Abstract
Background:
Traumatic spinal cord injury (tSCI) is a debilitating neurological condition often associated with lifelong disability. Despite this, there are limited data on pediatric tSCI epidemiology in the United States.
Objectives:
Our primary objective was to estimate tSCI hospitalization rates among children, including by age, sex, and race. Secondary objectives were to characterize tSCI hospitalizations and examine associations between sociodemographic characteristics and tSCI etiology.
Methods:
We used the 2016 Kids’ Inpatient Database to examine tSCI hospitalizations among children (<21 years). Descriptive statistics were used to report individual and care setting characteristics for initial tSCI hospitalizations. We used Census Bureau data to estimate tSCI hospitalization rates (number of pediatric tSCI hospitalizations / number of US children) and logistic regression modeling to assess associations between documented sociodemographic characteristics and injury etiology.
Results:
There were 1.48 tSCI admissions per 100,000 children; highest rates of hospitalization involved older (15–20 years), male, and Black children. Hospitalization involving male (adjusted odds ratio [AOR] 0.43; 95% CI, 0.33–0.58) or Black (AOR 0.37; 95% CI, 0.25–0.55) children were less likely to involve a motor traffic accident. Hospitalizations of Black children were significantly more likely to have a diagnosis of tSCI resulting from a firearm incident (AOR 18.97; 95% CI, 11.50–31.28) or assault (AOR 11.76; 95% CI, 6.75–20.50) compared with hospitalizations of White children.
Conclusion:
Older, male, and Black children are disproportionately burdened by tSCI. Implementation of broad health policies over time may be most effective in reducing pediatric tSCI hospitalizations and preventable injuries.
Keywords: epidemiology, hospitalization, pediatric, traumatic spinal cord injury
Introduction
Traumatic spinal cord injury (tSCI) is a devastating condition that may adversely impact health and quality of life. Historically, the annual incidence of tSCI among children has ranged from 14 to 25 per million in the United States.1–3 Individuals living with tSCI often have complex health care needs, which in turn may be associated with multimorbidity. Moreover, nearly half of individuals living with tSCI may experience one or more mental health disorders, such as depression, anxiety, or posttraumatic stress disorder.4 Oftentimes, economic costs of tSCI and future rehabilitation are borne by the public.5 Therefore, despite the low incidence of pediatric tSCI, the burden of such injury may be substantial.
Prior research suggests differences in pediatric tSCI incidence by sociodemographic factors.6,7 In particular, varied incidences of pediatric tSCI by race were observed, with Blacks reported to experience higher rates of tSCI than native Americans, Hispanics, and Asians.7 Black children are also reported as having more severe spinal injuries relative to other children, with a greater proportion experiencing firearm injuries.8
Although research has explored descriptive analyses of tSCI among children, few studies have examined characteristics associated with injury etiology. A US study using 2009 data modeled associations between sociodemographic factors and mortality from spinal injuries8; however, investigators did not assess relationships between examined factors and spinal injury cause. Given the preventable nature of most tSCIs,5 this is a critical gap in the literature that may inform future policies.
Despite the considerable personal and economic burden of tSCI, there remains limited data on tSCI hospitalizations among pediatric populations in the United States. Therefore, our primary study objective was to estimate tSCI hospitalization rates among US children, including by age, sex, and race. Our secondary objectives were to characterize hospitalizations for tSCI among pediatric populations in the United States and examine associations between select demographic characteristics (age, sex, and race) and tSCI etiology, focusing on accidental and violent causes of injury.
Methods
Ethics statement
Our study included analyses of publicly available datasets that do not contain any direct patient identifiers, and therefore it was exempt from University of British Columbia research ethics board review. Furthermore, it complies with conditions outlined in the United States Agency for Healthcare Research and Quality (AHRQ) Healthcare Cost and Utilization Project (HCUP) Data Use Agreement. Our study also complies with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.9
Study design and setting
Using inpatient pediatric data from the United States between January 1, 2016, and December 31, 2016, we performed a cross-sectional study to examine tSCI hospitalization rates and to assess whether age, sex, and race were associated with tSCI etiology.
Data source
Our study used the 2016 Kids’ Inpatient Database (KID), which is the largest publicly available pediatric (<21 years of age) inpatient care database in the United States that comprises administrative claims data for all payers.10 At the time of our study, the 2016 dataset was the most recent and only KID that used International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) coding. The KID is one of many databases and analytical tools developed and managed by the HCUP, which is a federal-state-industry partnership sponsored by the AHRQ. Data within the 2016 KID were sampled from 4200 US community hospitals (short-term, non-Federal, general and specialty hospitals; excluding hospital units of other institutions and rehabilitation hospitals) and includes weights for estimating national hospitalizations. When weighted, there are more than seven million pediatric hospitalizations within the 2016 KID. Compared with other HCUP datasets (such as the Nationwide Emergency Department Sample [NEDS] and the Nationwide Readmissions Database [NRD]), an advantage of the KID is that it includes race data in addition to in-depth sociodemographic information.
Clinical data in the 2016 KID are documented using ICD-10-CM codes. These ICD-10-CM codes are aggregated into clinically meaningful categories (530 clinical categories across 21 body systems) using HCUP’s clinical classifications software refined (CCSR).11 Diagnosis codes (ICD-10-CM and/or CCSR), including tSCI and injury etiology codes, examined within our study are outlined in eTables 1 and 2. Datasets analyzed for this study are available from the HCUP (www.hcup-us.ahrq.gov).
Study hospitalizations
Eligible study hospitalizations included hospitalizations where an incident tSCI (initial spinal cord injury encounter; CCSR INJ009) was recorded as a primary or secondary diagnosis. From these admissions, sacral spinal cord (ICD-10-CM: S34131A, S34132A, S34139A) or cauda equina-only (ICD-10-CM: S343XXA) injuries were excluded. We then excluded hospitalizations (1) where patient sociodemographic information (sex, race, income quartile, or payer status) was missing; (2) with invalid death data; (3) that were for elective reasons or where elective data was missing; and (4) that had no evidence of emergency department care or transfer from another health care facility.
Patient and hospital characteristics
It is unlikely that unique children experienced multiple incident tSCIs within the same calendar year; however, some hospitalizations included in our study may be readmissions. We reported the following sociodemographic characteristics for each tSCI hospitalization: patient age, sex, race, median household income, and primary payer. Clinical factors that were examined included length of stay, SCI location (cervical, thoracic, and/or lumbar), inpatient death, and discharge disposition, while hospital characteristics assessed included census region, number of beds, location/teaching status, and control/ownership.
Study outcomes
Our primary outcome was incident tSCI hospitalization. Secondary outcomes included tSCI etiology, categorized as accidental or violent. Accidental injuries were grouped into two categories: (1) all unintentional tSCIs, including but not limited to falls, sport-related injuries, occupational injuries, and traffic accidents; and (2) traffic accidents resulting in injury to drivers, passengers, or pedestrians. Violent injuries were defined as (1) involving a firearm, including but not limited to accidents, assaults, self-harm, and legal intervention; and (2) assaults, which included child neglect and abuse, assaults involving firearms, and other external causes of harm.
Statistical analyses
Survey weighting methods and HCUP-provided weights were used to estimate the total number of pediatric hospitalizations for tSCI in 2016. Patient and hospital characteristics, as well as injury etiologies, were reported using descriptive statistics, including weighted counts and pertinent measures of variability.
We used 2016 US Census Bureau data to compute the overall annual rate of tSCI hospitalizations, expressed as hospitalizations per 100,000 children, and rates of tSCI hospitalization by age (0–4 years, 5–9 years, 10–14 years, and 15–20 years), sex, and race (White, Black, and other).12,13 For secondary analyses, we used unconditional multivariable logistic regression modeling to examine associations between sociodemographic characteristics (age, sex, and race) and injury etiology. All models incorporated survey weights and the strata and clustering of patients and hospitals to generate precise variance estimates. Age, sex, and race variables were included in all statistical models, as these covariates were presumed a priori based on clinical knowledge to confound modeled relationships.
All analyses were completed using SAS v9.4 (SAS Institute Inc., Cary, NC), with an alpha of 0.05 as the threshold for significance. We used GraphPad v8.3.1 (GraphPad Software, San Diego, CA) to visually present tSCI hospitalization rates. As required, small cell counts (≤10) were suppressed following the HCUP Data Use Agreement.
Results
Traumatic SCI hospitalizations
Approximately 0.03% (n = 1673) of all pediatric hospitalizations between January 1, 2016, and December 31, 2016, were for incident tSCI (Figure 1). After applying study-specific exclusions, including the removal of elective hospitalizations (n = 69) and admissions with no evidence of emergency department presentation or transfer from another care facility (n = 130), 1281 distinct tSCI hospitalizations were retained for our analyses.
Figure 1.
Selection of pediatric traumatic spinal cord injury hospitalizations. KID = Kids’ Inpatient Database; SCI = spinal cord injury. aThere were no hospitalizations with injuries documented to both the cauda equina and the sacrum.
Most tSCI hospitalizations were among older children (median age [interquartile range]: 17.0 [4.0] years), males (72.6%), and children identified as White (51.8%) (Table 1). A large proportion of the cohort had a median household income in the first quartile ($1 to 42,999; 34.7%), and almost half had private medical insurance (48.3%).
Table 1.
Sociodemographic, clinical, and hospital characteristics documented during pediatric spinal cord injury hospitalizations, 2016 (N = 1281)
| Characteristic | Overall n (%) |
|---|---|
| Sociodemographic characteristics | |
| Age (years), median (IQR) | 17.0 (4.0) |
| 0–4 | 114 (8.9) |
| 5–9 | 64 (5.0) |
| 10–14 | 132 (10.3) |
| 15–20 | 971 (75.8) |
| Sex | |
| Female | 351 (27.4) |
| Male | 930 (72.6) |
| Race | |
| White | 663 (51.8) |
| Black | 312 (24.4) |
| Hispanic | 221 (17.2) |
| Othera | 85 (6.6) |
| Median household income ($) | |
| First quartile (1–42,999) | 444 (34.7) |
| Second quartile (43,000–53,999) | 311 (24.3) |
| Third quartile (54,000–70,999) | 294 (23.0) |
| Four quartile (71,000+) | 231 (18.0) |
| Primary Payer | |
| Medicare | ## |
| Medicaid | 511 (39.9) |
| Private insurance | 619 (48.3) |
| Self-pay | ## |
| Otherb | 82 (6.4) |
| Clinical characteristics | |
| Length of stay (days), median (IQR) | 9.0 (15.0) |
| Location of SCIc | |
| Cervical | 622 (48.5) |
| Thoracic | 461 (36.0) |
| Lumbar | 262 (20.4) |
| Died during hospitalization | 55 (4.3) |
| Discharge disposition | |
| Discharged home/self care | 482 (37.6) |
| Transfer: short-term hospital | 86 (6.7) |
Table 1.
Sociodemographic, clinical, and hospital characteristics documented during pediatric spinal cord injury hospitalizations, 2016 (N = 1281) (cont.)
| Characteristic | Overall n (%) |
|---|---|
| Transfer: other type of facility | 617 (48.1) |
| Home health care | ## |
| Against medical advice | ## |
| Died in hospital | 55 (4.3) |
| Hospital characteristics | |
| Hospital census region | |
| Northeast | 188 (14.6) |
| Midwest | 262 (20.4) |
| South | 568 (44.3) |
| West | 264 (20.6) |
| Hospital bed sized | |
| Small | 88 (6.9) |
| Medium | 238 (18.6) |
| Large | 955 (74.6) |
| Hospital location/teaching status | |
| Rural | ## |
| Urban nonteaching | ## |
| Urban teaching | 1190 (92.9) |
| Hospital control/ownership | |
| Government, non-federal | 257 (20.1) |
| Private, not-for-profit | 945 (73.8) |
| Private, investor-owned | 79 (6.2) |
Note: IQR = interquartile range; KID = Kids’ Inpatient Database; SCI = spinal cord injury.
Includes Asian/Pacific Islanders, Native Americans, and other races.
Includes no charge.
Segment of spinal cord where injury was documented; 4.8% of hospitalizations had injuries documented to two or more spinal cord segments.
Hospital bed size describes overall a hospital’s size based on its number of short-term acute care beds, location, and teaching status. There is no fixed threshold used to denote small, medium, and large bed size hospitals. Additional details on hospital bed size are available on the HCUP website (www.hcup-us.ahrq.gov).
## Data suppressed; 10 or fewer observations in some cells.
The majority of tSCI hospitalizations involved the cervical spinal cord (48.5%); few hospitalizations resulted in death (4.3%). The median length of inpatient stay was relatively short (9.0 days; IQR 15.0 days), with nearly all discharges occurring within 30 (88.3%) or 60 (97.5%) days. Discharge to another facility (48.1%), which included inpatient rehabilitation and skilled nursing facilities, was more common than discharges to home (37.6%). A higher proportion of hospitalizations occurred in hospitals in the South (44.3%), hospitals with a large number of short-term acute care beds (74.6%), and private, not-for-profit hospitals (73.8%). Almost all tSCI hospitalizations were to urban academic medical centers (92.9%) (Table 1).
Hospitalization rates
The overall rate of tSCI hospitalization (total number of pediatric tSCI hospitalizations / total number of US children) within our study was 1.48 admissions per 100,000 children. Stratified tSCI hospitalization rates by age, sex, and race are shown in Figure 2. The highest rates of tSCI hospitalization involved older (15–20 years: 3.82 per 100,000 children), male (2.11 per 100,000 children), and Black (2.39 per 100,000 children) children. The rate of tSCI hospitalization involving White children (1.48 per 100,000 children) was equal to the overall tSCI hospitalization rate.
Figure 2.
Traumatic spinal cord injury hospitalizations per 100,000 children by (A) age, (B) sex, and (C) race.
Accidental injuries
Most tSCI hospitalizations were documented as resulting from an accidental cause (73.3%), while more than a third (34.1%) of admissions were attributed to motor traffic accidents (Table 2). Relative to hospitalizations of the youngest children (0–4 years), hospitalizations of children between the age of 10 and 14 years were less likely to be documented as related to a motor vehicle traffic accident (adjusted odds ratio [AOR] 0.35; 95% CI, 0.17–0.72). Hospitalizations resulting from motor vehicle traffic accidents were less likely to involve male (AOR 0.43; 95% CI, 0.33–0.58) or Black (AOR 0.37; 95% CI, 0.25–0.55) children. Accidental tSCI hospitalizations were also less likely to involve a Black child (AOR 0.29; 95% CI, 0.20–0.41).
Table 2.
Associations between sociodemographic characteristics documented during traumatic spinal cord injury hospitalizations and accidental injury etiology
| Accidental | Motor vehicle traffic | |||||
|---|---|---|---|---|---|---|
| Yes (n = 939) | No (n = 342) | Yes (n = 437) | No (n = 844) | |||
| Characteristic | n (%) | n (%) | AOR (95% CI) | n (%) | n (%) | AOR (95% CI) |
| Age, years | ||||||
| 0–4 | 84 (8.9) | 29 (8.5) | Reference | 45 (10.3) | 69 (8.2) | Reference |
| 5–9 | 51 (5.4) | 13 (3.8) | 1.40 (0.59 – 3.32) | 29 (6.6) | 35 (4.1) | 1.15 (0.56 – 2.36) |
| 10–14 | 106 (11.3) | 26 (7.6) | 1.30 (0.57 – 3.00) | 25 (5.7) | 108 (12.8) | 0.35 (0.17 – 0.72)* |
| 15–20 | 697 (74.2) | 274 (80.1) | 0.85 (0.44 – 1.65) | 338 (77.3) | 633 (75.0) | 1.03 (0.62 – 1.70) |
| Sex | ||||||
| Female | 266 (28.3) | 85 (24.9) | Reference | 167 (38.2) | 184 (21.8) | Reference |
| Male | 673 (71.7) | 257 (75.1) | 1.01 (0.73 – 1.39) | 270 (61.8) | 660 (78.2) | 0.43 (0.33 – 0.58)** |
| Race | ||||||
| White | 538 (57.3) | 125 (36.5) | Reference | 254 (58.1) | 410 (48.6) | Reference |
| Black | 173 (18.4) | 139 (40.6) | 0.29 (0.20 – 0.41)** | 57 (13.0) | 255 (30.2) | 0.37 (0.25 – 0.55)** |
| Hispanic | 168 (17.9) | 53 (15.5) | 0.73 (0.49 – 1.08) | 93 (21.3) | 128 (15.2) | 1.15 (0.81 – 1.63) |
| Othera | 60 (6.4) | 25 (7.3) | 0.57 (0.31 – 1.03) | 32 (7.3) | 52 (6.2) | 1.04 (0.59 – 1.84) |
Note: AOR = adjusted odds ratio; CI = confidence interval.
Includes Asian/Pacific Islanders, Native Americans, and other races.
*p < .01.
**p < .001.
Violent injuries
Firearms (18.7%) and assault (12.9%) were commonly documented in pediatric tSCI hospitalizations (Table 3). The likelihood of a firearm-related SCI increased with age, with hospitalizations of children between the ages of 15 and 20 years having 16.06 times the odds (95% CI, 3.88–66.38) of having documentation of a firearm-related SCI compared with hospitalizations of children ages 0 to 4 years. Moreover, hospitalizations of Black children were significantly more likely to have a diagnosis of tSCI as a result of an incident involving a firearm (AOR 18.97; 95% CI, 11.50–31.28) or assault (AOR 11.76; 95% CI, 6.75–20.50) compared with hospitalizations of White children.
Table 3.
Associations between sociodemographic characteristics documented during traumatic spinal cord injury hospitalizations and violent injury etiology
| Firearm | Assault | |||||
|---|---|---|---|---|---|---|
| Yes (n = 239) | No (n = 1042) | Yes (n = 165) | No (n = 1116) | |||
| Characteristic | n (%) | n (%) | AOR (95% CI) | n (%) | n (%) | AOR (95% CI) |
| Age | ||||||
| 0–4 | ## | ## | Reference | ## | ## | Reference |
| 5–9 | ## | ## | 1.11 (0.09 – 13.28) | ## | ## | 0.27 (0.03 – 2.39) |
| 10–14 | 14 (5.9) | 118 (11.3) | 7.24 (1.42 – 36.96)* | ## | ## | 0.45 (0.11 – 1.90) |
| 15–20 | 220 (92.1) | 751 (72.1) | 16.06 (3.88 – 66.38)** | 149 (90.3) | 822 (73.7) | 2.24 (0.84 – 5.95) |
| Sex | ||||||
| Female | 38 (15.9) | 313 (30.0) | Reference | 30 (18.2) | 321 (28.8) | Reference |
| Male | 200 (83.7) | 730 (70.1) | 1.48 (0.91 – 2.40) | 135 (81.8) | 795 (71.2) | 1.19 (0.72 – 1.96) |
| Race | ||||||
| White | 32 (13.4) | 631 (60.6) | Reference | 24 (14.5) | 639 (57.3) | Reference |
| Black | 145 (60.7) | 167 (16.0) | 18.97 (11.50 – 31.28)** | 95 (57.6) | 218 (19.5) | 11.76 (6.75 – 20.50)** |
| Hispanic | 41 (17.2) | 180 (17.3) | 5.37 (3.01 – 9.60)** | 32 (19.4) | 189 (16.9) | 4.98 (2.55 – 9.73)** |
| Othera | 20 (8.4) | 64 (6.1) | 6.58 (3.44 – 12.59)** | 14 (8.5) | 71 (6.4) | 5.14 (2.18 – 12.12)** |
Note: AOR = adjusted odds ratio; CI = confidence interval.
Includes Asian/Pacific Islanders, Native Americans, and other races.
*p < .05.
**p < .001.
## Data suppressed; 10 or fewer observations in some cells.
Discussion
Because tSCI among children is often preventable and associated with lifelong disability, it is imperative to understand pediatric tSCI epidemiology so that injury risk mitigation strategies may be successfully implemented. Using a national pediatric inpatient database from the United States, we characterized hospitalizations for incident tSCI in 2016 and examined whether select sociodemographic characteristics (age, sex, and race) were associated with tSCI etiology, focussing on accidental and violent causes of injury. We found that the overall tSCI hospitalization rate was 1.48 admissions per 100,000 children. Notable differences in tSCI hospitalization rates were observed by age, sex, and race, particularly for hospitalizations involving older, male, and Black children. Key secondary findings were that (1) nearly half of hospitalizations were for cervical SCIs (48.5%); (2) discharge to rehabilitation facilities and specialty care was common (48.1%); (3) tSCI hospitalizations due to motor traffic accidents were less likely to involve male or Black children; and (4) tSCI hospitalizations due to incidents involving firearms or assault were more likely to involve non-White children.
The most recent US estimates of pediatric tSCI incidence by Piatt and Imperato revealed a downward trend among children younger than 18 years of age between 1997 and 2012, with the 2012 annual incidence of SCI estimated to be 1.4 hospitalizations per 100,000 children.1 Their study utilized KID data and thoroughly summarized prior literature on this topic. Despite their well-designed study, an incidence estimate by Vitale and colleagues of tSCI among children younger than 19 years of age (1.99 cases per 100,000 US children per year) based on KID data from 1997 and 2000 is still commonly reported.14 As pointed out by Piatt and Imperato, the earlier overall incidence estimate, as well as previously reported estimates of tSCI incidence by race, may be imprecise due to investigators’ assessment and handling of patient age and missing race data. Irrespective of these potential issues, the estimates are outdated and therefore warrant further investigation. Our study differs from prior investigations in that our tSCI case definition was based on ICD-10-CM (implemented in 2015) instead of ICD-9-CM coding and that we examined pediatric tSCI hospitalizations among children younger than 21 years of age. Furthermore, we excluded elective hospitalizations from our analyses to avoid counting repeat hospitalizations of the same patient, as elective admissions were presumed to be transfers to inpatient rehabilitation facilities following an initial hospitalization for tSCI. Investigators of other studies did not address this issue or did so via different methods (excluded discharges to other acute care facilities). Although not directly comparable to the prior reports, our findings provide evidence to support that the overall incidence of tSCI among children has continued to decline since 2012. Moreover, we show that a greater number of emergent tSCI hospitalizations involve Black children compared with hospitalization involving White children, whereas Vitale and colleagues reported the opposite. A likely explanation for this discrepancy is selection bias attributed to differences in missing race data across KID years (% missing race: 1997, 22.2%; 2000, 16.4%; 2016, 9.1%). Further, we are unable to account for readmission or lower quality of care resulting in readmission, which is more likely to affect Black children15,16; therefore, our findings are likely biased. Nevertheless, our findings suggest that injury prevention and postacute care interventions tailored to older (15–20 years), male, and Black children may be most effective in reducing tSCI hospitalizations.
Consistent with prior US reports,3,17,18 we observed that accidental injuries, motor traffic accidents, and incidents involving firearms and assault were leading causes of tSCI among children. Further, we found that hospitalizations involving males were associated with a lower likelihood of reporting motor traffic accident as tSCI cause. This finding is consistent with a 12-year patient-based study of facial fractures among US children (<19 years) that found that assault and motor vehicle traffic were the leading etiologies of injury for males and females, respectively.19 It is also supported by a 7-year National Trauma Data Bank study in the United States that found that 85.1% of pediatric firearm-related ocular injuries occurred in males.20 Our observed increased association between hospitalizations involving older children (15–20 years) and firearm injury (AOR 16.06; 95% CI, 3.88–66.38) likely highlights a critical age threshold where preventable tSCI injury risk increases. Moreover, observed associations between hospitalization injury etiology and race likely reflect underlying inequities in the determinants of health that contribute to differences in health service utilization.21,22 Hospitalizations due to firearm and other violent injuries may therefore be reduced over time by intervening upon the determinants of health and implementing broad health policies, such as increasing the federal minimum wage and increasing subsidies for health care.
There are a number of strengths to our study. To our knowledge, this study is the first national analysis of pediatric tSCI hospitalizations in the United States following the introduction of ICD-10-CM23 coding, with the last annual population-based incidences of pediatric hospitalization for SCI being reported based on 2012 data and ICD-9-CM diagnostic codes.1 It therefore provides updated benchmark data on the epidemiology of pediatric tSCI upon which future studies may be based. In a limited capacity, it may also be compared to reports of SCI incidence among children from prior years. Exclusion of hospitalizations for elective reasons (n = 69) from our study sample is presumed to minimize bias in reported estimates by reducing the possibility of double counting repeat hospitalizations for the same child. With the decline in percentage of missing race data in the 2016 KID compared with prior KID years (~7% missing race data in our sample), our analyses of tSCI hospitalization and injury etiology by race provide meaningful, although potentially biased, baseline data for additional investigations of racial disparities in pediatric tSCI. Collectively, our study serves to advance knowledge of pediatric tSCI epidemiology and to highlight intervention opportunities that may lead to a decrease in the risk of preventable injury among children.
Certain limitations should be considered when interpreting our findings. Rates reported within our study are of children hospitalized for tSCI; they do not reflect emergency department visits, account for children who died at the scene of trauma, or account for emergent readmission. As such, reported hospitalization rates may under- or overestimate the population-based incidence of tSCI among US children. Furthermore, with the adoption of ICD-10-CM coding, it is challenging to make direct comparisons between our findings and those from studies prior to 2015, as tSCI case definitions vary across studies. Other challenges in comparing our findings to prior reports are the inconsistent categorization of age across studies examining the incidence of tSCI among US children1,2,24–27 and the inappropriate use of HCUP data to measure tSCI incidence by other investigators.28 The effects of possible selection bias must also be contemplated when interpreting our findings, specifically with respect to missing race, emergency department care, and transfer data. Race data are known to be missing in a nonrandom manner within the KID, and the same may be true for missing emergency department care and transfer data.29 Although the proportion of missing data for these variables was relatively low (<10%) within our sample, effects of these missing data on our reported estimates remain unknown. Furthermore, we accounted for presumed confounders in our multivariable models examining associations between sociodemographic characteristics and tSCI etiology, but the possibility of residual bias still exists. Notwithstanding these limitations, our study provides meaningful benchmark data for future investigations into pediatric tSCI.
Conclusion
Overall, we found that tSCI hospitalizations commonly involved older, male, and Black children and that sociodemographic characteristics documented during tSCI hospitalizations were associated accidental and violent injury etiologies. Implementation of broad health policies may therefore be valuable public investments in reducing tSCI hospitalizations. Future studies should examine whether our hospitalization-based findings are comparable with individual-level estimates of pediatric tSCI incidence and injuries associated with pediatric tSCI not resulting in hospitalization and whether similar tSCI hospitalization rates are observed in future years of KID data.
Supplementary Material
Footnotes
Conflicts of Interest
The authors report no conflicts of interest.
REFERENCES
- 1.Piatt J, Imperato N. Epidemiology of spinal injury in childhood and adolescence in the United States: 1997–2012. J Neurosurg Pediatr . 2018;21(5):441–448. doi: 10.3171/2017.10.PEDS17530. [DOI] [PubMed] [Google Scholar]
- 2.Vitale MG, Goss JM, Matsumoto H, Roye DP., Jr Epidemiology of pediatric spinal cord injury in the United States: Years 1997 and 2000. J Pediatr Orthop . 2006;26(6):745–749. doi: 10.1097/01.bpo.0000235400.49536.83. [DOI] [PubMed] [Google Scholar]
- 3.Selvarajah S, Schneider EB, Becker D, Sadowsky CL, Haider AH, Hammond ER. The epidemiology of childhood and adolescent traumatic spinal cord injury in the United States: 2007–2010. J Neurotrauma . 2014;31(18):1548–1560. doi: 10.1089/neu.2014.3332. [DOI] [PubMed] [Google Scholar]
- 4.Migliorini C, Tonge B, Taleporos G. Spinal cord injury and mental health. Aust N Z J Psychiatry . 2008;42(4):309–314. doi: 10.1080/00048670801886080. [DOI] [PubMed] [Google Scholar]
- 5.Tyroch AH, Davis JW, Kaups KL, Lorenzo M. Spinal cord injury. A preventable public burden. Arch Surg . 1997;132(7):778–781. doi: 10.1001/archsurg.1997.01430310092020. [DOI] [PubMed] [Google Scholar]
- 6.National Spinal Cord Injury Statistical Centre Spinal cord injury (SCI) 2016 facts and figures at a glance. J Spinal Cord Med . 2016;39(4):493–494. doi: 10.1080/10790268.2016.1210925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Devivo MJ. Epidemiology of traumatic spinal cord injury: trends and future implications. Spinal Cord . 2012;50(5):365–372. doi: 10.1038/sc.2011.178. [DOI] [PubMed] [Google Scholar]
- 8.Piatt JH., Jr Pediatric spinal injury in the US: Epidemiology and disparities. J Neurosurg Pediatr . 2015;16(4):463–471. doi: 10.3171/2015.2.PEDS1515. [DOI] [PubMed] [Google Scholar]
- 9.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. Ann Intern Med . 2007;147(8):573–577. doi: 10.7326/0003-4819-147-8-200710160-00010. [DOI] [PubMed] [Google Scholar]
- 10.Moore BJ, Freeman WJ, Jiang HJ. Costs of pediatric hospital stays 2016 . HCUP Statistical Brief #250. August 2019. www.hcup-us.ahrq.gov/reports/statbriefs/sb250-Pediatric-Stays-Costs-2016.pdf. [PubMed]
- 11. Healthcare cost and utilization project user guide Clinical classifications software refined (CCSR) for ICD10CM diagnoses v20212 . March 2021. https://www.hcup-us.ahrq.gov/toolssoftware/ccsr/DXCCSR-User-Guide-v2021-2.pdf.
- 12.Barrett M, Coffey R, Levit K. Population denominator data for use with the HCUP databases (updated with 2016 population data) www.hcup-us.ahrq.gov/reports/methods/methods.jsp HCUP Methods Series Report # 2017-04. October 17, 2017.
- 13.US Census Bureau Explore census data . 2021. https://data.census.gov/cedsci/
- 14.American Spinal Injury Association Facts on pediatric spinal cord injury . 2021. https://asia-spinalinjury.org/committees/pediatric/pediatric-committee-news-and-resources/pediatric-spinal-cord-injury-facts/
- 15.Dotson JL, Kappelman MD, Chisolm DJ, Crandall WV. Racial disparities in readmission, complications, and procedures in children with Crohn’s disease. Inflamm Bowel Dis . 2015;21(4):801–808. doi: 10.1097/MIB.0000000000000325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Parikh K, Berry J, Hall M et al. Racial and ethnic differences in pediatric readmissions for common chronic conditions. J Pediatr . 2017;186:158–164. doi: 10.1016/j.jpeds.2017.03.046. e151. [DOI] [PubMed] [Google Scholar]
- 17.Chen Y, Tang Y, Vogel LC, Devivo MJ. Causes of spinal cord injury. Top Spinal Cord Inj Rehabil . 2013;19(1):1–8. doi: 10.1310/sci1901-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Odetola FO, Gebremariam A. Variation in patterns of hospitalization and associated resource use among children with spinal cord injury in the U.S. Injury . 2016;47(5):1123–1127. doi: 10.1016/j.injury.2016.01.031. [DOI] [PubMed] [Google Scholar]
- 19.Hoppe IC, Kordahi AM, Paik AM, Lee ES, Granick MS. Age and sex-related differences in 431 pediatric facial fractures at a level 1 trauma center. J Craniomaxillofac Surg . 2014;42(7):1408–1411. doi: 10.1016/j.jcms.2014.04.002. [DOI] [PubMed] [Google Scholar]
- 20.Weiss R, He C, Gise R, Parsikia A, Mbekeani JN. Patterns of pediatric firearm-related ocular trauma in the United States. JAMA Ophthalmol . 2019;137(12):1363–1370. doi: 10.1001/jamaophthalmol.2019.3562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Krause JS, Dismuke CE, Acuna J et al. Race-ethnicity and poverty after spinal cord injury. Spinal Cord . 2014;52(2):133–138. doi: 10.1038/sc.2013.147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Krause JS, Broderick LE, Saladin LK, Broyles J. Racial disparities in health outcomes after spinal cord injury: Mediating effects of education and income. J Spinal Cord Med . 2006;29(1):17–25. doi: 10.1080/10790268.2006.11753852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Price C, Makintubee S, Herndon W, Istre GR. Epidemiology of traumatic spinal cord injury and acute hospitalization and rehabilitation charges for spinal cord injuries in Oklahoma, 1988–1990. Am J Epidemiol . 1994;139(1):37–47. doi: 10.1093/oxfordjournals.aje.a116933. [DOI] [PubMed] [Google Scholar]
- 24.Mendoza-Lattes S, Besomi J, O’Sullivan C et al. Pediatric spine trauma in the United States--Analysis of the HCUP Kid’s Inpatient Database (KID) 1997–2009. Iowa Orthop J . 2015;35:135–139. [PMC free article] [PubMed] [Google Scholar]
- 25.Bracken MB, Freeman DH, Jr, Hellenbrand K. Incidence of acute traumatic hospitalized spinal cord injury in the United States, 1970–1977. Am J Epidemiol . 1981;113(6):615–622. doi: 10.1093/oxfordjournals.aje.a113140. [DOI] [PubMed] [Google Scholar]
- 26.Griffin MR, Opitz JL, Kurland LT, Ebersold MJ, O’Fallon WM. Traumatic spinal cord injury in Olmsted County, Minnesota, 1935–1981. Am J Epidemiol . 1985;121(6):884–895. doi: 10.1093/oxfordjournals.aje.a114058. [DOI] [PubMed] [Google Scholar]
- 27.Jain NB, Ayers GD, Peterson EN et al. Traumatic spinal cord injury in the United States, 1993–2012. JAMA . 2015;313(22):2236–2243. doi: 10.1001/jama.2015.6250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Fingar KR, Owens PL, Barrett ML, Steiner CA. Using the HCUP databases to study incidence and prevalence . HCUP Methods Series Report #2016-06. December 6, 2016. http://www.hcup-us.ahrq.gov/reports/methods/methods.jsp.
- 29.Healthcare Utilization and Cost Project Introduction to the HCUP Kids’ Inpatient Database (KID) September 2018. https://www.hcup-us.ahrq.gov/db/nation/kid/KID_2016_Introduction.pdf.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.