Biomedical subjects
Publications and source records attributed to S R Millis.
The Recognition Memory Test is frequently used to assess memory; however, one of the commonly cited limitations is a lack of data on reliability. The current study was undertaken to estimate the internal consistency reliability of the test with a sample of 72 persons with traumatic brain injury. Acceptable estimates of internal consistency for both subtests were obtained.
OBJECTIVE: To investigate the relation between duration of posttraumatic amnesia (PTA) and functional outcome in a traumatically brain injured population. PATIENTS: Two hundred seventy-six patients with traumatic brain injury (TBI) who were admitted to a Level I university trauma center and required inpatient rehabilitation. MEASURES: Duration of PTA was assessed by serial administrations of the Galveston Orientation Amnesia Test (GOAT). Functional Independence Measure (FIM) total scores, FIM cognitive and motor subscores, and Disability Rating Scale (DRS) scores were obtained at admission and discharge from inpatient rehabilitation. RESULTS: Duration of PTA was a significant predictor of all admission and discharge DRS and FIM scores. Duration of PTA and age at the time of injury, in combination, contributed significantly to the prediction of the DRS score and FIM total, cognitive, and motor scores at discharge. CONCLUSION: Duration of PTA appears to be a useful variable in predicting specific functional outcome in the TBI population receiving inpatient rehabilitation services. The use of age as a factor in addition to duration of PTA enhances the prediction of functional outcome.
OBJECTIVE: Acute inpatient traumatic brain injury (TBI) rehabilitation has seen a jump in complexity of medical patient care over the past several years, often necessitating transfer back to an acute care facility. The purpose of this study was to determine the association between selected clinical variables and transfer from inpatient rehabilitation to an acute care facility. DESIGN: A retrospective review of cases from 1992 to 1994. SETTING: A TBI unit in a freestanding rehabilitation hospital. PATIENTS: Twenty-two patients were identified as having received acute care transfer. This group was compared with 78 patients, admitted in the same interval, who did not require acute care transfer. The variables evaluated included recent surgery, pneumonia, fracture, intracranial blood, tracheostomy use, percutaneous feeding tube use, deep venous thrombosis, focal neurological examination, following simple commands, serum sodium level of < 135 mmol/L, serum white blood cell count of > 11,000 cells/microL, and serum hemoglobin level of < 10.0 g/dl. ANALYSIS: Chi-square analysis was performed on the association between acute care transfer and the noted variables. RESULTS: History of pneumonia (p < .03) and history of recent surgery (p < .02) were both associated with acute care transfer, and serum hemoglobin of < 10.0 g/dL had a trend towards association (p < .10). CONCLUSION: Physiatrists caring for the TBI patient may warrant more acute observation of individuals with these parameters to prevent the problems necessitating acute care transfer.
Violent injuries have become an increasingly prevalent cause of traumatic brain injury (TBI). These injuries can be classified as either penetrating or non-penetrating in nature. While much of the research on violence has been within a military population, there exists a marked difference between military and civilian injuries. Prior work has reported relatively poor outcomes for those individuals who have suffered penetrating TBIs, but little has been done to assess specific functional outcome parameters in survivors. We examined 25 subjects that had sustained blunt injuries and 25 cases with penetrating injuries as a result of a violent act. Cases were matched by initial Glasgow Coma Scale (GCS), age and educational level. Mean GCS for this study sample was 8.8. The following outcome variables were assessed at rehabilitation admission and discharge and at 1 year post injury: Disability Rating Scale (DRS), Rancho Los Amigos Scale (LCFS), Functional Independence Measure (FIM) (ambulation, expression items), length of stay, and cost of care. Student's t-tests were performed to assess for differences between the two groups. No significant differences were noted between the groups for any of the outcome variables. Although penetrating injuries may have a higher initial mortality, those who survive to come to rehabilitation appear to have similar outcomes to those patients with non-penetrating violence related injuries.
The ability of neuropsychological tests to predict later psychosocial outcome on the Community Integration Questionnaire (CIQ; WiIler, Rosenthal, Kreutzer, Gordon, & Rempel, 1993) was examined. A previous investigation by Millis, Rosenthal, and Lourie (1994) found that the Trail Making Test (TMT; Lezak, 1995) and the Rey Auditory Verbal Learning Test (RAVLT; Lezak, 1995) were related to performance on the CIQ at 1 year postinjury. The RAVLT and the TMT Parts A and B (TMT-A, TMT-B) in conjunction with age at time of injury significantly predicted psychosocial outcome as measured by individual self-report. When self-reported functioning was examined by domain, these variables significantly predicted home and social adaptation. These variables were applied to significant others' reports of community integration, where age and the TMT remained significant predictors of psychosocial outcome. These results underscore the utility of neuropsychological measures in the prediction of later community integration after traumatic brain injury.
Traumatic brain injury (TBI) has been associated with memory impairments, but the severity and qualitative aspects of such impairment do not appear homogeneous across patients. This study sought to replicate an earlier investigation that found distinct verbal learning subtypes in TBI using the California Verbal Learning Test (CVLT). CVLT data from 88 acute rehabilitation inpatients with mild, moderate, and severe traumatic brain injuries were analyzed with multiple cluster analytic techniques. Cluster analyses yielded five learning subtypes, three of which appeared similar to the subtypes previously identified as Active, Disorganized, and Passive subtypes, and two that appeared similar to the Deficient subgroup. Traumatic brain injury appears to be characterized by heterogeneous, but fairly reliable, verbal learning subtypes that can be detected early postinjury.
The present study investigated visuospatial functioning in 35 right-hand dominant patients with striatal, frontal white matter, or posterior thalamic infarction, in the right cerebral hemisphere. On tasks of visual detection and discrimination, the groups performed within normal limits and comparable to one another. On tasks of visual synthesis and spatial analysis, each group was significantly impaired, with greatest impairment following infarction of deep grey structures. Correlational analyses indicated that in the striatal and frontal white matter groups, visual synthesis and spatial analysis task performance were strongly correlated with performance on neuropsychological tests of executive control. This relationship was not observed in the posterior thalamic group. The results support the hypothesis of differential patterns of visuospatial performance depending on the anatomical locus of subcortical lesions. The findings are explained in terms of differing subcortical circuitry and suggest that multiple substrates mediate the sequelae of subcortical lesions.
The relationship between depression and cognition in older adults has become a topic of extensive research and clinical interest. Results of previous research are limited, however, by sampling and methodological problems. Small sample sizes and the treatment of depression as a dichotomous variable may be major factors that limit data interpretation in previous studies. The present sample of 220 geriatric medical patients was split into a validation and a cross-validation sample specifically to test the hypothesis that depression is an independent predictor of cognition. Multiple regression analyses, controlling for the direct effects of demographic factors, found that Geriatric Depression Scale scores were consistently a significant predictor of Dementia Rating Scale and Logical Memory scores. Overall, GDS scores accounted for approximately 8% of unique variance for both measures of cognition. Demographic factors, particularly age and race, were also significant predictors of cognitive test scores. These data were interpreted as providing support for the hypothesis that level of depression predicts level of cognitive functioning.
A factor analysis with principal components extraction and varimax rotation was performed on 19 California Verbal Learning Test age-uncorrected scores from a sample of 75 moderate and severe closed-head-injured patients. A 6-factor solution was obtained that accounted for 80% of the total variance and was similar to those reported previously. The factors were labeled General Verbal Learning, Response Discrimination, Learning Strategy, Proactive Effect, Self-monitoring, and Serial Position. Factor scores were derived and correlated with other neuropsychological measures. The General Verbal Learning factor scores were significantly correlated with those on measures of memory, complex attention, and strategy induction.
Verbal memory and learning patterns, as measured by the California Verbal Learning Test (CVLT), following penetrating head injury (PHI) from gunshot wound were studied in 10 acutely injured patients (mean age 25.3 years) at a mean of 2.1 months post-injury. Primary impairment was found on measures of free recall of new verbal information which appeared to be related to deficits in organisational and retrieval functions: (1) the group's learning characteristics were marked by disorganization and an underutilization of active learning strategies; (2) rate of acquisition also appeared to be mildly decreased; (3) nevertheless, the PHI group did not show severe disruption in all aspects of learning and memory. In fact, the group showed a relatively intact capacity to store new information in memory.
Warrington's Recognition Memory Test (RMT) was used in the assessment of memory and motivation in a study of 63 adult subjects. Subjects reporting mild head trauma who claimed disability and were seeking financial compensation (MT-Comp) obtained significantly lower scores on both subtests of the RMT than subjects with moderate and severe brain trauma (ST) and subjects with mild head injuries who had returned to work (MT-Work). Of the MT-Comp, 47% and 29% scored below chance on the Faces and Words subtests, respectively, compared to 10% and 6% in ST group and 0% and 0% in the MT-Work group. A discriminant function that consisted of both RMT subtests discriminated between the MT-Work and MT-Comp groups with an overall accuracy of 93%.
Previous studies that have examined performances on the California Verbal Learning Test (CVLT) among individuals with traumatic brain injury (TBI) have found differing levels of performance. Differential patterns of performance, however, have only been inferred. The present investigation sought to determine empirically if differential patterns of performance could be discerned in a TBI sample of 65 subjects with CVLT variables. The CVLT variables were selected based on the instrument's factor structure. Cluster analysis yielded four distinct subtypes of brain-injured individuals. The Active subtype demonstrated impaired unassisted retrieval, but used active encoding strategies and showed relatively intact ability to store novel information. The Disorganized subtype demonstrated an inconsistent, haphazard learning style along with deficits in encoding. The Passive subtype was marked by an overreliance on a serial clustering strategy as well as impaired encoding and/or consolidation. The Deficient subtype was the most impaired of all groups, exhibiting a slowed rate of acquisition, passive learning style, and significant impairment in encoding. Implications for rehabilitation are discussed.
The current investigation replicated the findings from a 1992 study by Millis. 20 subjects with claimed mild head injury who were seeking financial compensation obtained significantly lower scores on both subtests of the Recognition Memory Test than did 66 rehabilitation inpatients with documented moderate and severe traumatic brain injuries. The direct discriminant function derived by Millis was cross-validated on the current sample and yielded an improved over-all correct classification rate of 83%, with 85% and 82% of the mild and severe traumatic brain-injured subjects classified correctly, respectively. Over-all correct classification rate in the original study was 76%.
A neuropsychological test battery was administered to 23 subjects with traumatic brain injury during initial inpatient rehabilitation who were participating in the national Traumatic Brain Injury Model Systems study. Trails B and the Rey Auditory Verbal Learning Test total scores obtained during inpatient rehabilitation were significantly correlated with the Community Integration Questionnaire total score that was administered at one year postinjury. These findings suggest that performances on tests of cognitive speed and flexibility, complex attention, and memory during the acute phase of recovery may be useful for the later prediction of "real world" behavior and psychosocial outcome after one year postinjury.
Incontinence in an elderly person is frequently an important factor in deciding whether to institutionalize the patient. This study examined the association between cognition and continence in geriatric patients admitted to an inpatient rehabilitation unit. It was hypothesized that admission cognitive status would help predict discharge bowel and bladder continence and would account for significant unique variance beyond that explained by admission continence status. The Dementia Rating Scale (DRS) was administered to 103 consecutive subjects admitted to a geriatric rehabilitation unit. Admission and discharge bowel and bladder status ratings were made using the Functional Independence Measure. Multiple stepwise regression analysis found admission incontinence status coupled with the DRS Initiation/Preservation (IP) subtest significantly predicted bowel and bladder management status at discharge. The IP subtest accounted for a significant but modest amount of unique variance beyond the admission continence status.
The Glasgow Coma Scale (GCS) is routinely used in the acute care setting after traumatic brain injury (TBI) to guide decisions in triage, based on its ability to predict morbidity and mortality. Although the GCS has been previously demonstrated to predict mortality, efficacy in prediction of functional outcome has not been established. The purpose of this study was to assess the value of the acute GCS in predicting functional outcome in survivors of TBI. This study used the Multicenter National Institute on Disability and Rehabilitation Research TBI Model Systems database of 501 patients who had received acute medical care and inpatient rehabilitation within a coordinated neurotrauma program for treatment of TBI. Initial and lowest 24 hr GCS scores were correlated with the following outcome measures: the Disability Rating Scale (DRS), Rancho Los Amigos Levels of Cognitive Functioning Scale (LCFS), and cognitive and motor components of the Functional Independence Measure (FIM(SM)-COG and FIM(SM)-M). Outcome data were collected at admission to and discharge from the inpatient TBI rehabilitation unit. Correlation analysis revealed only modest, but statistically significant, relationships between initial and lowest GCS scores and outcome variables. Initial and lowest GCS score comparison with outcome demonstrated the following correlation coefficients: admission DRS, -0.25 and -0.28; discharge DRS, -0.24 and -0.24; admission LCFS, 0.31 and 0.33; discharge LCFS, 0.27 and 0.25; admission FIM-COG, 0.36 and 0.37; discharge FIM-COG, 0.23 and 0.23; admission FIM-M, 0.31 and 0.31; discharge FIM-M, 0.25 and 0.21. The GCS as a single variable may have limited value as a predictor of functional outcome.
Recent studies have clarified the role of hypoxic-ischemic damage as a secondary factor in traumatic brain injury (TBI). Many trauma centers are now consistently using the Revised Trauma Score (Glasgow Coma Scale, systolic blood pressure, and respiratory rate) to assist with triage of multitrauma patients. This study investigated the predictive power of the Revised Trauma Score (RTS) instead of the Glasgow Coma Scale (GCS) in determination of disability as measured by the Disability Rating Scale (DRS). Data were obtained as part of the National Institute for Disability and Rehabilitation Research TBI Model Systems database on 501 patients receiving acute medical care and inpatient rehabilitation within a coordinated neurotrauma program for treatment of TBI. Initial RTS and GCS were obtained on admission to the emergency department, along with the lowest GCS measured in the first 24 h. Analysis of initial RTS and GCS demonstrated modest, but statistically significant Pearson's correlations with DRS at rehabilitation admission (-0.18 and -0.25, respectively) and discharge (-0.22 and -0.24, respectively). Lowest GCS within the first 24 h postinjury also failed to show a strong relationship with DRS at rehabilitation admission (-0.28) and discharge (-0.24). Multiple regression analysis performed on RTS subsets for systolic blood pressure and respiratory rate did not reveal an added predictive value. Although RTS may be important in emergency triage for its ability to predict mortality, this study indicates its limited usefulness in prediction of disability.
The suitability of Warrington's Recognition Memory Test (RMT) for detecting memory impairment after traumatic brain injury was investigated in a pilot study of 36 patients with moderate and severe brain injuries. The patients showed significant memory deficits on both RMT subtests, Words and Faces, with greater impairment on the Faces subtest. No significant gender differences were found. The left hemisphere lesion group obtained the lowest mean score on the Words subtest, and the right hemisphere lesion group had the lowest mean score on the Faces subtest, although these differences did not reach statistical significance. At the 90-94% specificity level the sensitivity of the Faces and Words subtests for the detection of memory impairment after traumatic brain injury was 81% and 63%, respectively. The RMT appears to be a promising instrument for measuring and describing patterns of memory impairment after brain injury.