Comparison of logistic regression and neural networks to predict rehospitalization in patients with stroke.

CONTEXT: Rehospitalization following inpatient medical rehabilitation has important health and economic implications for patients who have experienced a stroke. OBJECTIVE: Compare logistic regression and neural networks in predicting rehospitalization at 3-6-month follow-up for patients with stroke discharged from medical rehabilitation. DESIGN: The study was retrospective using information from a national database representative of medical rehabilitation patients across the US. SETTING: Information submitted to the Uniform Data System for Medical Rehabilitation from 1997 and 1998 by 167 hospital and rehabilitation facilities from 40 states was examined. PARTICIPANTS: 9584 patient records were included in the sample. The mean age was 70.74 years (SD = 12.87). The sample included 51.6% females and was 77.6% non-Hispanic White with an average length of stay of 21.47 days (SD = 15.47). MAIN OUTCOME MEASURES: Hospital readmission from 80 to 180 days following discharge. RESULTS: Statistically significant variables (P <.05) in the logistic model included sphincter control, self-care ability, age, marital status, ethnicity and length of stay. Area under the ROC curves were 0.68 and 0.74 for logistic regression and neural network analysis, respectively. The Hosmer-Lemeshow goodness-of-fit chi-square was 11.32 (df = 8, P = 0.22) for neural network analysis and 16.33 (df = 8, P = 0.11) for logistic regression. Calibration curves indicated a slightly better fit for the neural network model. CONCLUSION: There was no statistically significant or practical advantage in predicting hospital readmission using neural network analysis in comparison to logistic regression for persons who experienced a stroke and received medical rehabilitation during the period of the study.

Applications of functional assessment to musculoskeletal disability evaluation.

A review of the history of the disability determination for LBP suggests that the early focus upon disability as a direct "effect" of impairment has given way to an updated approach acknowledging the multitude of factors, many of which are subjective, that contribute to an individual's future return to work and disability status. In addition to the evaluation of the underlying impairment and pathology, physicians performing disability evaluations for LBP will be best served when they can assess the individual's functional status, and incorporate that information into their decision regarding the individual's current limitations and prognosis. As discussed above, many of the factors contributing to the evaluation of functional status, including pain, physical functioning, and affective status, are necessarily subjective. Nonetheless, the rehabilitation literature suggests that inclusion of this information into the disability determination using a patient-centered approach may provide the factors that are most potent with respect to patient prognosis and eventual return to work.

Biologic patterns of disability.

We describe the use of a mathematical/statistical method (i.e., Rasch analysis) to elucidate biological patterns of disability present in the functional ability of persons undergoing medical rehabilitation. Two measures chosen for illustration are the FIM Instrument for inpatients and the Body Movement and Control (BMC) measure for outpatients. In order to meet the assumptions necessary for application of linear statistics to clinical measurement studies, Rasch analysis was used to transform ordinal scales into linear measures. Another unique feature of Rasch analysis is that it allows evaluation of the difficulty of items and the abilities of persons being tested, separately, on the same metric. Also, the difficulty represented by each item may be arranged along a hierarchy from easy to hard. The hierarchies of functional ability items are dependent upon the specific patterns of disability related to underlying pathophysiology. For inpatients, initial analyses of the 18 items of the FIM Instrument demonstrated separate hierarchies for the 13 motor items and for the 5 cognition items. Subsequent analyses demonstrated five distinct patterns for the 13 motor items of: brain dysfunction, orthopedic conditions, pain conditions, ambulatory spinal cord dysfunction, and wheelchair users with spinal cord dysfunction. Two patterns were identified for cognition: stroke with right body hemiparesis and all others. For outpatients, the BMC measure of physical functioning is used to demonstrate that pathophysiologic conditions are expected to affect the hierarchial pattern of items differently. This was noted to be the case for persons with lower body dysfunction, low back pain, and neck pain/upper limb dysfunction. Based upon the item responses, sitting, reaching and standing appear to represent items most useful for discriminating between the three conditions in terms of the functional consequences. Rasch analysis, among other advantages, enables investigation of the subtle relationships among items and is a useful method to evaluate underlying biological patterns of disability. A clinician, using a map that shows the expected relationships between item scores, may observe that a particular patient matches or does not match the expected pattern. Such insights may help the clinician in monitoring the responses of the patient to treatment efforts.

Status of functional outcomes for stroke survivors.

In the past, it has taken several years to accumulate a sufficient number of subjects in the community-based studies to arrive at generalized conclusions. With an ongoing database, such as resides at CFAR-UDSMR, it is possible to collect a large number of cases within a relatively short period of time. Further, it is relatively easy to perform continuous monitoring in order to determine trends that may be occurring as a result of the changing scenes in health care delivery. It has been possible to uncover patterns of scoring the functional status of patients that reveal a consistent picture of an underlying biology of disability and predictable characteristics in patients' courses through the rehabilitation process. Newer analytic methods have allowed one to predict expected recovery patterns. This new information will allow one to better measure and manage outcomes, to improve the quality of care, to improve cost-effectiveness, and to better manage financial risk. These are the tools necessary for clinicians to incorporate into practice, as expectations with respect to outcomes and reimbursement for health care are changing. All of the marks associated with FIM and UDSMR belong to the Uniform Data System for Medical Rehabilitation, a division of UB Foundation Activities, Inc.

Does the functional assessment measure (FAM) extend the functional independence measure (FIM) instrument? A rasch analysis of stroke inpatients.

Adding the items of the Functional Assessment Measure (FAM) to the Functional Independence Measure (FIM instrument) has been proposed as a method to extend the range of the FIM, particularly when assessing functional status in rehabilitation patients with brain injury, including stroke. It has been proposed that this approach is especially helpful in ameliorating ceiling effects when brain-injured patients have reached the end of their inpatient rehabilitation stay or are being seen in outpatient settings. In the present study, 376 consecutive stroke patients on a Canadian inpatient rehabilitation unit were concurrently administered the FIM and the FAM. Rasch analysis was used to evaluate how well the FAM items extended the difficulty range of the FIM for both the Motor and Cognitive domains. Within the Motor domain, only the FAM item assessing Community Access was found to be more difficult than extant FIM items, and this item showed some tendency to misfit with the other motor items. In the Cognitive domain, the only FAM item with a higher difficulty level than the FIM items was that assessing Employability. Notably, strict adherence to scoring guidelines for these two FAM items requires taking patients out into the community to evaluate their actual performances, a practice unlikely in the typical inpatient stroke rehabilitation unit. Results indicate that use of the entire FAM as an adjunct to the FIM reduces test efficiency while providing only minimal additional protection against ceiling effects.

Admission balance and outcomes of patients admitted for acute inpatient rehabilitation.

The objective of the study contained herein was to evaluate the clinical use of the Berg Balance Scale in a heterogeneous acute inpatient rehabilitation population and to assess the relationship between balance scores at admission and rehabilitation outcomes, including functional gain and length of stay. This was a prospective study of 45 patients with diagnoses including stroke (n = 15), traumatic brain injury (n = 19), and other impairments (n = 11) who were admitted for acute inpatient rehabilitation. Functional ability was evaluated with the Functional Independence Measure (FIM) instrument, and balance was measured using the 14-item Berg Balance Scale. These measures were assessed both at admission to and discharge from inpatient rehabilitation. Correlation and multiple regression analyses were used to determine the relationship between balance and functional ability scores at admission and rehabilitation outcomes at discharge, including length of stay, functional gain, and efficiency. Analyses of variance comparing impairment groups indicated that patients with traumatic brain injury were younger than the stroke and other groups, but there were no group differences on FIM or balance scores at admission or discharge. There were no significant group differences for any of the outcome measures. When the impairment groups were collapsed into a single heterogeneous group, multiple regression analyses demonstrated that the sitting unsupported item score at admission accounted for 27% of the variation in length of stay, once demographic influences were controlled. The FIM efficiency score was predicted best by the total Berg Balance Scale score, with 22% of the variance accounted for. In contrast, balance scores did not provide predictive information about the FIM gain score beyond that already provided by the FIM-Total score at admission, which accounted for 10% of the variance once demographics were controlled. Balance scores collected at admission to inpatient rehabilitation, in whole and in part, were shown to account for moderate amounts of variation in length of stay and the FIM efficiency score. For several of the rehabilitation outcomes, balance scores at admission accounted for more variation than scores on the FIM instrument. These findings suggest that routine assessment of balance at admission to inpatient rehabilitation may enhance the ability to predict rehabilitation outcomes beyond that provided by assessment of functional status alone.

Rasch analysis of the Functional Independence Measure (FIM) Mastery Test.

OBJECTIVE: To evaluate the range, distribution, and unidimensionality of the items included in Version Five of the Functional Independence Measure (FIM) Mastery Test, a 54-item written examination used to determine FIM rating proficiency by professionals who administer the instrument. STUDY DESIGN: Rasch analysis of the test results of 6,645 clinicians who took the FIM Mastery Test in 1996. RESULTS: Overall, clinicians performed well on the test. When the mean item difficulty was centered at 50, and the test scores were calibrated to range from 0 to 100 logits, the mean person ability level was 77 logits. Fit statistics indicated that the test was unidimensional, measuring a common theme. While the item separation index value was 17, 22 items were not efficient in determining the clinicians' ability levels. Analysis of the logit values by the correct response showed that descriptions of Total Assistance (Level 1) and Complete Independence (Level 7) were easiest for clinicians to score correctly. CONCLUSION: The examination measures the common theme of FIM knowledge in a consistent manner. Further, the results suggest that the FIM Mastery Test may be shortened, perhaps by eliminating one of the three case studies, and improved by adding more challenging questions. In addition, FIM training should focus on distinguishing among levels 2 through 6, since clinicians had more difficulty identifying these levels.

Central auditory dysfunction, cognitive dysfunction, and dementia in older people.

OBJECTIVES: To determine in older people the relation between auditory dysfunction and cognitive dysfunction, and if central auditory test abnormalities predict the onset of clinical dementia or cognitive decline. DESIGN: Prospective population-based cohort study. SETTING: Framingham Heart Study outpatient biennial examinations 18 and 21. PARTICIPANTS: Members of the Framingham Heart Study cohort with normal findings from cognitive screening tests at biennial examination 18. MEASUREMENTS: Peripheral audiometric thresholds and word recognition in quiet; Synthetic Sentence Identification with Ipsilateral Competing Message (SSI-ICM); Mini-Mental State Examination; and detailed neuropsychological testing of subjects with abnormal findings from the Mini-Mental State Examination. Relative risk of dementia was determined using age-adjusted Cox proportional hazards regression models. RESULTS: Hearing loss significantly lowered performance on the verbal parts of the Mini-Mental State Examination. The relative risk of subsequent clinical dementia or cognitive decline was 6 in subjects with very poor scores (< 50%) in one ear on the SSI-ICM (P = .02); the relative risk was 12.5 if the poor scores were present in both ears (P = .001). CONCLUSIONS: Central auditory dysfunction precedes senile dementia in a significant number of cases and may be an early marker for senile dementia. Hearing tests should be included in the evaluation of persons older than 60 years and in those suspected of having cognitive dysfunction.

The 'preclinical phase' of probable Alzheimer's disease. A 13-year prospective study of the Framingham cohort.

OBJECTIVE: To evaluate the interval between the onset of detectable cognitive impairment and clinical diagnosis in individuals with probable Alzheimer's disease (AD), and to identify the pattern of the earliest changes in cognition in probable AD. DESIGN: Longitudinal follow-up of a community-based cohort sample. In 1976 through 1978, a screening neuropsychological examination was administered to Framingham Study participants. These subjects were then followed up prospectively for development of probable AD for up to 13 years. SETTING: This study was conducted at a community-based center for epidemiologic research. PARTICIPANTS: The surveillance sample consisted of 1045 participants in the Framingham Study aged 65 to 88 years who were free of dementia at the time of the neuropsychological screening examination. MAIN OUTCOME MEASURES: Scores on a group of neuropsychological tests were entered into a series of age- and education-adjusted multiple regression procedures, with the presence or absence of probable AD as the outcome variable. RESULTS: Considered individually, most of the screening neuropsychological measures were significantly related to later AD diagnosis. When stepwise regression procedures were employed, only measures of verbal memory and immediate auditory attention span remained significantly related to AD diagnosis. Of note, subjects later diagnosed with probable AD performed at higher levels than normal subjects on the Digit Span test at initial screening. Regression results were essentially unchanged even when the AD sample was restricted to those individuals for whom the screening examination preceded the clinical onset of dementia by 7 years or more. CONCLUSIONS: These findings support previous contentions that a "preclinical phase" of detectable cognitive deficits can precede the clinical diagnosis of probable AD by many years, and they also support the hypothesis that problems with secondary verbal memory are among the first signs of AD.

Are multiple cerebral infarcts synergistic?

OBJECTIVE: The goal of this study was to characterize the cumulative effects of multiple strokes on cognition. DESIGN: We conducted a prospective, longitudinal case study with neuropsychological, neurological, and radiological evaluations. SETTING: Research was conducted at the Boston (Mass) Veterans Administration Medical Center, Neurology Service, on successive inpatient hospital admissions. PATIENT: We followed up a 66-year-old right-handed man with multiple subcortical lacunae during a 3.5-year period during which he suffered two additional cortical infarctions. MAIN OUTCOME MEASURES: Each evaluation included approximately 3 hours of neuropsychological testing spanning a range of cognitive domains (attention, language, memory, visuospatial functions, response inhibition, and mental flexibility), full neurological examination, and computed tomographic scan. RESULTS: The patient's stepwise cognitive decline was characterized by unexpected exacerbation of "frontal" neurobehavioral features following the occurrence of two posterior cortical lesions. At initial evaluation, the computed tomographic scan showed bilateral subcortical lacunae in basal ganglia and periventricular white matter, and symptoms included dysarthria and perseveration. The second evaluation, following a left posterior parietal lesion, revealed a range of new frontal features, including impulsivity, pull-to-stimulus, and difficulty shifting set. Following a subsequent right occipital infarct, further frontal lobe impairments emerged: forced grasp reflex and incontinence. CONCLUSIONS: We hypothesize that the cumulative effects of infarcts were synergistic. That is, the posterior cortical infarcts elicited frontal features that would not be expected from a simple sum of these lesions' effects.

Affective symptoms in the chronic stage of traumatic brain injury: a study of married couples.

A convenient sample of 60 brain-injured subjects and their spouses was evaluated cross-sectionally in the chronic stage of recovery on self-rated measures of sensory, motor, cognitive and behavioural disability and on the SCL-90-R depression and anxiety subscales. A majority of the brain-injured subjects, who were on average almost 6 years post-injury, demonstrated elevations on the affective symptom scales, with almost 70% showing depression and 50% showing anxiety. The spouses of the brain-injured individuals also demonstrated significantly elevated affective symptom scales, with 73% acknowledging symptoms of depression and 55% demonstrating symptoms of anxiety. For the individuals with brain injury, those with higher self-ratings of cognitive disability and social aggression had higher self-ratings of depression and anxiety. In contrast, spouse gender appeared to have the greatest association with the presence of elevation affective responses, with female spouses having higher levels of depression and anxiety than male spouses. These findings are discussed in terms of the reactionary nature of affective disturbances in the chronic stage of recovery.

The effects of parental traumatic brain injury on the behaviour of parents and children.

Little is known about the effects of a parent's brain injury and subsequent disabilities on the children in the family. This study examines 24 families in which one parent is brain injured. In each family the children were born before the parent's injury and still lived at home at the time of interview. Reports of the uninjured parent indicate that most of the children experienced some degree of negative behavioural change after the parent's injury. In 10 of the families, significant and problematic changes occurred. Types of problems included poor relationship with the injured parent, acting-out behaviour and emotional problems. Correlates of poor outcomes for the children were: (1) injured parent's gender, (2) compromised parenting performance of the injured parent, (3) compromised parenting performance of the uninjured parent and (4) depression in the uninjured parent. This study points to the importance of recognizing traumatic brain injury as a major family stressor.

Incidence of dementia and probable Alzheimer's disease in a general population: the Framingham Study.

OBJECTIVE: To determine the incidence of dementia and Alzheimer's disease (AD) in a general population sample. BACKGROUND: Utilizing subjects in the Framingham Study cohort determined to be free of dementia in 1976 to 1978, or on biennial examination 17 in 1982, all new cases of dementia arising in this cohort over a maximum of 10 years of follow-up were ascertained. METHODS: On biennial examination 14/15, a screening neuropsychologic examination was administered to 2,117 subjects, and cases of probable prevalent dementia were identified. Beginning on examination 17 and on all successive biennial examinations, a Mini-Mental State Examination was administered. Subjects previously free of dementia and falling below age-education levels were evaluated by a neurologist and neuropsychologist to determine if dementia was present and to ascertain the dementia type using standard criteria. RESULTS: Five-year incidence of dementia increased with age, doubling in successive 5-year age groups. Dementia incidence rose from 7.0 per 1,000 at ages 65 to 69 to 118.0 per 1,000 at ages 85 to 89 for men and women combined. Incidence of probable AD also doubled with successive quinquennia from 3.5 at ages 65 to 69 to 72.8 per 1,000 at ages 85 to 89 years. Incidence of dementia and of probable AD did not level off with age and was not different in men and women. CONCLUSIONS: In a general population sample, we determined incidence of dementia and of probable AD and will use these incident cases for study of precursors and natural history in this elderly cohort, which has been under close surveillance for over 40 years.

Practical issues in behavior management for family members of individuals with traumatic brain injury.

Changes in behavior, both cognitive and emotional, are common following traumatic brain injury (TBI). Typically, these changes have a greater impact on family members than the physical limitations that often accompany brain injury. In this article, we present an overview of the behavioral techniques that have been used by family members of individuals with brain injury to deal with these problematic behavior changes. Clinical vignettes are presented that describe the actual solutions that family members have used in the home to help deal with behavior changes. The importance of determining the source of the behavior changes is highlighted, and the need to apply behavior intervention programs with consistency is discussed. In addition, this article stresses the need for continued interaction between rehabilitation staff and family members beyond the period of inpatient rehabilitation.

Relationship of normal serum vitamin B12 and folate levels to cognitive test performance in subtypes of geriatric major depression.

This retrospective study evaluated the relationships between normal serum vitamin B12 and folate levels and neuropsychologic measures in a sample of 60 geriatric inpatients with psychotic depression, nonpsychotic depression, bipolar disorder, and dementia--all consecutively referred for cognitive testing. The psychotic depression subgroup demonstrated numerous significant positive correlations between B12 and cognitive subtests not seen in other diagnostic subgroups, especially those of IQ, and verbal and visual memory. Metabolic factors including vitamin B12 may play specific roles in the cognitive dysfunctions of different geropsychiatric disorders.

Responses to cortical injury: I. Methodology and local effects of contusions in the rat.

In order to develop some understanding of the evolution of cortical contusions, interdisciplinary studies including behavior, morphology and histochemistry were conducted at varying intervals after standardized injuries. A method for producing graded and reproducible focal cortical contusions in the rat is described. When these impact injuries are made in the "hindpaw cortical area,' specific trauma dose dependent behavioral deficits can be readily observed in the contralateral hindlimb. While most functional recovery occurs in the first two weeks after trauma, with severe contusions, deficits persist beyond 90 days. Morphologically these injuries progress from hemorrhages in white matter directly under contused cortex during the first hours after injury to the development of a necrotic cavity by 24 hours. The cavitation appears to expand over the subsequent two weeks and by 15 days is lined with fibroblast-like elements and macrophages. Intense acid phosphatase activity is seen on the borders of the area of necrosis. This lysosomal enzyme may participate in autolysis and development of focal cavitation following cortical contusion.

Responses to cortical injury: II. Widespread depression of the activity of an enzyme in cortex remote from a focal injury.

As a part of a broader study of the reaction of the brain to injury, we report here an interesting loss of the activity of an enzyme in areas quite remote from the site of direct injury. At 36 h following a laceration or contusion injury to the hindpaw area of the motor cortex, a peculiar loss of staining for the enzyme alpha glycerophosphate dehydrogenase (alpha-GPDH) was noted. alpha-GPDH activity was markedly depressed in cortical layers II and III throughout the hemisphere on the side of the injury. The depression of alpha-GPDH activity extended far laterally across the rhinal fissure into the pyriform cortex. The decrease in alpha-GPDH staining was prominent 4 days after the injury: however, the staining pattern had returned to normal at 9 days. Enzyme changes in animals lesioned in the occipital cortex paralleled that seen in animals with a lesion in the motor cortex. Animals which had received an undercut lesion in the motor cortex 56 days earlier were contused in the occipital cortex. The old injury site presented the same sequelae of changes as seen in other lesioned animals. Additionally, a suction ablation injury involving only a small part of motor cortex resulted in the same widespread reduction of staining for alpha-GPDH in layers II and III. The derangement in energy metabolism suggests that cells in layers II and III of the cerebral cortex may be particularly vulnerable to perturbations induced by cortical trauma. These findings may be related to the diffuse and transient functional losses observed after head injury in man.