Characteristics and outcomes for delayed diagnosis of concussion in pediatric patients presenting to the emergency department

Abstract

BACKGROUND

Concussions are common pediatric injuries. Previous studies have shown concussed youth may be underdiagnosed in the emergency department (ED), but outcomes for those with delayed diagnosis have yet to be describes.

OBJECTIVE

Compare visit characteristics and outcomes of patients who present to the ED with head injury who receive immediate vs. delayed diagnosis.

METHODS

Retrospective chart review of patients age 6–18 years diagnosed with concussion on their first ED/urgent care (UC) visit and patients requiring a second visit for diagnosis between 7/1/2017 and 6/30/2019. We compared demographic information, ED/UC visit features, and recovery outcomes using chi-square tests, student T-test, and Wilcoxon rank-sum tests.

RESULTS

Overall, we included 85 subjects with delayed concussion diagnosis and 159 subjects with immediate diagnosis.. Those with immediate diagnosis had more symptoms inquired at initial visit (5 vs. 4, p=0.003) and a higher likelihood of receiving concussion-specific physical examinations (80% vs 36.5%, p<0.001); 76.5% of delayed diagnosis patients had at least 1 symptom present at follow-up visit not inquired about at initial visit. Those with delayed diagnosis had more medical visits during recovery (3 vs. 2, p<0.001), longer average time to symptom resolution (21 vs. 11 days, p=0.004), and a higher likelihood of having persistent concussion symptoms (OR 2.9, 95% CI 1.4–5.9).

CONCLUSION

Concussed children evaluated for head injury acutely who do not receive an immediate diagnosis may be at risk for persistent symptoms. Performance of a concussion-specific physical examination, and use of a standardized symptom scale, may aid in identification of concussed youth acutely.

INTRODUCTION

Concussion is a common pediatric injury evaluated in the emergency department (ED) setting. Nearly 2 million concussions occur annually in the pediatric population.¹ Over 800,000 ED visits occur each year for traumatic brain injury in children,² and nearly 200,000 children each year receive the diagnosis of concussion from the ED.³ Diagnosis of concussion relies mainly on subjective symptom reporting, with limited objective measures of injury available to the general emergency medicine practitioner.^4,5 The heterogeneity of the presentation, and the fact that signs and symptoms can evolve and develop well after the initial visit,^6,7 makes acute diagnosis even more challenging.⁸ The difficultly in acute diagnosis was highlighted in a study from a single tertiary care pediatric ED, which found less than 50% of patients who met international consensus criteria for concussion received an ED concussion diagnosis.⁹ Looking beyond traditional symptoms scales, a visio-vestibular examination (VVE) can aid in the diagnosis, as one study found approximately 10% of patients diagnosed from the ED with minimal symptoms have abnormalities on this testing.¹⁰

Persistent post-concussion symptoms (PPCS), defined as symptoms lasting for greater than 28 days after the injury,¹¹ carry significant morbidity, including physical, cognitive, and mental health sequelae.¹² The direct healthcare utilization cost for PPCS has been estimated to exceed $3500 per patient, yielding a total cost of over $200 million annually for patients initially seen in the ED.¹³ Traditionally, the management strategies for concussion have relied on passive rest. More recently, studies have shown that active therapies, including aerobic exercise protocols¹⁴ and visio-vestibular rehabilitation,^15,16 can improve recovery times. Access to specialists within the first week of injury, who can prescribe such therapies, has been shown to improve recovery times,^17,18 suggesting that early diagnosis can potentially mitigate the effects of PPCS. However, currently there is limited knowledge of the effect of a delayed concussion diagnosis in patients initially seen in the acute setting. No study has specifically evaluated the population of children who are seen in the ED and discharged without a diagnosis of concussion, who then return to medical care and are subsequently diagnosed. This knowledge may guide refinement of concussion evaluation in the ED so that accurate diagnoses can be made and appropriate management initiated earlier in the injury time course. Thus, the objective of this study was to compare diagnostic features (including use of a VVE), subsequent healthcare utilization, and recovery outcomes in patients who receive a diagnosis of concussion in the ED on their initial presentation, as compared with those who were seen for head trauma in the ED but were not diagnosed until a later time.

MATERIALS AND METHODS

Study Design and Patient Population

We performed a retrospective study via chart review of children age 6–18 years old who were evaluated for head injury in the ED and affiliated urgent care (UC) centers of our tertiary care children’s hospital over a two year period (July 1, 2017-June 30, 2019). Annually, the ED (located at the children’s hospital in an urban setting) and 4 UC centers (located at suburban community sites) have a combined volume of 140,000 patient visits, with approximately 1,100 concussions diagnosed per year (700 in the ED, 400 in the UC centers). For this study, concussion was defined as the utilization of an International Classification of Diseases, 10^th Revision (ICD-10) code consistent with concussion, including concussion with and without loss of consciousness and postconcussion syndrome (S06.0X0, S06.0X1, S06.0X9, and F07.81). Previously, our institution has standardized the use of ICD diagnoses, a visio-vestibular examination, and anticipatory guidance via decision support tools in the electronic health record (EHR) that align with definitions in consensus statements.⁴ The ICD code appears on the discharge paperwork the family receives from the ED/UC, in addition to discharge instructions (the vast majority of which are pre-written). Patients with concussion diagnoses generally receive pre-written discharge instructions regarding concussion symptoms, graduated return to play and return to learn, and information on the benefit of early, light aerobic exercise. These instructions are standardized across specialties at our institution.¹⁹ Patients with head injury who are not diagnosed with concussion generally receive more generic pre-written discharge instructions regarding minor head injuries and contusions, focusing on warning signs for increased intracranial pressure.

For our comparison groups, we defined a “delayed diagnosis” of concussion as a patient evaluated for head injury in the ED or UC setting and discharged without an ICD-10 concussion diagnosis, who was subsequently diagnosed with concussion at a follow-up visit with in our care network (either primary care, UC, ED, or specialty care, including sports medicine, orthopedics and neurology). We identified these cases by querying the EHR for all patients with an ICD-10 concussion code seen in the care network in the specified time period, and then identifying those with an in-network ED or UC visit in the 28 days prior. We then manually reviewed charts to determine those patients whose ED or UC visit was related to the injury. Two authors (RH and KP) reviewed all charts to determine eligibility, with discrepancies resolved by a third author (DC). We defined subjects withan “immediate diagnosis” of concussion as a patient evaluated for head injury in the ED or UC during the same time period and discharged with an ICD-10 concussion diagnosis. We only included those patients whose medical home was within our care network in either group to maximize the likelihood for complete follow-up. Given the large number of controls eligible for the sample (764), we selected a 20% random sample for in-depth manual review. Exclusion criteria for both groups included Glasgow Coma Scale score of less than 13 (indicative of a moderate or severe traumatic brain injury), patients with underlying neurologic disorders suggesting vestibular or visual dysfunction (including, but not limited to, cerebral palsy, benign paroxysmal positional vertigo, bilateral or unilateral vestibular hypofunction, strabismus, and diplopia), patients with developmental delay that precluded VVE performance, patients who were within 30 days of clearance from a previous concussion, or patients admitted to a hospital from their ED or UC visit. Our institution’s Institutional Review Board approved this study.

Exposure Variables

One of two authors (RH and KP) reviewed each chart to abstract exposure and outcome variables. These authors received standardized training regarding electronic health record (EHE) documentation within the care network, and each chart review was verified by a third author (DC). We first abstracted demographic information, including age, sex, race, ethnicity, and insurance status. We then recorded patient-specific characteristics, including past medical history elements documented in the ED/UC record that are known risk factors for persistent symptoms, including a history of prior concussion, depression, anxiety, attention deficit hyperactivity disorder, learning disability, somatization, migraine headaches, seizures, and motion sickness.^20–22 We also abstracted mechanism of injury and initial visit location (ED or UC) from the chart. We categorized symptoms using the standardized Post-Concussion Symptom Inventory (PSCI).²³ We recorded symptoms present during both the initial ED/UC encounter and the first follow-up visit, as well as symptoms inquired about and documented (both present and absent) during the initial ED/UC visit. We also abstracted performance of a VVE at the initial and first follow-up visit, and any abnormalities on that examination. The VVE has been standardized across specialties at our institution, including primary care, specialty care, and acute emergency care.^10,19 The examination includes testing for smooth pursuits, horizontal and vertical saccades, horizontal and vertical gaze stability (the vestibulo-ocular reflex), binocular convergence, monocular accommodation, and a complex tandem gait.²⁴ EHR documentation is generally performed at our institution via a charting template, which provides the ability to document a VVE if selected. The template is identical between ED and UC settings, and similar (though not identical) templates exist in primary and specialty care. We collected and managed study data using Research Electronic Data Capture (REDCap).²⁵

Outcome Variables

The primary recovery outcome of interest was the proportion of patients in each cohort who developed persistent post-concussion symptoms (PPCS), defined as symptoms that persisted beyond 28 days following injury.¹¹ We abstracted symptom duration from chart review. As some patients did not complete follow-up through symptom resolution, we were unable to calculate days until symptom resolution for all patients. We classified those patients with incomplete symptom resolution data, but who were evaluated and remained symptomatic past 28 days from injury, as having PPCS. To avoid a bias towards over-diagnosis of PPCS in this retrospective sample, we performed a sensitivity analysis where we classified all patients who did not have complete follow-up, but whose final visit occurred within 10 days of injury, as not having PPCS (attributing the lack of subsequent visits to the resolution of concussion symptoms). We evaluated total number of follow-up visits (as a proxy of healthcare utilization)13 and time until symptom resolution as secondary outcomes.

Statistical Analysis

We used standard descriptive statistics to summarize demographic data. We performed comparisons between proportions of those with immediate and delayed diagnoses, including patients with a symptom present at the follow-up visit but not documented during the initial ED/UC visit, VVE performance during the initial visit, and proportion of subjects developing PPCS using chi-square testing. We compared medians for number of symptoms present at initial visit, positive or negative symptoms documented at initial visit, symptoms present at follow-up visit, symptoms not documented at initial visit but present at follow-up, and time until symptom resolution using a Wilcoxon rank sum test. In order to control for known associations of persistent concussion symptoms, including sex, prior concussion history, past medical history (depression, anxiety, attention deficit hyperactivity disorder, learning disability, somatization, migraine headaches, seizures, or motion sickness),^19–21 as well as differences between the cases and controls in mechanism of injury and initial visit location, we performed a logistic regression for the development of PPCS.

RESULTS

In total, 85 subjects with a delayed concussion diagnosis were present in the evaluation period; 797 total subjects with an immediate concussion diagnosis were identified, with 159 random patients (20% sample) selected for chart review (see Figure 1). There was no significant difference in age, sex, race/ethnicity, insurance status, history of concussion, history of a comorbidity between the groups, or time between injury and initial visit between the groups (Table 1). Those with an immediate concussion diagnosis were more likely to have a sports mechanism of injury and be seen in UC centers initially. Those with a delayed concussion diagnosis were seen and diagnosed at their follow-up visit a median of 3 days (interquartile range [IQR] 2 to 6 days) after their initial ED/UC visit.

Figure 1.

Flow diagram of patients included in the analysis

Table 1.

Demographic characteristics

	Delayed Diagnosis (N=85)	Immediate Diagnosis (N=159)	p-value
Median age in years (IQR)	13 (10, 16)	13 (10, 15)	0.165
Male, N (%)	42 (49.4%)	84 (52.8%)	0.611
Race/ethnicity, N (%)			0.655
Non-Hispanic white	35 (41.2%)	75 (47.2%)
Non-Hispanic black	36 (42.4%)	64 (40.3%)
Hispanic	6 (7.1%)	11 (6.9%)
Other / unknown	8 (9.4%)	9 (5.7%)
Private insurance (vs. Public), N(%)	54 (63.5%)	89 (56.0%)	0.254
History of concussion, N(%)	14 (16.5%)	39 (24.5%)	0.146
History of co-morbidity, N(%)*	7 (8.2%)	8 (5.0%)	0.321
Sports mechanism of injury (vs. Non-sports), N(%)	30 (35.2%)	78 (49.1%)	0.039
Initial location in ED (vs. UC), N(%)	62 (72.9%)	81 (50.9%)	0.001
First follow-up visit location, N(%)			0.132
Primary care physician	63 (74%)	72 (77%)
Sports medicine specialist	13 (15%)	18 (19%)
Other	9 (11%)	3 (3%)
Time from injury to ED/UC visit, N(%)			0.667
0–<48 hours	69 (81.2%)	122 (76.3%)
48 hours-7 days	13 (15.3%)	32 (20.1%)
>7 days	3 (3.5%)	5 (3.1%)
Median days from ED/UC to 1st follow-up visit (IQR)	3 (2,6)	4 (3,8)	0.035
Final diagnosis at ED/UC visit, N(%)
Superficial head injury	29 (34.1%)
Face/scalp contusion	12 (14.1%)
Headache	12 (14.1%)
Face/scalp laceration	10 (11.8%)
Syncope/seizure	7 (8.2%)
Extremity injury	5 (5.9%)
Minor head injury	5 (5.9%)
Other	5 (5.9%)

ED = emergency department; IQR = interquartile range UC = urgent care

Delayed diagnosis = patients diagnosed with concussion at follow-up visit; Immediate Diagnosis = patients diagnosed with concussion at initial visit

^*Co-morbidity includes depression, anxiety, ADHD, learning disability, somatization, migraine history, epilepsy, and motion sickness

Those with immediate concussion diagnoses had significantly more positive symptoms documented at initial visit (median [IQR] 2 [1,3] vs. 1 [1,2], p<0.001), however they also had more total symptoms (present or absent) documented by the treating provider at the initial visit (5 [3,6] vs. 4 [3,6], p=0.003; Table 2). In total, 76.5% of those with delayed concussion diagnoses had at least one symptom present at their follow-up visit that was not documented (positive or negative) at their initial visit, significantly higher than those with immediate concussion diagnosis (58.1%, p=0.009). For those with delayed concussion diagnoses, the symptoms most commonly present at follow-up, but not documented at the initial visit, included fatigue, light and noise sensitivity, difficulty concentrating, feeling foggy, and irritability (Figure 2). Significantly more subjects with immediate concussion diagnoses (79.9%) received a VVE at the initial visit more frequently compared with those with delayed diagnoses (36.5%, p<0.001). In those with a VVE performed, there was a significantly higher percentage of patients with at least one abnormality in those with an immediate diagnosis (82.7%) vs. delayed diagnosis (25.8%, p<0.001). Of the patients with delayed diagnosis who had a normal VVE examination, 63.6% had an abnormality at their first follow-up visit. Of those with delayed concussion diagnosis who did not receive a VVE at the initial visit, 68.1% had an abnormal examination at follow-up.

Table 2.

Symptom and Exam Characteristics

	Delayed Diagnosis (N=85)	Immediate Diagnosis (N=159)	P-value
Median symptoms present at initial visit (IQR)	1 (1,2)	2 (1,3)	<0.001
Median symptoms asked at initial visit (IQR)	4 (3,6)	5 (3,6)	0.003
Median symptoms present at follow-up visit (IQR)	4 (2,6)	0 (0,3)	<0.001
Symptoms not asked at ED/UC, present at follow-up, N(%)	65 (76.5%)	54 (58.1%)	0.009
Median number of symptoms not asked at ED/UC, present at follow-up (IQR)	2 (1,4)	1 (0,3)	0.011
Present with at least one symptom resolving prior to visit, N(%)	21 (24.7%)	53 (33.3%)	0.172
Visio-vestibular exam performed at ED/UC visit, N (%)	31 (36.5%)	127 (80.0%)	<0.001
At least one abnormality on ED/UC visio-vestibular exam, N(%)	8 (25.8%)	105 (82.%)	<0.001

Delayed diagnosis = patients diagnosed with concussion at follow-up visit; Immediate Diagnosis = patients diagnosed with concussion at initial visit

ED = emergency department; IQR = interquartile range; UC = urgent care

Figure 2.

Most frequent symptoms present at follow-up but not documented at initial visit in those with delayed diagnoses

Recovery outcomes are presented in Table 3. Those with delayed concussion diagnoses had significantly more visits to medical care on average (median [IQR] 3 [2,5] vs. 2 [1,3], p<0.001) and a significantly longer average time until symptom resolution (21 days vs. 11 days, p=0.04). A survival curve of symptom resolution stratified by diagnosis type (immediate vs. delayed) is presented in Figure 3. Overall, 131 of the 244 patients (54%) had a documented PPCS classification, and 226 patients (93%) either had a documented PPCS classification or had their last visit occur within 10 days of the injury. Of subjects with known symptom status at 28 days, a significantly higher proportion of subjects with delayed diagnoses developed PPCS (56.9%) as compared with those with immediate concussion diagnoses (31.5%, p=0.004). This difference remained in our sensitivity analysis of classifying those whose last visit to medical care occurred within 10 days of injury as negative for PPCS (resulting in a classification of 45.2% of those with delayed diagnoses as PPCS vs. 15.0% for immediate diagnoses, p<0.001). The unadjusted odds ratio (OR) of developing PPCS for the delayed diagnosis group was 2.9 (95% confidence interval [CI] 1.4, 5.9) compared to immediate diagnosis. After adjusting for sex, concussion history, past medical history, and total initial symptoms (known risk factors for PPCS), as well initial visit location and mechanism (different between cases/controls), the OR increased to 4.4 (95% CI: 1.7, 10.9).

Table 3.

Recovery Outcomes

	Delayed Diagnosis (N=85)	Immediate Diagnosis (N=159)	P-value
Median total number of follow-up visits (IQR)	3 (2,5)	2 (1,3)	<0.001
Median time to symptom free (IQR)	21 (6.5,52.5)	11 (4,26)	0.040
Prolonged symptoms of concussion, N (%)	33 (56.9%)	23 (31.5%)	0.004
Prolonged symptoms of concussion sensitivity	33 (45.2%)	23 (15.0%)	<0.001

Delayed diagnosis = patients diagnosed with concussion at follow-up visit; Immediate Diagnosis = patients diagnosed with concussion at initial visit

IQR = interquartile range

Figure 3.

Survival curve for time until symptom free, stratified by delayed vs. immediate diagnosis

DISCUSSION

This study evaluated visit characteristics and recovery outcomes for patients with delayed concussion diagnoses (patients seen in our emergency department [ED] or urgent care [UC] centers for head injury, discharged without a concussion diagnosis, who then returned to care and subsequently were diagnosed with concussion) when compared to a sample of patients with immediate concussion diagnoses made on their first ED/UC visit. These data show that those with delayed diagnoses utilized medical care more frequently, had prolonged overall duration of symptoms, and were at nearly 3 times higher risk of developing persistent post concussion symptoms (PPCS), a risk that increased to over 4 times after adjusting for known factors associated with PPCS.

The interquartile range for those in our delayed diagnosis cohort to receive their concussion diagnosis was between 2 and 6 days after their initial visit. This first week after injury has shown to be critical in concussion recovery. Lishchynsky et al, in a study of 30 youth hockey players, found those participating in vigorous physical activity in the first week after injury were more likely to have prolonged recovery when compared with those engaging in low to moderate physical activity.²⁶ In a study of 335 children and adolescents, Brown et al. showed that those who engaged in the highest quartile of cognitive activity in the first 3 weeks after injury were at highest risk for prolonged symptoms.²⁷ While our patients’activity levels during the time period between their initial and follow-up visits is unknown, it is possible that those who did not receive a concussion diagnosis returned to normal levels of activity and were not following the graduated return to physical or cognitive activity currently recommended in expert guidelines.^4,5 At our institution, children who do not receive concussion diagnoses generally do not receive the written instructions on graduated return to activity that our concussion patients receive; additionally, beyond provider anticipatory guidance, the actual diagnosis of concussion may prompt more graduated return to activity. Multiple studies have found an association with subspecialist evaluation within the first week of injury and improved recovery times. Desai et al., in evaluating gender differences in recovery, found that in 192 concussed youth presenting to a specialty clinic, the association between gender and time to symptom resolution disappeared when controlling for time from injury to specialty evaluation (dichotomized to within 1 week and greater than 1 week).¹⁷ Kontos et al. found that children seen in a sports medicine clinic more than 7 days from injury were 5.8 times more likely to have prolonged symptoms compared to those seen within a week.¹⁸ The importance of early specialist evaluation is likely multifactorial, including individualized rest and return to activity guidance, and the delivery of targeted assessments and rehabilitation therapies that have been shown to be beneficial early after injury, including supervised aerobic exercise and vestibular and oculomotor rehabilitation exercises.^14–16 In our care network, we have attempted to standardize this graduated return to activity from areas outside of specialty settings (including the ED, UC, and primary care) with success.¹⁹ While the majority of patients in our cohort did not receive early subspecialty care, the inclusion of specialist recommendations in ED/UC anticipatory guidance may have impacted the recovery times for our immediate diagnosis group.

The ED/UC evaluation, including physical examination techniques performed and symptoms documented, differed between our delayed and immediate diagnosis groups. We found a significantly higher proportion of those patients who received immediate diagnoses received a visio-vestibular examination (VVE) in the ED. The majority of the patients in the delayed diagnosis group who had VVE examinations performed at their index visit had no abnormalities, with the majority of those patients having abnormalities at follow-up visit. In addition, over 2/3 of our patients with delayed diagnoses who did not receive VVE at their index visit had abnormal findings at their first follow-up visit (with the possibility that these exams were also abnormal at the ED/UC visit, if they had been performed). Taken together, this suggests the VVE has a role in both immediate and subacute diagnosis. Visio-vestibular findings can evolve over time,⁷ and while prior studies have postulated multiple phenotypes of concussion exist, stratified by symptom type (for example cognitive, migraine, etc),²⁸ our findings also suggest that there may be a phenotype of concussion marked not by the type of findings, but rather the time to onset of findings (in this case, delayed onset and delayed recovery). If this were to be the case, it emphasizes further the important role of the VVE both in diagnosis and prognosis for pediatric concussion. Importantly, previous studies have shown the VVE is feasible for implementation by both emergency medicine physicians and pediatricians, particularly with the assistance of clinical support tools in the EHR.^10,19

A higher number of concussion symptoms, in a list of 22 symptoms adopted from the postconcussion symptom inventory (PCSI),²³ were documented in those with immediate diagnoses compared to those with delayed diagnoses. Of the 6 most common symptoms present at a follow-up visit but not documented at the initial visit in those with delayed diagnoses, 4 (difficulty concentrating, feeling foggy, fatigue, and irritability) were in cognitive, sleep, and emotional symptom categories; the other 2, photophobia and phonophobia, may not be considered “classic” physical symptoms of concussion. Very few of patients with delayed diagnoses reported visual problems, headache, or vomiting at the follow-up visit without having those symptoms documented at the initial visit, as these physical symptoms are likely considered more “classic” concussion manifestations. These findings support the use of standardized concussion symptom scales in order to ensure that a comprehensive list of concussion-related symptoms is systematically assessed and documented.^5,29 A prospective study by Eisenberg et al found that cognitive, sleep, and emotional symptoms were more likely to be present on follow-up questioning compared to initial presentation.³⁰ Therefore, similar to our hypothesis as to a phenotype of delayed presentation of VVE deficits, there may similarly be a phenotype of concussion that presents with delayed onset and delayed resolution of symptoms.

Finally, we found our subjects with immediate diagnoses were more likely to have a sports mechanism of injury when compared to those with delayed diagnoses. Multiple previous studies have shown children who present to the ED with a sports injury mechanism are more likely to receive a concussion diagnosis,^9,31 and children with a sports mechanism of injury are more likely to receive concussion-specific testing in the ED setting.¹⁰ Interestingly, previous studies have shown that non-sports injury mechanisms, particularly assault-related injuries and injuries sustained in motor vehicle collisions, are more strongly associated with persistent concussion symptoms, further emphasizing the need for accurate diagnoses in concussions sustained outside of the competitive sport setting.³²

Prior studies have attempted to quantify the proportion of children who receive an accurate diagnosis of concussion from the ED setting. Boutis et al. found that only 45% of 443 children who presented to the ED of a tertiary care medical center, who met criteria for concussion diagnosis per the International Consensus Statement on Concussion in Sport at the time,³³ received a concussion diagnosis.⁹ More recently, in a multicenter study of nearly 3,000 concussed youth, Boutis et al. found 79% of children meeting similar guidelines received an ED concussion diagnosis (though in this subsequent study, physicians were informed that the subject met concussion diagnostic criteria, likely having some influence on the proportion ultimately diagnosed).³¹ It should be noted that the multicenter study by Boutis et al. found that those not receiving an ED diagnosis of concussion were less likely to develop persistent symptoms, in contrast to our study. This is likely a result of study design, in that their comparison was between patients enrolled from the same cohort prospectively; in their study, those who did not receive a concussion diagnosis in the ED were likely truly not concussed, compared to our group of delayed diagnosis patients who, by design, were concussed but did not receive the diagnosis at the initial ED/UC visit.

In our current study, it is important to note that we did not include the cohort of children who were evaluated in the ED/UC for a head injury, were not diagnosed with a concussion, and who then did not seek any follow-up care (ie were never diagnosed with a concussion). This population is also of important consideration, as they could be at risk for overly restrictive anticipatory guidance if criteria for concussion diagnosis in the ED setting is intentionally liberal.³⁴ These concerns are not unfounded, as recent literature has emphasized the deleterious effect of prolonged strict rest immediately after injury, both in terms of symptom resolution time as well as the mental health of concussed youth.^12,35 This further emphasizes the need for accurate diagnosis, including limiting over-diagnosis, from the acute setting.

Limitations

There are several limitations to this study. First, as this was a retrospective study, only data recorded in the EHR were available for abstraction. It is possible symptoms were inquired about or examinations performed that were not documented in the medical record, and, while our written discharge instructions are relatively standardized, we cannot be certain what type of anticipatory guidance was verbally given to families. Secondly, not all patients had complete data on recovery outcomes, though the vast majority of those lost to follow-up had their final visit within 10 days of the injury. As we only included patients followed within our health system, while it is possible they sought follow-up care with another provider during recovery, it is more likely that these patients recovered without needing further follow-up. Our sensitivity analysis classifying the patients whose last follow-up visit was within 10 days of injury as not having persistent symptoms (rather than PPCS status unknown) found an even stronger association between delayed diagnosis and persistent symptoms. Finally, as our data were collected from the emergency departments, urgent care centers, primary care sites, and specialty sites of our tertiary care academic center, they may not be generalizable to the general population cared for in community practices. However, we note the 4 urgent care centers included in this study are all located at suburban, community sites.

CONCLUSION

In conclusion, our study found children evaluated for head injury who do not receive an immediate concussion diagnosis in the acute setting, but are diagnosed with concussion at a later date, may be at risk for persistent concussion symptoms. Performance of a visio-vestibular examination and the utilization of a standardized concussion symptom scale to ensure all relevant symptoms are queried may aid in the comprehensive identification of concussed youth shortly after the injury. While there are several symptom scales available, consistently having access to a list of standardized symptoms which one systematically assesses is likely more important than the specific scale used. Future, prospective studies exploring the etiology of prolonged symptoms in those with a delay to diagnosis, and whether there are modifiable factors present, are needed to further elucidate the effect of early interventions on mitigating persistent concussion symptoms.

ARTICLE SUMMARY

1). Why is this topic important?

Concussion is a common pediatric injury in the emergency department setting, with a risk for under-diagnosis acutely. However, the impact of delayed diagnosis from this setting on recovery outcomes is unknown.

2). What does this study attempt to show?

We compared visit characteristics and outcomes for children with immediate concussion diagnosis from the ED setting with those whose diagnosis was delayed (patients seen for head injury and discharged without a concussion diagnosis who returned to care and subsequently received the diagnosis).

3). What are the key findings?

Those with an immediate diagnosis had more symptoms documented at the initial visit and a higher likelihood of receiving a concussion-specific physical examination. Those with a delayed diagnosis had significantly more medical visits during recovery, significantly longer average time to symptom resolution, and a significantly higher likelihood of having persistent concussion symptoms.

4). How is patient care impacted?

Concussion-specific physical examination techniques and standard symptom scales could aid in accuracy of pediatric concussion diagnosis from the acute setting, with a potential to improve outcomes and minimize persistent symptoms.

Abbreviations

CI

confidence interval

ED

emergency department

EHR

electronic health record

ICD-10

International Classification of Diseases, 10th Revision

IQR

interquartile range

OR

odds ratio

PCSI

Post Concussion Symptom Inventory

PPCS

persistent concussion symptoms

UC

urgent care

VVE

visio-vestibular examination