Poster 21: Postacute mortality after moderate to severe traumatic brain injury

Objective: To establish very long term mortality rates and predictors of mortality for persons after moderate to severe traumatic brain injury (TBI) in adults. Design: A retrospective cohort study design. Setting: A large inpatient rehabilitation hospital in Pennsylvania. Participants: Consecutive records of persons with moderate to severe TBI who were discharged from the hospital in the years 1974-1984, 1988, and 1989 (N=642). Interventions: Not applicable. Main Outcome Measures: Eligible participants were traced and mortality was ascertained up to 24 years postinjury. Results: Poisson regression analyses revealed at least a 2-fold increased risk for mortality in relation to the general population. Preinjury characteristics and levels of disability at discharge from inpatient rehabilitation were among the strongest predictors of mortality. Conclusions: Using rigorous methodology, this study quantifies the increased risk for premature death in the postacute TBI population after a moderate to severe head injury and has clinical relevance for the prevention of premature death.     

Article 1: Repetitive traumatic brain injury: a population-based risk analysis of hospital discharge and emergency department visits among persons who had more than 1 brain injury in south carolina, 1996-2001

Objectives: To quantify the incidence of repetitive traumatic brain injury (TBI), to identify risk factors, and to determine frequently associated injury mechanisms. Design: Cross-sectional. Setting: Hospital discharges and emergency department (ED) visits. Participants: 70,671 persons with TBI unduplicated with personal identifiers. Interventions: Not applicable. Main Outcome Measures: Repetitive TBI. Results: We identified 945 people with >1 occurrence of TBI; 35% were of higher severity than the preceding ones. Mean number of days elapsing between episodes was 431±390 days. The risk of repetitive TBI was significantly associated with chronic health problems (OR=1.76; 95% CI, 1.22-2.54); residing in counties where the poverty rate was >25% (OR=1.62; 95% CI, 1.39-1.89); and being a Medicare-insured person (OR=1.66; 95% CI, 1.39-1.99). The most common external causes of injury accounting for repetitive TBI were falls (30%), struck by object (20%), and motor vehicle (9%). Conclusion: Repetitive TBI is common in South Carolina. Persons with chronic health problems have the highest risk of repetitive TBI. This suggests the need to evaluate the effect of polypharmacy and disease-specific risk patterns for prevention.     

Article 12: The prevalence of persistent postconcussive symptoms and self-reported functioning in an active duty population

Objectives: To measure (1) the prevalence of postconcussive syndrome (PCS) symptoms associated with prior traumatic brain injury (TBI) and (2) their effects on the functioning of soldiers. Design: Population survey. Setting: Army paratrooper base. Participants: Convenience sample of 2055 soldiers. Interventions: Not applicable. Main Outcome Measures: Self-reported PCS symptoms and role functioning. Results: 947 soldiers reported prior TBI, nearly all of which was mild (loss of consciousness, ≤20min), and most of which (60%) occurred within the last 5 years. Soldiers with prior mild TBI reported more moderate to severe symptoms than soldiers denying prior TBI (mean symptoms, 4.2 vs 2.1), but fewer symptoms than soldiers reporting more severe TBI (mean symptoms, 5.4). Confirmatory factor analysis replicated the clusters of symptoms previously identified in a small, clinical sample by Cicerone and Kalmar (1995). Functioning correlated weakly with symptoms (with cognitive and affective symptom clusters: r range, .09-.19; P<.001). Conclusions: Even highly functioning individuals may have sequelae from prior TBI that can affect functioning.     

Poster 22: Loss to follow-up and potential systematic bias in survivors of traumatic brain injury 12 months postinjury

Objectives: To predict which survivors of traumatic brain injury (TBI) are more likely to be lost to follow-up and to examine the possibility of systematic bias in TBI research. Design: Characteristics of participants initially enrolled in a prospective, longitudinal follow-up study were compared with those lost and found for the 12-month follow-up using logistic regression. Setting: A prospective, longitudinal database. Participants: 1153 adult survivors of TBI. Interventions: Not applicable. Main Outcome Measures: Lost participants were those with no 12-month follow-up interview, excluding the deceased. Results: Follow-up was attempted with 1153 survivors of TBI. Initial data were collected from hospital charts. The 12-month follow-up was a telephone interview. At follow-up, 21 had died (1.8%), 574 were lost to follow-up (49.8%), and 558 (48.4%) were located and interviewed. Variables related to follow-up availability included injury severity (critical injuries vs moderate injuries) with critical injuries associated with follow-up failure (OR=.24; 95% CI .14-.38). Conclusions: Differences in information-gathering techniques led to substantial loss of participants. Differential loss to follow-up potentially biases TBI outcome research.     

Poster 36: Neuropsychiatric symptoms in patients with traumatic brain injury: 2 years postdischarge

Objective: To assess the prevalence of neuropsychiatric symptoms in traumatic brain injury (TBI) 2 years postdischarge from inpatient rehabilitation and its associated caregiver distress. Design: Follow-up study. Setting: Inpatient acute rehabilitation hospital and community. Participants: 39 consecutive patients with TBI. Interventions: Not applicable. Main Outcome Measure: The Neuropsychiatric Inventory. Results: Neuropsychiatric symptoms were present in 75% of our sample 2 years postdischarge from acute rehabilitation. Among these symptoms, the most frequently encountered were: irritability, 48.7%; depression, 43.6%; appetite and eating changes, 25.6%; night-time and sleep behavior, 23.1%; apathy, 20.5%; anxiety, 17.9%; agitation and aggression, 17.9%; and disinhibition, 10.3%. The least frequently observed behaviors included: hallucinations, 5.1%; euphoria and elation, 2.6%; and delusions, 0%. At 2-year follow-up, anxiety, agitation and aggression, apathy, and irritability caused the most emotional distress to the caregivers, while hallucination and euphoria caused the least. Conclusions: Neuropsychiatric symptoms are prevalent in patients with TBI after discharge from acute rehabilitation. These symptoms are associated with significant caregiver distress. These findings further support the continued need for adequate psychologic assessment and possible intervention postdischarge.     

Poster 41: Objectively measured personality traits in the prediction of early neurobehavioral symptoms after mild traumatic brain injury

Objective: To determine if objectively measured preinjury personality predicts early neurobehavioral symptoms after mild traumatic brain injury (TBI) compared with orthopedic injury. Design: Cohort study with matched controls; participants and significant others completed questionnaires of preinjury personality and early postinjury symptoms. Setting: Inpatient hospital and outpatient follow-up. Participants: 87 people hospitalized with mild TBI and 82 with orthopedic injury. Interventions: Not applicable. Main Outcome Measures: The NEO Personality Inventory-Revised (NEOPI-R) and Neurobehavioral Functioning Inventory (NFI). Results: For both mild TBI and orthopedic injury groups, personality traits were generally normal; concordance rates between self and significant others personality and symptoms scales were moderate. Self and significant others’ ratings of conscientiousness on the NEOPI-R were higher in the orthopedic injury than in the mild TBI group. Significant others’ reports of NFI somatic symptoms were higher in the orthopedic injury than in the mild TBI group. Stepwise linear regression showed that, in both groups, NEOPI-R agreeableness and neuroticism accounted for a significant proportion of aggressive symptoms (orthopedic injury group, R²=.56; mild TBI group, R²=.46; P<.01). In the mild TBI group, NEOPI-R neuroticism accounted for a significant proportion of depressive symptoms (R²=.44, P<.01). Unfortunately, significant findings appear to be the result of substantial content overlap between the NEOPI-R and the NFI, rather than reflecting predictive relationships. Conclusion: Objectively measured personality holds limited predictive value in understanding early neurobehavioral symptoms after mild TBI.     

Poster 30: Relationship between p300 and neuropsychologic test performance during early recovery from traumatic brain injury

Objective: To examine the relationship between P300 latency and amplitude and neuropsychologic test scores in persons with traumatic brain injury (TBI) during acute inpatient rehabilitation. Design: Prospective, consecutive sample. Setting: Traumatic Brain Injury Model Systems (TBIMS) acute inpatient rehabilitation center. Participants: 52 patients with TBI were evaluated using a standardized neuropsychologic battery and P300 classical oddball paradigm after obtaining 2 consecutive Galveston Orientation and Amnesia Test scores within the target range. Data of 27 patients (age, 33±15y; admission Glasgow Coma Scale score, 8.0±3.9) satisfied electrophysiologic criteria (at least 20 correct rare responses) and were further analyzed. Interventions: Not applicable. Main Outcome Measures: Standard TBIMS neuropsychologic test battery and P300 latencies and amplitudes. Results: As expected, analyses revealed significant associations between Cz and Pz amplitudes and neuropsychologic measures of sustained attention (Symbol Digit Modalities Test, Trail-Making Test). However, relationships between Cz and Pz latencies and neuropsychologic test performance were less evident. Conclusions: During early recovery from TBI, P300 amplitudes more than latencies reflect cognitive processes associated with speed of information processing and attention.     

Poster 23: Loss to follow-up and potential systematic bias in survivors of traumatic brain injury 60 months postinjury

Objectives: To predict which survivors of traumatic brain injury (TBI) are more likely to be lost to follow-up and to examine the possibility of systematic bias in TBI research. Design: Characteristics of participants found for a 12-month follow-up were compared with those lost and found for a 60-month follow-up using logistic regression. Setting: A prospective, longitudinal database. Participants: 588 adult survivors of TBI. Interventions: Not applicable. Main Outcome Measures: Lost participants were those with no 60-month follow-up interview, excluding the deceased. Results: Follow-up was attempted with 558 survivors of TBI identified 12 months postinjury. At the 60-month follow-up, 34 had died (6.1%), 220 were lost to follow-up (39.4%), and 304 (54.5%) were located and interviewed. Deceased study participants were excluded from analysis. Inpatient rehabilitation was associated with a slight increase in probability of being available for follow-up at 60 months (OR=.56; 95% CI, .32-.98). Injury severity was also associated with loss to follow-up, with the critically injured more likely to be found at 60 months than the moderately injured (OR=4.1; 95% CI,1.13-19.6). Age was a significant predictor of follow-up status (OR=.86; 95% CI, .77−.95). Conclusions: The likelihood of being located at follow-up increased by 92% for every 5-year increase in age.     

Poster 6: Traumatic brain injury as a relevant cause of growth hormone deficiency in adults. Experience from KIMS (pharmacia international metabolic database)

Objective: To compare baseline clinical characteristics and 1-year growth hormone (GH) replacement results in patients with adult onset growth hormone deficiency (GHD) caused by traumatic brain injury (TBI) versus nonfunctioning pituitary adenoma (NFPA). Design: Pharmacoepidemiologic survey of hypopituitary adults with GHD. Setting: Records were selected from the KIMS database, which contains information on >8500 patients with GHD, for 168 of whom TBI was identified as a cause. Participants: Both groups (NFPA group, n=207; TBI group, n=29) were age- (at pituitary disorder onset and entry into the KIMS database) and sex-matched (60% men, 40% women), previously not irradiated, and had not received GH. Interventions: Not applicable. Main Outcome Measures: Values given as mean ± SE. Results: The age at GHD diagnosis was 38.8±2.0 years for the TBI group and 41.5±0.5 years for the NFPA group. In both groups, the most frequent additional hypopituitary deficiency was luteinzing hormone/follicle-stimulating hormone, followed by adrenocorticotropic hormone and thyroid-stimulating hormone. The mean GH peak at diagnosis was 1.25±0.42ng/mL in the TBI group, which was significantly lower than that of the NFPA group (2.38±0.7ng/mL). There were no significant statistical differences in medical history, glucose level, lipids, waist circumference, or body composition measurements. Interestingly, patients with TBI were significantly shorter (168.2±1.5cm) than the NFPA patients (172.5±0.6cm). After 1 year of GH treatment, differences were shown in waist, lean mass, heart rate, glucose levels, quality of life as measured by the Quality of Life Assessment in Growth Hormone Deficient Adults and insulin-like growth factor I. Conclusions: Although hypopituitarism secondary to TBI was described more than 50 years ago, it is only now evident that a considerable number of patients experience severe GHD after TBI. It is suspected that a large number of patients after TBI have undiagnosed GHD. The present results confirm that clinical characteristics and GH treatment effects in GHD caused by TBI are indistinguishable from those in GHD caused by NFPA.     

Article 5: A multidimensional measure of acute confusion after traumatic brain injury: the confusion assessment protocol

Objective: To provide preliminary validation of a new measure of posttraumatic confusional state (PTCS), the Confusion Assessment Protocol (CAP), that assesses 7 key symptoms of PTCS (disorientation, cognitive impairment, restlessness, fluctuation in presentation, nighttime sleep disturbance, decreased daytime arousal, psychotic-type symptoms). Design: Criterion standard investigation. Setting: Inpatient traumatic brain injury (TBI) rehabilitation program. Participants: 62 consecutive patients with moderate or severe TBI admitted for inpatient rehabilitation. Interventions: Not applicable. Main Outcome Measure: Clinical diagnosis of delirium based on Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) criteria. Results: 38 of 44 (86%) patients who were classified as confused on the CAP met the DSM-IV criteria for delirium while only 2 of 18 (11%) patients classified as nonconfused on the CAP met the DSM-IV criteria. Overall agreement between the CAP and the DSM-IV classification was 87%. Conclusions: The CAP is a brief, structured, repeatable measure of multiple aspects of PTCS. Classification of patients as confused versus nonconfused with the CAP showed excellent agreement with clinical diagnosis of delirium using DSM-IV criteria.     

Abstract 24: Occupational categories and return to work after traumatic brain injury.

Objectives: To further evaluate determinants of return to work (RTW) after traumatic brain injury (TBI) by examining the relationship between preinjury occupational category and RTW outcome. Design: Prospective collaborative cohort study. Setting: To be eligible for this study, patients had to have been hospitalized with a diagnosis of TBI and have received both acute neurotrauma services and inpatient rehabilitation services at any of the 17 Traumatic Brain Injury Model Systems centers. Participants: Consecutive sample of 1173 patients with TBI who consented to participate, were employed prior to injury, and had completed 1-year follow-up assessment. Intervention: Included an inpatient interdisciplinary brain injury rehabilitation program. Main Outcome Measure: Competitive employment 1 year after rehabilitation. Results: Participants (N=1173) were categorized into 1 of 3 groups, depending on preinjury occupational title: high decision making (n=170), service related (n=622), or manual labor (n=381). Chi-square analysis showed an association between these categories and RTW at 1 year (P<.005). The chance of successful RTW was greatest for high decision making (58.8%), less for service related (42.8%), and lowest for manual labor (32.5%). Of those with successful RTW, the majority did so within the same occupational category grouping. Preinjury manual labor jobs were most likely to shift to a different category postinjury (39.1%), whereas service-related jobs were least likely to shift (25.5%). Conclusions: Prior research has shown that preinjury employment status (employed vs unemployed) greatly influences the odds of successful RTW after TBI. The current study convincingly demonstrates that the type of occupation also influences RTW outcome, with the best prospect for RTW being among persons with high decision-making jobs. Occupational category should be examined in the future development of predictive models for RTW after TBI.     

Poster 15: Factors associated with unemployment 1 year after traumatic brain injury: a follow-up study in South Carolina, 1999-2002

Objective: To describe the effects of potential confounders associated with unemployment 1 year after traumatic brain injury (TBI) of participants in the South Carolina Traumatic Brain Injury Follow-Up Registry (SCTBIFR). Design: Statewide, population-based, pragmatic study that utilizes a longitudinal survey of people hospitalized with TBI in South Carolina. Setting: South Carolina. Participants: 938 persons, aged 19 to 70 years, of the first 1800 SCTBIFR participants. Interventions: Not applicable. Main Outcome Measure: Unemployment (working <20h/wk in a job for which one was paid). Results: Approximately 44% of persons selected for the study were unemployed 1 year after injury. Preliminary adjusted findings revealed that unemployment after TBI is associated with fair or poor general health (OR=5.58; 95% CI, 4.26-7.84), age between 45 and 70 years (OR=2.14; 95% CI, 1.48-3.09), severe injury (OR=1.93; 95% CI, 1.34-2.76), less than a high school-level education (OR=1.62; 95% CI, 1.07-2.46), being nonwhite (OR=1.60; 95% CI, 1.15-2.21), female gender (OR=1.51; 95% CI, 1.11-2.06), and being unmarried (OR=1.47; 95% CI, 1.06-2.04). Conclusion: General health, age, injury severity, level of education, race, gender, and marital status are significant predictors of unemployment in persons with TBI.     

Article 10: Effect of cognitive retraining on reorganization of attention network in patients with traumatic brain injury

Objective: To identify the areas of reorganization in the neural network for attention after cognitive retraining in the patients with traumatic brain injury (TBI) using functional MRI (fMRI). Design: Controlled clinical trials. Setting: Department of physical medicine and rehabilitation in general hospital. Participants: Ten TBI patients and 12 healthy volunteers participated. Interventions: Patients had received computer-assisted cognitive retraining to practice diverse attention skills for 3 to 4 weeks. Main Outcome Measures: fMRI was performed on a 3T ISOL Forte scanner with single-shot echoplanar imaging sequences before and after the intervention using a modified Posner paradigm as an activation task. Data were analyzed using Statistical Parametric Mapping 99. Results: The accuracy was lower and reaction time was slower in the patient group (P<.001). fMRI results showed more activation in the right prefrontal and less activation in the cingulate, occipitotemporal, and cerebellar regions in the patients than in the normal group. After cognitive retraining, reaction time improved significantly (P<.05). fMRI analysis showed a decrease in activation of the bilateral frontal and an increase in activation of the medial frontal areas after cognitive retraining. Conclusion: The effect of cognitive retraining was accompanied by reorganization of attention networks in frontal areas.     

Abstract 22: Acute neuroradiograhic predictors of rehabiliation costs.

Objective: To examine acute neuroradiographic and injury predictors of rehabilitation costs in a large traumatic brain injury (TBI) population. Design: Multiple regression-based within-group design. Setting: Urban university-based neurotrauma center and rehabilitation hospital. Participants: 293 persons presenting to a level 1 trauma center with a primary diagnosis of moderate or severe TBI who required inpatient rehabilitation. Interventions: Not applicable. Main Outcome Measures: Multiple variables derived from demographics, injury characteristics, ratings of various computed tomography (CT) scan indicators of neuropathology, and charges for rehabilitation services. Results: Several key variables that are predictive of rehabilitation outcome (including age, cause of injury, admission Glasgow Coma Scale [GCS] score) and several CT-derived neuroradiographic variables were entered into a multiple regression model to predict the total dollar charges for all rehabilitation services. The variables that emerged as statistically significant were (in order of amount of variance accounted for in the predictive model: presence of subarachnoid hemorrhage (SAH) (F change=17.89; P<.0001), admission GCS total score (F change=13.59; P<.0001), presence of frontal lobe contusion (F change=8.26; P<.004), presence of left parietal contusion (F change=7.15; P<.008), presence of right epidural hemorrhage (F change=4.51; P=.035), and presence of a punctate hemorrhage (F change=4.89; P=.028). Conclusions: Charges for TBI rehabilitation are an important consideration. The ability to predict the relative cost of rehabilitation can facilitate planning and may be helpful in more accurately determining the allocation of resources. The presence of SAH was an important predictor of charges and may reflect a risk factor for secondary brain injury not captured by other measures.     

Poster 7: Impact of balance deficits after traumatic brain injury on inpatient rehabilitation functional outcome: a multicenter study

Objective: To evaluate how sitting and standing balance ratings of patients with traumatic brain injury (TBI) on admission to rehabilitation impacts functional outcome, rehabilitation charges, and rehabilitation length of stay (LOS). Design: Multicenter analysis of consecutive admissions to designated Traumatic Brain Injury Model Systems (TBIMS). Setting: TBIMS centers. Participants: 908 adults with TBI were included in the study. Interventions: Not applicable. Main Outcome Measure: FIM™ instrument admission and discharge scores, FIM change and efficiency; rehabilitation LOS and charges; and discharge disposition. Results: Persons with grossly impaired sitting balance on admission to rehabilitation had lower discharge FIM scores and greater rehabilitation charges than those with normal or mildly impaired sitting balance. Persons with mildly impaired sitting balance had lower discharge FIM scores and greater rehabilitation charges than those with normal sitting balance. Subjects with grossly impaired standing balance compared with normal or mildly impaired standing balance had greater rehabilitation charges and LOS and lower discharge FIM scores and FIM efficiency. Subjects with grossly impaired balance on admission to rehabilitation had significantly lower motor FIM items compared with those who had normal or mildly impaired balance. Subjects with grossly impaired sitting balance were discharged from rehabilitation to someplace other than their home more than twice as often as those with normal sitting balance (11.7% vs 23.3%). Subjects with grossly impaired standing balance were discharged from rehabilitation to someplace other than their home more than 3 times as often as subjects with normal standing balance (6.7% vs 21.5%). Conclusion: In patients with TBI, sitting and standing balance ratings on admission to rehabilitation are useful indicators of functional impairment and not returning home at discharge from rehabilitation.     

Poster 14: Cost savings of pulmonary vest therapy

Objectives: To determine cost savings from using pulmonary vest therapy to prevent hospitalizations related to pneumonia. Design: Quantitative research using before-during treatment data. Setting: 2 pediatric skilled nursing facilities for children with severe cerebral palsy (CP). Interventions: Not applicable. Main Outcome Measures: The average hospitalization for pneumonia was 5 days, 4 of which were in the intensive care unit (ICU). The daily hospitalization rates were obtained from the business offices of 3 regional hospitals with pediatric ICU units that serve our patients. Results: During our 1-year study, the number of hospitalizations was decreased by 10. The average hospitalization charge was $8225. Thus, the cost savings for 10 hospitalizations was $82,550. The cost of a vest therapy unit was $15,000 and a total of 3 units were used. The vest therapy units continue to be in use for the third year, without mechanical failure or maintenance requirements. Conclusion: In the first year alone, there was a very significant cost savings. It should be noted that these sums do not include hospital laboratory testing, pharmacy costs, or physician costs, and so significantly underestimate the actual savings.     

Poster 20: Community integration after traumatic brain injury

Objective: To determine the level of community integration after traumatic brain injury (TBI) and its association with clinical inpatient outcome measures. Design: The Community Integration Questionnaire (CIQ) was used to conduct telephone interviews with either the patient or a proxy who lived with the patient. Setting: Community. Participants: 77 patients with TBI who were 5 to 34 months postdischarge from an acute rehabilitation hospital. Interventions: Not applicable. Main Outcome Measure: The CIQ. Results: The CIQ total score had an inverse correlation with age (r=−.300, P<.008) and length of stay (r=−.290, P<.011). There were low but statistically significant correlations with discharge cognitive skills on the FIM™ instrument (r=.451, P<.0001), discharge FIM total score (r=.366, P<.001), and FIM efficiency (r=.367, P<.001). There were significant differences between subjects who were not currently working or attending school versus those who were. The latter group scored higher on home integration (5.6±3.0 vs 3.6±2.9), social integration (8.9±2.1 vs 6.6±2.8), productivity (5.4±0.8 vs 1.5±1.1), and CIQ total scores (20.0±3.7 vs 11.8±5.8). Conclusion: Consistent with previous findings, community integration correlated with functional outcome at discharge from rehabilitation. Patients who return to work or school exhibit better social and home integration.     

Poster 28: Comparison of hemi-field fresnel prisms versus patching for visual neglect after stroke

Objective: To assess the effect of bilateral hemi-field prisms versus patches for treatment of visual neglect after stroke. Design: Randomized, prospective, controlled trial. Setting: Inpatient stroke unit. Participants: 38 patients with stroke, visual neglect defined by the Behavioral Inattention Test (BIT), and visual acuity ≥20/200. Interventions: The prism group received 15 diopter Fresnel prisms over the affected hemi-field bilaterally; patch group received black tape of the unaffected visual hemi-field bilaterally; controls received no visual intervention. The interventions lasted 14 days. Main Outcome Measures: 3 subtests of BIT were recorded at baseline and on days 2, 7, 14, and 16. Results: Admissions demographic were similar for all 3 groups. There were no significant differences in BIT scores while wearing the optical devices. The prism group showed a trend for improvement in all 3 BIT scores 2 days after removal of the optical device. This reached statistical significance for the line bisection scores (prism group ± SEM, 1.5±0.3; patch group, .12±.12; controls, .33±.33; P=.004). Dropouts were highest for prism group (10/16) versus patch group (4/12) and controls (1/10) (P=0.2). Conclusion: Hemi-field Fresnel prisms showed a beneficial effect on visual neglect 2 days after their removal.