Chapter 5 - Conclusions

In 2000, we estimate that $127.5 million was spent for inpatient rehabilitation of injuries in other motor vehicle crashes and $16.3 million inpatient rehabilitation of injuries in motorcycle crashes. Public funds paid for 26.1 percent of the other motor-vehicle-related costs and 19.5 percent of the motorcycle-crash-related costs. Inpatient rehabilitation accounted for an estimated 3 percent of inpatient care costs for motor vehicle injury costs and 4 percent for motorcycle injuries.

Motor vehicle injuries generate other costs related to functional losses and the resulting impacts on social and role functions. Although this study did not estimate those costs, it showed the losses for some injuries can be quite significant. Employment drops among TBI patients were notable.

Chapter 1 describes findings from UDSMR, which provides the most comprehensive medical rehabilitation data available. This not only is the sole source for primary payer across diagnosis groups but provides follow-up data on patient living status, vocational status, and FIM scores. These data are often missing cause codes; and some facilities refused us access to their data. Because of this, cause-coded cases accessible to us may not be representative.

This chapter shows that Other Motor Vehicle crash and Motorcycle crash victims are very similar in the distribution of impairments requiring rehabilitation, sharing the same top five categories: TBI; other multiple trauma; TBI with fracture and amputation; other orthopedic; and hip/knee replacement. These categories account for 81 percent of the Motorcycle cases and 79 percent of the Other Motor Vehicle cases. The Motorcycle group was slightly more likely to have TBI rehabilitation services than the Other Motor Vehicle and the Suicide patients, but less likely than the Assault cases.

The most frequent payer for inpatient rehabilitation of motor vehicle (excluding motorcycle) crash injuries was private insurance (40.2%). Public funding accounted for 26.1 percent of the cases, followed by no-fault automobile insurance (19.9%), other/unknown (5.6%), self-pay (2.0%), unreimbursed (2.5%), andWorkers’ Compensation (3.7%).

Payer distributions for inpatient rehabilitation of motorcycle crash injuries were similar to that of other motor vehicle, although we see more reliance on private insurance (63%) and less on public funding (19.5%). Other payers included no-fault automobile insurance (6.9%), other/unknown (5.5%), self-pay (2.4%), unreimbursed (1.6%), and Workers’ Compensation (1%).

Patients were asked at time of discharge where they would be living. Of those injured in motorcycle crashes, 97 percent of those with known pre- and post-measures lived in private homes both before and after rehabilitation (n= 328).

Across all injury categories, more than 50 percent of patients in the workforce changed their vocational status to nonworking or disabled at the time of rehabilitation discharge. Of the previously employed people injured in Other Motor Vehicle crashes, 64 percent were not working or disabled, (54.1% and 9.9%, respectively) at the time of discharge. Of those injured in Motorcycle crashes, we see a similar pattern: 62 percent were not working or disabled, (51.1% and 10.6%, respectively) at the time of discharge.

Chapter 2 analyzes perhaps the most comprehensive TBI disability outcome database. The data tracked a large sample of Colorado TBI cases for four years. The cases are predominantly white males 16 to 64. But, individuals 65 or older accounted for approximately 15 percent of the cases. Unlike in other chapters of this report, the database contained motor vehicle injury, including motorcycle injury. Among causes, the average initial hospital length of stay was greatest for motor vehicle crashes. Head injuries of moderate severity were more commonly associated with motor vehicle crashes and less so with other injury categories. Life-threatening head injuries were most commonly associated with unintentional falls, though this may in part reflect the average older age of this group.

What are most remarkable about the analysis of functional impacts is not the differences between injury cause groups but the strong similarities between causes. For example, total CHART and HSQ scores were approximately similar for all injury cause groups. Our analyses suggest that HSQ scores may have been slightly more effective in identifying differences between groups. In addition, with the exception of HSQ scores, the analysis did not reveal noticeable change over time. For example, the FIM cognitive, motor and total scores changed little from Year 1 to Year 2.

Finally, our analyses suggest that although considerable numbers of TBI victims return to work after an injury, permanent or temporary disability or extended medical care prevents many individuals from returning to a productive life.

Chapter 3 provides Traumatic Brain Injury data from a self-selected sample of 17 TBI model systems that chose to pool their data. The charges and duration for care at other rehabilitation providers may vary. So may the outcomes. Nevertheless, These data include follow-up at one year and is by far the largest case series available. The ratios of rehabilitation charges to acute care charges provide a credible basis for costing TBI rehab care from known acute care hospitalization costs.

Of the TBI cases in rehab, 81 percent were tested for BAC, including 85 percent of motorcyclists and 82.5 percent of other motor vehicle groups (table 8). The proportion of injured motorcycle riders in rehabilitation who tested positive for alcohol in the emergency department was 48.1 percent. By comparison, an analysis of 2001 FARS data found that 37 percent of motorcycle riders killed in crashes tested positive for alcohol. (Shankar, 2003b). The percentage of BAC-positive cases at emergency department admission was virtually identical for TBI rehabilitation patients injured in motorcycle and other motor vehicle crashes.

The marital status of TBI patients did not change noticeably from before the injury to the time of one-year follow-up. In the first year post-TBI, families largely stayed together.

In comparing the four injury categories, Motorcycle TBI patients were most likely to be employed before the injury (80.2%), compared to Other Motor Vehicle (60.7%) and violence (self or other) (51.8%). Employment status changed dramatically one-year post-injury. Overall, the proportion of employed patients fell 34.2 percentage points, from 59.8 percent to 25.6 percent. For motorcyclists, the drop was from 80.2 percent to 44.7 percent and for other motor vehicle crash injuries employment fell from 60.7 percent to 26.1 percent. Unemployment rose 27.6 percentage points overall (from 17.1% to 44.7%), and nearly tripled among motorcyclists (from 10.8% to 31.9%). Those on disability or in sheltered employment more than quadrupled, rising from 1.3 percent to 5.7 percent. The drop in employment may be due to some loss of aptitude or changes in personality. It may also be due to patients still being out of work or finding job search difficult after losing jobs during the months they spent recovering from their TBI.

Although this data set does not address the issue, the employment status of caregivers also may change. (Hall et al., 1994) interviewed 51 caregivers of TBI inpatients by telephone at 12- and 24-months post-injury. Forty-seven percent of caregivers had altered or given up their jobs at one-year post injury and 33 percent at two years post-injury.

Chapter 4 combines prospective payment rates for inpatient rehabilitation with data on the severity of injuries requiring rehabilitation and the probability of requiring rehabilitation to estimate average costs per inpatient rehabilitation and total cost in 2002 of inpatient rehabilitation for motor vehicle and motorcycle injuries. PPS appears to have contained costs, holding them roughly to the levels in the 1999 AMRPA data. Overall, inpatient rehabilitation of motor vehicle injuries cost an estimated $127.5 million in year 2000 dollars and for motorcycle crash injuries cost an estimated $16.3 million.
Appendix B reports that AMRPA data shows that rehabilitation hospital costs for all injury causes in 1999 ranged from $7,613 for the replacement of a lower-extremity joint to $29,495 for a traumatic spinal cord injury. Average cost per day ranged from $716 for a hip fracture to $991 for burns. Traumatic spinal cord injuries required the longest length of stay–on average 34.3 days, while replacement of a lower-extremity joint had the shortest – on average 10.5 days. These findings are relatively consistent among age groups.

Length of stay has been dropping since 1999 due to the implementation of a Medicare prospective payment system. With PPS, rehabilitation hospitals are paid a set fee for each patient, providing an incentive to shorten hospital stay to conform to the available payment. Chapter 4 presents 2002 length of stay and cost data under PPS.

Limitations and unresolved problems
One of the largest problems in an analysis of this sort is mapping the diagnosis codes from one dataset to another. AMRPA’s data tables used rehabilitation impairment categories (RICs). Diagnoses coded using the International Classification of Diseases (ICD) were limited in this dataset and thus unusable. Data from the UDSMR used ICD groups and listed co-morbidities. HCUP data also was coded in the ICD codes, as was the three-state hospital discharge dataset. However, the new Prospective Payment System used diagnostic Case-Mix Groups (CMGs) that considered both diagnostic and impairment information. With UDSMR data on impairment levels, we collapsed the PPS data into RIC-like categories. The diagnostic categories had to be mapped from one system to another, providing the potential for inexact classification. Further imprecision arose in using modest numbers of cases to collapse the data into the common categories. Some data sets lacked cause codes completely or made cause coding voluntary for rehabilitation, leading to many uncoded cases. Other data sets did not distinguish motorcycle injury patients from other motor vehicle injury patients. Thus, the motorcycle cases available could be a biased sample, and some data were not differentiated by motor vehicle user type. This study was not able to separately analyze rehabilitation costs of on-road motorcycle crashes involving off-road motorcycles.

This study analyzed only inpatient rehabilitation costs. Unlike acute-care hospital payments, professional services generally are bundled into rehabilitation payments. We did not study post-discharge rehabilitation costs for physical therapy, speech therapy, occupational therapy and counseling, chiropractic services, doctor visits, etc. The Databook on Nonfatal Injury (Miller, Pindus, Douglass, and Rossman, 1995) gives post-discharge rehabilitation cost estimates by injury diagnosis drawn from the Workers’ Compensation System, which had a vested interest in providing rehabilitation to the point of maximum medical improvement.

Despite its limitations, this study provides a good picture of inpatient rehabilitation costs for injuries in motor vehicle and motorcycle crashes. It also substantially increases our knowledge of the longer-term impacts of motor vehicle and motorcycle injury on functioning, work, and marriage.