RACIAL AND ETHNIC disparities in health care in the

Transcription

1 1712 ORIGINAL ARTICLE Black-White Differences in Patient Characteristics, Treatments, and Outcomes in Inpatient Stroke Rehabilitation Susan D. Horn, PhD, Daniel Deutscher, MSc, PT, Randall J. Smout, MS, Gerben DeJong, PhD, Koen Putman, PT, PhD ABSTRACT. Horn SD, Deutscher D, Smout RJ, DeJong G, Putman K. Black-white differences in patient characteristics, treatments, and outcomes in inpatient stroke rehabilitation. Arch Phys Med Rehabil 2010;91: Objective: To describe racial differences in patient characteristics, nontherapy ancillaries, physical therapy (PT), occupational therapy (OT), and functional outcomes at discharge in stroke rehabilitation. Design: Multicenter prospective observational cohort study of poststroke rehabilitation. Setting: Six U.S. inpatient rehabilitation facilities. Participants: Black and white patients (n 732), subdivided in case-mix subgroups (CMGs): CMGs 104 to 107 for moderate strokes (n 397), and CMGs 108 to 114 for severe strokes (n 335). Interventions: Not applicable. Main Outcome Measure: FIM. Results: Significant black-white differences in multiple patient characteristics and intensity of rehabilitation care were identified. White subjects took longer from stroke onset to rehabilitation admission and were more ambulatory prior to stroke. Black subjects had more diabetes. For patients with moderate stroke, black subjects were younger, were more likely to be women, and had more hypertension and obesity with body mass index greater than or equal to 30. For patients with severe stroke, black subjects were less sick and had higher admission FIM scores. White subjects received more minutes a day of OT, although black subjects had significantly longer median PT and OT session duration. No black-white differences in unadjusted stroke rehabilitation outcomes were found. Conclusions: Reasons for differences in rehabilitation care between black and white subjects should be investigated to understand clinicians choice of treatments by race. However, we did not find black-white differences in unadjusted stroke rehabilitation outcomes. Key Words: African Americans; European continental ancestry group; Occupational therapy; Physical therapy modalities; Rehabilitation; Rehabilitation centers; Stroke by the American Congress of Rehabilitation Medicine From the Institute for Clinical Outcomes Research, International Severity Information Systems, Salt Lake City, UT (Horn, Smout); Physical Therapy Department, Maccabi Healthcare Services Health Maintenance Organization, Tel-Aviv, Israel (Deutscher); Center for Post-acute Studies, National Rehabilitation Hospital, Washington, DC (DeJong, Putman); Department of Medical Sociology and Health Sciences, Vrije Universiteit Brussel, Brussels, Belgium (Putman). Supported by the National Institute for Disability and Rehabilitation Research, Department of Education (grant no. H133G050153). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Susan D. Horn, PhD, ISIS, Inc, 699 E South Temple, Ste 300, Salt Lake City, UT 84102, shorn@isisicor.com /10/ $36.00/0 doi: /j.apmr RACIAL AND ETHNIC disparities in health care in the United States have been a major focus of biomedical and health services research in recent years, and with good reason. Evidence over several decades has found that black subjects in particular and ethnic minorities in general have disparities across numerous health conditions in severity of disease and quality of care. Black subjects compared with white subjects are more likely to be in poor health, less likely to receive a range of services and procedures, and more likely to receive discriminatory treatment in care settings. Disparities in care frequently are associated with subsequent outcomes. 1-5 We make a distinction between differences and disparities among black and white subjects in patient characteristics, provision of care, and outcomes. Thus, we use the term difference to describe variations in patient characteristics or provision of care not necessarily associated with difference in quality of care or in outcomes, and the term disparity to describe differences in health status at intake and in quality of care or outcomes favoring black or white subjects. Disparities in Stroke Rehabilitation and Outcomes Despite unambiguous evidence for racial disparities in severity of diseases, evidence for racial disparities between black and white subjects in rehabilitation use and outcomes has been both limited and contradictory. 5,6 A recent study suggested that racial/ethnic differences may exist in use of poststroke rehabilitation services, 7 while others found no consistent racial disparities in rehabilitation care. 8,9 In a prospective study of over 800 elders in California, race appeared as the most significant single predictor of PT and OT use. Being white was the most significant predictor of rehabilitation use (odds ratio 4.25) with clear findings: Minority groups, less-well educated individuals, and the oldest old were significantly less likely to have PT or OT services, even when disability and rehabilitation diagnosis were controlled for. 10 Evidence for racial disparities in rehabilitation outcomes also is mixed. Some recent reports suggest racial disparities exist in postacute care outcomes for persons with stroke, 11,12 with inconclusive findings reported by others. 9,13 Challenges to Understanding Racial Disparities in Rehabilitation As stroke mortality declines, duration of survival increases, and costs of long-term care increase, we need more highquality evidence about potential racial disparities in stroke rehabilitation processes and outcomes. Three challenges to understanding racial disparities in rehabilitation treatments and CSI OT PSROP PT List of Abbreviations Comprehensive Severity Index occupational therapy Post-Stroke Rehabilitation Outcomes Project physical therapy

2 RACE DIFFERENCES IN STROKE REHABILITATION, Horn 1713 outcomes are the following: (1) limited data or standardized measures to characterize patient acuity or severity of illness while in rehabilitation, (2) limited data to characterize rehabilitation interventions and actual processes of care by disassembling all components of rehabilitation care and how they contribute individually and collectively to rehabilitation outcomes, and (3) inadequate numbers of minority participants to examine potential racial disparities in effectiveness of stroke interventions. Our research addressed each of these challenges. First, our study included a comprehensive method to characterize severity of illness (principal diagnosis, complications, comorbidities) apart from functional status. This enabled better characterization of clinical differences that patient groups present in rehabilitation and may account for some differences in rehabilitation care and outcome. Second, our study included a detailed method to characterize actual interventions and treatments provided in rehabilitation. This enabled better characterization of treatment differences between black and white subjects while in rehabilitation. Third, our study was drawn from a large, geographically dispersed, multicenter database of 1161 U.S. poststroke rehabilitation patients with a large minority representation, which enabled overcoming the small numbers problem for some comparisons. Physical and Occupational Therapy Activities and Interventions The importance of recording PT and OT interventions for stroke rehabilitation research in a reliable and valid manner has been described previously We make a distinction between therapy activities and therapy interventions. Therapy activities are aimed at specific functional activities such as bed mobility, transfers, gait, and community mobility. Therapy interventions are the methods, modalities, or means by which patients participate in therapy activities for example, strengthening exercises, stretching, use of assistive devices, and gait with body weight support. Each intervention can be provided within different activities. Other studies have described the use of activities and nontherapy interventions such as medications and nutrition. 19,20 However, it is not known whether differences exist in the way these activities and nontherapy interventions are provided between black and white subjects. Furthermore, to our knowledge, no previous studies have investigated how specific interventions provided within these activities are used in black and white patients during stroke rehabilitation. Thus, the purposes of this article are to describe and compare black and white patients who received inpatient rehabilitation after stroke in terms of (1) their demographic and health characteristics; (2) use of nontherapy ancillaries, PT, and OT activities; and (3) use of PT and OT interventions provided within different therapy activities. METHODS Design and Patient Selection We conducted secondary analyses of the database from the PSROP, a multicenter prospective observational cohort study. 19 This database contains 1161 patients with stroke from 6 U.S., geographically dispersed, hospital-based rehabilitation centers. Patient inclusion criteria included the following: rehabilitation diagnosis of International Classification of Diseases 9th Revision codes or , age older than 18 years, and first rehabilitation admission after current stroke; the principal reason for admission was stroke. The patient may have had previous strokes and previous rehabilitation admissions for previous strokes, but this was the first admission for the current stroke. If patients were transferred to another setting of care, such as an acute hospital, and returned to the rehabilitation center/unit within 30 days, they remained study patients. If patients were transferred to another setting of care and returned to the rehabilitation center/unit after 30 days, participation in the study ended on the day the patient was transferred. Study sites had varying numbers of black and white patients in their patient sample. The ratios of blacks to whites for each of the 6 sites were as follows: 96 to 46, 11 to 104, 20 to 78, 13 to 127, 1 to 106, and 98 to 32. Patients were excluded from the analyses here if they were of races other than black or white (n 187), had a mild stroke (n 95), had incomplete admission FIM forms (n 66), were discharged to a hospital (n 34) or other acute rehabilitation (n 13), were discharged with tube feeding (n 29), or died during rehabilitation (n 5). Hence, our study sample was 732 patients. Discharge motor FIM was missing for 3 white patients with moderate strokes and 4 patients with severe strokes (2 black and 2 white patients). Patients were subdivided in case-mix subgroups in effect at the time of the study: case-mix subgroups 104 to 107 for moderate strokes (n 397), and case-mix subgroups 108 to 114 for severe strokes (n 335). Study Group and Clinical Sites We analyzed data from 239 black and 493 white patients with moderate and severe stroke. The PSROP database carefully disassembled each component of the poststroke rehabilitation process. This approach, called practice-based evidence, 21 involved several well defined steps that are described in detail elsewhere. 19 Practice-based evidence was designed to encourage new findings including those that may challenge conventional wisdom and long-standing practice. Consecutive patients from 6 not-for-profit hospital-based rehabilitation centers contributed to the PSROP database. 19 The facilities were a convenience sample based on their diverse geographic locations and their willingness to participate. Each site enrolled up to 200 consecutive patients with stroke and obtained institutional review board approval for the protection of human subjects. There were no issues with patient refusals in this study. Patients did not have to give consent to be part of the initial study because there was no intervention beyond existing practice involved and no patient identifiers were used. Existing practices were measured from existing or newly created documentation forms and records. Data Collection For each patient, we collected detailed information from medical records and from point-of-care forms (scannable standardized documentation forms) completed by clinicians as they treated the patient. We collected additional information on some aspects of their acute hospital stay from the patient s acute hospital medical record or from documentation that accompanied the patient into rehabilitation. Data Collection Instruments and Data Collection Processes Three types of data were collected for study patients. These have been described before. 19 Briefly, these include the following: 1. Disease-specific severity of illness data (signs and symptoms) for all of patient s diagnoses obtained from chart review. Severity measures were based on the most abnormal findings for each disease criterion during each of 3 time frames: admission (first 24 hours), during entire

3 1714 RACE DIFFERENCES IN STROKE REHABILITATION, Horn Table 1: Patient Variables by Stroke Severity and Race Age (y), mean SD Women (%) * * Payer Medicaid (%) <.001* * Payer Medicare (%) * * Hemorrhagic stroke (%) * * Patient employed prior to admission (%) * * Patient lived alone prior to stroke (%) * * Patient had previous stroke (%) * * Patient s ADLs independent prior to stroke (%) * * Patient s ambulation independent prior to stroke (%) * * Urinary tract infection during rehabilitation (%) * * BMI (%).048*.137* Underweight ( 18.5) Normal ( ) Overweight (25 29) Obese ( 30) Brain location of stroke (%).003*.001* Lobar Brainstem/cerebellum Subcortical Subcortical and brainstem/cerebellum Unknown Brain side of stroke (%).026*.657* Bilateral Left Right Unknown Comorbidities (%) Congestive heart failure in rehabilitation or hospital admission * * Hypertension in rehabilitation or hospital admission * * Ischemic heart disease in rehabilitation or hospital admission * * Diabetes in rehabilitation or hospital admission <.001* <.001* Dysphagia group (%).001*.019* Unable to swallow solids Unable to swallow liquids Not otherwise specified Normal swallowing Missing Motor impairment (%).027*.846* Incomplete low paraplegia or worse Complete hemiplegia Incomplete hemiplegia Monoplegia None Neurobehavioral impairment (%).244*.022* Mood/behavioral disturbances Cognitive dysfunction Both mood/behavioral disturbances and cognitive dysfunction Psychiatric medications but no reported mood/ behavioral disturbances or cognitive dysfunction Mood/behavioral disturbances No. of days from stroke symptom onset to rehabilitation admission, mean SD BMI at admission, mean SD < Rehabilitation admission severity CSI, mean SD Admission total FIM score <.001

4 RACE DIFFERENCES IN STROKE REHABILITATION, Horn 1715 Table 1: Patient Variables by Stroke Severity and Race (Cont d) Admission Motor FIM score, mean SD Admission Motor FIM score components, mean SD Bathing Bed to chair transfer Bladder management Bowel management Dressing lower Dressing upper Eating <.001 Grooming Stairs Toilet transfer Toileting Tub transfer Walk Admission cognitive score, mean SD <.001 Admission Cognitive score components, mean SD Comprehension Expression Memory Problem solving <.001 Social interaction NOTE. All P values are based on 2-sample t tests except those marked with an asterisk. Significant race disparities (P.05) for higher values in black subjects are boldface. Significant race disparities (P.05) for higher values in white subjects are in italics. Abbreviations: ADLs, activities of daily living; BMI, body mass index. *P values from 2 tests. rehabilitation stay (maximum), and discharge (last 24 hours). For each time frame, the CSI produced a continuous measure (0 to no predetermined upper limit) for each patient diagnosis and an overall continuous score that reflected interactions of diseases. Higher scores indicated higher severity. 2. Rehabilitation point-of-care intervention data were documented by clinicians for each therapy session during the patient s rehabilitation stay. Subcommittees for different therapy disciplines created forms to include descriptions of activities and interventions they deemed necessary to capture the contribution of their discipline to rehabilitation care (this did not duplicate what was contained typically in standard documentation). Activities and interventions have been described and analyzed previously Briefly, most PT interventions were derived from the guide to PT practice, 26 and some were added and defined by participating physical therapists. Physical therapists could select from 10 activities, including prefunctional, bed mobility, sitting, transfers, sit to stand, wheelchair mobility, pregait, gait, advanced gait, and community mobility; and 59 interventions within an activity, including 27 assistive devices and 1 other intervention. For this study, the latter 2 categories were not included in analyses, leaving 31 possible PT interventions (8 neuromuscular, 5 musculoskeletal, 2 cardiopulmonary, 4 cognitive/perceptual/sensory, 3 educational, 4 equipment application or ordering, 3 modalities [electric stimulation, ultrasound, biofeedback], 2 pet therapy interventions). Occupational therapists could select from 52 interventions, including 20 assistive devices and 1 other intervention. As in PT, the latter 2 categories were not included in analyses, leaving 31 possible OT interventions (7 neuromuscular, 4 adaptive/compensatory, 4 musculoskeletal, 2 cardiopulmonary, 4 cognitive/perceptual/sensory, 3 educational/training, 4 equipment, 3 modalities interventions). Again, each intervention could be provided within specific activities, which included prefunctional, bed mobility, sitting balance, upper extremity control, transfers, wheelchair management, bathing, grooming, dressing, toileting, feeding, functional mobility, home management, community integration, and leisure. Therapists recorded the amount of time spent on each activity in 5 minute increments. Up to 5 different interventions could be recorded as provided within each activity. 3. Patient, process, and outcome data from medical records/ auxiliary data module. The CSI system allows for creation of auxiliary data modules, which are sets of studyspecific data elements that are collected along with patient severity information from each patient s rehabilitation medical record. Most variables contain date and time fields so that they can be associated with other variables in time sequence. All patient, process, and outcome variables in the auxiliary data module were determined through an iterative process with the study s clinical advisory panel. The auxiliary data module contained over 200 variables, and many of these variables had numerous data entries. Training and Reliability Each site s data collector attended a 4-day training session during which efficient collection of accurate data was explained and practiced. After the training session, each data collector underwent a rigorous manual reliability testing pro-

5 1716 RACE DIFFERENCES IN STROKE REHABILITATION, Horn cess to ensure complete and accurate data collection that went beyond internal data editing features of CSI software (eg, features that prohibit the entry of nonsensible values). Reliability monitoring was conducted at 4 points throughout the PSROP to ensure that data accuracy was maintained. An agreement rate of 95% at the criterion level between each data collector and the project training-team reliability staff was required for each reliability test. If 95% agreement was not obtained, the project trainer conducted a review session with the data collector. Follow-up reliability tests were conducted until 95% agreement at the criterion level was achieved. Analysis We used multiple statistical techniques to analyze the PSROP black-white differences data. Descriptive statistics were used to examine frequencies of categoric patient and treatment variables, and average and amount of variation (SD) for continuous measures. Chi-square tests were used for analysis of categoric data, and t tests or analyses of variance for continuous data. Significance alpha was set at.05. All analyses were performed with SPSS statistical software, version 15. a We used 2 sets of process variables to describe PT and OT. The first set represented intensity of activities (number of minutes a day for each activity), and the second set represented intensity of interventions provided within activities (number of minutes a day for each intervention within activity). Because time intervals were originally assigned to activities in the point-of-care forms, time intervals needed to be reassigned to the combination of activities and interventions. Therapists told us that different interventions were provided within each activity simultaneously, allowing us to assign time values recorded for each activity to each activity-intervention combination. We analyzed data of patients with moderate strokes separately from patients with severe strokes. RESULTS Using bivariate analyses, we found significant race differences in (1) patient characteristics (table 1); (2) general process variables, nutrition, and nontherapy ancillaries used during rehabilitation (table 2), and PT or OT treatment processes (tables 3 and 4, respectively); and (3) outcomes of patients with severe stroke (table 5). For clarity, we highlighted race differences that were significant at the.05 level in boldface for variables with higher values among black subjects and used italics for variables with higher values in white subjects. For patients with either moderate or severe stroke, white patients took longer from stroke onset to rehabilitation admission, had significantly more hemorrhagic and lobar strokes and Table 2: Nontherapy Ancillary Process Variables by Stroke Severity and Race Rehabilitation length of stay, mean SD Evidence of O 2 use during rehabilitation (%) * * Evidence of C/BI PAP (%) * * Evidence of sleep apnea or C/BI PAP (%) * * Nutrition (%) Tube-feeding use during rehabilitation * * Tube-feeding groups:.345*.013* Tube-feeding 1 24% of stay Tube-feeding 25% of stay No tube-feeding Medication (%) Other antidepressants * * Analgesic: muscle relaxants * * Opioid analgesics * * Allergy/antihistamine (sedating) medication * * Benzodiazepines * * Gastrointestinal: 5-hydroxytriptamine 3 (5-HT3) receptor agonist * * Gastrointestinal: nausea/vomiting * * New SSRIs * * Old SSRIs * * Atypical antipsychotics * * Typical antipsychotics * * Neurologic: Other (modafinil) * * Neurostimulant: anorexiant * * Neurostimulant: antiparkinson * * New anticonvulsant * * Old anticonvulsant * * Anxiolytic * * Anxiolytic/hypnotic * * Tricyclic antidepressant * * NOTE. All P values are based on 2-sample t tests except those marked with an asterisk. Significant race disparities (P.05) for higher values in black subjects are boldface. Significant race disparities (P.05) for higher values in white subjects are in italics. Abbreviations: C/BI PAP, continuous/bilevel positive airway pressure; SSRI, selective serotonin reuptake inhibitor. *P values from 2 tests.

6 RACE DIFFERENCES IN STROKE REHABILITATION, Horn 1717 Table 3: Physical Therapy Process Variables by Stroke Severity and Race Physical therapy process variables No. of min/d Median session duration, (min/d) mean SD < Family education (sessions/d) Patient education (sessions/d) Formal assessment (min/d) Total time for activities/length of stay (min/d), mean SD Bed mobility Sitting Transfers Sit to stand Wheelchair mobility Pregait Gait Advanced gait Community mobility Total time for interventions within activities/length of stay (min/d), mean SD Advanced gait: balance training Advanced gait: motor control Advanced gait: motor learning Advanced gait: postural awareness Bed mobility: motor control Bed mobility: motor learning Bed mobility: strengthening Community mobility: balance training Community mobility: motor learning Gait: aerobic conditioning exercises Gait: cognitive training Gait: family caregiver education Gait: gait with body weight support Gait: motor control Prefunction: motor control <.001 Sitting: motor control Sitting: NDT NA NA NA Sit to stand: motor control Transfers: perceptual training Transfers: strengthening Wheelchair: patient education NOTE. All P values are based on 2-sample t tests. Significant race disparities (P.05) for higher values in black subjects are boldface. Significant race disparities (P.05) for higher values in white subjects are in italics. Abbreviations: NA, not applicable because of frequencies 10%; NDT, neurodevelopmental treatment. dysphagia, and were more ambulatory prior to stroke. Black patients had more subcortical strokes and almost double the frequency of diabetes. In addition, for patients with moderate stroke, black patients were younger, were more likely to be female and under Medicaid coverage, and had more hypertension and obesity with body mass index greater than or equal to 30. Also, for patients with severe stroke, blacks were less sick (lower admission CSI scores) and had higher admission total, motor, and cognitive FIM scores (see table 1). For patients with either moderate or severe stroke, white patients received significantly more oxygen use during rehabilitation, and white patients with severe stroke received overall more nutrition supplements. Other nontherapy ancillaries variables with racial differences included white patients having greater use of muscle relaxants, opioid analgesics, 5-hydroxytryptamine receptors inhibitors, and anxiolytics/hypnotics. In addition, for patients with severe stroke, white patients had a slightly longer rehabilitation length of stay, used more continuous/bilevel positive airway pressure, used more tube feeding, and took more new selective serotonin reuptake inhibitors, other antidepressants, atypical antipsychotics, modafinil, and anorexiant neurostimulants (see table 2). Comparisons of PT and OT process variables between black and white patients (see tables 3 and 4) showed that for patients with either moderate or severe stroke, white patients received overall more intense (min/d) occupational therapy, although blacks had significantly longer median session duration for both PT and OT. Also, whites had more intense PT and OT family education. Racial differences were found for PT and OT intensity of activities, with black patients receiving more intense PT wheelchair mobility and OT transfers, leisure performance, and sitting balance/trunk control activities. White pa-

7 1718 RACE DIFFERENCES IN STROKE REHABILITATION, Horn Table 4: Occupational Therapy Process Variables by Stroke Severity and Race OT process variables No. of min/d Median session duration, (min/d) mean SD < Family education (sessions/d) Patient education (sessions/d) Formal assessment (min/d) Total time for activities/length of stay (min/d), mean SD Bathing Dressing Grooming Toileting Feeding/eating Transfers < Bed mobility Functional mobility Home management Community integration Leisure performance Upper-extremity control Wheelchair management Sitting balance/trunk control < Total time for interventions within activities/length of stay (min/d), mean SD Dressing: cognitive training Dressing: perceptual training Dressing: postural awareness Dressing: strengthening Dressing: visual training Feeding/eating: motor learning NA NA NA Functional mobility: motor learning Home management: balance training Home management: patient education Home management: perceptual training Prefunctional: balance training Prefunctional: visual training Sitting balance: cognitive training NA NA NA Toileting: balance training Upper-extremity control: cognitive training Upper-extremity control: motor learning Upper-extremity control: patient education Upper-extremity control: sensory training NOTE. All P values are based on 2-sample t tests. Significant race disparities (P.05) for higher values in black subjects are boldface. Significant race disparities (P.05) for higher values in white subjects are in italics. Abbreviation: NA, not applicable because of frequencies 10%. tients had more intense PT transfers and OT bathing, dressing, grooming, and toileting activities. For patients with moderate stroke, white patients had more intense PT community mobility activities and sit to stand exercises, along with more OT community integration and home management activities and OT patient education. For patients with severe stroke, white patients had more than 2 days longer rehabilitation length of stay, more intense OT feeding/eating activities, and more intense wheelchair management exercises overall. Blacks received more PT patient education, more PT formal assessment time, and more intense OT functional mobility overall. Because of a very large number of combination variables (interventions within activities), we included in tables 3 and 4 only those found in at least 1 subsequent regression model (presented in our second article within this issue), which investigates associations between processes of care and discharge motor FIM. From those variables, there were 3 PT therapy variables that had significant racial differences among patients with moderate and severe stroke. Two variables had higher values in white patients (family education within gait and perceptual training within transfers), and 1 was provided more intensely among black patients (motor control within prefunctional activities). For patients with moderate and severe stroke, we found 7 OT therapy vari-

8 RACE DIFFERENCES IN STROKE REHABILITATION, Horn 1719 Table 5: Outcome Variables by Stroke Severity and Race Severity (CSI) during rehabilitation, mean SD Maximum CSI Discharge CSI Increase in CSI (maximum admission) Total FIM, mean SD Discharge total FIM Increase in total FIM (discharge admission) Motor FIM, mean SD Discharge motor FIM Increase in motor FIM (discharge admission) Cognitive FIM, mean SD Discharge cognitive FIM Increase in cognitive FIM (discharge admission) Discharge destination (%).351*.636* Home/community Skilled nursing facilities NOTE. All P values are based on 2-sample t tests except those marked with an asterisk. Significant race disparities (P.05) for higher values in black subjects are boldface. Significant race disparities (P.05) for higher values in white subjects are in italics. *P values from 2 tests. Discharge motor FIM was missing for 3 white patients with moderate stroke and 4 patients with severe stroke (2 black and 2 white patients). ables that had significant racial differences, all with higher intensity among white patients. There were no racial differences in bivariate outcomes for patients with moderate stroke (see table 5). However, for patients with severe stroke, whites had higher maximum CSI (sicker), lower discharge cognitive FIM, and greater increases in CSI, total FIM, and cognitive FIM scores. DISCUSSION Our main purpose was to assess black-white differences in patient characteristics and provision of PT and OT rehabilitation care. We examined PT and OT care at the level of therapy activities and of interventions provided within these activities. Comparisons of patient demographic characteristics (see table 1) revealed some differences between black and white patients, mainly in the moderate stroke group. Blacks were younger, had a higher percentage of women, were less independent or ambulatory prior to their stroke, and were more likely to be on Medicaid, suggesting different socioeconomic backgrounds compared with white patients. In addition, for patients with moderate and severe stroke, black patients were admitted to rehabilitation (from day of stroke onset) significantly sooner than white patients, which might be related to their type of medical coverage, stroke characteristics, illness, or other demographic characteristics. It may be that acute care hospitals discharge self-pay and Medicaid patients to rehabilitation sooner because of lower reimbursement for care than for other payers. Racial differences in health characteristics (see table 1) were found on admission to rehabilitation. White patients with severe strokes were sicker on admission with higher CSI scores and lower admission FIM scores. However, rate of diabetes was about double for black patients with severe stroke compared with white patients, similar to previous findings. 27 In the moderate group, black and white patients had different health status at intake in some health characteristics like body mass index and diabetes (higher in black patients) or dysphagia (higher in white patients). However, in moderate strokes, overall health status seemed similar between the races, with no differences in CSI and admission FIM scores, even though black patients were approximately 4 years younger. Comparisons of general processes of care between black and white patients (see table 2) showed longer length of stay for white patients with severe stroke, possibly because of white patients with severe stroke being sicker than black patients with severe stroke. However, table 2 reveals that for patients with moderate and severe stroke, almost all significant differences in general process variables and nontherapy ancillaries indicated more care for white patients. These differences do not appear to be related to admission health status of the 2 groups because white patients with moderate stroke were not sicker than black patients with moderate stroke, but still white patients received more care. This result suggests the need to study the clinical reasoning processes of health care practitioners when selecting different care for black or white patients. One possible explanation is that black patients started rehabilitation earlier than white patients and as a result, may have needed less rehabilitation to achieve the same functional outcome. Previous research has shown that patients with stroke who start rehabilitation earlier have better outcomes. 20 Those who start later may need more rehabilitative care in order to mitigate the debilitating effects of prolonged bed rest in the interval between stroke onset and the start of rehabilitation. We analyzed racial differences in provision of 9 different activities for PT and 14 different activities for OT, as well as 31 interventions, each possibly being provided within each activity (see tables 3 and 4). Racial differences were found, with white patients usually getting more intense therapy than black patients, a finding that supports past 10 and more recent 7 reports. However, the magnitude of these differences in minutes a day was usually small. Differences were much more apparent in OT, where many therapy variables had higher intensity among white patients, with only a few activities and no interventions within activities having higher intensity in black patients. Here again, this suggests future investigation of the reasons for these differences in choice of treatments between black and white patients. This will help us understand whether it was the

9 1720 RACE DIFFERENCES IN STROKE REHABILITATION, Horn patient s health condition, or possibly the patient s race, associated with choice and intensity of therapy. We did not find any racial differences in unadjusted outcomes for patients with moderate stroke but did find statistically significant differences in discharge cognitive FIM scores favoring black patients with severe stroke (see table 5). Although outcomes that are not adjusted for patient and process variables cannot be used to draw conclusions regarding riskadjusted outcome comparisons between black and white patients, our result of no apparent differences in functional status at discharge (ie, discharge motor FIM) suggests that no practical racial disparities in outcomes exist. Study Limitations Our study included 6 facilities, each with a different frequency of black and white patients with stroke as noted in the Methods section. For example, 1 site had 1 black and 106 white patients, while another site had 98 black and 32 white patients. This may have led to confounding of facility policy with racial differences in provision of care. We did not have enough cases of other races, such as Asian or Hispanic, to include them in separate comparisons. We acknowledge that there is a risk of racial misclassification. Another possible limitation is related to the way we chose to assign time values to interventions within activities. Because time values were initially assigned to activities, assignment of activity time to each intervention provided within that activity (because of therapists stating that most interventions within an activity were done simultaneously) might have added noise to the data, possibly affecting our results. CONCLUSIONS Significant racial differences in intensity of rehabilitation care were identified, with white patients usually getting more intense therapy than black patients in OT. Reasons for these differences should be investigated to understand better clinical reasoning that leads clinicians in their choice of treatments by race. However, we did not find racial differences in unadjusted stroke rehabilitation motor outcomes, raising the possibility of white or black subjects receiving more intense therapies associated with both better and worse outcomes. This suggests the need to study associations of PT and OT care with functional outcomes, adjusting for patient characteristics, to enable optimization of treatment choices for all patients attending stroke rehabilitation. Acknowledgments: We acknowledge the role and contributions of the collaborators at each of the clinical sites represented in the Post-stroke Rehabilitation Outcomes Project: Brendan Conroy, MD (Stroke Recovery Program, National Rehabilitation Hospital, Washington, DC); Richard Zorowitz, MD (Department of Rehabilitation Medicine, University of Pennsylvania Medical Center, Philadelphia, PA); David Ryser, MD (Neuro Specialty Rehabilitation Unit, LDS Hospital, Salt Lake City, UT); Jeffrey Teraoka, MD (Division of Physical Medicine and Rehabilitation, Stanford University, Palo Alto, CA); Frank Wong, MD and LeeAnn Sims, RN (Rehabilitation Institute of Oregon, Legacy Health Systems, Portland, OR); and Murray Brandstater, MD (Loma Linda University Medical Center, Loma Linda, CA). References 1. Adler NE, Rehkopf DH. U.S. disparities in health: descriptions, causes, and mechanisms. Ann Rev Public Health 2008; 29: Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med 1996;335: Oddone EZ, Petersen LA, Weinberger M, Freedman J, Kressin NR. Contribution of the Veterans Health Administration in understanding racial disparities in access and utilization of health care: a spirit of inquiry. Med Care 2002;40(1 Suppl): I Schneider EC, Zaslavsky AM, Epstein AM. Racial disparities in the quality of care for enrollees in Medicare managed care. JAMA 2002;287: Smedley BD, Stith AY, Nelson AR, Institute of Medicine (U.S.). Committee on Understanding and Eliminating Racial and Ethnic Disparities in Health Care. Unequal treatment: confronting racial and ethnic disparities in health care. Washington, DC: National Academy Press; Ellis C, Breland HL, Egede LE. Racial/ethnic differences in utilization of post-stroke rehabilitation services: a systematic review. Ethn Dis 2008;18: Ellis C, Egede LE. Racial/ethnic differences in poststroke rehabilitation utilization in the USA. Exp Rev Cardiovasc Ther 2009; 7: Ross JS, Halm EA, Bravata DM. Use of stroke secondary prevention services: are there disparities in care? Stroke 2009; 40: Goldstein LB, Matchar DB, Hoff-Lindquist J, Samsa GP, Horner RD. Veterans Administration Acute Stroke (VASt) Study: lack of race/ethnic-based differences in utilization of stroke-related procedures or services. Stroke 2003;34: Mayer-Oakes SA, Hoenig H, Atchison KA, Lubben JE, De Jong F, Schweitzer SO. Patient-related predictors of rehabilitation use for community-dwelling older Americans. J Am Geriatr Soc 1992; 40: Bhandari VK, Kushel M, Price L, Schillinger D. Racial disparities in outcomes of inpatient stroke rehabilitation. Arch Phys Med Rehabil 2005;86: Ottenbacher KJ, Campbell J, Kuo YF, Deutsch A, Ostir GV, Granger CV. Racial and ethnic differences in postacute rehabilitation outcomes after stroke in the United States. Stroke 2008;39: Stansbury JP, Jia H, Williams LS, Vogel WB, Duncan PW. Ethnic disparities in stroke: epidemiology, acute care, and postacute outcomes. Stroke 2005;36: De Wit L, Kamsteegt H, Yadav B, Verheyden G, Feys H, De Weerdt W. Defining the content of individual physiotherapy and occupational therapy sessions for stroke patients in an inpatient rehabilitation setting: development, validation and inter-rater reliability of a scoring list. Clin Rehabil 2007;21: Pomeroy VM, Cooke E, Hamilton S, Whittet A, Tallis RC. Development of a schedule of current physiotherapy treatment used to improve movement control and functional use of the lower limb after stroke: a precursor to a clinical trial. Neurorehabil Neural Repair 2005;19: Tyson SF, Selley A. The development of the Stroke Physiotherapy Intervention Recording Tool (SPIRIT). Disabil Rehabil 2004;26: Wade DT. Barriers to rehabilitation research, and overcoming them. Clin Rehabil 2003;17: Wittwer JE, Goldie PA, Matyas TA, Galea MP. Quantification of physiotherapy treatment time in stroke rehabilitation criterionrelated validity. Aust J Physiother 2000;46: Gassaway J, Horn SD, DeJong G, Smout RJ, Clark C, James R. Applying the clinical practice improvement approach to stroke rehabilitation: methods used and baseline results. Arch Phys Med Rehabil 2005;86(12 Suppl 2):S Horn SD, DeJong G, Smout RJ, Gassaway J, James R, Conroy B. Stroke rehabilitation patients, practice, and outcomes: is earlier and more aggressive therapy better? Arch Phys Med Rehabil 2005;86(12 Suppl 2):S

10 RACE DIFFERENCES IN STROKE REHABILITATION, Horn Horn SD, Gassaway J. Practice-based evidence study design for comparative effectiveness research. Med Care 2007;45(10 Suppl 2):S Hatfield B, Millet D, Coles J, Gassaway J, Conroy B, Smout RJ. Characterizing speech and language pathology outcomes in stroke rehabilitation. Arch Phys Med Rehabil 2005;86(12 Suppl 2):S Jette DU, Latham NK, Smout RJ, Gassaway J, Slavin MD, Horn SD. Physical therapy interventions for patients with stroke in inpatient rehabilitation facilities. Phys Ther 2005;85: Latham NK, Jette DU, Slavin M, et al. Physical therapy during stroke rehabilitation for people with different walking abilities. Arch Phys Med Rehabil 2005;86(12 Suppl 2):S Richards LG, Latham NK, Jette DU, Rosenberg L, Smout RJ, DeJong G. Characterizing occupational therapy practice in stroke rehabilitation. Arch Phys Med Rehabil 2005;86(12 Suppl 2):S American Physical Therapy Association. Guide to physical therapist practice: second edition. American Physical Therapy Association. Phys Ther 2001;81: Bravata DM, Wells CK, Gulanski B, et al. Racial disparities in stroke risk factors: the impact of socioeconomic status. Stroke 2005;36: Supplier a. SPSS Inc, 233 S Wacker Dr, Chicago, IL