Comparative Outcomes from Severe
Traumatic Brain Injury without Inpatient Rehabilitation
FINAL REPORT
There has been little research comparing the different types and approaches
to rehabilitation following brain injury. The current project proposed
to compare initial status and one-year outcomes for persons with severe
traumatic brain injuries who received acute rehabilitation, those who
received care in other institutional settings, and those who were discharged
directly to home. It was hypothesized that clinical differences in readiness
for acute rehabilitation, as well as differences in severity of functional
abilities, drive differences in trauma service discharge disposition for
these three groups. We further hypothesized that functional abilities
at one year would not differ significantly; however, community integration
and subjective well-being were expected to be better for those individuals
who received rehabilitation in an inpatient program. Results of this study
can provide useful information for examining the structure and definition
of the TBI Model System.
Background and preliminary studies
The continuum of rehabilitation services provided immediately post-trauma
care has been undergoing change over the past decade (DeJong, Wheatley,
& Sutton, 1995). Changes in health care economics driven by the use
of diagnostic related groups (DRG's) for reimbursement in federally-funded
programs, as well as the penetration of managed care into third party
reimbursement has shortened the time of the initial, acute hospitalization
(Harrison-Felix et al., 1996), resulting in decreased readiness for the
typical three hours per day of therapy provided in acute rehabilitation.
At the same time, the effort to reduce overall health costs has resulted
in a new type of provider--the subacute rehabilitation unit--that is presumably
less costly than hospital-based, acute rehabilitation. In 1995, the Commission
on Accreditation of Rehabilitation Facilities (CARF, 1997) expanded its
definition of Comprehensive Inpatient Rehabilitation services to include
discrete categories of programs based on the medical acuity of the patient
and whether the program was housed in a hospital or skilled nursing facility.
One year later, new standards for home-based rehabilitation were developed
to address the increasing number of interdisciplinary rehabilitation programs
providing services in the patient's own home.
Despite these treatment alternatives, there has been very little research
comparing the effectiveness of different methods of comprehensive rehabilitation.
One exception, a study of patients with stroke conducted by Keith, Wilson,
and Gutierrez (1995) found subacute rehabilitation to be twice as cost-effective
as acute-based services. The systematic study of alternative treatment
approaches has been hampered by the absence of clinical guidelines regarding
what type of service may be most effective for whom. Indeed, economic
factors related to an individual's reimbursement, combined with the structure
of the delivery system within a particular community, appear to dictate
the course of rehabilitation more than patients' clinical needs.
Descriptive statistics of admissions to the OSU trauma service in 1996
were compiled. Ninety-five individuals were admitted with a severe TBI
(GCS score 8) and 20 individuals sustained a TBI of moderate severity
(8< GCS score <13). As would be expected, 46% of those individuals
with severe injuries did not survive, while 5% of those with moderate
injury did not. Of those who survived, only 10% of those with severe injuries
and 5% of those with moderate severity were discharged to extended care
facilities, skilled nursing facilities, or other acute hospitals. On the
other hand, only 70% of those surviving with severe injuries were discharged
to acute rehabilitation; and only 27% of those with moderate injury. The
remainder, 20% with severe injury and 68% with moderate, were discharged
from trauma service directly to home. While comparative data from a time
that preceded the current economic influences on healthcare were not available,
these statistics suggested that to gain a complete picture of outcomes
from severe traumatic brain injury would require investigation of those
individuals who did not receive acute rehabilitation, particularly those
being discharged to home. This question is consistent with the issues
raised in the final priorities for the NIDRR TBI Model Systems program,
in which programs were invited to examine alternative pathways to acute
rehabilitation.
The current research project proposed to compare initial status and one-year
outcomes for persons with severe traumatic brain injuries who receive
acute rehabilitation, those with severe injuries who receive care in other
institutional settings, and those with severe injuries who are discharged
directly to home. It was hypothesized that clinical differences in readiness
for acute rehabilitation, as well as differences in severity of functional
abilities, drive differences in trauma service discharge disposition for
these three groups. We further hypothesized that functional abilities
at one year will not differ significantly among the groups; however, we
posited that community integration and subjective well-being will be better
for those individuals who received an earlier, more thorough course of
rehabilitation in an inpatient program.
Objectives
The hypotheses for this study were based on the presumption that clinical
differences rather than economic factors determine the patterns of service
delivery following initial trauma care. As a result, patients who received
post-trauma care in institutional settings other than inpatient rehabilitation
had medical reasons that preclude inpatient rehabilitation. We also hypothesized
that patients going home were those with fewer functional limitations
as assessed by the Functional Independence Measure (FIM). Thus, because
referral patterns reflect clinical conditions, systematic differences
will be evident between those receiving inpatient rehabilitation, those
receiving care in other institutional settings, and those going home.
1. Patients with severe brain injuries receiving post-trauma care in institutional settings other than inpatient rehabilitation will have (a) more medical complications, (b) lower functional physical ability, and (c) lower functional cognitive ability at discharge from acute care than those referred to inpatient rehabilitation.
2. Patients with severe brain injuries discharged to home will (a) fewer medical complications, (b) better functional physical ability, and (c) better functional cognitive ability at discharge from acute care than those referred to inpatient rehabilitation.
Systematic clinical differences that are present will cause selection
biases when comparing later outcomes. To the extent that variance due
to these differences could be accounted for, we hypothesized that specialized,
acute inpatient rehabilitation will result in better long-term community
integration and subjective well-being, though functional independence
may not differ:
3. After differences in injury severity, other medical complications, and acute functional status are accounted for, functional independence (measured by FIM scores) one year post-injury will not differ among patients who received inpatient rehabilitation, those who received treatment in other institutional settings, and those who went directly home.
4. After differences in injury severity, other medical complications, and acute functional status are accounted for, community integration (measured by the Community Integration Questionnaire (CIQ), Willer, et al., 1993) and subjective well-being (measured by the Satisfaction With Life Scale; Pavot & Diener, 1993) one year post-injury will be better for patients who received inpatient rehabilitation, than for those who received treatment in other institutional settings, and those who went directly home.
Method
Subjects. The proposed sample for the current study was to consist
of three groups: (1) 25 patients with an initial GCS 8 who received inpatient
acute rehabilitation and otherwise met Model Systems inclusion criteria
(2) 25 patients who had an initial GCS 8 and met the inclusion criteria
of the Model Systems dataset except they were discharged from the trauma
service to an institutional setting other than acute rehabilitation; and
(3) 25 patients who had an initial GCS 8 and met the inclusion criteria
of the Model Systems dataset except they were discharged from the trauma
service to home.
Instrumentation and procedures. Data collection procedures for
those eligible and enrolled in the Model System dataset did not differ
from the standard procedures. Persons receiving alternative pathways of
care were identified by trauma unit staff upon discharge and enrollment
in the study would be sought. Data to be collected within 72 hours of
discharge included demographics and premorbid history, functional status,
as well as information about injury severity and medical complications.
While hypotheses related to the one year follow-up, subjects were also
contacted at 3 and 6 months post-injury to obtain information about services
being received, verify residence, and monitor work re-entry.
Analyses. Two phases of data analysis were to be conducted: the
first, primarily descriptive, examines what alternate pathways, taken
for what reasons, with what potential selection biases. Analysis of variance
was to be used to identify significant differences in initial GCS score
and number of medical complications. In the second phase, outcomes at
one year were to be compared using analysis of covariance to minimize
the effects of selection biases identified in the first phase. The proposed
sample sizes would have allowed 80% power for detecting differences of
4.0 or more on the Community Integration Questionnaire total score, and
6.5 or more on the Satisfaction With Life Scale total score.
Progress/major findings
Considerable difficulties were encountered with recruiting a sufficient
number of subjects. Within the first 18 months of the study, only two
subjects who did not receive inpatient rehabilitation were enrolled. Discussions
with the trauma coordinator did not suggest that any subjects were being
missed. In order to increase sample size, the decision was made to collaborate
with Charlotte Institute of Rehabilitation, however the combined efforts
still resulted in a recruitment rate that was not sufficient to obtain
the required sample size. Extensive retrospective chart reviews were conducted
to determine the source of the low recruitment rate. In the process, it
was found that one subject from OSU was missed. While there were 11 additional
patients with an initial GCS less than or equal to 8, clinical data suggested
that they had not sustained a severe TBI. Seven of these individuals were
chemically paralyzed, and had a GCS of 15 when the agent wore off. Four
were intubated, but had a Motor Score of 5.
It was concluded that the original assumption that a significant number
of individuals with severe TBI were being discharged to home was incorrect.
Due to an admission policy allowing individuals with pending Medicaid
to be admitted, nearly all persons who are admitted to OSU and survive
a severe TBI receive inpatient rehabilitation. These findings also have
important implications regarding the use of the GCS score in outcome studies,
suggesting that an alternate measure of severity be used to either replace
or supplement the GCS score.