Utilization and Cost of Health Services in Individuals With Traumatic Brain Injury

Traumatic Brain Injury (TBI) has gained attention in the past decade as a “signature injury” in the conflicts in Iraq and Afghanistan. TBI is a major burden for both the military and civilian population in the US and worldwide. It is a leading cause of death and disability in the US and a major health services resource burden. We seek to answer two questions. What is the evidence regarding the association of TBI with health services utilization and costs in the US and worldwide? What is the evidence regarding racial/ethnic, gender, geographic, socio-economic and other disparities in health services utilization and cost in the US and worldwide? To attain this goal we searched several databases using key words to perform a systematic review of the literature since 2000. We found 36 articles to be eligible for inclusion in the review. The evidence demonstrates a wide variation in health services utilization and costs depending on population of study and severity of TBI. The evidence also supports the existence of racial/ethnic, gender, insurance, geographic disparities in the US as well as other unique disparities worldwide.


Introduction
Traumatic Brain Injury (TBI) has gained attention in the past decade as a "signature injury" in the conflicts in Iraq and Afghanistan (Egede, Dismuke, Echols, 2012). TBI is defined as a physical force to the brain sufficient to cause structural alteration or physiological disruption of brain function that results in altered consciousness, amnesia, change in mental state, neurological deficits, or intracranial lesions. TBI is classified by severity as mild, moderate or severe. Mild TBI is the most common type among veterans who sustain a TBI (Morgan et al., 2012). Approximately 313,816 military service members have been diagnosed with TBI since 2000 (Defense and Veterans Brain Injury Center, 2015). However, TBI is also a major health burden in the US civilian population with approximately 1.7 million TBI injuries annually (Faul et al., 2010). Moreover, TBI is a leading cause of death and disability in the US with approximately 53,000 persons (18.4 per 100,000) dying from TBI related injuries annually (Coronado, Xu, Basavaraju et al., 2011). Fire-arms (34.8%), motor-vehicle accidents (31.4%) and falls (16.7%) are the leading causes of TBI related death (4).
TBI is a major health services resource burden with approximately 1,365,000 (80.7%) emergency department (ED) visits and 275,000 (16.3%) hospitalizations annually (Faul et al., 2010). More recently, evidence has shown that TBI may be accompanied by mental health co-morbidities in the military population, especially PTSD (Taylor et al., 2012). Given the large health and resource burden of TBI, we conducted a systematic review of the literature on health services utilization and costs associated with TBI. We sought to answer the following questions: 1) What is the evidence for TBI associated health services cost and utilization in the US and worldwide? 2) What is the evidence for racial/ethnic, gender, geographic, socio-economic and other disparities in health services cost and utilization in the US and worldwide? We were specifically interested in post-injury cost, rather than initial hospitalization cost.

Information Sources, Eligibility Criteria and Search
In order to answer these two questions, three databases (Medline, PsychInfo, and CINAHL) were searched for articles published between January 2000 through June 2013 using a reproducible strategy. Four searches with broad search terms were performed in each database using MeSH headings search. The first search used the terms traumatic brain injury and cost, the second used traumatic brain injury and utilization, the third used traumatic brain injury and outpatient, and the fourth used traumatic brain injury and ambulatory care.
The following inclusion criteria were used to determine eligible study characteristics: (1) must be published in English, (2) must include health services cost or utilization related to TBI (3) must include adults. Exclusion criteria included: (1) did not focus on clinical efficacy, (2) was not limited to describing cost-effectiveness of an intervention, and (3) did not focus on pediatric TBI.

Study Selection and Data Collection
The process used to screen the citations is shown in Figure 1. Titles were eliminated if they were obviously ineligible, for instance describing cost-effectiveness of an intervention or including children with TBI. Full articles were read and reviewed using a standardized check-list by two independent reviewers (CD, RW). A third independent reviewer (LE) was asked to make the final decision regarding eligibility in the case of disagreement. Data collected from the eligible articles is shown in Tables 1, 2 and 3. A summary of the evidence in each article is presented specific to health services costs (Table 1), health services utilization (Table 2) and disparities in cost and utilization (Table 3). We separated our review and tables by disparities due to our findings of a number of articles encountering racial/ethnic, geographic and other socio-economic disparities in the US and around the world. A narrative review was performed because of the heterogeneous nature of the information, which precluded conducting a meta-analysis.

Study Selection
review. See Table 1 for process details. Fourteen articles provided information on cost of TBI alone (Table 1), seven provided information on utilization in TBI alone (Table 2), four provided information on both cost and utilization (Tables 1 and 2) and eleven provided information on disparities in cost and utilization (Table 3).

Cost and Utilization Analysis
Type of cost and utilization are included in Tables 1 through 3 for each paper reviewed. Cost referred to monetary value either measured in real cost or charges when real cost was not available. Cost included both direct and indirect cost. Monetary value was measured in dollars for the US and the appropriate currency for international studies. Variability in cost can be due to provider differences, third party payer differences, and country differences. For this reason, utilization is a much better standard for comparing resource use, especially between health systems and countries.   Ethnic and racial disparities in emergency department care for mild traumatic brain injury Bazarian et al., 2003 In a national sample of ED visits, after controlling for confounders, Hispanics more likely to receive nasogastric tube, nonwhites more likely to receive care by a resident and less likely to be sent back to a referring physician after ED discharge showed that the ability to obtain a waiver was associated with social-medical-political climate, similarity to other waivers, ability to strengthen access and reduce barriers, and expenditure of resources. Managing waivers was associated with cost effectiveness, developmental process of waiver implementation, ability to improve access and reduce barriers, and expenditure of resources.

Medicaid Waiver Utilization in TBI.
National estimates of hospitalization charges Schootman M. et al., 2003 Mean and median acute care charges were only slightly higher for males compared with Inpatient hospital charges in TBI. Non-inpatient utilization in TBI in Australia.

Rehabilitation
Outcomes of Terror Victims with Multiple Traumas Schwartz et al., 2008 Terror victims with TBI had higher rates of brain surgery but no difference in length of stay in hospital relative to non-terror victims in Israel.
Surgical utilization in TBI in Israel.

Health Services Cost Associated With TBI in the United States
Costs associated with TBI vary widely depending on the population studied, severity of injury and time period.
In an early review of mild TBI without surgical intervention, indirect costs were much higher than direct costs, accounting for 92% of total costs in 1981 dollars (Borg et al., 2004). Admission and radiologic policies were found to be determining factors in the level of direct costs (Borg et al., 2007). A study conducted between 1997 and 1999 found that costs of hospitalization ranged from an average of $8,189 for moderate to $33,537 for critical TBI (McGarry et al., 2002). Those due to falls averaged $15,860, while those due to gunshot wounds averaged $20,084 and motor vehicle accidents averaged $20,522 (McGarry et al., 2002). In a study from 1990-1996, acute care daily charges showed routine increases, while lengths of stay generally decreased. (Kreutzer et al., 2001) Rehabilitation charges were about 10% higher than medical care prices, and offset the corresponding decreases in lengths of stay. (Kreutzer et al., 2001) Since 2000, there have been a number of studies in the US and world wide of health services costs associated with TBI in various populations with costs ranging from $9,000 to $103,667. (Borg et al., 2004;Hu et al., 2013;Thompson et al., 2012;Davis et al., 2007;Wei et al., 2005;McGarry et al., 2002;Kreutzer et al., 2001;Thompson, 2001;Rockhill et al., 2011;Leibson et al., 2012;Kayani et al., 2009;Rochette et al., 2009;Vangel et al., 2005). Most recent, a study from 2004 to 2009 found that median hospital costs for adults were $13,000 for the 18-64 age group and $9,000 for the 65 and older age group. (Hu et al., 2013) In a 2005 study, unadjusted total mean one year costs were $77,872 for those aged 55-64 years, $76,903 for those aged 65-74 years and $72,733 for those aged 75-84 years in 2005 dollars (Thompson et al., 2012). In a study of the managed care population, initial hospitalization charges ranged from $32,627 for TBI alone to $103,667 for TBI along with other trauma (Davis et al., 2007). Another study found mean cost of treating a TBI to be $96,612, with inpatient rehabilitation accounting for on average $43,212, not including physician charges (Thompson, 2001).
Costs are known to vary due to severity. One case controlled found total health services ranged from $12,990 for mild TBI vs. $42,4441 for moderate/severe in 2009 dollars (Rockhill et al., 2011). As a comparison, total health services costs for non-TBI matched controls were $7,377 (Rockhill et al., 2011). Mean costs were 76% higher in the three years after injury for the mild TBI group and 5.75 times greater for the moderate/severe group compared to controls (Rockhill et al., 2011). Another case control study for that for definite and probable TBI, most incremental costs occurred within the first six months while significant long-term incremental medical www.ccsenet.org/gjhs Global Journal of Health Science Vol. 7, No. 6; costs were not apparent among one-year survivors (Leibson et al., 2012). Cost differences between possible TBI cases and matched controls were not as great in the first 6 months but were substantial among one-year survivors (Leibson et al., 2012).
Studies also found significant variation in cost by geographic region, type of service, and comorbidity. One study found the highest costs occurring in the states of California and Washington (Hu et al., 2013). State specific studies found varying costs but often categorized costs differently, making comparisons difficult (Kayani et al., 2009;Rochette et al., 2009;Vangel et al., 2005). A study specific to the geriatric population found hospitalization and inpatient rehabilitation costs significantly lower in the 75-84 age category while outpatient care costs and nursing home costs were lower in the younger age categories (Thompson et al., 2012). In a study of the Medicaid population in four states, the presence of co-morbid mental illness increased the cost of care for those with TBI (Wei et al., 2005). Total expenditures ranged by state from $6,093 to $10,907 for those without SMI, compared to $8,723 to $21,924 for those with SMI (Wei et al., 2005). Another study showed presence of psychiatric illness was associated with more than doubling of total costs for both inpatient and outpatient non-mental health care (Rockhill et al., 2011).
Finally, a national study of US Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) veterans treated in the VA found that OEF/OIF veterans with TBI had 4 times higher median health services costs relative to OEF/OIF veterans without TBI ($5,831 vs. $1,547) (Taylor et al., 2012). Mental health co-morbidities in the military population increase health services costs associated with TBI, similarly to civilian populations, with those with TBI and PTSD having a median cost of $5,053 relative to those with TBI alone being $2,391 (Taylor et al., 2012).

International Health Services Costs Associated With TBI
Health services costs associated with TBI have also been estimated for other countries. In China, hospitalization costs were found to be highest for traffic accidents and lowest for blows to the head (Yuan et al., 2012). Multiple factors were associated with higher acute hospitalization costs with doubling of length of stay associated with a 61% higher hospital cost (Yuan et al., 2012). The latest estimate for the cost of TBI in Europe, based on data from 30 European countries in 2010 was €2,697 in direct health care costs (Olsen et al., 2012). A study of England and Wales found that length of stay in critical care accounted for 51% of mean total costs, regular ward 38% and travel costs 5%. (Morris et al., 2008) Those aged 45-64 years had the highest costs. (Morris et al., 2008) In another European study based on 2004 data, average inpatient cost for TBI in Germany was €2,529, compared with €2,833 in Spain and €3,024 in Sweden. Average inpatient cost for concussion was €1,071 for Germany, €987 for Spain, and €927 for Sweden. Average inpatient cost for severe brain injury was €6,647 for Germany, €6,362 for Spain and €6,045 for Sweden (Berg, 2004).

Health Services Utilization Associated With TBI in the United States
TBI injuries are associated with approximately 275,000 hospitalizations and 1,365 million ED visits in the US annually (Faul, 2010). One study found that 81% used medical and allied health services, 66% used transport and 40% used vocational rehabilitation (Hodgkinson et al., 2000). In Georgia, during the first three months of post rehabilitation discharges, at least 80% saw an MD, 42% reported four or more MD visits, over 50% attended day rehabilitation programs, 42% had physical therapy, 36% occupational therapy, 33% speech pathology, and 11% had psychological counseling. (Philips et al., 2004) Re-hospitalization rates declined between the one and five year follow up period from 23% to 17% (Marwitz et al., 2001). While during the first year, orthopedic and reconstructive surgery were the primary reason (25%), infections accounted for 10% of readmissions at 1 year follow-up and 8% at 5 years after injury (Philips et al., 2004). Incidence of rehospitalizations relating to seizures ad psychiatric disorders were 12% at one year and rose to 19% at 5 years (Philips et al., 2004).
Financial incentives may influence end of life care for severe TBI (Holloway et al., 2010). Hospitals have been found to have a strong incentive and insurers a strong disincentive to pay for performing tracheotomies since doing so quintuples a hospital's DRG associated reimbursement rate (Holloway et al., 2010). If a tracheotomy is not performed, the hospital has a financial incentive to discontinue aggressive treatment due to the lower fixed DRG reimbursement rate (Holloway et al., 2010). In another sample of 273 patients with TBI, only 26% were found to have received a neuropsychology evaluation. (NPE) (Schatz et al., 2001). These patients were younger, more likely to be involved in liability claims, achieved a higher functional rehabilitation and attended multiple rehabilitation facilities (Schatz et al., 2001).
A number of patient characteristics are significantly associated with utilization of various services. In a national study of the ED population diagnosed with isolated mild TBI, 44.3% were found to have received CT, 23.9% other non-extremity, non-chest x-rays, 17.1% wound care and 14.1% IV fluids (Bazarian et al., 2005). However, only 43.8% of the ED population were assessed for pain and of those with documented pain, only 45.5% received analgesics while in the ED (Bazarian et al., 2005). Almost 38% of the ED population was discharged from the ED without recommendations for follow-up (Bazarian et al., 2005). State specific studies again focused on different categories, thus are difficult to compare. In an Ohio study, the odds of requiring a ventilator (OR=3.66) and being admitted to the ICU (OR=2.51) were significantly higher for a TBI roadway injury when compared to non-TBI roadway injury (Rochette et al., 2009). In Michigan, more than 4/5 of study subjects received a CT scan while MRI and electroencephalography were used less frequently (Vangel et al., 2005). In Michigan, the most frequent prescriptions filled were anticonvulsant/mood stabilizers (Vangel et al., 2005), however, nationally the most frequently used medicines were analgesics and anxiolytics as well as anticonvulsants in more severe injuries (McGarry et al., 2002).

International Health Services Utilization Associated With TBI
Few studies were found in this review addressing utilization outside the United States. A study of TBI rehabilitation in Italy found that the mean LOS was 87.31, and 40.4% of patients had access to rehabilitation facilities after a month (Zampolini et al., 2011). A study of CT use and TBI in some European countries found that CT use in Spain was much lower than Sweden and Germany with only 6% of patients receiving a CT (Berg, 2004).

Disparities in Cost and Utilization Associated With TBI in the United States
Disparities in health services cost and utilization have been found in the US based on race. Non-White race and lack of insurance were found to be associated with lower likelihood of placement for rehabilitation post-discharge (Heffernan et al., 2011). In South Carolina, uninsured Black females had a lower likelihood of hospitalization (Selassie et al., 2004). A national study of the ED population diagnosed with mild TBI found that Hispanics were more likely to receive a nasogastric tube, while non-Whites were more likely to receive care by a resident and less likely to be sent back to a referring physician after discharge (Bazarian et al., 2003). American Indians and Alaskan Natives (AI/AN) had the highest age adjusted rates of hospitalization for TBI relative to other race/ethnicities (Rutland-Brown et al., 2005) In one study, private insurance was associated with a shorter length of stay (Heffernan et al., 2011). However, in another study insured patients had a longer length of stay in the intensive care unit (ICU), and lower mortality (Alban et al., 2010). Additionally, in a national study mean and median acute care charges were only slightly higher for males compared with females, however mean and median charges were highest for the western region of the US and for persons covered by Medicaid and those treated in urban teaching hospitals (Schootman et al., 2003).
In a study of unmet needs of individuals with TBI, the most prevalent unmet needs were for memory (50.5%), job skills (46.3%), and increasing income (50.5%) (Heinemann et al., 2002). Black, younger and single individuals as well as those dependent in one or more daily activities and with more recent injuries had greater unmet needs (Heinemann et al., 2002). A review of TBI Medicaid Waivers in six states showed that the ability to obtain a waiver was associated with the social-medical-political climate, ability to strengthen access and reduce barriers, and expenditure of resources (Spearman et al., 2001).

International Disparities in Cost and Utilization Associated With TBI
Among 79 patients with moderate/severe TBI in Netherlands, those with a high locus of control were more likely to visit medical specialists and use supportive care than those with lower levels of internal locus of control. (Willemse-van, 2009). In Australia, rural TBI patients were found to be more likely treated in a non-inpatient setting than urban TBI patients, though this did not appear to significantly impact functional outcomes (O'Callaghan et al., 2009).. Finally, terror victims with TBI were found to have higher rates of brain surgery but no difference in length of stay relative to non-terror victims in Israel (Schwartz et al., 2008).

Conclusions
We reviewed a decade of literature regarding the utilization and cost of health services for TBI in civilian and military adults in the US as well as other countries. We found that cost estimates vary depending on the study population, TBI severity and presence of other mental health co-morbidities. We also found there was evidence of significant racial/ethnic disparities in utilization. However, there is need for further research on cost of TBI over time and better understanding of racial/ethnic, geographic and socioeconomic variations in cost and utilization among individuals with TBI in both civilian and military/Veteran populations in the US and worldwide.
as some analyses are not published.
We can conclude, however, that the cost and utilization of TBI is an important area of research for the next 5 years, especially considering the growing recognition of the impact of mental health on health costs and outcomes. A potential concern, which has been addressed only marginally in the literature, is racial/ethnic, geographic, and socio-economic disparities in utilization of health services for TBI and its associated co-morbid conditions. There is also a need to control for severity and follow patients over time to truly compare costs and utilization. Future research should investigate the change in cost during the first year vs. over the lifetime, and costs for those who receive mental health services vs. those who do not.