is the difference in concussion management in children as compared with adults? A systematic review. Br J Sports Med. 2017;51(12):949-957.

21.

Dewan

Roger James

Kumar

Sawyer

SF.

Kinematic validation of postural sway measured by Biodex Biosway (force plate) and SWAY Balance (accelerometer) technology. Biomed Res Int. 2019;2019:8185710.

22.

Dharnipragada

Naik

Denduluri

, et al. Diagnostic predictive values for sport-related concussions: a systematic review and diagnostic meta-analysis. J Neurosurg. 2024;140(2):560-569.

23.

Downey

Hutchison

Comper

Determining sensitivity and specificity of the Sport Concussion Assessment Tool 3 (SCAT3) components in university athletes. Brain Inj. 2018;32(11):1345-1352.

24.

Dunn

Bay

Cárdenas

Anastasi

McLeod

TCV

Williams

. Reliability of the Sway Balance Mobile Application: a retrospective analysis.

25.

Echemendia

Brett

Broglio

, et al. Introducing the Sport Concussion Assessment Tool 6 (SCAT6). Br J Sports Med. 2023;57(11):619-621.

26.

Echemendia

Bruce

Bailey

Sanders

Arnett

Vargas

The utility of post-concussion neuropsychological data in identifying cognitive change following sports-related MTBI in the absence of baseline data. Clin Neuropsychol. 2012;26(7):1077-1091.

27.

Echemendia

Burma

Bruce

, et al. Acute evaluation of sport-related concussion and implications for the Sport Concussion Assessment Tool (SCAT6) for adults, adolescents and children: a systematic review. Br J Sports Med. 2023;57(11):722-735.

28.

Echemendia

Meeuwisse

McCrory

, et al. The Sport Concussion Assessment Tool 5th Edition (SCAT5): Background and rationale. Br J Sports Med. 2017;51(11):848-850.

29.

Echemendia

Thelen

Meeuwisse

, et al. Neuropsychological assessment of professional ice hockey players: a cross-cultural examination of baseline data across language groups. Arch Clin Neuropsychol. 2020;35(3):240-256.

30.

Elbin

Sufrinko

Schatz

, et al. Removal from play after concussion and recovery time. Pediatrics. 2016;138(3):e20160910.

31.

Feddermann-Demont

Echemendia

Schneider

, et al. What domains of clinical function should be assessed after sport-related concussion? A systematic review. Br J Sports Med. 2017;51(11):903-918.

32.

Ferris

Kontos

Eagle

, et al. Utility of VOMS, SCAT3, and ImPACT baseline evaluations for acute concussion identification in collegiate athletes: findings from the NCAA-DoD Concussion Assessment, Research and Education (CARE) Consortium. Am J Sports Med. 2022;50(4):1106-1119.

33.

Field

. Intraclass correlation. In: Everitt

Howell

, eds, Encyclopedia of Statistics in Behavioral Science. John Wiley & Sons; 2005:948-954.

34.

Harmon

Whelan

Aukerman

, et al. Diagnosis of sports-related concussion using symptom report or standardized assessment of concussion. JAMA Netw Open. 2024;7(6):e2416223.

35.

Hinton-Bayre

Geffen

McFarland

Friis

Concussion in contact sports: reliable change indices of impairment and recovery. J Clin Exp Neuropsychol. 1999;21(1):70-86.

36.

Hou

Zhang

Fei

Zhou

Sports-related concussion affects cognitive function in adolescents: a systematic review and meta-analysis. Am J Sports Med. 2023;51(13):3604-3618.

37.

Hunt

Ferrara

MS.

Age-related differences in neuropsychological testing among high school athletes. J Athl Train. 2009;44(4):405-409.

38.

Iverson

GL.

Interpreting change on the WAIS-III/WMS-III in clinical samples. Arch Clin Neuropsychol. 2001;16(2):183-191.

39.

Iverson

Lovell

Collins

MW.

Interpreting change on ImPACT following sport concussion. Clin Neuropsychol. 2003;17(4):460-467.

40.

Iverson

Schatz

Advanced topics in neuropsychological assessment following sport-related concussion. Brain Injury. 2015;29(2):263-275.

41.

Iverson

Schatz

Brief iPad-based assessment of cognitive functioning with ImPACT pediatric. Dev Neuropsychol. 2019;44(1):43-49.

42.

Jacobson

Truax

Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. J Consult Clin Psychol. 1991;59(1):12-19.

43.

Kis

Reliability of a new test of balance function in healthy and concussion populations. J Funct Morphol Kinesiol. 2020;5(1):13.

44.

Konstantinides

Murphy

Whelan

, et al. Influence of preexisting conditions and concussion history on postconcussion symptom severity and recovery time in collegiate athletes. J Neurotrauma. 2025;42(1-2):57-70.

45.

Koo

MY.

A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med. 2016;15(2):155-163.

46.

Lau

Collins

Lovell

MR.

Sensitivity and specificity of subacute computerized neurocognitive testing and symptom evaluation in predicting outcomes after sports-related concussion. Am J Sports Med. 2011;39(6):1209-1216.

47.

Louey

Cromer

Schembri

, et al. Detecting cognitive impairment after concussion: sensitivity of change from baseline and normative data methods using the CogSport/Axon cognitive test battery. Arch Clin Neuropsychol. 2014;29(5):432-441.

48.

Lovell

Collins

MW.

Neuropsychological assessment of the college football player. J Head Trauma Rehabil. 1998;13(2):9-26.

49.

Lovell

Solomon

GS.

Psychometric data for the NFL neuropsychological test battery. Appl Neuropsychol. 2011;18(3):197-209.

50.

Maddocks

Dicker

Saling

MM.

The assessment of orientation following concussion in athletes. Clin J Sport Med. 1995;5(1):32-35.

51.

McCrea

Guskiewicz

Marshall

, et al. Acute effects and recovery time following concussion in collegiate football players: the NCAA Concussion Study. JAMA. 2003;290(19):2556-2563.

52.

McCrea

Iverson

Echemendia

Makdissi

Raftery

Day of injury assessment of sport-related concussion. Br J Sports Med. 2013;47(5):272-284.

53.

McGraw

Wong

SP.

Forming inferences about some intraclass correlation coefficients. Psychological Methods. 1996;1(1):30-46.

54.

Mummareddy

Brett

Yengo-Kahn

Solomon

Zuckerman

SL.

Sway Balance Mobile Application: reliability, acclimation, and baseline administration. Clin J Sport Med. 2020;30(5):451-457.

55.

NCAA. Concussion Safety Protocol Management. National Collegiate Athletic Association. Published January 15, 2024. Accessed February 26, 2026. https://www.ncaa.org/sports/2016/7/20/concussion-safety-protocol-management.aspx

56.

NFL Concussion Diagnosis and Management Protocol (PDF) . National Football League. Published September 1, 2024. Accessed February 26, 2026. https://www.nfl.com/playerhealthandsafety/resources/fact-sheets/nfl-head-neck-and-spine-committee-s-concussion-diagnosis-and-management-protocol

57.

NHL Public Relations. NHL Concussion Evaluation and Management Protocol for 2022-23 Season. National Hockey League. Published July 15, 2022. Accessed February 26, 2026. https://www.nhl.com/news/nhl-concussion-evaluation-and-management-protocol-for-2022-23-season-335002568

58.

Patricios

Schneider

Dvorak

, et al. Consensus statement on concussion in sport: the 6th International Conference on Concussion in Sport–Amsterdam, October 2022. Br J Sports Med. 2023;57(11):695-711.

59.

Patterson

Amick

Pandya

Hakansson

Jorgensen

MJ.

Comparison of a mobile technology application with the Balance Error Scoring System. Int J Athl Ther Train. 2014;19(3):4-7.

60.

Pellman

Lovell

Viano

Casson

IR.

Concussion in professional football: recovery of NFL and high school athletes assessed by computerized neuropsychological testing—part 12. Neurosurgery. 2006;58(2):263-274.

61.

Pellman

Lovell

Viano

Casson

Tucker

AM.

Concussion in professional football: neuropsychological testing—part 6. Neurosurgery. 2004;55(6):1290-1303; discussion 1303-1305.

62.

Prien

Junge

Brugger

Straumann

Feddermann-Demont

Neurocognitive performance of 425 top-level football players: sport-specific norm values and implications. Arch Clin Neuropsychol. 2019;34(4):575-584.

63.

Putukian

Echemendia

Dettwiler-Danspeckgruber

, et al. Prospective clinical assessment using Sideline Concussion Assessment Tool-2 testing in the evaluation of sport-related concussion in college athletes. Clin J Sport Med. 2015;25(1):36-42.

64.

Randolph

Baseline neuropsychological testing in managing sport-related concussion: does it modify risk?

Curr Sports Med Rep. 2011;10(1):21-26.

65.

Resch

Brown

Schmidt

, et al. The sensitivity and specificity of clinical measures of sport concussion: three tests are better than one. BMJ Open Sport Exerc Med. 2016;2(1):e000012.

66.

Rogers

Smith

Stephenson

Erik Everhart

A retrospective cross-sectional and longitudinal study of the effects of age on CNS Vital Signs scores in high-school athletes. Sports Med. 2017;47(9):1893-1899.

67.

Schatz

Pardini

Lovell

Collins

Podell

Sensitivity and specificity of the ImPACT Test Battery for concussion in athletes. Arch Clin Neuropsychol. 2006;21(1):91-99.

68.

Schatz

Robertshaw

Comparing post-concussive neurocognitive test data to normative data presents risks for under-classifying “above average” athletes. Arch Clin Neuropsychol. 2014;29(7):625-632.

69.

Schatz

Sandel

Sensitivity and specificity of the online version of ImPACT in high school and collegiate athletes. Am J Sports Med. 2013;41(2):321-326.

70.

Skinner

Armstrong

Carter

Brendan Clark

VanRavenhorst-Bell

HA.

Validation of the Sway Medical App as a neurocognitive assessment tool. Int J Athl Ther Train. 2025;30(6):1-7.

71.

Solomon

Lovell

Casson

Viano

DC.

Normative neurocognitive data for National Football League players: an initial compendium. Arch Clin Neuropsychol. 2015;30(2):161-173.

72.

Stephenson

Maietta

Cook

, et al. Examining for race-associated differences on Sway Medical System balance and cognitive tests used for sport-related concussion. Front Neurol. 2025;16:1547004.

73.

Sway Medical Inc. Computer software. Tulsa: OK, 2011.

74.

Terry

Büttner

Huebschmann

Gardner

Cook

Iverson

GL.

Systematic review of pre-injury migraines as a vulnerability factor for worse outcome following sport-related concussion. Front Neurol. 2022;13:915357.

75.

Van Patten

Iverson

Muzeau

VanRavenhorst-Bell

HA.

Test-retest reliability and reliable change estimates for four mobile cognitive tests administered virtually in community-dwelling adults. Front Psychol. 2021;12:734947.

76.

VanRavenhorst-Bell

Muzeau

Luinstra

Goering

Amick

RZ.

Accuracy of the SWAY Mobile Cognitive Assessment Application. Int J Sports Phys Ther. 2021;16(4):991-1000.

77.

Wallace

Schatz

Covassin

Iverson

GL.

Brief iPad-based assessment of cognitive functioning with ImPACT Quick Test: prevalence of low scores using multivariate base rates. Arch Clin Neuropsychol. 2020;35(8):8.

78.

Wilmoth

Brett

Emmert

, et al. Psychometric properties of computerized cognitive tools and standard neuropsychological tests used to assess sport concussion: a systematic review. Neuropsychol Rev. 2023;33(4):675-692.