INTRACEREBRAL HEMORRHAGE (ICH) is a major

968 Functional Recovery Following Rehabilitation After Hemorrhagic and Ischemic Stroke Peter J. Kelly, MB, MRCPI, Karen L. Furie, MD, MPH, Saad Shafqat, MD, PhD, Nikoletta Rallis, BA, Yuchiao Chang, PhD, Joel Stein, MD From the Department of Physical Medicine and Rehabilitation, Spaulding Rehabilitation Hospital, Boston (Kelly, Shafqat, Rallis, Stein); Stroke Service, Department of Neurology, Massachusetts General Hospital, Boston (Kelly, Furie, Shafqat); and Medical Practices Evaluation Center, Department of Medicine, Massachusetts General Hospital, Boston, MA (Chang). Supported by the Clinical Investigator Training Program, Harvard/MIT Health Sciences and Technology Beth Israel Deaconess Medical Center, in collaboration with Pfizer Inc; the American Heart Association, New England Affiliate; a Clinical Scientist Development Award, the Doris Duke Charitable Foundation; a Junior Faculty Award, Hartford Foundation for Excellence in Geriatric Research; and the Arthur Merrill and the Esther U. Sharp Memorial Fund. 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 Joel Stein, MD, Spaulding Rehabilitation Hospital, 125 Nashua St, Boston, MA 02114, e-mail: Jstein@partners.org. 0003-9993/03/8407-7620$30.00/0 doi:10.1016/s0003-9993(03)00040-6 ABSTRACT. Kelly PJ, Furie KL, Shafqat S, Rallis N, Chang Y, Stein J. Functional recovery following rehabilitation after hemorrhagic and ischemic stroke. Arch Phys Med Rehabil 2003;84:968-72. Objectives: To quantify recovery after rehabilitation therapy and to identify factors that predicted functional outcome in survivors of intracerebral hemorrhage (ICH) compared with cerebral infarction. Design: Retrospective study of consecutive ICH and cerebral infarction admissions to a rehabilitation hospital over a 4-year period. Setting: Free-standing urban rehabilitation hospital. Participants: A total of 1064 cases met the inclusion criteria (545 women, 519 men; 871 with cerebral infarction, 193 with ICH). Interventions: Not applicable. Main Outcome Measures: Functional status was measured using the FIM instrument, recorded at admission and discharge. Recovery was quantified by the change in FIM total score ( FIM total score). Outcome measures were total discharge FIM score and FIM total score. Univariate and multivariate analyses were performed. Results: Total admission FIM score was higher in patients with cerebral infarction than in patients with ICH (59 vs 51, P.0001). No difference in total discharge FIM score was present. Patients with ICH made a significantly greater recovery than those with cerebral infarction ( FIM total score, 28 vs 23.3; P.002). On multivariate analysis, younger age, longer length of stay, and admission FIM cognitive subscore independently predicted total discharge FIM and FIM total score. The severity of disability at admission, indicated by total admission FIM score, independently predicted total discharge FIM score, but not FIM total score. The ICH patients with the most severely disabling strokes had significantly greater recovery than cerebral infarction patients with stroke of similar severity. Conclusions: The patients with ICH had greater functional impairment than the cerebral infarction patients at admission, but made greater gains. Patients with the most severely disabling ICH improved more than those with cerebral infarction of comparable severity. Initial severity of disability, age, and duration of therapy best predicted functional outcome after rehabilitation. Key Words: Cerebral hemorrhage; Rehabilitation; Stroke; Treatment outcome. 2003 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation INTRACEREBRAL HEMORRHAGE (ICH) is a major cause of stroke-related morbidity and mortality, accounting for 5% to 10% of the 700,000 annual incident strokes in the United States. 1,2 Although the risk factors, pathophysiologic mechanisms, and acute complications of ICH are well understood, fewer data are available regarding outcome after ICH. 3 Compared with cerebral infarction, ICH has been associated with more severe neurologic impairment and higher mortality rates in the acute phase. 4-12 However, relatively few detailed data are available concerning functional outcome in survivors of ICH after discharge from the acute hospital setting, particularly after intensive rehabilitation therapy. To date, outcome after ICH has mainly been described in terms of survival, neurologic impairment or functional disability, as measured by the modified Rankin score or Barthel Index. 5-13 Poor outcome has been strongly associated with larger hemorrhage volume, initial measures of clinical severity, and hematomas located in the brainstem, cerebellum, and putamen. 3,14-16 Few studies with a large sample of ICH survivors have examined their trajectory of recovery and degree of disability after rehabilitation therapy compared with cerebral infarction patients, using a measurement instrument that responds to progressive increments in functional status. 5,17 This information is valuable to further define the natural history of ICH and cerebral infarction, and may be of practical use to physicians when attempting to determine the prognosis of ICH in both the acute and rehabilitative phases of care. To further study these issues, we retrospectively analyzed all patients with ICH and cerebral infarction consecutively admitted to an inpatient rehabilitation facility over a 4-year period. Our hypotheses were: (1) ICH is associated with greater functional disability than cerebral infarction at the onset of rehabilitation therapy; (2) ICH is associated with greater increments in functional recovery during rehabilitation therapy than cerebral infarction; (3) there is no difference between ICH and cerebral infarction patients in functional disability after rehabilitation therapy; and (4) clinical variables present at the time of discharge from the acute hospital may be predictive of postrehabilitation functional recovery. Our aims were: (1) to quantify functional impairment in ICH and cerebral infarction at the onset of rehabilitation therapy, measured by the total admission score on the FIM instrument; (2) to determine the change in functional ability during rehabilitation therapy in patients with ICH compared with those with cerebral infarction, measured by the change in total FIM scores ( FIM total score, ie, total discharge FIM score minus total admission FIM

RECOVERY AFTER HEMORRHAGIC AND ISCHEMIC STROKE, Kelly 969 Table 1: Baseline Characteristics of Study Sample All (N 1064) Cerebral Infarct (n 871) ICH (n 193) P Test Statistic ( 2 or t test) Sex, n (%) Male 519 (48.8) 416 (47.8) 103 (53.4).200 1.99 Female 545 (51.2) 455 (52.2) 90 (46.6) Mean age SD 69.8 13.9 70.5 13.7 66.5 14.7.0003 3.61 Location, n (%) Supratentorial 905 (85.1) 740 (85) 165 (85.5).900.035 Infratentorial 159 (14.9) 131 (15) 28 (14.5) Laterality, n (%) Left 475 (44.7) 383 (44) 92 (47.6) Right 434 (40.8) 348 (40) 86 (44.6) Bilateral 53 (5) 51 (5.8) 2 (1.0) Other/unknown 102 (9.5) 89 (10.2) 13 (6.8) Rehabilitation LOS: mean SD 33.7 25.6 32.9 25 37.3 28.3.050 1.946 Abbreviation: SD, standard deviation. score); (3) to quantify functional outcome of patients with ICH and cerebral infarction after rehabilitation therapy, measured by the total discharge FIM score; and (4) to determine what variables are predictive of functional outcome after cerebral infarction and ICH. METHODS Using electronic hospital admission and billing databases, we retrospectively identified all persons with a diagnosis of stroke who were consecutively admitted to Spaulding Rehabilitation Hospital (Boston, MA) between January 1, 1996 and December 31, 1999. All medical records were reviewed by a trained research assistant and stroke neurologist (PJK, SS), and relevant clinical, demographic, and FIM information was extracted. The stroke category (infarct, hemorrhage), stroke location (supratentorial, infratentorial), and side of the stroke (right, left, bilateral, midline) were categorized based on reports from information recorded during referral clinical exams and neuroimaging (brain computed tomography [CT] or magnetic resonance imaging [MRI]). If a patient required acute hospital transfer for evaluation of medical complications, that transfer was considered as an interruption of the rehabilitation course when the patient was readmitted within 7 days, and as a new rehabilitation course when the patient was readmitted after 7 days. Criteria for inclusion in the study were: (1) a primary indication for admission for inpatient rehabilitation of new ICH or cerebral infarction, confirmed by either CT or MRI, and (2) complete FIM data available. Exclusion criteria were: (1) a primary indication for inpatient rehabilitation other than new ischemic or hemorrhagic stroke (including functional impairment related to previous stroke or other neurologic disease); (2) stroke due to subarachnoid, subdural, or epidural hemorrhage; (3) absence of neuroimaging data; and (4) death during inpatient rehabilitation therapy. Functional status was measured by the patients scores on the motor and cognitive components of the FIM, 18-21 which was prospectively obtained at the time of admission and discharge for all patients by clinicians trained in the use of the instrument. The FIM is an 18-item scale, which measures independence in tasks involved in feeding, grooming, dressing, toileting, mobility, and cognition. Subjects are scored from 7 (totally independent) to 1 (totally dependent or not testable) on each item, with a score of 126 indicating total functional independence. The FIM has proven content and construct validity, is responsive to small increments in functional status after stroke, and correlates highly with measures of neurologic impairment after stroke, such as the National Institutes of Health Stroke Scale. 18-21 The study was approved by the institutional review board at Spaulding Rehabilitation Hospital. Statistical Analysis Primary outcome measures were the patients FIM total score and total discharge FIM score. Secondary outcome measures were FIM efficiency ( FIM divided by rehabilitation length of stay [LOS]), motor and cognitive subscores of FIM total score ( FIM motor, FIM cognitive), and total discharge FIM score (discharge FIM motor, discharge FIM cognitive). Two-sample t tests and Fisher exact tests were used to compare the patient characteristics between the hemorrhage and infarct groups. Multiple regression analyses were performed separately for cerebral infarction and ICH subgroups, to examine the contribution of suspected predictor variables to FIM total score and total discharge FIM score. Independent variables included: (1) age; (2) total admission FIM score; (3) motor admission FIM subscore (admission FIM motor); (4) cognitive admission FIM subscore (admission FIM cognitive); (5) stroke location (supratentorial, infratentorial); (6) stroke laterality (right, left, bilateral, midline); and (7) rehabilitation LOS. To further examine the effect of disability severity on outcome, total admission FIM scores were stratified into 4 groups as defined previously by Alexander and colleagues 17,22 (total admission FIM strata: 40, 41 60, 61 80, 80), and FIM total score was compared for cerebral infarction and ICH for each stratum, using 2-sample t tests. To avoid the inflation of type I error due to multiple comparisons, a Bonferroni adjustment was applied, raising the significance threshold to.0125 for this part of the analysis. RESULTS Patient Sample Baseline characteristics of the study sample are presented in table 1. A total of 1064 cases were identified that fulfilled inclusion criteria (871 with cerebral infarction, 193 with ICH; 545 women, 519 men). For both cerebral infarction and ICH subgroups, 85% of strokes in each category were located in the supratentorial compartment. Overall, the mean age was significantly lower (P.001) and the mean rehabilitation hospital LOS significantly longer (P.05) in the ICH subgroup compared with the cerebral infarction subgroup.

970 RECOVERY AFTER HEMORRHAGIC AND ISCHEMIC STROKE, Kelly Table 2: FIM Scores of Study Sample FIM Instrument Cerebral Infarct ICH P Test Statistic (t test) Total admission FIM score 59 24.7 51.1 24.7.001 4.006 Total discharge FIM score 82.3 1.0 79.1 2.2.200 1.334 FIM total score 23.3 0.6 28 1.4.002 3.111 Admission FIM motor 37.3 17.4 32.5 18.2.001 3.399 Discharge FIM motor 57.5 0.8 55.2 1.7.200 1.197 FIM motor 20.2 0.5 22.7 1.1.026 2.224 Admission FIM cognitive 21.6 9.3 18.6 8.8.001 4.168 Discharge FIM cognitive 24.9 0.3 23.6 0.6.080 1.771 FIM cognitive 3.2 0.2 5.0 0.5.001 3.719 NOTE. Values are mean SD. Functional Outcome At admission to the rehabilitation facility, the ICH subgroup was significantly more functionally impaired than the cerebral infarction subgroup, as indicated by their admission FIM total score and their admission FIM motor and admission FIM cognitive subscores (P.001 for all comparisons) (table 2). However, no statistically significant differences in total discharge FIM, discharge FIM motor, or discharge FIM cognitive scores were present between the subgroups (see table 2). Overall, the ICH subgroup had larger increases in total FIM scores than the cerebral infarction subgroup during inpatient rehabilitation. These differences were highly significant, both for FIM total score (P.002) and for FIM motor (P.036) and FIM cognitive (P.001) subscores. At admission, 78% of subjects with ICH and 67% of subjects with cerebral infarction had total FIM scores less than 72 (P.003), corresponding on average to greater than minimal assistance needed for items measured. By discharge, 66% of subjects with ICH and 68% of subjects with cerebral infarction had FIM scores greater than 72 (corresponding on average to minimal or no assistance needed for items measured). At discharge, 50% of subjects with cerebral infarction and 42% of those with ICH had FIM scores greater than 90 (corresponding on average to functional independence). For both cerebral infarction and ICH subgroups, we performed multivariate analyses to examine predictors of functional outcome, with total discharge FIM and FIM total scores as the dependent variables (table 3). Consistent with previous reports, younger age and longer LOS independently predicted higher total discharge FIM and FIM total scores (P.001 for all), for both cerebral infarction and ICH subgroups. A higher admission FIM cognitive score was an independent predictor of total discharge FIM and FIM total scores (P.001) in the cerebral infarction but not in the ICH subgroup. Surprisingly, the initial severity of disability measured by total admission FIM was not a significant predictor of the degree of functional improvement attained during rehabilitation therapy ( FIM total score, P.15) for both cerebral infarction and ICH subgroups. As anticipated, total admission FIM score was strongly predictive of functional status at discharge (total discharge FIM score, P.001) for both subgroups. To further examine the effect of initial functional severity on recovery, we compared FIM in subjects with ICH and with cerebral infarction, the scores stratified into 4 groups based on initial stroke functional severity measured by their total admission FIM ( 40, 41 60, 61 80, 80), as previously described by Alexander and colleagues. 17,22 After correction for multiple statistical comparisons, when FIM was compared for cerebral infarction and ICH of similar initial stroke severity, ICH subjects with the most severe strokes (ie, those in the lowest total admission FIM strata) had significantly greater recovery compared with those with cerebral infarction (see table 3, fig 1). No significant difference in recovery was present between ICH and cerebral infarction in subjects with less severe strokes on admission. To evaluate the potential confounding effect of LOS on recovery, the rate of recovery (measured by FIM efficiency) was compared among subjects with cerebral infarction and ICH, stratified by initial stroke functional severity. No significant difference in FIM efficiency was present between ICH and cerebral infarction subgroups across all levels of stroke severity. Table 3: Mean Change in FIM Scores by Stroke Subtype, Stratified by Stroke Functional Severity, Measured by Total Admission FIM Stroke Category Total Admission FIM Score 40 41 60 61 80 80 Cerebral infarct 19.7 26.1 27 20 (n 226) (n 219) (n 234) (n 192) ICH 27.8 32.4 29.4 18.8 (n 78) (n 50) (n 37) (n 28) P.004.037.310.500 Fig 1. FIM score change during rehabilitation therapy for ICH and cerebral infarction (CI), stratified by stroke functional severity, measured by total admission FIM. * P<.05; P<.0125.

RECOVERY AFTER HEMORRHAGIC AND ISCHEMIC STROKE, Kelly 971 DISCUSSION The present findings support and extend previous studies on functional outcome after ischemic and hemorrhagic stroke. 5,17,22-24 The present study also provides new information on the degree of recovery in survivors of ICH, particularly those most severely affected by their stroke. Overall, when compared with patients with cerebral infarction, patients with ICH had more functional impairment at admission for inpatient rehabilitation therapy, made greater gains during rehabilitation, and their impairment at the time of discharge was similar to that in the cerebral infarction group. When stratified by initial severity of disability, ICH patients with the most severe strokes had greater functional improvement than cerebral infarction patients of similar severity. Perhaps surprisingly, we found that the initial severity of disability did not significantly predict the amount of recovery during rehabilitation therapy. However, initial severity was a strong independent predictor of functional status at discharge. Consistent with other studies, 5,17,22 we found that younger age and longer LOS were significant independent predictors of better functional outcome. Approximately four fifths of our ICH and two thirds of our cerebral infarction patients had FIM scores corresponding to a requirement for more than minimal assistance at admission. Although mean total discharge FIM scores did not differ between groups, 50% of the patients with cerebral infarction were likely to have attained FIM scores consistent with independence at discharge, while only 42% of those with ICH attained that functional level. The patients with ICH had more severe cognitive impairment on admission (measured by admission FIM cognitive), but made significantly greater cognitive gains during rehabilitation compared with the cerebral infarction group. Furthermore, cognitive status on admission was a strong independent predictor of overall functional outcome in the cerebral infarction but not in the ICH subgroup. The importance of the role of higher cognitive function in physical and functional recovery during stroke rehabilitation has been described by other researchers. 25-27 However, to our knowledge, no one has reported the impact that stroke subtype specific differences in cognition have on outcome. The underlying mechanisms for these observations are not clear. One possibility is that gradual resolution of edema surrounding subcortical hemorrhage in the postacute period may be associated with better global cognitive improvement in patients with ICH, compared with infarcts which more commonly involve large areas of cortex and are likely to improve more slowly. Further study is needed to clarify this issue. We found that stroke severity of disability was a major determinant of functional status at discharge from rehabilitation therapy a finding that is in agreement with previous studies. 5,13,17,22,24 However, for strokes of similar severity, we found significant differences in the degree of recovery between cerebral infarction and ICH. Specifically, compared with the patients who were most severely affected by cerebral infarction (ie, lowest total admission FIM stratum), the patients with equally severe ICH made greater gains during rehabilitation therapy (see table 3). Previous studies 12,22,28 have suggested that patients may be triaged for rehabilitation based on their level of functional impairment measured in the acute hospital at 5 to 7 days after stroke onset. In general, according to that model, the most impaired (FIM score 40) would either be most efficiently treated with basic rehabilitation at a skilled nursing facility or intensive therapy at a rehabilitation hospital, the choice of which would be largely determined by the patient s age. In contrast, our findings indicate that a subgroup of patients with the most severe strokes due to supratentorial ICH make greater functional gains than those with cerebral infarction, suggesting that these patients should be targeted for early intensive rehabilitation. Further research is required to define the factors that determine favorable outcome in severely impaired patients with ICH. The potential confounding effect of the duration of inpatient rehabilitation therapy, measured by LOS, is difficult to evaluate. In the present study, no difference existed in the rate of recovery (measured by FIM efficiency) between patients with ICH and those with cerebral infarction, but the patients with ICH had longer LOS. In the Stroke Rehabilitation Program from which the study sample was derived, the inpatient rehabilitation course is continued until functional gains are no longer made and a plateau in recovery has been attained. We believe that the longer LOS in the ICH subgroup is likely a consequence of a more prolonged period of continuous recovery observed by the treating specialists. However, we cannot exclude the possibility that the longer LOS in ICH patients was related to some other unidentified factor, thus exposing this subgroup to a more prolonged course of inpatient rehabilitation than the subgroup with cerebral infarction. Our findings are in agreement with those of Ween et al, 17 who reported greater FIM change after rehabilitation in a relatively small sample of severely impaired patients with ICH compared with patients with large-vessel cerebral infarction. However, their results differ from those of other communityand hospital-based studies, 5,6,23 which examined the interaction of stroke severity and stroke pathophysiologic mechanism on stroke outcome after rehabilitation. The reasons for this difference are unclear at this time. One explanation may relate to a lack of power in previous studies to detect differences in outcome between ICH and cerebral infarction, due to the relatively small number of ICH patients in cells after data were stratified for stroke severity. 5,29 An alternative explanation may be related to differences in the study samples. We report outcome in a select sample, who were considered capable of participating in a rehabilitation therapy program at the time of acute hospital referral. In the Copenhagen Stroke Study, Jorgensen et al 29 reported no difference in outcome in unselected cases admitted with ICH and cerebral infarction, after controlling for stroke severity. A greater proportion of patients with most severe strokes in Jorgensen s sample had strokes that were due to ICH. Compared with our sample, it is likely that more of these patients were too impaired to participate in rehabilitation efforts, which may partly account for the differences in the findings between the 2 studies. 5,29 Our study has some limitations, which must be considered when interpreting these results. Our findings may not be generalized to all patients with stroke due to cerebral infarction and ICH, but are representative of recovery only in patients who undergo inpatient rehabilitation therapy. Furthermore, we relied on referral information from the acute hospital for details of the stroke mechanism (ICH, cerebral infarction), side, and location (supratentorial, infratentorial), raising the possibility of misclassification in some cases. However, all cases were confirmed by neuroimaging, so we consider it unlikely that significant misclassification of stroke mechanism or location occurred. Despite these considerations, the present analysis has several strengths. It has the advantage of relatively large sample sizes, which provided adequate statistical power to detect clinically relevant and statistically robust differences in comparisons of subgroups stratified by stroke functional severity. Recovery described in terms of functional outcome may be more relevant to the patient than measures of neurologic deficit after stroke.

972 RECOVERY AFTER HEMORRHAGIC AND ISCHEMIC STROKE, Kelly While impairment correlates with disability after stroke, functional improvement during rehabilitation occurs even in the absence of reductions in neurologic impairment. 30 Because the FIM is more highly responsive to change over time, its use as the measurement instrument (rather than the modified Rankin score) allowed quantification of recovery with greater accuracy. CONCLUSION These results provide new information concerning the interaction between severity of functional impairment and stroke mechanism in determining recovery after stroke. Our findings may assist efforts to rationally triage specific subgroups of patients for different levels of rehabilitation therapy after acute hospital care. References 1. Wolf PA, D Agostino RB. Epidemiology of stroke. In: Barnett HJ, Mohr JP, Stein BM, Yatsu FM, editors. Stroke. Pathophysiology, diagnosis and management. 3rd ed. New York: Churchill Livingstone; 1998. p 3-28. 2. American Heart Association. Stroke statistics. Dallas: AHA; 2000. 3. Kase CS, Mohr JP, Caplan LR. Intracerebral hemorrhage. In: Barnett HJ, Mohr JP, Stein BM, Yatsu FM, editors. Stroke. Pathophysiology, diagnosis and management, 3rd ed. New York: Churchill Livingstone; 1998. p 649-700. 4. Helweg-Larsen S, Sommer W, Strange P, Lester J, Boyson G. Prognosis for patients treated conservatively for spontaneous intracerebral hematomas. Stroke 1984;15:1045-8. 5. Jorgensen HS, Nakayama H, Raaschou HO, Olsen TS. Intracerebral hemorrhage versus infarction: stroke severity, risk factors and prognosis. Ann Neurol 1995;38:45-50. 6. Franke CL, van Swieten JC, Algra A, van Gjin J. Prognostic factors in patients with intracerebral haematoma. J Neurol Neurosurg Psychiatry 1992;55:653-7. 7. Hankey GJ, Jamrozik K, Broadhurst RJ, et al. Five-year survival after first-ever stroke and related prognostic factors in the Perth Community Stroke Study. Stroke 2000;31:2080-6. 8. Chambers BR, Norris JW, Shurvell BL, Hachinski VC. Prognosis of acute stroke. Neurology 1987;37:221-5. 9. Abu-Zeid HA, Choi NW, Hsu PH, Maini KK. Prognostic factors in the survival of 1,484 stroke cases observed for 30 to 48 months. Diagnostic types and descriptive variables. Arch Neurol 1978;35: 121-5. 10. Westling B, Norrving B, Thorngren M. Survival following stroke. A prospective population-based study of 438 hospitalized cases with prediction according to subtype, severity and age. Acta Neurol Scand 1990;81:457-63. 11. Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project 1981-86. Incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 1990;53:16-22. 12. Anderson CS, Jamrozik KD, Broadhurst RJ, Stewart-Wynne EG. Predicting survival for one year among different subtypes of stroke. Results from the Perth Community Stroke Study. Stroke 1994;25:1935-44. 13. Jorgensen HS, Nakayama H, Raaschou HO, Vive-Larsen J, Stoier M, Olsen TS. Outcome and time course of recovery in stroke. Part I: Outcome. The Copenhagen Stroke Study. Arch Phys Med Rehabil 1995;76:399-405. 14. Hier DB, Kavis KR, Richardson EP, Mohr JP. Hypertensive putaminal hemorrhage. Ann Neurol 1977;1:152-9. 15. Kase CS, Williams JP, Wyatt DA, Mohr JP. Lobar intracerebral hematomas: clinical and CT analysis of 22 cases. Neurology 1982;32:1146-50. 16. Broderick J, Brott TG, Duldner JE, Tomsick T, Huster G. Volume of intracerebral hemorrhage: a powerful and easy-to-use predictor of 30-day mortality. Stroke 1993;24:987-93. 17. Ween JE, Alexander MP, D Esposito M, Roberts M. Factors predictive of stroke outcome in a rehabilitation setting. Neurology 1996;47:388-92. 18. Keith RA, Granger CV, Hamilton BB, Sherwin FS. The Functional Independence Measure: a new tool for rehabilitation, In: Eisenberg MG, Grzesiak RC, editors. Advances in clinical rehabilitation. New York: Springer; 1987. p 6-18. 19. Kidd D, Stewart G, Baldry J, et al. The Functional Independence Measure: a comparative validity and reliability study. Disabil Rehabil 1995;17:10-4. 20. Hamilton BB, Laughlin JA, Granger CV, Kayton RM. Interrater agreement of the seven level Functional Independence Measure (FIM) [abstract]. Arch Phys Med Rehabil 1991;71:790. 21. Pallicino P, Snyder W, Granger C. The NIH Stroke Scale and the FIM in stroke rehabilitation [letter]. Stroke 1992;23:919. 22. Alexander MP. Stroke rehabilitation outcome. A potential use of predictive variables to establish levels of care. Stroke 1994;25:28-134. 23. Chae J, Zorowitz RD, Johnston MV. Functional outcome of hemorrhagic and nonhemorrhagic stroke patients after inpatient rehabilitation. A matched comparison. Am J Phys Med Rehabil 1996;75:177-82. 24. Kelly PJ, Stein J, Shafqat S, et al. Functional recovery following rehabilitation for cerebellar stroke. Stroke 2001;32:530-4. 25. Galski T, Bruno RL, Zorowitz R, Walker J. Predicting length of stay, functional outcome, and aftercare in the rehabilitation of stroke patients. The dominant role of higher-order cognition. Stroke 1993;24:1794-800. 26. Mysiw WJ, Beegan JG, Gatens PF. Prospective cognitive assessment of stroke patients before inpatient rehabilitation. The relationship of the Neurobehavioral Cognitive Status Examination to functional improvement. Am J Phys Med Rehabil 1989;68:168-71. 27. Hyman MD. Social psychological determinants of patients performance in stroke rehabilitation. Arch Phys Med Rehabil 1972; 53:217-26. 28. Asberg KH, Nydevik I. Early prognosis of stroke outcome by means of Katz index of activities of daily living. Scand J Rehabil Med 1991;23:187-91. 29. Jorgensen HS, Reith J, Nakayama H, Kammersgaard LP, Raaschou HO, Olsen TS. What determines good recovery in patients with the most severe strokes? The Copenhagen Stroke Study. Stroke 1999;30:2008-12. 30. Roth EJ, Heinemann AW, Lovell LL, Harvey RL, McGuire JR, Diaz S. Impairment and disability: their relation during stroke rehabilitation. Arch Phys Med Rehabil 1998;79:329-35.