Epidemiology of Traumatic Spinal Cord Injury Among Persons Older Than 21 Years: A Population-Based Study in South Carolina, 1998–2012

Abstract

Background:

A gap exists in the current knowledge regarding the epidemiology of traumatic spinal cord injury (TSCI) in a statewide population.

Objective:

To describe population-based epidemiology and trend of TSCI in persons 22 years and older in South Carolina over a 15-year period from 1998 through 2012.

Methods:

Data on patients with TSCI were obtained from ongoing statewide TSCI surveillance and follow-up registry. Deaths were ascertained by linking surveillance files and the multiple cause-of-death dataset. Descriptive analyses were completed, and incidence and mortality rates were calculated based on the civilian adult population of the state.

Results:

Over the 15 years, 3,365 persons with incident TSCI were discharged alive from acute care hospitalization, of whom 555 died during the period of observation. Age-standardized cumulative mortality rate was 14 per million, and the average incidence rate was estimated at 70.8 per million population per year. Age-standardized incidence rate of TSCI increased significantly from 66.9 in 1998 to 111.7 per million in 2012. Standardized incidence rates were significantly higher among non-Whites and males. Motor vehicle crashes and falls were the leading causes, accounting for nearly 70% of TSCI.

Conclusions:

Standardized incidence and mortality rates of TSCI in South Carolina are higher than reported rates for the US population. Motor vehicle crashes and falls are the leading causes of TSCI. There was a significant increase in the overall trend of the incidence rates over the 15 years. A well-coordinated preventive strategy is needed to reduce incidence and improve survival of persons with TSCI.

The public health importance of traumatic spinal cord injury (TSCI) as a major source of long-term disability and health care cost is well recognized.¹ The incidence of TSCI in the United States has been reported to vary from 25 to 59 per million population per year, with an average of 40 per million.² A few isolated reports of incidence rates of TSCI vary by state and source of data, ranging from 51.0 to 93.8 per million per annum.^3–7 A meta-analysis of 64 articles published from 1950 through 2012 provides a global incidence of 8.0 to 246.0 per million population per annum.⁸ Estimates derived from these studies are difficult to compare because of irregularities in representativeness of the numerator to the base population and variations in case identification. Most recently, estimates derived from Nationwide Emergency Department Sample indicate a 3-year cumulative incidence of 56.4 per million for persons age 18 years and older⁵ and 17.5 per million per annum for children and adolescents age 17 years and younger.⁹ This recent estimate is hardly suggestive of a long-term trend of TSCI and yields a lower annualized incidence rate of 36.4 per million population than the average estimate of 40 per million population per annum.

Estimates on the prevalence of TSCI in the United States are widely divergent, ranging from 50¹⁰ to 906 per million population.¹¹ However, those estimates are calculated on various assumptions of life expectancy after TSCI, which contribute to the variability. Current estimates from the National Spinal Cord Injury Statistical Center (NSCISC) indicate 265,000 (range, 238,000 to 332,000) people living with TSCI.¹² This translates to a prevalence rate of approximately 855 per million population.¹² Recent advances in emergency trauma care have led to increased survival of persons with TSCI.^13,14 Thus, the number of Americans living with TSCI has increased over the past decade. The leading causes of TSCI remain relatively constant, with motor vehicle crashes (MVCs) and falls accounting for more than two-thirds of TSCIs.

The TSCI Model Systems have shown a trend toward older age at injury. From 1973 to 1979, 61.9% of those injured were aged 16 to 30 years; but from 2000 to 2003, only 40.7% of injuries occurred among that age group.¹⁵ Conversely, during the same time frames, the proportion of persons injured in the 65+ age group increased from 4.8% to 12.1%. Blacks and males have a significantly higher risk of incurring TSCI than Whites and females.^16,17 Males collectively are 3 to 4 times more likely than females to sustain TSCI.^2,18 Despite some differences in the magnitude of these estimates, the epidemiological characteristics and mechanisms of injury are consistent across various studies in the United States.

Although the NSCISC continues to release annual incidence and prevalence estimates, evaluation of the actual TSCI trends has not been possible because of the changes in the composition of NSCISC-participating registries.² We found only 1 published report from Olmstead County, Minnesota, for the years 1935 through 1981, in which the long-term trend and epidemiology of TSCI were examined in a defined population.⁶ However, this study’s generalizability to the current trends of TSCI in the US population is limited by the lack of population diversity (Caucasians, 98.7%) and contemporaneity with recent epidemiology. Hence, population-based incidence of acute TSCI in a well-defined population in the United States that is reflective of current trends and covering a wider time span is unavailable in the literature. The purpose of this study is to fill the gap in the current knowledge regarding the epidemiology of TSCI in a statewide population in South Carolina with emphasis on incidence and trend over a 15-year period from 1998 through 2012.

Methods

Data

The Medical University of South Carolina Institutional Review Board approved this study. Data for this study were collected through the Spinal Cord Injury Surveillance Registry (SCISR) as described in a comparable study by Cao, Selassie, and Krause.¹⁹ All nonfederal hospitals in South Carolina are legally required to report uniform billing discharge data (UB-92/04) to the State Budget and Control Board. SCISR validates the accuracy of the information from randomly selected medical charts and verifies the data quality regularly. Military and veterans hospitals are not included in this analysis, making this cohort representative of the civilian population with TSCI in South Carolina. Thus, when we use census population estimates as denominators, we will exclude the noncivilian population. Data are 99% accurate and complete.²⁰

Inclusion/exclusion criteria

TSCI cases were selected according to the Centers for Disease Control and Prevention (CDC) case ascertainment criteria of TSCI.²¹ The CDC case ascertainment criteria of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes of 806.0–806.9 (fracture of vertebral column with spinal cord lesion) or 952.0–952.9 (spinal cord lesion without evidence of spinal bone injury) were used to identify persons with TSCI. Observations were unduplicated using personal identifiers. We excluded out-ofstate residents, late effects, and nonincident cases. After TSCI case identification, we linked the data with the multiple causes of death data (MCDD) to identify all deaths occurring among those with incident TSCI. Based on the American Academy of Pediatrics Council on Child Health statement of pediatric age limits,²² individuals included in this study were 22 years old and older.

Multiple causes of death data

After we identified incidence cases of TSCI, data were linked with the MCDD obtained from the Division of Vital Records at the South Carolina Department of Health and Environmental Control (SCDHEC) to identify deaths after discharge from acute care. MCDD provide underlying cause of death and up to 20 contributing causes, along with information on the circumstances and manner of death. The Office of Health and Demography in South Carolina links the UB-04 files with MCDD files using each person’s unique identification. The final merged data for analysis provide demographic information, dates of discharge and death, underlying and contributing causes of death, external cause of injury codes (E-codes), principal payers, up to 10 ICD-9-CM diagnosis codes, and other variables available at the Web site of the TSCI registry (http://academicdepartments.musc.edu/sctbifr/).

Definitions

The operational definition of TSCI was an acute traumatic lesion of the neural elements in the spinal canal (spinal cord or cauda equina) resulting in temporary or permanent sensory deficit, motor deficit, or bowel or bladder dysfunction.²¹ Age was categorized into 6 groups: 22 to 34, 35 to 44, 45 to 54, 55 to 64, 65 to 74, and 75+ years. Race and gender were combined and classified into 4 categories: White male, non-White male, White female, and non-White female. The non-White category includes less than 4% of races other than African American. State counties were classified as urban and rural based on the metropolitan statistical area designation. Health insurance coverage was used to approximate socioeconomic status based on the primary payer recorded at the time of discharge and was grouped as shown in Table 1. Length of hospital stay (LOS) was calculated as days elapsing from date of admission to date of discharge. Charge refers to all services billed during acute care and does not reflect the revenue received by the providers, but it is a true reflection of actual consumption of resources for acute care that falls on society.²³ External cause of injury was classified as MVC/transportation injuries (road traffic, motorcycle, other transportation, and bicycle injuries), falls, and all other causes of injury, which include cutting/piercing injuries, drowning, burn injuries, injuries caused by machinery, and unspecified. Approximation of neurological lesion was based on the number of extremities involved and designation of intensity of neural damage as complete, incomplete, and unspecified based on the fifth digit of ICD-9-CM diagnosis codes as proposed by Jones et al.²⁴ Level of injury was defined as cervical, thoracic, lumbar, or sacrococcygeal. Because mortality is coded with the International Classification of Diseases, 10th Revision (ICD-10) rubric, a coding scheme that does not require the fifth digit, we were unable to quantify the clinical parameters of TSCI for deaths.

Table 1. Characteristics of persons with traumatic spinal cord injury in South Carolina from 1998 through 2012.

Characteristics	Survival status by December 31, 2012
	Deceased (n = 555)		Alive (n = 2,810)		Total, %
	%	95% CI	%	95% CI	(N = 3,365)
Years of event
1998–2002	40.1	36.0–44.2	25.7	24.1–27.3	28.1
2003–2007	30.6	26.8–34.4	33.9	32.1–35.6	33.3
2008–2012	29.3	25.5–33.1	40.4	38.6–42.2	38.6
Age group, years
22–34	6.1	4.1- 8.1	24.4	22.8–26.0	21.4
35–44	7.8	5.5–10.0	19.0	17.5–20.5	17.1
45–54	13.2	10.4–16.0	20.8	19.3–22.3	19.5
55–64	19.2	15.9–22.5	16.3	14.9–17.7	16.8
65–74	20.6	17.2–24.0	10.4	9.2–11.5	12.0
75+	33.1	29.2–10.3	9.2	8.1–10.3	13.1
(Mean)	(64.40)	(63.0–5.8)	(48.90)	(48.2–49.5)	(51.40)
Race and gender
White male	39.1	35.0–43.2	40.5	38.7–42.3	40.3
Non-White male	29.1	25.3–32.9	34.8	33.0–36.6	33.9
White female	25.3	21.7–28.9	17.5	16.1–18.9	18.8
Non-White female	6.5	4.4- 8.6	7.1	6.1- 8.0	7.0
Place of residence
Urban	56.6	52.4–60.7	62.4	60.6–64.2	61.4
Rural	43.4	39.3–47.5	37.6	35.8–39.4	38.6
Health care insurance
Commercial	20.6	17.2–24.0	38.6	36.8–40.4	35.4
Medicare	60.6	56.5–64.7	22.2	20.6–23.7	28.5
Medicaid	10.5	7.9–13.1	13.1	11.8–14.3	12.7
Indigent care	2.4	1.1- 3.7	9.6	8.5–10.7	8.4
Uninsured/self-pay	6.0	4.0- 8.0	16.5	15.1–17.9	14.7
Causes of injury
MVC/transport	20.6	17.2–24.0	38.3	36.5–40.1	35.4
Falls	45.4	41.2–49.5	29.6	27.9–31.3	32.2
Assault/firearm	2.7	1.3- 4.1	6.1	5.2- 7.0	5.6
Struck by/against	6.7	4.6- 8.8	7.1	6.2- 8.0	7.0
Sport	1.8	0.7- 2.9	3.7	3.0- 4.4	3.4
All other causes	4.5	2.7- 6.2	4.2	3.4- 4.9	4.3
Unknown/missing	18.3	15.1–21.5	11.0	9.8–12.1	12.2
Neurological lesion
Tetraplegia, complete	8.1	5.8–10.4	5.9	5.0- 6.8	6.3
Tetraplegia, incomplete	54.4	50.2–58.5	56.3	54.5–58.1	56.0
Paraplegia, complete	3.4	1.9- 4.9	4.2	3.4- 4.9	4.1
Paraplegia, incomplete	25.1	21.4–28.7	25.3	23.7–26.9	25.2
Unspecified	8.9	6.5–11.3	8.3	7.3- 9.3	8.4
Level of injury
Cervical	62.7	58.7–66.7	62.2	60.4–64.0	62.3
Thoracic	19.0	15.7–22.3	18.8	17.3–20.2	18.8
Lumbar	7.1	5.0- 9.2	9.4	8.3–10.5	9.0
Sacrococcygeal	11.2	8.6–13.8	9.6	8.5–10.7	9.9
Type of injury
Closed	96.0	94.3–97.6	95.9	95.1–96.6	95.9
Open	1.4	0.4- 2.4	2.0	1.5- 2.5	1.9
Undetermined	2.5	1.2- 3.8	2.1	1.6- 2.6	2.2

Note: CI = confidence interval; MVC = motor vehicle crashes.

Statistical analysis

Analyses were conducted using SAS version 9.4 statistical software (SAS Institute, Cary, NC). Frequencies and proportions were used to describe categorical variables. For descriptive statistics of continuous variables, means and standard deviations were used. The 2 continuous variables of interest were approximately normally distributed, although there was some evidence of overdispersion (coefficient of variation = 42% for charge and 34% for LOS). Age-adjusted mean values for acute care charge and LOS were effected with least square means (LSM) in the generalized linear model. In contrast to arithmetic mean, LSM are adjusted for covariates and are better estimates of the true population mean.²⁵ Proportions corresponding to demographic and clinical characteristics were compared among the groups by constructing 95% confidence intervals (CIs) for binomial proportions under the assumption of independence and normal approximation.²⁶ Incidence rates were adjusted for age, gender, and race using direct standardization methods with the 1990 South Carolina census population as the reference. To assess the temporal trend of the adjusted rates, we used linear regression (ProcReg in SAS), with P < .05 indicating a significant trend.²⁷ Regression technique with smoothing effect (spline regression) was used to plot the trend and construct the 95% CI. Graphic displays of the trend lines were constructed with Gplot procedure (an SAS procedure for producing 2-dimensional graphs). Smoothing was effected to get narrower and conservative confidence width and dampen the effect of sudden jump between data points for visual display of the trend lines. Chi-square test of trend was used to determine the statistical significance of the trend lines. Age-adjusted probability of death after hospital discharge from acute care facilities was estimated using the Life Test procedure (an SAS procedure for analyzing survival data).

Results

From 1998 through 2012, there were 4,042 South Carolina residents discharged with a diagnosis of TSCI from acute care hospitals. Of these, 481 (11.9%) were younger than 22 years and 196 (4.9%) were repeat admissions for the same acute TSCI. After exclusion of these observations, 3,365 (83.2%) persons constituted the total number of persons with acute TSCI discharged alive from nonfederal hospitals in South Carolina.

Table 1 displays the demographic and clinical characteristics of persons age 22 years and older with acute TSCI among South Carolina residents. During the 15 years of follow-up, 555 deaths (16.5%) occurred after hospital discharge. The mean age of all adults was 51.5 years (95% CI, 50.9–52.1). Males accounted for 74.2%, urban residents for 61.4%, and the uninsured for 14.7% (95% CI, 13.5–15.9) of the hospital discharges. The uninsured population in the state was 15%. The most common neurological lesion was incomplete tetraplegia, accounting for 54.4% (95% CI, 50.2–58.5), followed by incomplete paraplegia with 25.1% (95% CI, 21.4–28.7). Among the external causes of injury, 35.3% of TSCI resulted from motor vehicle traffic injuries and 32.2% from falls, accounting for more than two-thirds of the injuries. In regard to survivorship, the proportion of deaths was higher with advancing age, with 53.7% of the deaths occurring in persons age 65 years and older; this age group accounted for 25% of the total hospital discharges.

The highest proportion of TSCI deaths in both Whites and non-Whites resulted from fall injuries (Table 2). TSCI caused by assault or firearms accounted for 9.4% of injuries among non-Whites and 2.75% of injuries among Whites. This makes assault- or firearm-related TSCI 3.4 times more common among non-Whites compared with Whites. The proportions of TSCI attributable to MVCs or transport-related injuries were highly comparable between non- Whites and Whites with comparable influence on mortality. The rest of the causes of injuries showed comparable distribution between non- Whites and Whites.

Table 2. Comparison of causes of traumatic spinal cord injury by race and survivorship.

Cause of injury	Survival status by December 31, 2012				Total (N = 3,365)
	Deceased		Survivors
	Non-White (n = 201)	White (n = 354)	Non-White (n = 1,180)	White (n = 1,630)
Column percent (95% CIs)
MVC/transport	21.6 (15.9–27.3)	19.9 (15.7–24.1)	39.7 (36.9–42.5)	37.2 (34.8–39.5)	35.3 (33.0–37.6)
Falls	43.2 (36.3–50.1)	46.6 (41.4–51.8)	22.8* 20.4–25.2)	34.5 (32.2–36.8)	32.2 (29.9–34.5)
Assault/firearm	4.0 ( 1.3- 6.7)	2.0 (0.5–3.5)	10.3* (8.6–12.0)	3.1 (2.3–3.9)	5.6 (4.5–6.7)
Struck by/against	6.0 (2.7–9.3)	7.0 (4.3–9.6)	8.2 (6.6–9.8)	6.4 (5.2–7.6)	7.1 (5.6–8.3)
Sport	3.1 (0.7–5.5)	1.1 (0.0–2.2)	4.0 (2.9–5.1)	3.4 ( 2.5–4.3)	3.4 (2.5–4.3)
All other causes	5.0 (2.0–8.0)	4.2 (2.1–6.3)	4.2 (3.1–5.3)	4.2 (3.2–5.2)	4.2 (3.2–5.2)
Unknown/missing	17.1 (11.9–22.3)	19.1 (15.0–23.2)	10.9 (9.1–12.7)	11.1 ( 9.6–12.6)	12.2 (10.6–13.8)

Note: CIs = confidence intervals; MVC = motor vehicle crashes.

^*Significantly different at the alpha = 0.05 level.

The mortality experience of the cohort as a function of time since hospital discharge is shown in Table 3. Fifty-two percent of the deaths occurred within the first year, and 21.2% occurred within the second and third years. Taken together, the first 3 years after hospital discharge accounted for nearly three-fourths (73.5%) of the deaths. Analysis based on Kaplan-Meier product-limit probability of death indicated that risk of dying after hospital discharge was 8.6% within the first year, 3.8% between the first and third years, and 3.0% between the third and sixth years. This suggests that the mortality experience of the cohort was comparable to that of the general population after the fifth year post hospital discharge.

Table 3. Summary of age-adjusted probability of death, 1998–2012.

Follow-up (in years)	Cohort size	Subjects deceased	Subjects alive (censored)	Probability of death, %
0 to 1	3365	290	3,075	8.61
1 to 3	3075	118	2,957	3.84
3 to 6	2957	90	2,867	3.04
6 to 10	2867	54	2,813	1.88
10–15	2813	3	2,810	0.11
Total	3365	555	2,810	17.48

Incidence rates and trends of TSCI

Annual incidence rate, which includes those alive and deceased, was highest among persons age 75 years and older with 173.5 (101.2 alive + 72.3 deceased) per million population (Figure 1). The ratio of mortality to survivorship is the highest in this age group, with 1 death for 1.4 survivors. Incidence rate was the lowest in the 22 to 34 age group (55.7 per million/year), with a ratio of 1 death for 20.4 survivors. Incidence rate increased with advancing age with a slight decline to 97.1 per million/year in the 65 to 74 age group, although it had the second highest ratio of 1 death for 2.6 survivors. The overall incidence rate was 84.9 per million with a death to survivorship ratio of 1:5. Mortality rates by age group showed a progressive and monotonic increase with advancing age. The rate increased from 2.6 in the 22 to 34 age group to 72.3 per million per annum in persons 75 years and older.

Figure 1.

Annualized traumatic spinal cord injury (TSCI) survival and death rates by age groups, South Carolina 1998–2012.

The temporal trend of age-adjusted overall TSCI hospital discharge rate over the course of 15 years is shown in Figure 2. The rate increased steadily from 66.9 in 1998 to 83.8 in 2003, almost leveled off through 2007, and increased sharply from 84.2 per million population in 2008 to 111.7 per million population in 2012 (P < .0001). Although the plots are smoothed with spline regression, nearly all the observed data points were within the 95% CI of the smoothed line, suggesting that the smoothing effect did not depart from the actual age-adjusted rates. The increasing trend of age-adjusted TSCI incidence rate from 66.9 per million in 1998 to 111.7 per million in 2012 suggests an increase of 4.5% per annum, while the rate of increase in the referent population of age 22 and older was 1% per annum.

Figure 2.

Overall temporal trends of traumatic spinal cord injury (TSCI) hospital discharge rates by year, 1998–2012. Smooth line = smoothed trend line of rates; broken lines = upper and lower 95% confidence limits.

Figure 3 shows a panel of trend lines plotted on the same scale by demographic characteristics. Figures 3A and 3B show the age-standardized incidence rate of TSCI in females and males. While both show a significant increase, the rate of increase was higher in males than in females, notwithstanding the wide 3:1 rate ratio between the rates throughout the period of observation. Trend lines comparing Whites and non-Whites show different patterns of TSCI rates over time. The rates in 1998 were comparable between Whites and non-Whites, 65 and 71 per million, respectively, with the non-White rate being higher by 9%. In 2012, the rates were 90 and 155 per million in Whites and non-Whites, respectively, with a rate difference of 72%. Similarly, the increases in the rates noted in 15 years were 38.5% and 121.4% for Whites and non-Whites, respectively.

Figure 3.

Temporal trends of traumatic spinal cord injury (TSCI) hospital discharge rates by gender and race, 1998–2012. Smooth line = smoothed trend line of rates; broken lines = upper and lower 95% confidence limits.

Acute care charge and length of hospital stay

Age-adjusted mean acute care LOS was 17.9 days (high-low, 14.9–22.2 days) with no significant increase or decrease over the course of the study (P value for trend = .2243). In 1998, the age-adjusted mean LOS was 17.8 days; in 2012, the mean LOS was 20.0 days. There was a significant increase in the age-adjusted average acute care charge of TSCI hospitalization ranging from a low of $47,000 in 1998 to $189,000 in 2012, a 302% increase in 15 years.

Discussion

The results from this extensive population-based study spanning 15 years indicate the significance of TSCI as a public health problem in South Carolina. The statewide hospital discharge and mortality data provide very useful information on the incidence, mortality, and demographic, clinical, and etiological characteristics of TSCI to describe its epidemiology. The strength of the study was the representativeness of the data sources to provide robust estimates on incidence of TSCI and subsequent mortality as described elsewhere.^19,28 The findings demonstrate that (a) the hospital discharge rate for persons with TSCI in South Carolina is higher than rates previously reported in the United States,² (b) the rate significantly increased over the span of 15 years, (c) significant differences existed among population subgroups, and (d) the rate of increase in acute care charge exceeded the inflation-adjusted consumer price index by nearly 8-fold while LOS remained constant.

Our findings of an increasing trend of SCI incidence are also observed in traumatic brain injury research, which has shown significant increases in both hospitalizations and emergency department visits in the United States.²⁹ The increase is attributable to falls among the elderly with the increase in the aging population; increase in motor vehicle and motorcycle injuries, although mortality did not increase as a result of motor vehicle or motorcycle injuries; and increase in sport-related and recreational injuries due to the popularity of all-terrain vehicle use, roller skating, and football, although the increase was not profound.³⁰ Additionally, our trend analysis demonstrated a widening racial disparity in the trend of TSCI. Although it is unsettling to note a positive trend in both race and gender, the rate of increase among non-Whites and males exceeded the rates in their counterparts. A similar epidemiological trend in the incidence of TSCI has been noted nationally where non- Whites accounted for 24% before 2000 but 37% since.¹² The disproportionate increase in the rate of non-Whites is probably due to the increasing rate of motor vehicle and transport injuries. It appears that expansion of urban areas and motor vehicle and motorcycle ownership in minority communities could have contributed to increased incidence and mortality rates in non-Whites.

The increased occurrence of TSCI among males is expected and is consistent with results of other studies.^2,18 One possible explanation for this difference is the propensity of males to be engaged in high-risk activities such as speeding, driving under the influence, assaults, and firearm use. However, the persistence of the 3-fold gap across time is quite surprising, because the number of women age 75 years and older is increasing faster than the number of men, and the risk of fall-related TSCI is much higher in this age group. The increased incidence rate of TSCI with advancing age, despite the greater likelihood of trauma in the younger age groups as result of high-risk behavior, suggests that the effect of age on increasing susceptibility to serious injuries is mostly due to vertebral fractures.³¹

TSCI mortality rates increased at a faster rate with advancing age. This finding is consistent with results of other studies of TSCI examining the association of age and mortality.^32,33 Life expectancy in the United States is increasing, which has increased the number of people living to older ages.³⁴ Additionally, older adults are more likely to have comorbidities, including those that increase the risk of falls resulting in injury such as a gait or balance disorder, a visual disorder, or cognitive impairment.^19,29,35,36 This trend indicates that we will continue to see more older adults experience SCI as more adults continue to hit aging milestones. This is particularly troubling because older adults who experience SCI are more likely to experience secondary conditions compared with their younger counterparts.³⁷

The high age-adjusted probability of death proximal to the date of hospital discharge from acute care has 2 important implications. First, advances in clinical care of trauma patients have enabled critically injured patients to survive through the time of hospital discharge because of the availability of coordinated trauma and intensive care, which are difficult to maintain in the community,³⁸ and transitional care facilities (Medicare and Medicaid swing beds).³⁹ The goal of acute care hospitals is to keep patients alive to the point of discharge by exerting the maximum life sustenance effort such that inpatient death is reduced to the lowest possible rate. Because these patients experience accelerated probability of death after hospital discharge,¹⁹ their families and health care professonals need to intensively monitor their health status immediately after their discharge. Second, policymakers should be concerned that many critically injured persons with TSCI without adequate health insurance do not get advanced postacute rehabilitation after hospital discharge.

The consistent increase in acute care charge without a corresponding increase in LOS may reflect the influence of market forces in the delivery of medical services. Average fee-for-service (FFS) spending for hospital inpatient care has increased at an average rate of 3%.⁴⁰ If acute care charge were to follow the FFS trend, the expected charge would have increased to $71,050 in 2012, a 45% change. However, in reality, it is a 302% increase. This suggests that hospital charge has increased nearly 8-fold above the FFS rate. A similar rate of increase is reported by the Consumer Price Index for Healthcare for the southeastern United States. The increase has been much greater since 2010 (Figure 4). The difference is most likely explained by the severity of spinal cord trauma, which often necessitates longer inpatient and intensive care unit days, especially for perons with high cervical injuries who need tracheostomy and intubation. However, the increased acute care charges place increased financial burden on a population that already has high costs both at the time of injury and during the years following.^41,42

Figure 4.

Age-adjusted mean acute care charge and length of hospital stay by year.

Limitations

Finally, despite the important findings reported in our study on a population-based epidemiology of TSCI, our study has several limitations. First, our analysis relied on UB-04 administrative data to classify the types and levels of TSCI. The accuracy of the fourth and fifth digits of the codes could be unreliable to ascertain the specificity of the diagnosis, especially from rural and underresourced hospitals. However, the South Carolina surveillance system routinely evaluates the quality of the data through chart review, and the positive predictive value rate of the data is high. Second, data from federal health care establishments on the noncivilian population are not available to fully account for populationwide rates; this is also true for all other states in the union. However, we minimized dilution of the rates by excluding the noncivilian population from the denominator based on the US Census data. Third, the information available to us to quantify the economic impact of TSCI is acute care charge, which is different from the revenue collected by the hospitals.

Conclusion

This population-based analysis of hospital discharge and death shows the importance of TSCI as a significant health problem in South Carolina because the discharge rate is higher than the national average. With the increase in the aging population in South Carolina, TSCI from falls is likely to increase, leading to higher mortality rates than those associated with TSCI from MVCs. The charge for acute care has tripled in just 15 years, accompanied by an increasing number of uninsured and underinsured persons sustaining TSCI. A well-coordinated strategy addressing the whole spectrum of primary, secondary, and tertiary prevention appears to be necessary to reduce the socioeconomic impact of TSCI.