IQ After Pediatric Concussion

Abstract

OBJECTIVES

This study investigated IQ scores in pediatric concussion (ie, mild traumatic brain injury) versus orthopedic injury.

METHODS

Children (N = 866; aged 8–16.99 years) were recruited for 2 prospective cohort studies from emergency departments at children’s hospitals (2 sites in the United States and 5 in Canada) $\le$48 hours after sustaining a concussion or orthopedic injury. They completed IQ and performance validity testing postacutely (3–18 days postinjury; United States) or 3 months postinjury (Canada). Group differences in IQ scores were examined using 3 complementary statistical approaches (linear modeling, Bayesian, and multigroup factor analysis) in children performing above cutoffs on validity testing.

RESULTS

Linear models showed small group differences in full-scale IQ (d [95% confidence interval] = 0.13 [0.00–0.26]) and matrix reasoning (0.16 [0.03–0.30]), but not in vocabulary scores. IQ scores were not related to previous concussion, acute clinical features, injury mechanism, a validated clinical risk score, pre- or postinjury symptom ratings, litigation, or symptomatic status at 1 month postinjury. Bayesian models provided moderate to very strong evidence against group differences in IQ scores (Bayes factor 0.02–0.23). Multigroup factor analysis further demonstrated strict measurement invariance, indicating group equivalence in factor structure of the IQ test and latent variable means.

CONCLUSIONS

Across multisite, prospective study cohorts, 3 complementary statistical models provided no evidence of clinically meaningful differences in IQ scores after pediatric concussion. Instead, overall results provided strong evidence against reduced intelligence in the first few weeks to months after pediatric concussion.

What’s Known on This Subject:

Pediatric concussion (ie, mild traumatic brain injury) is highly prevalent, but whether concussion results in lower IQ scores is controversial.

What This Study Adds:

This study revealed no evidence of clinically meaningful differences in IQ scores postacutely (ie, 2 weeks) or at 3 months post-injury between pediatric concussion and mild orthopedic injury, suggesting that pediatric concussions do not alter IQ.

Intellectual functioning, as measured by IQ, is predictive of global functioning across multiple domains, including academic achievement and quality of life.¹ Concussion (ie, mild traumatic brain injury [TBI]) is highly prevalent, affecting millions of children in North America each year.² In contrast to early and persistent intellectual impairment after moderate to severe TBI in children,^3–7 the effect of concussion on IQ is unclear.⁷ Evidence for IQ differences after pediatric concussion is mixed in individual studies.^8–12 Meta-analyses suggest slightly reduced IQ several months after pediatric concussion.^7,8 However, current studies are limited in number and in methodology, with small samples with restricted age ranges, unbalanced male to female ratios, varying assessment times postinjury, restriction to hospitalized children, the use of healthy children for comparison, and reliance on suboptimal statistical methods for testing group differences.^9,10 Moreover, it is not known whether injury characteristics or environmental factors, such as socioeconomic status (SES), influence children’s risk for lower IQ after concussion. Overall, these shortcomings hamper conclusions about whether concussion negatively affects IQ and prevents clinicians from accurately identifying which children may be at risk.

This study addressed this clinically important knowledge gap and the methodological limitations of previous studies. Specifically, we combined data from 2 multisite, prospective cohort studies to capitalize on the methodological strengths of both studies, (eg, prospective recruitment, large sample sizes [ie, N >200], broad age ranges [8–16.99 years], balanced sex ratios [∼60% male], and mild orthopedic injury [OI] comparison group) and used 3 complementary statistical approaches to investigate IQ score differences in children with concussion versus mild OI. Demographic and injury characteristics were also investigated as potential moderators of IQ after concussion, to identify any subgroups of children at increased risk of lower postinjury IQ. We expected to find no evidence of a clinically meaningful difference in IQ after concussion relative to OI, and that demographic and clinical characteristics would not moderate difference in IQ scores. However, SES was expected to be associated with IQ in both groups given research on typical development.^11,12

Methods

General Study Design

Data for this study were drawn from 2 prospective cohort studies of the longitudinal outcomes of pediatric concussion: the Mild Injury Outcomes Study (MIOS)^13,14 and the Advancing Concussion Assessment in Pediatrics (A-CAP) study.^15–17

Participants and Recruitment

Children between 8.00 and 16.99 years of age were recruited ≤24 hours of injury from the emergency department (ED) of 2 children’s hospitals in the United States for MIOS (N = 214) and ≤48 hours of injury from 5 children’s hospitals across Canada that are all members of the Pediatric Emergency Research Canada group¹⁸ for the A-CAP study (N = 967). In both studies, injury information, including acute clinical signs and symptoms, was collected during the initial ED visit. Enrolled participants returned for postacute (ie, ∼1–2 weeks postinjury), 3-month, and 6-month postinjury follow-up assessments. This study included children who completed IQ testing at the postacute assessment (ie, 3–18 days postinjury: MIOS) or the 3-month postinjury assessment (A-CAP).

Concussion

Children in the concussion group sustained a blunt head trauma resulting in at least 1 of the following criteria, consistent with the World Health Organization definition of mild TBI¹⁹: an observed loss of consciousness (LOC), a Glasgow Coma Scale (GCS) score of 13 to 14, or a GCS score of 15 with at least 1 (A-CAP) or 2 (MIOS) acute signs or symptoms of concussion as noted by ED medical personnel.¹⁵ Children were excluded if they demonstrated delayed neurologic deterioration (eg, GCS <13), required neurosurgical intervention, or had LOC >30 minutes or posttraumatic amnesia >24 hours.

Mild OI

Children in the OI group sustained an upper or lower extremity fracture for both studies, or sprain or strain (A-CAP), because of blunt force trauma, associated with an Abbreviated Injury Scale score ≤4.²⁰ Children were excluded if they had head trauma or signs or symptoms of concussion at the time of recruitment, or any injury requiring surgical intervention or procedural sedation. Children with OI were chosen for comparison because they are like children with concussion demographically and in terms of predisposing risk factors for injury. Their inclusion also helps to control for the general effects of trauma.^15,16

Exclusion Criteria

Both groups were subject to the following exclusion criteria: any injury with an Abbreviated Injury Scale score >4; hypoxia, hypotension, or shock during or after the injury; previous concussion within 3 months or before TBI requiring hospitalization; premorbid neurologic disorder; severe neurodevelopmental disability; injury resulting from nonaccidental trauma; or severe psychiatric disorder requiring hospitalization within past year.¹⁵

Demographic Information

MIOS

Demographic information was collected during the postacute assessment. An SES composite index was computed by averaging sample z scores for years of maternal education, median family income for census tract based on the 2010 Census, and the Duncan Socioeconomic Index, a measure of occupational prestige.²¹

A-CAP

A demographic questionnaire was completed by parents at the initial ED visit.¹⁵ SES was estimated using the Pampalon Material Deprivation Index, whereby higher percentiles correspond with lower SES.^22,23 Derived percentiles were inversely z scored to provide comparable measures across studies.

Injury Characteristics

The Predicting and Preventing Postconcussive Problems in Pediatrics (5P) clinical risk score²⁴ was calculated on the basis of age at injury, sex, previous concussion history, migraine history, mental slowing on exam, balance, and 3 specific symptoms. This validated risk score can be used to predict likelihood of persistent symptoms after concussion. Other clinical features recorded at the time of injury included LOC, GCS score, and injury mechanism.

Symptoms

The Health and Behavior Inventory, a core measure in the Common Data Elements for Pediatric TBI,^25–27 was used to assess cognitive and somatic symptoms. Parents rated preinjury symptoms during the ED visit (MIOS) and postacute visit (A-CAP), and parents and children rated postacute and 1-month postinjury symptoms (both studies).²⁸ A reliable change index (z score) comparing total 1-month postinjury symptom scores to preinjury scores was calculated on the basis of regression analyses in the OI group,²⁷ and used to classify children with concussion into 2 groups (critical z score >1.65, 1-tailed P < .05): (1) concussion with persisting symptoms, and (2) concussion without persisting symptoms.^29,30

Intellectual Functioning and Test Validity

Children completed the Wechsler Abbreviated Scale of Intelligence, Second Edition (WASI-II) 2-subtest version. Full-scale IQ scores were based on the vocabulary (ie, crystalized knowledge) and matrix reasoning (ie, fluid reasoning) subtests and scored using WASI-II manual norms.³¹ The Medical Symptom Validity Test (MSVT) was used as a measure of performance validity.³² The WASI-II and the MSVT were completed at the postacute visit (MIOS) or the 3-month postinjury visit (A-CAP).³¹

Statistical Analyses

Sample demographic data were analyzed in RStudio v1.4.993 (R v4.0.3), using t test for continuous variables and χ² techniques for categorical variables.^33,34

Group differences in IQ were examined in participants who scored above standard cutoffs on the MSVT (ie, valid performance) using 3 complementary statistical approaches: linear modeling (ie, frequentist), Bayesian, and multigroup factor analyses. Frequentist approaches, such as linear modeling, test the likelihood of a group mean difference under the assumption of no difference (ie, null hypothesis [H₀]), yielding traditional tests of statistical significance (P values). However, frequentist approaches provide no information about the probability of H₀ being true and are vulnerable to sampling bias.^35,36 In contrast, Bayesian statistics, a nonfrequentist approach, directly assess the probability of accepting H₀ and are robust to common limitations of frequentist approaches.^35,36 Finally, multigroup factor analysis examines whether the IQ test is measuring the same construct across groups by testing the underlying factor model for measurement invariance, including the equivalence of latent group means. In summary, these approaches tested for:

1. statistically significant differences between group means (linear modeling);

2. the probability that the group means did not differ (Bayesian statistics); and

3. the comparability of IQ measurement between groups.

Linear Models

Linear models were conducted using the stats package in R.^33,34 Initial models included the independent variables group, sex, age at injury, and SES, as well as all 2-way group interactions (ie, group × sex, group × age, and group × SES) with covariate study (MIOS, A-CAP). No moderators of group differences in IQ scores were significant (Supplemental Table 3), so the interaction terms were removed from the final models. Age at injury was also removed because IQ scores were already normed for age. Therefore, final linear models examined group differences in IQ scores, controlling for sex, SES, and study. Linear models also were used to test for univariate associations between IQ scores and clinical characteristics (ie, premorbid symptom ratings, previous concussion, LOC, GCS, injury mechanism, 5P score, litigation, postacute symptom ratings, symptomatic status at 1-month postinjury) in the concussion group, controlling for SES and study.

Bayesian Analysis

Bayesian modeling was conducted using R and the rstanarm v.2.21.1 package.^33,34,37 Following the Sequential Effect eXistence and significance testing framework, the median (highest density interval and 89% confidence interval [CI]) of the posterior was computed, along with the probability of direction, probability of significance, and probability of the effect size within the context of the Region of Practical Equivalence range.³⁸ The Region of Practical Equivalence range was calculated as the standardized parameter using the 90% CI reliable change index values (full-scale IQ/vocabulary/matrix reasoning = 8.70/7.20/10.30).³⁹ Convergence and stability of the Bayesian sampling was assessed using R-hat <1.01⁴⁰ and effective sample size >1000.⁴¹ On the basis of the higher posterior probability model and median probability model, sex was excluded but SES and study were retained as covariates. Bayes factors were interpreted using Lee and Wagenmakers (2014) criteria.⁴²

Measurement Invariance

Multigroup factor analyses were conducted using Mplus v8.8⁴³ and its robust maximum likelihood estimator. A multiple-indicator, multiple-cause (MIMIC) model was employed.⁴⁴ Several invariance models with increasingly restrictive parameter constraints were used to assess equivalence^45,46 of IQ across the concussion and OI groups.

First, the configural invariance model, with no equality constraints across groups, was specified. Then, metric invariance was tested by placing equality constraints on factor loadings. Next, scalar invariance was examined by constraining the subtest means to equality. Strict invariance was tested by holding the respective residual variances to equality across groups. Additionally, invariance of the latent variances, structural relations (ie, equal MIMIC regressions), and latent mean were tested across groups.

The MIMIC model used a single latent IQ factor (initially fixed latent variance at 1.00 and latent mean at 0.00), with factor loadings for the vocabulary and matrix reasoning subtests freely estimated. Following the Bayesian results, sex was excluded from the invariance analyses, although results with sex included were not substantively different, with direct effects of covariates study and SES included to avoid classification issues. Nonsignificant changes in χ² and the comparative fit index <0.010 across the sequence of tests were used as indications of invariance.^47,48 The Bayesian information criterion was used to inform fit of successive invariance models, with lower values indicating better fit.^49,50

Results

Sample

Figure 1 summarizes the derivation of the final sample from each study. Of the 1282 (MIOS/A-CAP = 315/967) children who were recruited for the studies, 936 (MIOS/A-CAP = 217/719) children returned for assessment. Of those, 933 (MIOS/A-CAP = 214/719) completed the WASI-II and 914 (MIOS/A-CAP = 214/700) also had SES data. Children with WASI-II and SES data who scored below cutoffs on the MSVT were excluded from final analysis (n = 48, MIOS/A-CAP = 18/30). The final sample therefore included 866 children (concussion/OI = 566/300).

FIGURE 1.

Flowchart summarizing how the final sample was derived for MIOS and the A-CAP study. MIOS participants who returned for the postacute assessment did not significantly differ from those who did not in terms of acute symptoms, sex, premorbid postconcussive symptom ratings, or age at time of injury.^13,14 However, SES (based on the 2010 Census tract median family income in the ED) was higher among children who returned (n = 217) than who did not return (n = 98) for the MIOS postacute assessment, and a greater number of white and Asian American children returned for the postacute assessment than children of another race. Children in the MIOS study with completed WASI-II/SES data who scored below cutoffs on the MSVT (n = 18) did not differ from those who scored above the cutoffs in terms of age at injury, sex, race, or SES. For the A-CAP study, the children who returned (n = 719) at the 3-month follow-up did not differ significantly from those who did not return in age, sex, race, or parental education. Of those, 700 completed the WASI-II and also had completed SES data. However, 48 children with completed WASI-II/SES data scored below cutoffs on the MSVT and were excluded from final analyses; those who scored below cutoffs did not differ from those who scored above cutoffs in sex, SES, race, or injury mechanism, but were younger and more likely to have sustained the injury during nonsport-related activities.

The final sample is summarized in Table 1. Groups did not differ in age at injury, sex, race, parent education, days to assessment, or whether the injury was sport-related. However, children with concussion had higher SES, higher postacute and 1-month postinjury symptom ratings, and different injury mechanism, than children with OI.

TABLE 1.

Demographic Sociodemographic and Injury Characteristics for the Final Sample of Children With Completed WASI-II and Socioeconomic Who Scored Above Cutoffs on the MSVT

Variable	Concussion	OI	P
	n = 566	n = 300
Study, n (%)			.008
MIOS (United States)	125 (22.08)	71 (23.67)
A-CAP (Canada)	441 (77.92)	229 (76.33)
Age, y, mean (SD)	12.39 (2.53)	12.60 (2.22)	.24
Sex, male, n (%)	350 (61.84)	168 (56.00)	.11
Race, white, n (%)	419 (74.03)	206 (68.67)	.23
SES, z score, mean (SD)	0.29 (0.68)	0.18 (0.68)	.03
Parent education, n (%)			.67
No certificate, diploma, or degree	18 (3.18)	9 (3.00)
High school diploma or equivalent	70 (12.36)	45 (15.00)
Partial college or training program	182 (32.16)	97 (32.33)
Bachelor’s degree	182 (32.16)	83 (27.67)
Graduate degree	90 (15.90)	44 (14.67)
Days to assessment, mean (SD)	73.98 (36.48)	75.33 (35.80)	.60
Mechanism of injury, n (%)			<.001
Fall	192 (33.92)	123 (41.00)
Bicycle-related	19 (3.36)	14 (4.67)
Motor vehicle collision	13 (2.29)	2 (0.67)
Struck object	136 (24.02)	54 (18.00)
Struck person	92 (16.25)	31 (10.33)
Other	5 (0.88)	25 (8.33)
Unknown/not documented	7 (1.23)	14 (4.67)
Injury during sport/recreational play, n (%)	361 (63.78)	194 (64.67)	.12
Previous concussion history, n (%)			.764
Yes	156 (27.56)	77 (25.67)
Unknown	1 (0.18)	1 (0.33)
Litigation involved, n (%)	10 (1.77)	0 (0)	.06
GCS 15, n (%)	509 (89.93)	—	—
LOC, n (%)			—
Yes/suspected	107 (18.90)	—
Unknown	27 (4.77)	—
Hospitalized, n (%)	15 (2.65)	4 (1.33)	—
5P total score	5.90 (1.93)	3.47 (1.58)	<.001
Symptom ratings, mean (SD)
Parent preinjury somatic	2.91 (3.72)	2.20 (3.09)	.005
Parent preinjury cognitive	9.24 (7.76)	7.92 (7.42)	.02
Parent postacute somatic	6.72 (5.32)	1.77 (2.83)	<.001
Parent postacute cognitive	11.26 (8.29)	6.97 (7.40)	<.001
Child postacute somatic	9.49 (5.57)	3.21 (3.65)	<.001
Child postacute cognitive	13.28 (8.16)	6.47 (6.42)	<.001
Symptomatic,a n (%)			—
Postacute, parent	217 (38.34)	—
Postacute, child	241 (42.58)	—
1-mo postinjury, parent	78 (13.78)	—
1-mo postinjury, child	93 (16.43)	—

—, not applicable.

^aRelative to preinjury parent ratings calculated on the basis of reliable change index z score as detailed previously.²⁷

Power Analysis

A power analysis conducted using G*Power v3.1⁵¹ indicated that the final sample size (n = 866) was sufficiently powered (1-β = .95) to detect small effects (Cohen’s f² = 0.07) for a full factorial model with multiple predictors (eg, group, SES, age, sex, and the 2-way interactions with group) at an α = .05.

IQ Scores

Linear Models

Linear model results are summarized in Fig 2 and Table 2. There were very small group differences in full-scale IQ (Cohen’s d [95% CI] = 0.13 [0.00–0.26], concussion/OI mean [SD] = 104.95 [14.10]/106.08 [14.39]) and matrix reasoning scores (d [95% CI] = 0.16 [0.03–0.30], concussion/OI = 52.58 [10.10]/53.81 [9.87]), but not vocabulary scores (d [95% CI] = 0.005 [−0.08 to 0.19], concussion/OI = 53.25 [8.86]/53.27 [9.36]). IQ scores did not differ by sex but were lower in the United States versus Canada study (ie, MIOS < A-CAP; full-scale IQ d [95% CI] = −0.41 [−0.56 to −0.26], vocabulary d [95% CI] = −0.38 [−0.54 to −0.23], matrix reasoning d [95% CI] = −0.33 [−0.48 to −0.17]) and were positively associated with SES.

FIGURE 2.

Graphs illustrating IQ scores. Graphs illustrate postacute and 3 months postinjury full-scale IQ, matrix reasoning t, and vocabulary t scores for children with concussion and mild OI.

TABLE 2.

Linear Model Results for IQ Scores

	Full-Scale IQ, F(4–861) = 34.46, P < .001, R2 = 0.14			Vocabulary, F(4–861) = 31.71, P < .001, R2 = 0.13			Matrix Reasoning, F(4–861) = 19.51, P < .001, R2 = 0.08
Predictor	β	t	P	β	t	P	β	t	P
Group (OI, concussion)	1.87	1.97	.05	.49	0.81	0.42	1.61	2.33	.02
Sex (male, female)	−.43	−0.47	.64	−0.04	−0.07	0.95	−.49	−0.73	.47
SES (z score)	6.13	9.13	<.001	3.83	8.94	<.001	3.21	6.57	<.001
Study (MIOS, A-CAP)	−5.79	−5.28	<.001	−3.47	−4.95	<.001	−3.28	−4.11	<.001

IQ scores were also examined in relation to clinical characteristics in the concussion group. Preinjury cognitive and somatic symptom scores, previous concussion history, LOC, GCS, injury mechanism, the 5P score, and postacute symptom scores were not related to full-scale IQ, vocabulary, or matrix reasoning scores, all P > .09. IQ scores did not differ between children involved in litigation relative to children not involved in litigation, P > .11, or between children with versus without persisting symptoms 1-month postinjury, P > .24.

Bayesian Analysis

Results for the Bayesian analysis are provided in Supplemental Table 4. Results provided moderate to very strong evidence for the H₀ for full-scale IQ, vocabulary, and matrix reasoning scores, indicating equivalent group IQ scores. In contrast, there was extreme evidence for effects of SES and strong to extreme evidence for effects of study for all 3 IQ scores.

Invariance Analysis

Model fit indices for each multigroup MIMIC model are summarized in Fig 3 (for statistics, see Supplemental Table 5). Successive tests of increasingly restrictive constraints supported strict invariance (ie, equivalent factor loadings, means, and residual variances across groups). Tests of latent factor variance across groups and of the regression of the latent factor onto the study covariate were not significant, across SES. Finally, the latent mean invariance model was also supported, indicating the means of the latent IQ factor were equal across groups, controlling for SES.

FIGURE 3.

Graphic illustration of the factor model underlying the MIMIC invariance analyses. In the configural invariance model, the latent factor variance was fixed to 1.00 in both the concussion and mild OI groups for model identification. In the metric invariance model, factor loadings were fixed to equality to freely estimate the latent variance in the concussion group, σ²_IQ = 0.98 (95% CI, 0.63–1.32). Finally, in the scalar invariance model, the intercepts were fixed to equality to freely estimate the latent factor mean in the concussion group, η_IQ= −0.59 (95% CI, −1.17 to −0.01). Estimates (95% CI) for the configural invariance model are summarized for each group (denoted in subscripts) for each parameter, including standardized factor loadings (λ_j), unstandardized intercepts (τ_j), residual variances (ε_j), and MIMIC regression coefficients (β_j).

Discussion

This study investigated differences in IQ scores between children with concussion (ie, mild TBI) relative to mild OI using data drawn from 2 of the largest, prospective pediatric studies of mild traumatic injury to date. Overall, the results provide no evidence that concussion in children is associated with clinically meaningful differences in IQ in the first several weeks to months postinjury. Although full-scale IQ and matrix reasoning subtest scores differed significantly in concussion relative to mild OI in linear models, mean scores were average in both groups and group differences were very small in magnitude. Moreover, group differences were not moderated by demographic factors, and IQ scores were not related to clinical characteristics in the concussion group. Additionally, Bayesian analyses provided strong evidence that IQ scores were equivalent for children with concussion and OI. The factor structure and latent means of IQ tests also did not differ between groups, indicating comparability of IQ measurement across groups. The results were consistent across the United States (MIOS; postacute) and Canadian (A-CAP; 3 months postinjury) studies, suggesting that the findings are robust and likely generalizable. Moreover, results are consistent with independent results in other large samples.⁵²

Full-scale IQ and subtest scores were well within the average range for both injury groups, across studies and time postinjury. This differs from previous meta-analytic findings of reduced full-scale and verbal IQ scores reported several months after pediatric concussion,^7,8 likely reflecting methodological limitations and inconsistencies in previous studies (eg, generally small sample sizes, poorly specified diagnostic criteria, variable definitions of injury severity, sample selection bias).¹⁵ The current study addresses the need for prospective research in pediatric concussion with large sample sizes that include a broad age range and more equitable sex ratio.¹⁵

The positive relation of SES to IQ scores, found across the groups and studies, is not surprising.^11,12 Children from families of greater social disadvantage are at risk for lower full-scale IQ.¹¹ Importantly, SES related to IQ scores despite the use of different measures of SES in the 2 studies. The A-CAP study was less diverse in terms of SES than the MIOS sample, and average IQ scores were higher in the Canadian cohort. Normative sample differences on cognitive tests are known to occur between children in Canada and the United States.⁵³ Thus, this finding was not unexpected given that the same test norms were used in both studies (ie, Canadian norms are available for some comprehensive IQ tests, but not for the WASI-II).

Limitations

Recruitment occurred in ED settings, which may limit generalizability to children who seek care in other settings or who do not seek medical care. Some enrolled participants had missing data and were not included. This could have introduced selection bias, although data loss was minimal and consistency of findings with other studies suggests minimal bias. IQ was only collected at a single time point in both studies. However, on the basis of existing meta-analyses,^7,8 we have no reason to expect changes in IQ scores to occur at later times postinjury. Psychometrics of the WASI-II 2-subtest version differ from comprehensive IQ tests, particularly those that include processing speed and/or working memory subtest scores during the calculation of full-scale IQ (eg, Wechsler Intelligence Scale for Children, Fifth Edition).³¹ Thus, previous studies that demonstrated differences in IQ scores on the basis of comprehensive tests that included additional subtests in the calculation of full-scale IQ could be capturing processing speed or working memory, which can be lowered after concussion in children.¹⁴ The current results do not rule out effects of concussion on IQ in some children or on other IQ measures, and it is important to consider differences in how IQ is operationalized across different studies. However, the WASI-II full-scale IQ (4 subtest version) and subtests (matrix reasoning and vocabulary) correlate between 0.59 and 0.87 with the Wechsler Intelligence Scale for Children, Fifth Edition, full-scale IQ and respective subtests.³¹ The small number of children involved in litigation (Table 1) had average full-scale IQ, but may not be representative of all children who are in litigation given they passed the MSVT. Also, although IQ scores were not related to clinical characteristics, we did not explore potential associations between IQ and specific symptom profiles.

Conclusions

These findings contribute novel information to research on the outcomes of pediatric concussion and could aid clinicians in providing meaningful guidance to children and caregivers about expected effects on IQ. Specifically, these results indicate that pediatric concussion does not negatively affect intellectual functioning early and up to 3 months postinjury, and hence suggest that using IQ tests to assess concussion outcomes is likely of limited utility.

Glossary

5P

Predicting and Preventing Postconcussive Problems in Pediatrics

A-CAP

Advancing Concussion Assessment in Pediatrics

CI

confidence interval

ED

emergency department

GCS

Glasgow Coma Scale

H₀

null hypothesis

LOC

loss of consciousness

MIMIC

multiple-indicator, multiple-cause

MIOS

Mild Injury Outcomes Study

MSVT

Medical Symptom Validity Test

OI

orthopedic injury

SES

socioeconomic status

TBI

traumatic brain injury

WASI-II

Wechsler Abbreviated Scale of Intelligence, Second Edition