Age-based differences in the disability of spine injuries in pediatric and adult motor vehicle crash occupants

Abstract

Objective:

The objective was to develop a disability-based metric for quantifying disability rates as a result of motor vehicle crash (MVC) spine injuries and compare functional outcomes between pediatric and adult subgroups.

Methods:

Disability rate was quantified using Functional Independence Measure (FIM) scores within the National Trauma Data Bank—Research Data System for the top 95% most frequent Abbreviated Injury Scale (AIS) 3 spine injuries (14 unique injuries). Pediatric (7–18 years), young adult (19–45 years), middle-aged adult (46–65 years), and older adult (66+ years) MVC occupants with FIM scores available and at least one of the 14 spine injuries were included. FIM scores of 1 or 2 at time of discharge were used to define disability and correspond to full functional or modified dependence in self-feeding, locomotion, and/or verbal expression. Disability rate was evaluated on a per injury basis for each AIS 3 spine injury and calculated as the proportion of cases associated with disability (i.e. FIM of 1 or 2) out of the total cases of that particular injury. Disability rates were calculated with and without the exclusion of cases with severe co-injuries (AIS 4+) to minimize bias from additional non-spinal injuries that could have contributed to disability. Associations between adjusted disability rates and existing mortality rates were investigated.

Results:

Locomotion impairment alone was the most frequent disability type for the top 14 AIS 3 spine injuries (7 cervical, 4 thoracic, and 3 lumbar) across all age groups and spine regions. Adjusted and unadjusted disability rates ranged from 0–69%. Adjusted disability rates increased with age: 14.8 ±10% (mean ±SD) in pediatrics to 16.2 ±6.6% (young adults), 29.2 ±10.9% (middle-aged adults), and 45.0 ±12.2% (older adults). Among all adult populations, adjusted mortality and disability rates were positively correlated (R²>0.24), with disability rates consistently greater than corresponding mortality rates.

Conclusions:

Older adults had significantly greater disability rates associated with MVC spine injuries across all spinal regions. MVC disability rates for pediatrics were considerably lower. Overall, rates of mortality were significantly lower than rates of disability. The adjusted disability rates developed can supplement existing injury metrics by accounting for age- and location-specific functional implications of MVC spine injuries.

Introduction

Disability and mortality rates associated with motor vehicle crash (MVC) injuries can vary based on the occupant’s age (Weaver et al. 2013; Doud et al. 2015, 2017; Gaffley et al. 2019). While injury-associated mortality and disability typically increase with age, varying levels of short- and long-term disability are observed for nonfatal injuries across all ages (Gaffley et al. 2019). Although older and middle-aged adults may have a higher incidence of disability following injury, disabling injuries in children and young adults are also concerning due to substantial loss of quality-of-life years.

Spine injuries sustained in MVCs can vary based on seat-belt usage, airbag deployment, vehicle type, and crash type. Despite the use of vehicle safety measures, thoracolumbar fractures are still common in restrained and unrestrained occupants (Ejima et al. 2018). MVCs are one of the most common causes of lumbar injuries and the most common cervical spine injury mechanism is rear-impact MVCs. Safety improvements have reduced the number of severe injuries in MVCs, but traumatic spine injuries are still frequent. Spine injuries often result in significant disability (Currens 2000; Zonfrillo et al. 2013); therefore, it is clinically significant to quantify disability rates associated with MVC-induced spine injuries.

The Abbreviated Injury Scale (AIS) is a consensus-derived and anatomy-based measure to classify and describe injury severity by various dimensions, some of which include threat-to-life, quality-of-life, mortality, tissue damage, and hospitalization. AIS quantifies an injury’s severity on a scale of 1–6, where 1 represents a minor injury unlikely to cause mortality and 6 represents a maximal injury highly likely to cause mortality. However, AIS scores may not fully capture an injury’s impact on quality-of-life, since injuries resulting in life-long impairments are not always life-threatening (Read et al. 2004). Functional Independence Measure (FIM) scores from discharged trauma patients are a viable supplement to AIS to assess disability associated with injuries (Gaffley et al. 2019). The objective of this study was to develop a disability-based metric for MVC spine injuries to support existing metrics by quantifying associated rates of an injury for four age groups: older adults, middle-aged adults, young adults, and pediatrics.

Methods

Spine injuries

With Institutional Review Board approval (IRB#00026698), the National Automotive Sampling System–Crashworthiness Data System (NASS-CDS) was used to identify the top 95% most frequent AIS 3 spine injuries in MVCs across all ages (Association for the Advancement of Automotive Medicine. AIS 1990 Revision, Update 98). NASS-CDS contains detailed data for a representative, random sample of thousands of minor to fatal tow-away crashes in the U.S. Injury data are weighted to provide population-based estimates. Data from years 2000–2011 were utilized since these years were used in a larger study examining the mortality, time sensitivity, predictability, and disability of MVC injuries in children and adults (Doud et al. 2015, 2017). Only AIS 3 spine injuries were analyzed since AIS 1–2 injuries are associated with minimal impairment risk and AIS 4–6 spine injuries were not among the 95% most frequent injuries. Both disabling and non-disabling injuries included as the most frequent AIS 3 spine MVC injuries were assessed to investigate disability rates. Additional references of relevance to this study can be found in the annotated bibliography (Appendix, Online supplementary information).

MVC occupants were stratified into 4 age groups: pediatric (7–18 years), young adult (19–45 years), middle-aged adult (46–65 years), and older adult (66+ years). Pediatrics younger than 7 were excluded from analysis as previous studies have only validated FIM scores for 7–18 year-old children (Dodds et al. 1993; Winthrop et al. 2005). Older adults were classified as 66+ years since 65 years is a common threshold for defining older adults (Gaffley et al. 2019; Schoell et al. 2018). The top 95% most frequent AIS 3 spine injuries across these age groups consisted of 14 injuries (unique AIS codes; Appendix 0003 Table A3, Online supplementary information). However, the 14 injuries were not common to each group individually. If the injury was not among the age group’s top 95% spine injuries, it was indicated as missing data and was not included in the analysis.

Disability rate

To quantify disability rate, the National Trauma Data Bank-Research Data Set (NTDB-RDS) was used which represents the most complete U.S. database of discharged trauma patients. Research databases such as NASS-CDS commonly use AIS, which provide information on injury severity, whereas hospital databases such as the NTDB-RDS use the International Classification of Diseases (ICD) lexicon. A robust algorithm developed by Barnard et al. (2013) was used to map codes from the AIS lexicon to the ICD-9 lexicon by analyzing the frequency of AIS and ICD-9 pairings. This algorithm allowed us to input the AIS codes corresponding to the most frequent AIS 3 spine injuries, identify MVC occupants by ICD codes and then match each occupant’s ICD to the corresponding AIS code.

Disability was defined with the FIM instrument - a widely accepted measure of disability based on the International Classification of Impairment, Disabilities, and Handicaps. This instrument typically contains 18 items (13 motor; 5 cognitive) with a score of 1 (total assistance) to 7 (complete independence) per item. NTDB-RDS provides a truncated form of FIM with three items (self-feed; locomotion; verbal expression) graded from 1 (full functional dependence) to 4 (full functional independence) at the time of patient discharge. Self-feeding encompasses an individual’s ability to perform mechanical movements of eating without assistance. Locomotion includes the ability to walk, use a wheelchair, and navigate stairs unassisted. Verbal expression includes comprehension as well as the ability to express oneself clearly. This shortened FIM has been validated for children ≥7 years (Winthrop et al. 2005), and has shown reliability in adults across raters, settings, and patients (Ottenbacher et al. 1996). A FIM score of 1 or 2 (full functional/modified dependence) on any of the three items was used to classify occupants as disabled.

For each of the top 95% most frequent AIS 3 spine injuries, a maximum AIS (MAIS)-adjusted disability rate (DR_MAIS) was calculated on a per injury basis using Eq. (1) to account for the influence of more severe associated co-injuries. For this calculation, occupants with AIS 4+ co-injuries were excluded. However, AIS ≤ 2 co-injuries were not controlled for, since it was assumed these injuries were less likely to result in disabling outcomes. In addition, MAIS-adjusted mortality rate (MR_MAIS) reported for these spine injuries and age groups in prior studies was compared to DR_MAIS to assess the association between disability and mortality (Weaver et al. 2013; Doud et al. 2015).

$$D{R}_{\mathit{\text{MAIS}}}=\frac{\#\mathit{\text{Disabled}}\mathit{\text{cases}}\mathit{\text{with}}\mathit{\text{MAIS}}=3\mathit{\text{that}}\mathit{\text{sustained}}a\mathit{\text{given}}\mathit{\text{injury}}}{\#\mathit{\text{Cases}}\mathit{\text{with}}\mathit{\text{MAIS}}=3\mathit{\text{that}}\mathit{\text{sustained}}a\mathit{\text{given}}\mathit{\text{injury}}}$$ (1)

Unadjusted disability rates, without exclusion of severe co-injuries, were also calculated (Appendix Tables A1, A3, Online supplementary information). Disability rates could range from 0–1 (0–100% disability rate). Linear regression was used to compare disability rate between age groups and with mortality rate. Analysis of variance and pairwise comparisons (student’s t-tests) were used to examine differences in disability rate between age groups and fracture locations (cervical; thoracic; lumbar) with α = 0.05. Statistical analyses were performed using JMP Pro 13.0 (SAS Institute, Cary, NC).

Results

Adjusted disability rate (DR_MAIS)

A total of 30,855 occupants who were alive at discharge and 7 years or older sustained at least one of the top 95% AIS 3 spine injuries. FIM scores were available for 54.5% (n = 16,822) of these occupants. Demographic details on this cohort are provided (Appendix 0002 Table A2, Online supplementary information). Exclusion of AIS 4+ co-injuries resulted in a total sample size of 11,919 cases for DR_MAIS calculations. Among this sample, classification of disability using FIM scores revealed that 11.8% of pediatric (n = 124), 13.5% of young adult (n = 833), 20.4% of middle-aged adult (n = 609), and 39.1% of older adult (n = 670) occupant injuries resulted in disability. Sample sizes for each age group along with DR_MAIS mean, standard deviation, and IQR are reported (Table 1). DR_MAIS for the top 14 AIS 3 spine injuries ranged from 0% to 69% and, on average, increased with age. Overall, DR_MAIS of older adults was least correlated with DR_MAIS of pediatrics (R² <0.01), followed by young adults (R² = 0.06), and middle-aged adults (R² = 0.29). On average, DR_MAIS was higher in older adults compared to the other age groups as indicated by most injuries falling above the equivalency line in Figure 1. The most frequent type of disability for the top 14 AIS 3 spine injuries among each age group was locomotion impairment alone (Appendix 0004 Table A4, Online supplementary information). Locomotion was still the most common disability when stratifying the spine injuries by region (cervical; thoracic; and lumbar).

Table 1.

Sample size, mean, and standard deviation of MAIS-adjusted disability rate (DR_MAIS) with IQRs for top 95% AIS 3 spine injuries in each age group with FIM scores available.

Age group	NFIM	DRMAIS (%)
		Mean ± SD	IQR
Pediatric, 7–18 yrs	1051	14.8 ± 10.0	8.7–17.5
Young adult, 19–45 yrs	6170	16.2 ± 6.6	13.5–16.9
Middle-aged, 46–65 yrs	2983	29.2 ± 10.9	20.9–31.4
Older adult, 66+ yrs	1715	45.0 ± 12.2	37.0–46.0

Figure 1.

DR_MAIS for AIS 3 spine injuries for older adults versus younger occupants (<65 years). DR_MAIS for each spinal injury region in the older adult group is plotted against the DR_MAIS of that injury region for the pediatric (n = 8 injuries), young adult (n = 9 injuries), and middle-aged groups (n = 10 injuries). The equivalency line is plotted as a dashed line. 7–18 yrs = pediatric, 19–45 yrs = young adult, and 46–65yrs = middle-aged adult.

DR_MAIS by age and spine region

Seven cervical injuries were in the top 14 AIS spine injuries (Appendix 0003 Table A3, Online supplementary information). Among these, cervical vertebra major compression fracture was associated with the highest DR_MAIS for all age groups except for pediatrics. Cervical facet fractures resulted in the least amount of DR_MAIS among the cervical spine injuries across all age groups except older adults. Analysis of variance revealed that DR_MAIS for cervical spine injuries differed between all age groups (p < 0.0001), with higher DR_MAIS in older adults (μ = 36.7%), followed by middle-aged adults (μ = 19.7%) (Figure 2). Pair-wise comparisons revealed differences in cervical spine injury DR_MAIS between all age groups (all p < 0.03), except for the pediatric versus young adult comparison (p = 0.28).

Figure 2.

DR_MAIS by age group for cervical, thoracic, and lumbar spinal injuries. Of the three lumbar injuries included in the study, data is only available for one in the young adult age group. Disability is greatest for older adults and lowest for pediatric patients and young adults. Box plots span the first through third quartile ranges with the median DR_MAIS indicated by a horizontal line and the mean DR_MAIS indicated by an “x.” Whiskers designate the minimum and maximum DR_MAIS in cases where these values exceed the first and third quartiles.

Among the four thoracic spine injuries, the injury resulting in the highest DR_MAIS differed by age group. DR_MAIS for thoracic spine injuries differed between all age groups (p < 0.02), with significantly higher disability among older adults (μ=51%). Among the three lumbar injuries, lumbar lamina fractures resulted in the highest DR_MAIS among all age groups. Average DR_MAIS for lumbar injuries differed between all age groups (p < 0.01), with higher disability in older adults (μ=43.3%), followed closely by younger adults (μ=30.8%) and middle-aged adults (μ=30.8%). Due to the limited thoracic and lumbar data points, pair-wise comparisons among age groups were not performed for either of these spinal regions.

The pediatric age range of 7–18 years encompasses many different developmental stages. As a result, refining this age group into early- and late-stage adolescents for a sub-analysis of DR_MAIS is of clinical significance. The American Academy of Pediatrics classifies 7–14 year-olds as early-adolescents and 15–18 year-olds as late-adolescents. Within our pediatric group, 6.7% were early-adolescents and 93.3% were late-adolescents. Analysis of variance revealed no difference in DR_MAIS between early- and late-stage adolescents (p = 0.45), with an average DR_MAIS of 12% for early-adolescents and 15% for late-adolescents. Among the youngest occupants (7 year-olds), only 3 of the 14 spine injuries resulted in disability with an overall DR_MAIS of 33.3% (Appendix 0005 Table A5, Online supplementary information).

DR_MAIS correlation with MR_MAIS

The relationship between DR_MAIS and MR_MAIS was examined for the top 14 AIS 3 spine injuries across all age groups (Figure 3). MR_MAIS is the maximum-AIS adjusted mortality rate, which is the proportion of injuries resulting in death, excluding more severe co-injuries (Doud et al. 2015). Among pediatric spine injuries, MR_MAIS was <1%, with MR_MAIS = 0 occurring most frequently and DR_MAIS ranging from 0% to 40%. On average, older adults had higher associated mortality (MR_MAIS = 0–11%) and disability (DR_MAIS 32–69%) for these spine injuries. DR_MAIS versus MR_MAIS correlations were not feasible for pediatrics as less than 3 spine injuries had MR_MAIS >0.

Figure 3.

MAIS-adjusted disability rate (DR_MAIS) versus MAIS-adjusted mortality rate (MR_MAIS) for the top 95% AIS 3 spine injuries by age group. The equivalency line is plotted as a solid line, and mortality rates of zero are plotted on the y-axis.

For adults, correlations were performed with and without exclusion of MR_MAIS=0 injuries. In both cases, DR_MAIS and MR_MAIS had a positive association, with DR_MAIS consistently greater than MR_MAIS. With MR_MAIS = 0 injuries included, correlations were highest among young adults (R² = 0.21), followed by middle-age adults (R² = 0.12) and older adults (R² = 0.11). With exclusion of MR_MAIS = 0 injuries, correlations increased for young adults (R² = 0.32), middle-aged adults (R² = 0.49), and older adults (R² = 0.24). This indicates a stronger association between mortality and disability when focusing on life-threatening injuries. The lower correlations with inclusion of MR_MAIS = 0 provides strong support for consideration of DR_MAIS since there were a high number of cases with negligible mortality but non-negligible disability rates.

Discussion

MVC-induced spine injuries can result in varying levels of mortality and morbidity based on age, injury type, and possible co-injuries. Combining mortality and AIS severity information with disability and functional outcomes allows for better characterization of injury severity and resulting functionality. In this study, adjusted and unadjusted disability rates were quantified for the 14 most common AIS 3 MVC-induced spine injuries across the pediatric, young adult, middle-aged adult, and older adult populations using occupant data with FIM scores available from the NTDB-RDS. For this study, disability was defined by a FIM score of 1 or 2 associated with full functional or modified dependence in self-feeding, locomotion, or verbal expression.

Disability rates increased with age, with older adults having greater adjusted and unadjusted rates than the pediatric, young adult, and middle-aged adult groups across all regions (cervical; thoracic; lumbar). This higher disability rate among older adults following injury is likely attributable to age-related decreases in bone density and muscle strength. This impaired functionality and decreased physical resilience in older adults contributes to less successful rehabilitation and recovery following injury (Suetta et al. 2009). Therefore, spine injuries in the older adult population commonly lead to some form of disability which can result in loss of independence, decreased quality-of-life, and high health care costs.

Pediatric occupants usually have greater biological plasticity and increased capacity for healing from orthopedic trauma (Ho-Fung et al. 2017). As a result, pediatric occupants experience significantly lower disability and mortality compared to adults. In contrast to bone loss in older adults, the pediatric population is still in the process of gaining bone mass, with peak mass occurring around age 25–30 and declining thereafter. In the state of bone mass accrual, bone physiology in children is more adept to recover quickly and efficiently from orthopedic trauma. Although pediatric occupants may have an innate ability to recover more easily, their disability outcomes, along with those of adult occupants, can be strongly affected by rehabilitation efforts after sustaining a spinal injury.

Rehabilitation is vitally important in improvements of spine injury-associated disability. However, it can be challenging to determine which patients will benefit from rehabilitation. Patients that sustain a mildly disabling injury or an injury not considered to be severe by the AIS are often not referred to therapy. A previous study of inpatient rehabilitation for pediatric patients found that those sustaining severe injuries were most frequently referred (Zonfrillo et al. 2013). Although most pediatric patients may not be classified as severe, they are more than likely to receive some benefit from therapy, especially in an effort to help them (1) adjust to a ‘new normal’ following even a mild to moderately disabling injury, (2) optimize re-integration into society at a critical point in their care (even if they continue to improve with ongoing outpatient rehabilitation), and (3) maximize their quality-of-life years.

Spine injuries in pediatrics were associated with little to no mortality, but disability rates varied from 0% to 40%. With higher disability rates in childhood and adolescence, there is a possibility for significant quality-of-life years lost and long-term care/rehabilitation. Adult groups experienced increased mortality and disability compared to children, with a small, positive correlation between DR_MAIS and MR_MAIS. Across all age groups, DR_MAIS was consistently greater than MR_MAIS for a given spinal injury (Figure 3). This disparity between mortality and disability rates highlights the utility of DR_MAIS in identifying injuries that have significant potential for disability with little or no threat-to-life. Therefore, MR_MAIS and DR_MAIS should be considered independently when evaluating injury severity and functional outcomes since many MVC spine injuries may not result in death, but instead may result in long-term disability that necessitates costly, institutionalized care (Aitken et al. 2010).

Current impairment metrics have been established and include the Injury Impairment Scale (IIS), the Functional Capacity Index (FCI), and the predictive form of the FCI (pFCI), summarized in Appendix Table A6, online supplementary information. Overall, research indicates a lack of correlation between these impairment metrics and true functional outcomes, particularly for head and spine injuries (Barnes et al. 2009; MacKenzie et al. 1996). This is not unexpected as IIS, FCI, and pFCI are all designed to characterize expected impairment mapped to certain AIS injury codes (rather than an individual patient). Additionally, development of these three metrics relies on assumptions, including: that the injury was survived; it was the only injury sustained; timely and appropriate care was received; and the individual was within a younger age range (IIS: 25–40; FCI/pCFI: 18–34). These assumptions do not represent the entire trauma patient population. Age limitations have been addressed by some impairment metrics, such as the Swedish risk of permanent medical impairment (RPMI) (Malm et al. 2008), where injuries were assessed by medical doctors and given a degree of temporary or permanent medical impairment between 1% and 99%. This tool has been used to show that the risk of impairment differs significantly based on age and gender, yet it still relies on associations with AIS injury codes and has not been substantially validated (Gustafsson et al. 2015). The metric developed in this study, DR_MAIS, is a valuable addition to these existing metrics by providing information on location- and age-specific impairment which is explicitly associated with functional limitations assessed at the time of discharge.

A limitation of this study is that NTDB may not provide a generalizable set of data—it consists solely of data submitted by participating hospitals and is not a population-wide database. Patients who are dead upon arrival are excluded from NTDB (Barnard et al. 2013) which could contribute as a possible confounder to the overall reported mortality rates. Crash characteristics are not included in the NTDB, so they were not controlled for in the selection of spine injuries. It should also be noted that all 14 AIS 3 spine injuries identified in this study were not found in every age group’s top 95% list. The use of AIS as a selection criterion may limit the ability to study disability as an independent factor, and future studies could select injuries based on the most common disabling injuries as opposed to the most commonly occurring injuries.

We found no significant difference in adjusted disability rates between early- and late-stage adolescents. However, our sample was heavily skewed with 93.3% of children >14 years-old. This may be justified by prior studies that indicate children >14 are more likely to sustain spine injuries compared to younger children (Piatt and Imperato 2018). Additionally, this difference could be due to teens reaching driving age in late adolescence. Overall, further research is needed among a larger, evenly distributed sample of early/late-stage adolescents to investigate disability differences.

DR_MAIS is only a partial solution to addressing the influence of multiple injuries on functional outcomes. In this study, only individuals with AIS 4+ co-injuries were controlled for in the analyses. Therefore, individuals could have sustained AIS 1 or 2 co-injuries as well as multiple AIS 3 injuries that contributed to their functional limitations assessed at the time of discharge. As a result, some disability rates may be exacerbated by these individuals that sustained more than one of the top 95% most frequent AIS 3 spine injuries, although these cases were rare. To more explicitly assess the disabling effects of certain spinal injuries, future work may seek to exclude all individuals sustaining multiple spine injuries and focus on subjects with isolated spine injuries or develop a method to compound disability as a result of multiple injuries sustained.

Another consideration is that FIM scores were only collected by the NTDB at the time of discharge and were available for only 56.5% of the patients identified for inclusion in the study. The remainder of the patients with no FIM scores available were excluded from the analysis, which could introduce some bias in disability and injury distribution of the population. Pre-health status, preexisting comorbidities, and need for surgical intervention were not considered in the analysis, but could potentially affect a patient’s disability outcome. For example, premorbid psychosocial factors appear to have an impact on psychological difficulties associated with post-injury disability. Disability is commonly accompanied by psychological consequences such as depression that can delay recovery (MacKenzie et al. 1996; Read et al. 2004). In future studies, incorporation of these factors is necessary to further refine the disability metrics to improve diagnosis, treatment, and long-term management following injury.

In conclusion, disability rates for cervical, thoracic, and lumbar MVC spine injuries were all significantly greater in older adults compared to younger age groups. This difference highlights the importance of assessing injury and disability rates differently over the lifespan. DR_MAIS can supplement the existing injury metrics by accounting for age- and location-specific functional implications of MVC spine injuries. Additionally, DR_MAIS can provide a measure of disability based upon full functional or modified dependence in self-feeding, locomotion, or verbal expression assessed at discharge, giving clinicians a clearer view of the differing disability rates associated with common spine injuries that all have the same AIS 3 injury severity. In this regard, DR_MAIS provides valuable insight on quality-of-life post-injury and can help guide rehabilitation and treatment strategies.

Data availability statement

Preexisting datasets were utilized for the analyses conducted in this study. These included the NASS-CDS datasets (https://www.nhtsa.gov/research-data/national-automotive-sampling-system-nass) from years 2000–2011, which are publicly available, as well as NTDB-RDS (https://www.facs.org/quality-programs/trauma/tqp/center-programs/ntdb) datasets from years 2000–2011. The mortality metrics used are available in the Appendix of a prior publication (Weaver et al. 2013). The final, compiled datasets from these sources are available upon request from the authors.