Abstract
Testa JA, Malec JF, Moessner AM, Brown AW. Outcome after traumatic brain injury: effects of aging on recovery.
Objective
To identify differences in outcome after traumatic brain injury (TBI) compared with orthopedic injuries as a function of age.
Design
Longitudinal data analyses from an inception cohort.
Setting
Outpatient rehabilitation program.
Participants
Eighty-two orthopedic injury patients and 195 TBI patients.
Interventions
Not applicable.
Main Outcome Measures
Independent living, employment, and level of functioning 1 to 2 years after injury.
Results
Older patients and those with TBI were more likely to have increased dependence postinjury. Older TBI patients were more likely to have changes in employment status compared with orthopedic injury patients younger or TBI. The Mayo-Portland Adaptability Inventory and Disability Rating Scale were moderately predictive of level of functioning, return to employment, and independent living status 1 to 2 years postinjury. Injury severity was only mildly predictive of outcome.
Conclusions
The effect of age on outcome affects recovery from neurologic injuries and, to a lesser extent, orthopedic injuries. Outcome after TBI is best predicted by patients’ age and estimates of level of function at discharge. Findings suggest that older patients and those with TBI have a greater likelihood of becoming physically and financially dependent on others. Rehabilitation efforts should focus on maximizing levels of independence to limit financial and emotional costs to patients and their families.
Key Words
TRAUMATIC BRAIN INJURY (TBI) is a devastating injury, often resulting in death or chronic disability that disrupts family, community, and vocational ties. The annual incidence of TBI is estimated to be 2 million per year,
1
with similar annual incidence rates for rural areas (464/100,000) and urban communities (461/100,000).2
Annegers’s seminal work3
more than 20 years ago indicated a rate of 386 per 100,000 in Mayo Clinic’s home of Olmsted County, MN. The direct medical costs of TBI are estimated to be more than $4 billion annually.4
, 5
Because survivors of moderate or severe TBI have an otherwise normal life span, they will likely live with disability and be at increased risk for subsequent TBI.6
Thus, long-term societal costs are very likely even greater, especially as survivors of TBI age.Epidemiologic studies reveal a bimodal peak incidence of TBI, occurring between the ages of 15 to 24 years and in people 70 years old and older.
7
Falls are the most common cause of TBI in the elderly and are a major cause of morbidity and mortality. Falls are the leading cause of TBI-associated death among people 75 years of age or older.7
In the United States, falls are the leading cause of TBI hospitalizations in people 65 years or older1
and are associated with an increased rate of institutionalization.8
The increased risk for falls and subsequent TBI in older people is multifactorial. Recurrent falls were more likely in people with gait impairment, dizziness, history of stroke, poor mental status, and postural hypotension.
9
Poor visual acuity, high-risk medication, multiple medications, and depression are also risk factors for falls.10
, 11
Falls occur more frequently in people with cognitive impairment than in normal functioning older adults. The estimated annual incidence of falls for people with dementia ranges from 70% to 85%, nearly double the fall risk of cognitively normal older adults.8
, 12
One final important factor is that more older patients are surviving TBI and, as such, are left with residual impairments and long-term disability.13
, 14
, 15
There is a growing body of knowledge that suggests there are greater long-term effects and a poorer prognosis after TBI in older populations.
16
, 17
, 18
, 19
, 20
, 21
, 22
, 23
, 24
, 25
Elderly patients with TBI are known to have higher mortality and worse functional outcome than younger TBI patients, despite less severe injuries.17
, 23
, 24
, 26
Even when injury type and severity are comparable, elderly patients have longer rehabilitation stays, higher rehabilitation costs, and greater levels of disability.18
The risk of death after TBI rises as early as 30 years of age, increasing linearly across decades,17
and is most significant for those over age 70.16
Elderly patients with TBI are more likely to suffer declines in mood, psychosocial functioning, and cognition, and to have less complete recovery in psychosocial functioning after TBI compared with younger TBI patients.
18
, 21
, 22
, 27
, 28
, 29
, 30
Rothweiler et al22
reported that increased age at time of TBI was associated with increased dependence, change in living circumstances, and greater mortality 1 to 2 years after injury. In fact, 1 to 2 years after injury, 80% of their study population over the age of 60 years were moderately disabled, required assistance in basic daily activities, and were unable to resume preinjury activities. Also, 31% no longer lived independently and 75% of those previously gainfully employed were no longer working. There is also a strong relation between increased age and poorer cognitive outcome, especially in those with more severe brain injuries.19
, 21
, 29
, 31
Several studies32
, 33
, 34
, 35
, 36
, 37
have reported that a history of TBI increases the likelihood that a person will develop Alzheimer’s disease, and may actually shortened the time to symptom onset.Increased vulnerability of the aging brain to TBI has been conceptualized as being caused by decreased brain reserve (see review by Lye and Shores
38
). Goldstein and Levin39
suggested that “apart from injury severity, it is likely that a host of preinjury protective and vulnerability factors alter the brain reserve capacity of older adults to the effects of TBI, including medical comorbidities, neuropsychiatric disturbance, substance abuse, and possibly the apolipoprotein E-4 allele.”39
(p750) This has been shown experimentally with animal models of TBI, which found age-dependent molecular changes and greater neuropathology after TBI compared with young animals.40
, 41
In this study, we investigated functional outcome in older TBI patients compared with younger TBI patients and an age-matched orthopedic injury group approximately 1 to 2 years after hospitalization. Based on the literature, we hypothesized that older patients with TBI would have greater disability and poorer functional outcomes compared with younger TBI patients after controlling for injury severity. Additionally, we hypothesized that older TBI patients would have greater disability and poorer functional outcome compared with orthopedic injury patients who had no evidence of brain injury. Our primary goals were to extend the findings of previous studies and to gain a better understanding of the effect of age on recovery after injury.
Methods
Participants
Potential subjects were people admitted consecutively to emergency and acute care who consented to participate and had either TBI or orthopedic injury. The study sample consisted of 195 patients with TBI and 82 patients with orthopedic injury. Of the patients in the TBI group, 87 had mild TBI (worst Glasgow Coma Scale [GCS] score, >12; and absence of trauma-related intracranial neuroimaging abnormalities) and 106 had moderate or severe TBI (worst GCS score, <13; or presence of trauma-related intracranial neuroimaging abnormalities). The study protocol involved matching injury type and severity groups on age, sex, and education. For each patient with moderate or severe TBI, we attempted to recruit a patient with mild TBI and one with orthopedic injury who were of similar age (±5y), sex, and education by broad category (<8y, 9–11y, 12–15y, bachelor’s degree, master’s/doctorate degree). However, because of subject attrition and failure to complete questionnaires, the resulting samples were not perfectly matched. Nevertheless, the subjects for the mild TBI and orthopedic injury groups did not differ statistically from the moderate or severe TBI group in age, sex, and education. We combined mild and moderate or severe TBI patients into a single group after 2 preliminary analysis showed no significant differences in their ages, education, sex, or ratings of preinjury level of independence and employment.
Within the 2 study groups, 2 age levels were examined: (1) a younger group of patients (n=189), ages 18 to 49 years and (2) an older group of patients (n=76), ages 50 to 89. These age groups were selected for analysis based on previous studies, which found poor outcome to be associated with an age threshold of between 50 to 60 years.
18
, 23
Inclusion of 50 year olds in the older sample was considered after no significant differences were found between 50 to 65 year olds and 66 to 85 year olds in our sample on outcome variables. Table 1 presents subject and injury characteristics by age group and injury type.Table 1Subject and Injury Characteristics at Time of Hospital Admission
| Characteristics | TBI | Orthopedic Injury | ||
|---|---|---|---|---|
| Younger (16–49y) | Older (50–89y) | Younger (16–50y) | Older (50–89y) | |
| (n=146) | (n=49) | (n=54) | (n=28) | |
| Age (y) | 29.1±10.6 | 65.8±9.5 | 31.1±10.2 | 64.2±11.3 |
| Education (y) | 12.5±2.0 | 12.5±2.1 | 12.6±2.0 | 12.4±2.0 |
| Sex (% M:F) | 62:38 | 62:38 | 69:31 | 61:39 |
| Etiology (%) | ||||
| Fall | 10 | 51 | 39 | 68 |
| MVC | 70 | 39 | 39 | 25 |
| Other | 20 | 10 | 22 | 7 |
| Ethnicity (%) | ||||
| White | 95.5 | 98 | 98 | 100 |
| Hispanic | 1.5 | 0 | 2 | 0 |
| Black | 1.5 | 0 | 0 | 0 |
| Other | 1.5 | 2 | 0 | 0 |
| ILS-Pre score | 5.0±0.3 | 4.8±0.6 | 5.0±0.0 | 5.0±0.0 |
| VIS-Pre score | 4.7±0.9 | 3.5±2.0 | 4.6±1.0 | 3.3±2.0 |
| ISS score | 20.9±10.8 | 16.8±9.5 | 9.4±6.1 | 9.6±3.1 |
| PTA duration (%) | ||||
| None | 0 | 2 | 100 | 100 |
| <1h | 22 | 27 | ||
| <24h | 11 | 4 | ||
| <7d | 17 | 10 | ||
| <2wk | 7 | 8 | ||
| 2–4wk | 13 | 8 | ||
| >4wk | 10 | 14 | ||
| Unknown | 20 | 27 | ||
| GCS score | 11.3±4.3 | 12.5±3.8 | NA | NA |
NOTE. Values are mean ± standard deviation (SD) unless otherwise noted.
Abbreviations: F, female; ILS-Pre, ILS preinjury; ISS, Injury Severity Scale; M, male; MVC, motor vehicle collision; NA, not applicable; PTA, Posttraumatic amnesia; VIS-Pre, VIS preinjury.
Measures
A series of questionnaires were completed by patients, their significant others, and staff during or immediately after hospitalization and again approximately 1 to 2 years later (mean, 417d; range, 171–869d) in an attempt to measure quantitatively patients’ outcomes. Two of the questionnaires, the Disability Rating Scale (DRS)42 and Mayo-Portland Adaptability Inventory (MPAI), version 2.3,
43
assess impairment, activity, and participation, which we discuss as ratings of “functional outcome” in this article. These measures were obtained at or shortly after hospital discharge and again at long-term follow-up. Independent living and employment status preinjury and at long-term outcome were evaluated. We designate time of assessment, with each questionnaire followed by “pre,” “DC,” or “LT” to specify preinjury, discharge, or long-term follow-up, respectively. These measures are described below.Dependent Variables
Disability Rating Scale
The DRS,
42
a commonly used measure of TBI-related disability and handicap, has demonstrated reliability and validity, but is limited by significant ceiling effects in brain-injured patient populations.44
Higher scores indicate greater disabilities. Possible scores range from 0 to 30. The DRS was completed by hospital staff.MPAI, version 2.3
We used the MPAI, version 2.3, as completed by participants with TBI and a significant other at hospital discharge and at long-term follow-up, as an outcome measure.
43
The MPAI instruments provide ratings of emotions, behavior, functional abilities, and physical disabilities specific to people with brain injury. Three versions allow for assessment of identical areas of impairment, disability, and handicap from the perspectives of professional staff, survivors of brain injury, and their significant others. Versions of the MPAI have generated expected patterns of correlations with neuropsychologic measures and with other instruments used to assess recovery after TBI.45
, 46
, 47
The staff MPAI correlates highly (Spearman ρ=.81) with the DRS, but provides a broader band assessment at lower levels of disability.47
Higher scores indicate greater disability. Possible scores range from 0 to 130.This study extended over several years and the MPAI was concurrently revised.
48
, 49
MPAI 2.3 data were modified using information gained from these subsequent revisions. Specifically, the item for child rearing was eliminated. For the transportation item, levels 0 and 1 were combined and scored 0; levels 2, 3, and 4 were combined and scored 1; and level 5 was scored 2. For the work/school item, levels 1, 2, and 3 were combined and scored 1; levels 4 and 5 were combined and scored 2; level 0 remained 0. For all other items, levels 1 and 2 were combined and scored 1, with higher levels being reduced by 1 point, resulting in the 5-point scale that is used for most items in subsequent versions of the MPAI.Malec JF, Lezak MD. Manual for the Mayo-Portland Adaptability Inventory. 2003. Available at: http://tbims.org/combi/mpai/manual.pdf. Accessed March 8, 2005.
Ongoing research with MPAI suggests that combining significant others’ ratings of patients’ level of function with the patients’ self-report provides a more reliable estimate of outcome than rating each individually.
48
, 49
For this reason, participant and family ratings for all study groups were submitted to Rasch Facets analysis, using rating source as a facet in addition to items and subjects. Following initial Facets analysis of MPAI 2.3 data obtained at about the time of hospital discharge, 11 highly misfitting cases (Infit, >2.0) were eliminated. Subsequent final Facets analysis of 172 cases (51 mild; 73 moderate-severe; 48 orthopedic controls) showed person reliability equal to .92 and separation equal to 3.45 and item reliability equal to .97 and separation equal to 5.36. There was not a reliable difference between rating source (ie, participant vs significant other). We report combined patient and family full scale MPAI Rasch logit scores.Malec JF, Lezak MD. Manual for the Mayo-Portland Adaptability Inventory. 2003. Available at: http://tbims.org/combi/mpai/manual.pdf. Accessed March 8, 2005.
Independent Living Scale and Vocational Independence Scale
The Independent Living Scale (ILS) is a 5-point rating scale that was introduced by Johnston.
51
The Vocational Independence Scale (VIS) has been used in several studies to describe broad categories of independence in work activities.52
, 53
, 54
See Appendix 1, Appendix 2 for descriptions of these scales. For each scale, 1 level is selected that best describes a subject’s level of independence in living and work. People whose primary occupation was attending school were rated at level 4 on the VIS if they returned to school without special education, special accommodations, or supports. However, if they required special support, such as reduced hours, tutoring, or other special accommodations, they were rated at level 3. If they returned to a special education classroom or special school, they were rated at level 2. Those unable to return to school were rated at level 1.Independent Variables
Glasgow Coma Scale
The lowest GCS score obtained within the first 24 hours of admission to the emergency department was collected for most TBI patients. Those scores ranged from 3 to 15. TBI patients with GCS scores in the mild range (13–15) and who had trauma-related computed tomography scan abnormalities were assigned to the moderate to severe group. All patients in the mild TBI group had GCS scores between 13 to 15 and normal neuroimaging.
Posttraumatic amnesia
Posttraumatic amnesia (PTA) was estimated through serial administration of the Galveston Orientation and Amnesia Test (GOAT) or from chart review. Termination of PTA was determined either by scores greater than 70 on the GOAT on 2 consecutive evaluations, or after 2 consecutive evaluations at which the patient was able to follow commands and was oriented to person, time, and place. Evaluations in each case were typically conducted every day or 2. Because of the lack of precision in the measurement of PTA, estimates of PTA were converted to a 6-point ordinal scale (1, <1h; 2, 1–24h; 3, 1–7d; 4, 8–14d; 5, 2–4wk; 6, >4wk; 7, PTA unknown). Most of those with unknown PTA were TBI patients who did not have evidence of PTA or PTA was not documented during their hospitalization.
Injury Severity Scale and Abbreviated Injury Scale
The Injury Severity Scale (ISS) is derived from the Abbreviated Injury Scale (AIS), which ranks injury in 6 distinct body areas including the head. The ISS score is the square of the 3 highest AIS scores and indicates overall severity of multiple injuries.
Statistical Methods
We conducted 2-way analysis of variance (ANOVA) of continuous variables (eg, demographics, medical characteristics) to find group, age, and group by age interactions. We used chi-square tests to compare groups for categoric data, and multiple linear regression analysis to examine the association between injury type (TBI vs orthopedic), adjusted for potentially confounding or moderating variables. Moderating variables were chosen from 3 general categories, including severity of injury (ISS), level of function at discharge (DRS, MPAI), and premorbid factors (age, education, sex, ILS, VIS), if a significant correlation was observed in preliminary analysis. Injury type was forced into the model and a stepwise search was conducted through the remaining variables. A similar model was constructed using age group in place of injury type, and by necessity including only the TBI. As a final step, 2-way ANOVAs were conducted to examine group, age, and group by age interactions at long-term follow-up.
Results
Demographics of Study Groups
Table 1 contrasts basic demographic and injury-related variables for age and injury groups at discharge. Groups did not differ significantly in education, age, ethnicity, or sex. Etiology differed between injury groups and between age groups. At the time of admission, patients were rated on their preinjury independent living (ILS-Pre) and employment (VIS-Pre) status. The majority of patients reported on ILS-Pre that they lived independently before their injury. This did not differ by injury type (P=.06) or age (P=.16). Overall, older patients had lower mean VIS-Pre scores, indicating that more older patients were unemployed or retired at the time of their injury (P<.001), but this did not differ by injury type (P=.63). In the TBI group, 5% of the younger group and 39% of the older group indicated on the VIS-Pre that they were unemployed or retired before their injury. In the orthopedic injury group, 6% of the younger group and 42% of the older group were unemployed or retired before their injury.
Injury Characteristics
ISS scores differed significantly between groups, with TBI patients having more severe injuries overall compared with orthopedic injury patients (F=70.9, P<.001). In the TBI group, 87 patients met the criteria for mild TBI while 106 patients met the criteria for moderate or severe TBI. Fifty-two percent of mild TBI patients and 1% of moderate or severe TBI patients also had evidence of limb fractures, which explains the higher ISS score. The distribution of GCS scores for patients in the moderate or severe TBI group at discharge was 46.4% equal to less than 9; 17.4% equal to 9 to 12; and 36.2% equal to 13 to 15. Those with GCS scores greater than 12 were classified as moderate or severe, based on trauma-related neuroimaging abnormalities. There was no significant difference in GCS score between younger and older TBI patients (P=.115). The distribution of PTA for patients with mild TBI was 58.7% equal to 1; 13.1% equal to 2; and 28.2% equal to 7 (unknown). The distribution of PTA for patients with moderate to severe TBI was 5.5% equal to 1; 12.5% equal to 2; 19.4% equal to 3; 15.3% equal to 4; 23.6% equal to 5; 22.2% equal to 6; and 1.5% equal to 7 (unknown). There was no significant difference in known duration of PTA between younger and older TBI patients (χ2 test=8.49, P=.291). Eighty-two orthopedic injury patients did not have evidence of TBI, PTA, or loss of consciousness.
At or shortly after hospital discharge, the DRS-DC and MPAI-DC were completed to assess patients’ level of functioning. For the DRS-DC, there were significant main effects for group (F1,271=23.201, P<.001) and age (F1,271=6.70, P=.01), but no age by group interaction (F1,271=.153, P=.70). Orthopedic injury patients and older TBI patients had higher scores, indicating greater disability. For the MPAI-DC, a significant main effect for group was found (F1,168=24.85, P<.001), with TBI patients having more impaired scores. There was no significant age effect (F1,168=2.41, P=.12) or age by group interaction (F1,168=.12, P=.73).
Long-Term Outcome
Average scores on long-term outcome measures by age group and injury type are shown in table 2. Median elapsed time postinjury for questionnaire completion was 420.0 days for TBI patients and 405.5 days for orthopedic injury patients. Time to data collection did not differ across groups (F=.74, P=.39). ANOVAs were conducted to examine group, age, and group by age interactions. For the ILS-LT, significant main effects for group (F1,230=5.09, P=.025) and age (F1,230=7.31, P=.007) were found with older patients and those with TBIs more likely to be dependent on others postinjury. There was no age by group interaction (F1,230=1.48, P=.23). For the VIS-LT, there were significant main effects for group (F1,230=20.30, P<.001) and age (F1,230=8.02, P=.005). There was a significant age by group interaction (F1,230=3.95, P=.048). Older TBI patients had the lowest VIS-LT mean score, indicating that they were more likely to be unemployed or retired. For the DRS-LT, there were significant main effects for group (F1,229=4.231, P=.04) and age (F1,229=5.517, P=.02), but no age by group interaction (F1,229=.209, P=.65). For the MPAI-LT, there was a significant main effect for group (F1,152=14.00, P<.001), with TBI patients having more impaired scores 1 to 2 years after injury. There was no significant age effect (F1,152=.01, P=.93) or age by group interaction (F1,152=1.05, P=.31).
Table 2Long-Term Outcome Variables Postinjury
| Variables | TBI by Age Group | Orthopedic Injury by Age Group | Group Effect | Age Effect | Interaction | ||
|---|---|---|---|---|---|---|---|
| 16–49 | 50–89 | 16–49 | 50-89 | ||||
| ILS-LT | 4.9±0.5 | 4.5±1.2 | 5.0±0.0 | 4.8±0.8 | .025 | .007 | .226 |
| VIS-LT | 4.1±1.5 | 3.0±2.0 | 4.7±0.7 | 4.5±1.3 | .000 | .005 | .048 |
| DRS-LT | 0.4±1.5 | 1.1±2.5 | 0.03±0.2 | 0.5±1.5 | .041 | .020 | .648 |
| MPAI-LT | −1.7±1.3 | −2.01±1.3 | −2.97±1.6 | −2.7±1.5 | .000 | .925 | .308 |
NOTE. Values are mean ± SD.
Table 3 displays the percentage of people who endorsed level 5 on the ILS and VIS preinjury and at long-term follow-up, which suggests that they were able to live and work independently. A greater percentage of younger patients compared with older patients reported they lived independently on the ILS-Pre and worked independently on the VIS-Pre. At long-term follow-up, fewer older patients or TBI patients endorsed level 5 on the ILS-LT. Likewise, older TBI patients were less likely to endorse level 5 on the VIS-LT, which suggests that they required some degree of support postinjury.
Table 3The Percentage of Patients Reporting Level 5 Independence on the ILS and VLS
| Percentage Independent | TBI by Age Group | Orthopedic Injury by Age Group | ||
|---|---|---|---|---|
| 16–49 | 50–89 | 16–49 | 50–89 | |
| ILS-Pre (%) | 99.3 | 93.2 | 100.0 | 100.0 |
| ILS-LT (%) | 93.6 | 80.0 | 100.0 | 96.0 |
| VIS-Pre (%) | 83.1 | 61.4 | 80.8 | 73.1 |
| VIS-LT (%) | 68.8 | 48.9 | 79.5 | 88.0 |
Other Predictors of Long-Term Outcome
Once we determined that long-term outcome differed by injury and age group, additional analyses were conducted to assess predictors of long-term outcome. The strength of the relation between moderating variables and outcome measures was assessed with Pearson product moment correlations (table 4). ILS-Pre, and VIS-Pre, MPAI-DC scores, as well as patient’s age were generally moderate predictors of long-term outcome. Sex, education, and injury severity (ISS) had minimal to low effect on outcome. Next, we used multiple linear regression analyses to examine variables predictive of long-term outcome postinjury. Moderating variables with significant correlations were entered into the regression analysis. Long-term outcome was assessed via ILS-LT, VIS-LT, DRS-LT, and MPAI-LT scores as a function of age (table 5) and injury type (table 6). After forcing entry of the primary variable of interest (ie, age group or injury type), several moderating variables entered the model via a stepwise selection process. Overall, outcome differed by injury type and age level and was successfully predicted by several moderating variables. MPAI-DC and DRS-DC scores were the most consistent outcome predictors followed by ILS-Pre and VIS-Pre scores. Injury severity was not a powerful moderating factor for either regression model.
Table 4Correlation Between Moderating Variables and Outcome Variables for All Subjects
| Variables | Premorbid Characteristics | Injury Severity | Discharge Functioning | |||||
|---|---|---|---|---|---|---|---|---|
| Long-Term Outcome | Age | Sex | Education | ILS-Pre | VIS-Pre | IS | DRS-DC | MPAI-DC |
| ILS-LT | −.220 | .021 | .107 | .347 | .300 | −.081 | −.365 | −.269 |
| VIS-LT | −.239 | −.075 | .084 | .227 | .455 | −.078 | .100 | −.340 |
| DRS-LT | .172 | .003 | −.138 | −.500 | −.321 | .123 | .456 | .285 |
| MPAI-LT | −.040 | .044 | −.283 | −.089 | −.052 | .229 | .091 | .526 |
NOTE. Boldface values are significant at P<.01.
Table 5Regression Analysis of the Moderating Variables on Dependent Variables by TBI Age Groups
| Long-Term Outcome | R | R2 | SEE | R2 Change | F Change | df (1,2) | Signif F Change |
|---|---|---|---|---|---|---|---|
| ILS-LT | |||||||
| Age group | .241 | .058 | 0.76 | .058 | 6.941 | 1,113 | .010 |
| DRS-DC | .499 | .249 | 0.68 | .192 | 28.581 | 1,112 | .000 |
| ILS-Pre | .558 | .312 | 0.65 | .062 | 10.082 | 1,111 | .002 |
| MPAI-DC | .594 | .353 | 0.63 | .042 | 7.063 | 1,110 | .009 |
| VIS-LT | |||||||
| Age group | .288 | .083 | 1.66 | .083 | 10.254 | 1,113 | .002 |
| VIS-Pre | .585 | .343 | 1.42 | .260 | 44.217 | 1,112 | .000 |
| MPAI-DC | .652 | .413 | 1.34 | .070 | 13.225 | 1,111 | .000 |
| DRS-LT | |||||||
| Age group | .172 | .029 | 1.18 | .029 | 3.372 | 1,111 | .069 |
| DRS-DC | .557 | .310 | 1.53 | .281 | 44.745 | 1,110 | .000 |
| ILS-Pre | .693 | .480 | 1.34 | .170 | 35.707 | 1,109 | .000 |
| MPAI-DC | .717 | .514 | 1.30 | .034 | 7.514 | 1,108 | .007 |
| ISS | .731 | .535 | 1.28 | .020 | 4.687 | 1,107 | .033 |
| MPAI-LT | |||||||
| Age group | .101 | .010 | 1.34 | .010 | 0.892 | 1,86 | .348 |
| MPAI-DC | .463 | .214 | 1.20 | .204 | 20.031 | 1,85 | .000 |
| Education | .508 | .258 | 1.18 | .044 | 4.984 | 1,84 | .028 |
Abbreviations: SEE, standard error of the estimate; Signif, significant.
Table 6Regression Analysis of the Moderating Variables on Dependent Variables by Injury Groups
| Long-Term Outcome | R | R2 | SEE | R2 Change | F Change | df (1,2) | Signif F Change |
|---|---|---|---|---|---|---|---|
| ILS-LT | |||||||
| Injury group | .202 | .041 | 0.70 | 0.041 | 6.712 | 1,157 | .010 |
| DRS-DC | .390 | .152 | 0.66 | 0.111 | 20.382 | 1,156 | .000 |
| MPAI | .489 | .239 | 0.63 | 0.087 | 17.795 | 1,155 | .000 |
| ILS-Pre | .543 | .30 | 0.06 | 12.258 | 8.495 | 1,154 | .001 |
| VIS-LT | |||||||
| Injury group | .208 | .043 | 1.57 | 0.043 | 7.067 | 1,157 | .009 |
| VIS-Pre | .456 | .208 | 1.44 | 0.165 | 32.488 | 1,156 | .000 |
| MPAI-DC | .562 | .315 | 1.34 | 0.107 | 24.293 | 1,155 | .000 |
| DRS-LT | |||||||
| Age group | .162 | .026 | 1.63 | 0.026 | 4.218 | 1,157 | .042 |
| ILS-Pre | .511 | .261 | 1.43 | 0.235 | 49.559 | 1,156 | .000 |
| DRS-DC | .627 | .393 | 1.30 | 0.132 | 33.591 | 1,155 | .000 |
| MPAI-DC | .672 | .452 | 1.24 | 0.059 | 16.678 | 1,154 | .000 |
| VIS-Pre | .682 | .466 | 1.23 | 0.014 | 3.931 | 1,153 | .049 |
| MPAI-LT | |||||||
| Injury group | .057 | .003 | 1.49 | 0.003 | 0.407 | 1,126 | .525 |
| MPAI-DC | .526 | .277 | 1.27 | 0.274 | 47.312 | 1,125 | .000 |
| Education (y) | .566 | .320 | 1.24 | 0.043 | 7.899 | 1,124 | .000 |
Within the TBI group (see table 5), age of injury predicted ILS-LT and VIS-LT score at follow-up. Thus, even though the majority of older adults were gainfully employed and lived independently before their TBI, they were not as independent 1 to 2 years postinjury. As would be expected, whether a person returned to independent living or work was also predicted in part by their preinjury level of independence, assessed with the ILS-Pre and VIS-Pre, respectively. ILS-LT and VIS-LT scores were also partially predicted by level of function on the DRS-DC and MPAI-DC. Long-term level of function on the DRS-LT was predominantly predicted by DRS-DC, ILS-Pre, and MPAI-DC scores, and to a small degree by the ISS. Similarly, long-term level of function on the MPAI-LT was predominantly predicted by MPAI-DC score, which suggests that assessment of function at discharge is an important prognosticator of long-term function.
Next, we analyzed predictors of outcome as a function of injury type. As shown in table 6, ILS-LT, VIS-LT, and DRS-LT scores were in part predicted by injury type. ILS score was successfully predicted by level of function at discharge (DRS-DC, MPAI-DC) and premorbid level of independence. VIS-Pre and MPAI-DC scores predicted return to employment on the VIS-LT. Long-term level of function, as rated by the DRS-LT, was best predicted by ILS-Pre and VIS-Pre scores, and discharge level of function (on both the DRS-DC and MPAI-DC). Variance in the MPAI-LT was best explained by MPAI-DC score and years of education.
Discussion
Our primary purpose in this study was to investigate the impact of age on outcome after TBI. Several issues have traditionally confounded TBI outcome studies. First, etiologies of TBI differ across age groups. Young adults incur more injuries as a result of motor vehicle collisions while older adults are more prone to TBI as a result of falls. This is important because pathophysiologic mechanisms differ as a result of injury etiology, making controlled comparisons across age groups complicated, even in this study. Second, although many studies have focused on the effect of age on outcome, few have utilized a comparison group to determine the effect of neurologic insult on outcome. That is, the relative contribution of neurologic injury on recovery has not been teased apart from recovery from trauma in and of itself. Neurologic insult may have an additive negative effect on people over and above that caused by trauma alone.
To investigate the effect of age on recovery after trauma, we posed 2 hypotheses: (1) older patients with TBI would have greater disability and poorer functional outcome compared with younger TBI patients after controlling for injury severity, and (2) older TBI patients would have greater disability and poorer functional outcome compared with orthopedic injury patients with no evidence of brain injury. Our results indicate that both age and type of injury have an effect on recovery after trauma. Thus, our overall findings support these hypotheses.
TBI patients and older patients, regardless of injury type, lived with some degree of support. Similarly, older TBI patients were more likely to have a changed employment status after their injury compared with both younger TBI and orthopedic injury patients and older orthopedic injury patients. Although a greater percentage of older adults were unemployed or retired before their injury, an even greater proportion of older TBI patients failed to return to work after being injured compared with older and younger orthopedic injury patients (see table 3).
Interestingly, injury severity had little effect on outcome, as shown in the regression analysis, and was only important in explaining a small amount of variance on the DRS for TBI patients. This result is consistent with other studies that failed to find an association between injury severity and later return to employment
55
, 56
, 57
, 58
and independent living.22
Recovery from injury can be measured on many levels: level of disability, independent living, community integration, emotional functioning, or cognitive ability. Our focus in this study was on the effect of age on functional outcome after injury. Thus, we chose scales that assessed basic levels of function, including independent living, employment, and 2 measures of impairment, activity, and participation—the DRS and MPAI. These 2 scales differ in several ways. The DRS, which was completed by hospital staff, is used to quantify recovery of a patient after injury. It assesses 4 categories: arousability, awareness, and responsivity; cognitive ability for self-care activities; dependence on others; and employability. The DRS is known to be limited at 2 extremes, higher functional levels (DRS score, <3) and severe levels (DRS score, >25).42,44,59 Two items, employment and level of independent functioning, are directly relevant to community integration
60
, 61
and in our sample were responsible for most of the variability on the DRS. The MPAI, however, is sensitive to milder levels of impairment. The MPAI provides ratings of emotions, behavior, functional abilities, and physical disabilities and makes possible the assessment of identical areas of impairment, activity, and participation from the perspectives of TBI survivors and their significant others. Consequently, these measures differ in how functional outcome is defined and assessed.Family and patient ratings of functional outcome may differ from that of hospital staff members, thereby limiting the direct comparability of the DRS and MPAI. Malec et al
46
compared the responses of medical staff, patients with TBI, and significant others on the MPAI and found significant differences between groups. Ratings were biased by a number of factors that included lay versus professional use of terminology, values, observational opportunities, and the impact and burden of brain injury sequelae. Malec suggested that “the person with ABI [acquired brain injury] may tend to adopt a positive bias; whereas, staff and SO [significant others] may be biased toward negative ratings.”48
(p573) Thus, the use of measures that assess outcome from multiple perspectives (ie, patient, significant other, medical staff) will obtain the most representative view of how the patient is functioning. In our sample, staff reports of functioning on the DRS and combined patient and significant other reports on the MPAI at discharge were moderately predictive of long-term functional outcome, return to employment, and independent living status. Thus, the DRS and MPAI appear sensitive to the effects of TBI and are useful measures of patients’ residual impairment at follow-up.We should remark on several limitations inherent in this study. First, cross-sectional studies cannot detect differences in rate of change between age groups.
62
Our study was limited to investigating change between 2 rather large age groups. As a result, our findings are limited by lack of multiple time points found in other studies, particularly those with large trauma-based samples.16
, 24
, 63
Another important concern is the cutoff age used for the older group. In today’s society, 50 years of age is not considered “elderly.” However, many studies have found subtle or not-so-subtle cognitive and physical changes that begin in the fifth decade of life. By including a younger “elderly” population, any aging effects would likely be a conservative estimate of the effects of aging on recovery from TBI.A final limitation of our findings is that pre- or postinjury cognitive deficits were not assessed. Other research has suggested that elderly people who incur a TBI from a fall are more likely to have evidence of cognitive decline and, as such, may be functioning at a lower cognitive level at the time of their injury.
8
, 12
We did not investigate the connection between cognitive status and functional outcome in this study.Conclusions
Our results have several implications. First, increasing age negatively affects outcome after neurologic and, to a lesser extent, orthopedic injuries. When trauma affects the central nervous system, age becomes a risk factor for poorer prognosis. Furthermore, recovery from TBI is not strongly predicted by general injury severity, which is consistent with previous findings.
56
, 57
, 58
, 59
Instead, a combination of hospital staff, patient, and family estimates of level of function at discharge provide a more sensitive and important prognostication of long-term outcome. Finally, older people have a greater likelihood of becoming functionally dependent after a traumatic injury and, more specifically, of not returning to work after TBI. The implications of these findings are important for rehabilitation efforts, as well as for long-term planning. Functional gains in level of independence translate to increased quality of life for older TBI patients. As reported by Cifu et al,18
older TBI patients have slower rates of functional recovery, longer stays in rehabilitation at greater financial cost, and greater levels of disability than do younger TBI patients with comparable injuries. Thus, rehabilitation efforts with older TBI patients should focus on maximizing levels of independence to limit financial and emotional costs to patient’s and their families.Appendix 1. Vocational independence scale
- Level 5: Independent. Community-based work (at least 15h/wk) without external supports.
- Level 4: Transitional. Community-based work (at least 15h/wk) with temporary supports, such as, job coach, reduced hours OR enrollment in an educational or training program.
- Level 3: Supported. Community-based work with permanent supports or less than 15 hours per week OR volunteer work.
- Level 2: Sheltered. Work in a sheltered workshop.
- Level 1: Unemployed.
Appendix 2. Independent living scale
- Level 5: Independent and responsible for self. The person needs virtually no help or supervision from others.
- Level 4: Mostly independent. The person primarily cares for self and needs little or no supervision but does receive some help (eg, 0.5–1h/d).
- Level 3: Independent during much of the day. The person can be left alone for 8 hours or long enough for the caregiver to hold down a full-time job.
- Level 2: Contribution to self-care or supervision. The person can be left alone for 2 hours or long enough to go to the store or do other chores.
- Level 1: The person needs day and night (24h/d) supervision or care. No effective contribution to self-care or supervision.
References
- Centers for Disease Control and Prevention.Traumatic brain injury in the United States. CDC, Atlanta1999
- Incidence of mild and moderate brain injury in the United States, 1991.Brain Inj. 1996; 10: 47-54
- The incidence, causes, and secular trends of head trauma in Olmsted County, Minnesota, 1935-1974.Neurology. 1980; 30: 912-919
- Head injuries costs and consequences.J Head Trauma Rehabil. 1991; 6: 76-91
- The lifetime cost of injury.Inquiry. 1990; 27: 332-343
- Long-term survival after traumatic brain injury.NeuroRehabilitation. 2004; 19: 37-43
- Traumatic brain injury in the United States.J Head Trauma Rehabil. 1999; 14: 602-615
- Senile dementia of the Alzheimer’s type.J Gerontol. 1987; 42: 412-417
- Falls in the elderly.Am J Epidemiol. 1996; 143: 1129-1136
- Falls in cognitive impairment and dementia.Clin Geriatr Med. 2002; 18: 159-173
- Multifactorial intervention after a fall in older people with cognitive impairment and dementia presenting to the accident and emergency department.BMJ. 2003; 326: 73-79
- Risk factors for falls among elderly persons living in the community.N Engl J Med. 1988; 319: 1701-1707
- Traumatic brain injury in elderly individuals.Phys Med Rehabil Clin North Am. 1992; 3: 441-449
- Rehabilitation of the elderly crash victim.Clin Geriatr Med. 1993; 9: 473-483
- Geriatric rehabilitation.Arch Phys Med Rehabil. 1993; 74: S406-S412
- Mortality risk after head injury increases at 30 years.J Am Coll Surg. 2003; 197: 711-716
- Patient age and outcome following severe traumatic brain injury.J Neurosurg. 2003; 99: 666-673
- Functional outcomes of older adults with traumatic brain injury.Arch Phys Med Rehabil. 1996; 77: 883-888
- Cognitive and neurobehavioral functioning after mild versus moderate traumatic brain injury in older adults.J Int Neuropsychol Soc. 2001; 7: 373-383
- Cognitive and behavioral sequelae of closed head injury in older adults according to their significant others.J Neuropsychiatry Clin Neurosci. 1999; 11: 38-44
- Head injuries and cognitive decline among older adults.Neurology. 1999; 52: 557-562
- Aging effect on psychosocial outcome in traumatic brain injury.Arch Phys Med Rehabil. 1998; 79: 881-887
- Survival after severe brain injury in the aged.Arch Surg. 1993; 128 (discussion 93-4): 787-793
- Traumatic brain injury in the elderly.J Trauma. 2002; 53 (discussion 223-4): 219-223
- Traumatic brain injury outcome in older adults.J Clin Geropsychol. 2001; 7: 161-187
- Age and outcome of severe head injury.Acta Neurochir. 1979; 75: S37-S49
- Neurobehavioural consequences of closed head injury in the elderly.Brain Inj. 1997; 11: 871-875
- Age and functioning after mild traumatic brain injury.Brain Inj. 2001; 15: 857-864
- The effects of normal ageing on neuropsychological functioning following traumatic brain injury.Brain Inj. 1998; 12: 569-576
- The functional effects of head injury in the elderly.Brain Inj. 1987; 1: 183-188
- Neurobehavioral consequences of closed head injury in older adults.J Neurol Neurosurg Psychiatry. 1994; 57: 961-966
- Severe head injury hastens age of onset of Alzheimer’s disease.J Am Geriatr Soc. 1989; 37: 970-973
- Traumatic brain injury and time to onset of Alzheimer’s disease.Am J Epidemiol. 1999; 149: 32-40
- Alzheimer’s disease after remote head injury.J Neurol Neurosurg Psychiatry. 1997; 62: 119-124
- Head trauma and age of onset of dementia of the Alzheimer type.JAMA. 1987; 257: 2289-2290
- The association between head trauma and Alzheimer’s disease.Am J Epidemiol. 1990; 131: 491-501
- Head trauma as a risk factor for Alzheimer’s disease.Int J Epidemiol. 1991; 20: S28-S35
- Traumatic brain injury as a risk factor for Alzheimer’s disease.Neuropsychol Rev. 2000; 10: 115-129
- Cognitive outcome after mild and moderate traumatic brain injury in older adults.J Clin Exp Neuropsychol. 2001; 23: 739-752
- Age-dependent synuclein pathology following traumatic brain injury in mice.Exp Neurol. 2003; 184: 214-224
- Laser capture microdissection and analysis of amplified antisense RNA from distinct cell populations of the young and aged rat brain.Brain Res Mol Brain Res. 2004; 122: 47-61
- Disability rating scale for severe head trauma.Arch Phys Med Rehabil. 1982; 63: 118-123
- Refining a measure of brain injury sequelae to predict postacute rehabilitation outcome.J Head Trauma Rehabil. 2000; 15: 670-682
- Functional measures after traumatic brain injury.J Head Trauma Rehabil. 1996; 11: 27-39
- Factor analysis of the Mayo-Portland Adaptability Inventory.Brain Inj. 1997; 11: 469-482
- Differing problem perceptions of staff, survivors, and significant others after brain injury.J Head Trauma Rehabil. 1997; 12: 1-13
- Relationship of the Mayo-Portland Adaptability Inventory to functional outcome and cognitive performance measures.J Head Trauma Rehabil. 1994; 9: 1-15
- Comparability of Mayo-Portland Adaptability Inventory ratings by staff, significant others and people with acquired brain injury.Brain Inj. 2004; 18: 563-575
Malec JF, Lezak MD. Manual for the Mayo-Portland Adaptability Inventory. 2003. Available at: http://tbims.org/combi/mpai/manual.pdf. Accessed March 8, 2005.
- Facets for Windows. 3.41.2 ed.Winsteps, Chicago2003
- Outcomes of community re-entry programmes for brain injury survivors. Part 2.Brain Inj. 1991; 5: 155-168
- Maximizing vocational outcome after brain injury.Mayo Clin Proc. 1995; 70: 1165-1171
- A medical/vocational case coordination system for persons with brain injury.Arch Phys Med Rehabil. 2000; 81: 1007-1015
- Impact of comprehensive day treatment on societal participation for persons with acquired brain injury.Arch Phys Med Rehabil. 2001; 82: 885-894
- Acute predictors of return to employment after traumatic brain injury.Arch Phys Med Rehabil. 2002; 85: 635-641
- A comparison of acute and postdischarge predictors of employment 2 years after traumatic brain injury.Arch Phys Med Rehabil. 2001; 82: 435-439
- Comparison of the predictive power of socio-economic variables, severity of injury and age on long-term outcome of traumatic brain injury.Brain Inj. 2002; 16: 9-27
- Prediction of employment outcome one to three years following traumatic brain injury (TBI).Brain Inj. 1998; 12: 255-263
- Glasgow Outcome Scale and Disability Rating Scale.Arch Phys Med Rehabil. 1985; 66: 35-37
- Measuring psychosocial recovery after brain injury.Arch Phys Med Rehabil. 2004; 85: 538-545
- Assessing traumatic brain injury outcome measure for long-term follow-up of community-based individuals.Arch Phys Med Rehabil. 2001; 82: 367-374
- The need for multiple time points in aging studies.Neurobiol Aging. 2004; 25: 3-4
- The impact of aging on head injuries.J Trauma. 2002; 53: 391
Article Info
Footnotes
Supported by the Traumatic Brain Injury Model Systems, National Institute for Disability and Rehabilitation Research (grant no. H133A020507).
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated.
Identification
Copyright
© 2005 American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation. Published by Elsevier Inc. All rights reserved.
