Clinical Management Provided by Board-Certificated Physiatrists in Early Rehabilitation Is a Significant Determinant of Functional Improvement in Acute Stroke Patients: A Retrospective Analysis of Japan Rehabilitation Database Shoji Kinoshita, MD, Wataru Kakuda, MD, PhD, Ryo Momosaki, MD, PhD, Naoki Yamada, MD, Hidekazu Sugawara, MD, PhD, Shu Watanabe, MD, PhD, and Masahiro Abo, MD, PhD
Background: Early rehabilitation for acute stroke patients is widely recommended. We tested the hypothesis that clinical outcome of stroke patients who receive early rehabilitation managed by board-certificated physiatrists (BCP) is generally better than that provided by other medical specialties. Methods: Data of stroke patients who underwent early rehabilitation in 19 acute hospitals between January 2005 and December 2013 were collected from the Japan Rehabilitation Database and analyzed retrospectively. Multivariate linear regression analysis using generalized estimating equations method was performed to assess the association between Functional Independence Measure (FIM) effectiveness and management provided by BCP in early rehabilitation. In addition, multivariate logistic regression analysis was also performed to assess the impact of management provided by BCP in acute phase on discharge destination. Results: After setting the inclusion criteria, data of 3838 stroke patients were eligible for analysis. BCP provided early rehabilitation in 814 patients (21.2%). Both the duration of daily exercise time and the frequency of regular conferencing were significantly higher for patients managed by BCP than by other specialties. Although the mortality rate was not different, multivariate regression analysis showed that FIM effectiveness correlated significantly and positively with the management provided by BCP (coefficient, .35; 95% confidence interval [CI], .012-.059; P , .005). In addition, multivariate logistic analysis identified clinical management by BCP as a significant determinant of home discharge (odds ratio, 1.24; 95% CI, 1.08-1.44; P , .005). Conclusions: Our retrospective cohort study demonstrated that clinical management provided by BCP in early rehabilitation can lead to functional recovery of acute stroke. Key Words: Board-certificated physiatrist—acute stroke—early rehabilitation—Functional Independence Measure—home discharge. Ó 2015 by National Stroke Association
From the Department of Rehabilitation Medicine, The Jikei University School of Medicine, Tokyo, Japan. Received November 17, 2014; revision received December 20, 2014; accepted December 22, 2014. Address correspondence to Masahiro Abo, MD, PhD, Department of Rehabilitation Medicine, The Jikei University School of Medicine, 3-25-8, Nishi-Shimbashi, Minato-Ku, Tokyo 105-8461, Japan. E-mail: [email protected]. 1052-3057/$ - see front matter Ó 2015 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2014.12.026
The clinical importance of early rehabilitation for acute stroke has been recently emphasized in the field of Physical Medicine and Rehabilitation (PM&R) in addition to other medical fields. Several clinical studies have already confirmed the beneficial effects of early stroke rehabilitation.1,2 In addition, the guidelines for acute ischemic stroke published by the American Heart Association and the American Stroke Association also highlight the importance of early stroke rehabilitation.3
Journal of Stroke and Cerebrovascular Diseases, Vol. -, No. - (---), 2015: pp 1-6
1
S. KINOSHITA ET AL.
2
Provision of appropriate early rehabilitation to acute stroke patients requires active participation of physiatrists. Physiatrists are usually involved in the clinical management of multidisciplinary rehabilitation team that consists of nurses, physical therapists, occupational therapists, speech therapists, and medical social workers.4 In some countries, a legal body of the PM&R oversees the certification of physiatrists after an appropriate residency training program.5,6 Therefore, it is desirable to have board-certificated physiatrists (BCP) with sufficient knowledge and experience about stroke rehabilitation as the primary care providers in the early rehabilitation of patients after stroke. In Japan, however, early stroke rehabilitation at acute hospitals is not always provided by BCP. At some hospitals, physicians with other specialties head the team that provides early stroke rehabilitation. This is in part due to the shortage in the number of certified BCP. So far, there are no clinical data regarding the impact of BCP involvement in early rehabilitation management for acute stroke patients in Japan. We hypothesized that the involvement of BCP in early poststroke rehabilitation correlates with good functional recovery compared with management provided by other specialties. To test the hypothesis, we performed retrospective cohort study to clarify the impact of participation of BCP in early stroke management on functional recovery after acute stroke using the Japan Rehabilitation Database containing the clinical data of a large number of acute hospitals throughout Japan.
Subjects and Methods Data Source The Japan Rehabilitation Database was established with financial support from the Ministry of Health, Labor, and Welfare of Japan in 2012.7 The database contains detailed clinical data collected on patients who were discharged from the participating hospitals during the period between January 2005 and December 2013. The database is divided into different sections based on the diagnosis, such as stroke, femoral neck fracture, spinal cord injury, and other disorders. The database for stroke patients includes mainly the identifiers of the following patient characteristics: age, sex, Functional Independence Measure (FIM, scores range from 18 for totally dependent to 126 for totally independent),8 length of stay at the acute hospital, days from stroke onset, type of stroke, discharge destination, amount of rehabilitative exercise per day, 5-grade modified Rankin Scale (mRS level 1, no significant disability; level 5, severe disability),9 National Institutes of Health Stroke Scale (NIHSS) score, self-exercise, regular conferencing, and involvement of a BCP as the responsible physician. The medical staff (physicians, therapists, and nurses) at each participating hospital recorded the data and submitted them electronically through the
Internet to the office of Japan Association of Rehabilitation Database. Because of the anonymous nature of the data, informed consent was waived.
Subjects Among the acute hospitals that provided clinical data for the database, we selected 19 acute hospitals where both BCP and other physicians (non-BCP) provided clinical management for acute stroke patients. For this study, the clinical data of patients who were admitted to these 19 hospitals with a diagnosis of stroke during the period between January 2005 and December 2013 were collected from the Japan Rehabilitation Database. We only included patients for whom the following data were recorded: diagnosis of cerebral infarction or cerebral hemorrhage; admission within 3 days of onset; length of stay of less than 6 months; available information on attending physician (BCP or non-BCP); and FIM score at admission/ discharge at acute hospitals. Because the distinction of functional recovery between subarachnoid hemorrhage and other subtypes of stroke has been previously reported,10 we excluded patients with the diagnosis of subarachnoid hemorrhage from this study.
Board-Certificated Physiatrist The PM&R Associations in some countries define the requirement for board certification, which includes residency trainings for 2-6 years followed by examination.5,6,11 The requirements for BCP, which were set up by Japanese Association of Rehabilitation Medicine (JARM) include12,13 a 3-year residency program that covers the entire field of PM&R (eg, brain injury, spinal cord injury, cerebral palsy, musculoskeletal diseases, neurologic diseases, and amputation) at the institutions that were certificated by JARM, followed by submission of 2 abstracts at scientific PM&R meetings, 30 case reports, and a list of 100 cases, before receiving the final written and oral examinations. Currently, 80% of the trainees pass the examinations and are certified every year. In addition, JARM requires renewing of certification every 5 years after the last registration. In 2014, there are 1959 certified BCP in Japan. Because Japanese physicians are allowed to acquire more than 1 board certification, almost 70% of BCP hold board certification in another specialty.
Outcome Measures For this study, 2 widely accepted measures were applied, FIM effectiveness and discharge destination. Evaluation of FIM was performed and recorded at admission to and discharge from the acute hospitals. The outcome measure of this study, FIM effectiveness, was calculated using the following equation: ([discharge FIM – admission FIM]/[maximum FIM – admission FIM]).14 In addition, data regarding discharge destination
ROLE OF BOARD-CERTIFICATED PHYSIATRISTS IN EARLY STROKE REHABILITATION
3
Table 1. Patients’ characteristics Characteristics Gender Female Male Age at admission (y) Time to admission after onset (d) Daily exercise time (min/d) Duration of hospitalization (d) NIHSS at admission FIM at admission Prestroke disability (mRS, 2-5) Death Cerebral infarction Comorbidities Hypertension Diabetes Atrial fibrillation Past history of stroke
Total (n 5 3838)
BCP (n 5 814)
Non-BCP (n 5 3024)
P
1670 (43.5) 2168 (56.5) 74.4 6 12.6 1.23 6 .52 76.6 6 49.6 29.5 6 21.3 8.12 6 8.25 52.5 6 31.4 1163 (30.3) 25 (.7) 2578 (67.2)
344 (42.3) 470 (57.7) 73.5 6 12.5 1.18 6 .46 123.4 6 42.3 27.9 6 16.5 7.56 6 7.33 48.7 6 27.3 250 (30.7) 4 (.5) 555 (66.9)
1326 (43.9) 1698 (56.1) 74.6 6 12.5 1.24 6 .54 64.1 6 43.6 30.0 6 22.4 8.29 6 8.50 53.2 6 32.3 913 (30.2) 21 (.7) 2023 (66.9)
.41 ,.05 ,.005 ,.001 ,.005 ,.05 ,.001 .77 .52 .49
2412 (62.8) 771 (20.1) 682 (17.8) 106 (2.8)
521 (64.0) 160 (19.7) 154 (18.9) 23 (2.8)
1891 (62.5) 611 (20.2) 528 (17.5) 83 (2.7)
.44 .73 .33 .90
Abbreviations: BCP, board-certificated physiatrist; FIM, Functional Independence Measure; mRS, modified Rankin Scale score; NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation. Values are expressed as mean 6 SD or number of patients (%).
(discharge to own home or other institutions) were collected.
Statistical Analysis The demographic and clinical characteristics of the patients were compared between patients managed by BCP and those managed by other specialists by the chi-square test for categorical variables and by unpaired t test for continuous variables. Multivariate linear regression analysis for FIM effectiveness as the dependent variables was performed to calculate the correlation coefficients of the following independent variables: responsible physician (BCP or nonBCP), sex, age, preonset disability (mRS of $2), subtype of stroke, time to admission, daily exercise time, length of stay, NIHSS score at admission, hypertension, diabetes, atrial fibrillation, and history of stroke. Multivariate linear regression analysis was also performed for both cognitive and motor FIM effectiveness. Furthermore, 2 subanalyses were performed after dividing the patients into 2 groups based on the calculated mean of age at admission and FIM score at admission. Multivariate logistic regression analysis for the discharge to patients’ own home as the dependent variable was also performed to calculate the odds ratios of the following independent variables: same clinical characteristics applied into multivariate linear regression analysis for FIM effectiveness and at least 1 caregiver because the presence of caregiver was previously found to correlate with discharge distinction.15 In this study, we used generalized estimation equations to account for clustering
of observations within hospitals and provide more accurate confidence intervals (CIs). Statistical analysis was performed using the SPSS 19.0 (IMB SPSS Inc., Armonk, NY). The level of significance was set at P less than .05.
Results In total, 6332 stroke patients were admitted to 19 participating acute hospitals in Japan during the study period. Of these, 5666 patients were admitted within 3 days of stroke onset and discharged within 180 days from their admission. Among them, information about responsible physicians (BCP or non-BCP) and FIM at admission/discharge was available in 3838 patients who were considered eligible for the present study. Table 1 shows the clinical characteristics of the study patients. BCP cared for 814 patients (21.2%). The age, time to admission, length of stays, and NIHSS score at admission were significantly lower for patients managed by the BCP compared with those managed by non-BCP. The FIM at admission and duration of daily exercise time were significantly higher in patients managed by the BCP than those managed by non-BCP. The rate of self-exercise, regular conferencing, and social worker intervention was also significantly higher in the patients managed by BCP than in those managed by non-BCP. There was no significant difference in the mortality rate between the 2 groups. Multivariate linear regression analysis identified the involvement of BCP as the clinical provider as significant factor in FIM effectiveness (coefficient, .35; 95% CI, .012-.059; P 5 .003; Table 2). In addition, BCP involvement
S. KINOSHITA ET AL.
4
Table 2. Multivariate linear regression analysis for FIM effectiveness Variable
Coefficients (b)
95% CI
P value
Board-certificated physiatrist Age at admission (y) Female gender NIHSS at admission Time to admission after onset (d) Daily exercise time (min/d) Duration of hospitalization (d) Prestroke disability (mRS, 2-5) Cerebral infarction Hypertension Diabetes Atrial fibrillation Past history of stroke
.035 2.008 2.017 2.015 2.007 .000 2.002 2.181 .060 2.009 2.006 2.021 2.089
.012-.059 2.010 to 2.006 2.025 to 2.010 2.017 to 2.014 2.024 to .010 .000-.000 2.004 to 2.001 2.208 to 2.155 .034-.085 2.028 to .011 2.032 to .020 2.052 to .010 2.144 to 2.034
,.05 ,.001 ,.001 ,.001 .44 .87 ,.001 ,.001 ,.001 .38 .66 .19 ,.005
Abbreviations: CI, confidence interval; FIM, Functional Independence Measure; mRS, modified Rankin Scale score; NIHSS National Institutes of Health Stroke Scale.
was a significant positive factor for motor FIM effectiveness (coefficient, .037; 95% CI, .014-.060; P 5 .002), although the involvement did not significantly affect cognitive FIM effectiveness (coefficient, .019; 95% CI, .032-.070; P 5 .47). We divided patients into 2 groups based on the calculated mean of age at admission (mean, 74.4 6 12.6 years, ,75 and $75 years of age) and FIM score at admission (mean: 52.5 6 31.4, ,52 and $53 points). The results of subgroup analysis showed that the involvement of BCP was a significant factor for FIM effectiveness in the younger patient group (age ,75 years) and in patients with mild-to-moderate disability (admission FIM $53, Table 3). Multivariate logistic analysis identified clinical care by BCP as a significant correlate with home discharge (odds ratio, 1.24; 95% CI, 1.08-1.44; P 5 .003; Table 4).
Discussion The present study showed that the involvement of BCP in early rehabilitation management of poststroke patients was associated with good functional recovery. FurtherTable 3. Subgroup analysis of board-certified physiatrist for FIM effectiveness Subgroup
Coefficient (b)
Age at admission, y $75 .015 ,75 .073 FIM at admission $53 .072 ,52 .021
95% CI
P value
.001-.028 .036-.110
,.05 ,.001
.034-.109 2.003 to .042
,.001 .090
Abbreviations: CI, confidence interval; FIM, Functional Independence Measure.
more, acute stroke patients managed by BCP were likely to be discharged home from acute hospitals. Because randomized controlled trials to identify the impact of management by BCP are not feasible in real clinical setting, we performed retrospective cohort analysis based on a large database for this purpose. The database used in the present study, the Japan Rehabilitation Database, included data of a large number of patients under multicenter registry of several hospitals across Japan. Therefore, the strong statistical power of this study allows us
Table 4. Multivariate logistic regression analysis for discharge to patients’ own home Variable
OR
95% CI
P value
Board-certificated physiatrist Age at admission (y) Female gender NIHSS at admission Time to admission after onset (d) Daily exercise time (min/d) Duration of hospitalization (d) Cerebral infarction Prestroke mRS (score, 2-5) Hypertension Diabetes Atrial fibrillation Past history of stroke At least 1 caregiver
1.24
1.08-1.44
,.005
.97 .96 .86 .90
.97-.98 .77-1.18 .84-.87 .83-.96
,.001 .68 ,.001 ,.005
.99 .96
.99-1.00 .94-.98
.11 ,.001
2.00 .57 .98 .93 .93 .95 2.31
1.75-2.24 .49-.67 .83-1.17 .81-1.07 .77-1.13 .55-1.65 1.80-2.96
,.001 ,.001 .84 .29 .46 .85 ,.001
Abbreviations: CI, confidence interval; mRS modified Rankin Scale score; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio.
ROLE OF BOARD-CERTIFICATED PHYSIATRISTS IN EARLY STROKE REHABILITATION
to draw definite conclusions. In addition, generalized estimating equations were used to account for the potential correlation of outcome within hospitals in this study. In other medical fields, previous clinical studies identified the important role of services provided by boardcertificated physicians in facilitating the introduction of appropriate and aggressive therapeutic interventions, which often lead to better functional recovery.16-18 For example, clinical care provided by midcareer anesthesiologists with board certification was associated with less mortality within 30 days of admission than that provided by physicians without such certification. Furthermore, the mortality rate of critically ill patients admitted to the intensive care units and managed clinically by board-certified intensivists was lower than that of patients managed by other specialties.19 In the field of PM&R, Jeong et al7 published a study that described the positive impact of clinical management by BCP. The group investigated the extent of BCP involvement in 12 rehabilitation hospitals and showed that clinical outcome after the completion of rehabilitative program was better in patients admitted to hospitals with sufficient number of BCP than in those admitted to hospitals without BCP. The study was based on data from rehabilitation hospitals and involved patients with subacute-to-chronic stroke. To our knowledge, however, there are no studies that examined the relation between the involvement of BCP in early rehabilitative management and functional recovery of acute stroke. Thus, our study is the first clinical study that examined the direct relationship between BCP clinical management in early rehabilitation and good functional recovery of patients with acute stroke. What is the main reason for the significant association between BCP involvement and good functional recovery? Although our study design does not provide a definite answer to this question, several reasons can be speculated. Comparison of clinical data of patients managed by BCP and other specialties (Table 1) indicated that the duration of daily exercise was longer and frequency of regular meeting was significantly higher for patients managed by BCP than those managed by non-BCP. We speculate that this is because of the confidence of BCP in the safety of the prescribed rehabilitative training program even in patients with acute stroke compared with the hesitation of non-BCP in initiating early poststroke rehabilitation for fear of complication development. Furthermore, the better results with BCP could be because of the ability of these specialties to coordinate a multidisciplinary rehabilitative team through regular meeting. In this regard, a meta-analysis by Lohse et al20 indicated that high-dose rehabilitative training during the acute phase had a positive impact on clinical outcome of stroke patients. Although the high-dose exercise was not associated with the functional recovery in the current multivariable linear analysis (Table 2), the correlation between the
5
duration of exercise time and BCP who had tendency to prescribe high-dose training was thought to cause the impact to disappear. Thus, it seems that the ability of BCP to introduce early stroke rehabilitation and regular conferencing contributed to the achievement of good functional recovery. The improvement of FIM effectiveness in patients managed by BCP would contribute to the high likelihood of home discharge because our analysis was made after adjustment for FIM at admission, existence of caregiver, and diagnosis of stroke subtype. This finding indicates that BCPs were superior in terms of management of home discharge program with multidisciplinary coordination. Two subanalyses in this study found positive impact for BCP involvement in young patients and those with higher score of FIM at admission. These findings indicate that BCP involvement in early rehabilitation does not seem to be beneficial to clinical outcome in older patients and severely disabled acute patients. We consider that the lack of significant association between BCP management and functional improvement in the latter 2 groups is because of the small window available in training of these patients. In this regard, Denti et al21 reported that elderly stroke patients are less responsive to rehabilitative program compared with younger patients. Furthermore, Inouye et al22 reported markedly poor clinical outcome of acute stroke patients with low FIM score at admission despite the introduction of early stroke rehabilitation. Interestingly, the recovery of cognitive function was not associated with BCP management. What is the reason for this finding? In Japan, cognitive rehabilitation is seldom initiated in the acute phase of stroke and significant improvement of cognitive function is usually found in the subacute chronic phase of stroke. In addition to BCP involvement, the results of multivariate linear regression analysis and logistic regression analysis identified other clinical factors, such as female gender, NIHSS at admission, duration of hospitalization, prestroke mRS, and subtype of stroke, as significant determinants of functional recovery after rehabilitation program for acute stroke. This finding was, however, not novel as the prognostic impact of the same factors has already been identified.23-25 It is noteworthy that various medical and surgical interventions cannot have an impact on most of these factors. Thus, based on the finding of the important role of physiatrists, we recommend further development of the training system for physiatrist-in-training, with special emphasis on education and clinical experience on stroke rehabilitation. Currently, the number of BCP practicing in clinical settings differs markedly among countries. Although some developed countries in Europe have more than 1 BCP per 100,000 populations, most Asian developing countries have less than 1 BCP per 100,000 populations.4,5 Thus, there is a need to increase the number of BCP not only in Japan but also in other countries, together with
S. KINOSHITA ET AL.
6
the introduction of more sophisticated education system for better patient care. This study has certain limitations. First, because the database used was constructed without random sampling, we cannot generalize the study results. Second, information about acute medical and surgical treatment, such as intravenous administration of tissue plasminogen activator in the database, was incomplete to allow its inclusion in the multivariate analysis. Therefore, we were not able to assess the effect of acute treatment or the association between acute treatment and the involvement of BCP in the management of patients with acute stroke. Third, because no data were available after hospital discharge, the impact of BCP care in early phase on long-term outcome of stroke remains unknown.
Conclusions In this retrospective cohort study, analysis of the Japan Rehabilitation Database identified clinical management provision by BCP in the form of early rehabilitation for acute stroke patients as a significant predictor of good functional prognosis. Based on the positive results on the role played by BCP in the clinical management of such patients and the small number of BCP per capita in Japan, we recommend increasing the number of physiatrist trainees in Japan to provide better rehabilitation management for acute stroke patients. Acknowledgments: We would like to thank Japan Association of Rehabilitation Database for database provision. Japan Association of Rehabilitation Database did not provide the perspective in this study.
References 1. Bernhardt J, Dewey H, Thrift A, et al. A very early rehabilitation trial for stroke (AVERT): phase II safety and feasibility. Stroke 2008;39:390-396. 2. Bernhardt J, Thuy MN, Collier JM, et al. Very early versus delayed mobilisation after stroke. Cochrane Database Syst Rev 2009;21. 3. Jauch EC, Saver JL, Adams HP Jr, et al. Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2013;44:870-947. 4. Franceschini M, Caso V, Zampolini M, et al. The role of the physiatrist in stroke rehabilitation: a European survey. Am J Phys Med Rehabil 2009;88:596-600. 5. Ward AB. Rehabilitation medicine: the European perspective. Am J Phys Med Rehabil 2005;84:233-237. 6. Han TR, Bang MS. Rehabilitation medicine: the Asian perspective. Am J Phys Med Rehabil 2007;86:335-338. 7. Jeong S, Kondo K, Shiraishi N, et al. An evaluation of the quality of post-stroke rehabilitation in Japan. Clinical Audit 2010;2:59-66.
8. Chumney D, Nollinger K, Shesko K, et al. Ability of Functional Independence Measure to accurately predict functional outcome of stroke-specific population: systematic review. J Rehabil Res Dev 2010; 47:17-29. 9. Sulter G, Steen C, De Keyser J, et al. Use of Barthel index and modified Rankin scale in acute stroke trials. Stroke 1999;3:1538-1541. 10. Westerkam WR, Cifu DX, Keyser L. Functional outcome after inpatient rehabilitation following aneurysmal subarachnoid hemorrhage: a prospective analysis. Top Stroke Rehabil 1997;4:29-37. 11. Brandstater ME. International survey of training and certification in physical medicine and rehabilitation. Arch Phys Med Rehabil 2000;81:1234-1235. 12. Chino N. Certification and measuring competency in Japan, South Korea, and the Philippines. Arch Phys Med Rehabil 2000;81:1248-1249. 13. Izumi S. Postgraduate training and certification of physiatrists (rehabilitation physicians) in Japan. Jpn J Rehabil Med 2005;42:199-204. 14. Shah S, Vanclay F, Cooper B. Efficiency, effectiveness, and duration of stroke rehabilitation. Stroke 1990;21: 241-246. 15. Tanwir S, Montgomery K, Chari V, et al. Stroke rehabilitation: availability of a family member as caregiver and discharge destination. Eur J Phys Rehabil Med 2014; 50:355-362. 16. Sharp LK, Bashook PG, Lipsky MS, et al. Specialty board certification and clinical outcomes: the missing link. Acad Med 2002;77:534-542. 17. Prystowsky JB, Bordage G, Feinglass JM. Patient outcomes for segmental colon resection according to surgeon’s training, certification, and experience. Surgery 2002;132:663-722. 18. Silber JH, Kennedy SK, Even-Shoshan O, et al. Anesthesiologist board certification and patient outcomes. Anesthesiology 2002;96:1044-1052. 19. UP Iyegha, Asghar JI, Habermann EB, et al. Intensivists improve outcomes and compliance with process measures in critically ill patients. J Am Coll Surg 2013;216: 363-372. 20. Lohse KR, Lang CE, Boyd LA. Is more better? Using metadata to explore dose-response relationships in stroke rehabilitation. Stroke 2014;45:2053-2058. 21. Denti L, Agosti M, Franceschini M. Outcome predictors of rehabilitation for first stroke in the elderly. Eur J Phys Rehabil Med 2008;44:3-11. 22. Inouye M, Kishi K, Ikeda Y, et al. Prediction of functional outcome after stroke rehabilitation. Am J Phys Med Rehabil 2000;79:513-518. 23. Di Carlo A, Lamassa M, Baldereschi M, et al. Sex differences in the clinical presentation, resource use, and 3-month outcome of acute stroke in Europe: data from a multicenter multinational hospital-based registry. Stroke 2003;34:1114-1119. 24. Adams HP Jr, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Neurology 1999;53:126-131. 25. Ween JE, Alexander MP, D’Esposito M, et al. Factors predictive of stroke outcome in a rehabilitation setting. Neurology 1996;47:388-392.
Background: Early rehabilitation for acute stroke patients is widely recommended. We tested the hypothesis that clinical outcome of stroke patients who receive early rehabilitation managed by board-certificated physiatrists (BCP) is generally better than that provided by other medical specialties. Methods: Data of stroke patients who underwent early rehabilitation in 19 acute hospitals between January 2005 and December 2013 were collected from the Japan Rehabilitation Database and analyzed retrospectively. Multivariate linear regression analysis using generalized estimating equations method was performed to assess the association between Functional Independence Measure (FIM) effectiveness and management provided by BCP in early rehabilitation. In addition, multivariate logistic regression analysis was also performed to assess the impact of management provided by BCP in acute phase on discharge destination. Results: After setting the inclusion criteria, data of 3838 stroke patients were eligible for analysis. BCP provided early rehabilitation in 814 patients (21.2%). Both the duration of daily exercise time and the frequency of regular conferencing were significantly higher for patients managed by BCP than by other specialties. Although the mortality rate was not different, multivariate regression analysis showed that FIM effectiveness correlated significantly and positively with the management provided by BCP (coefficient, .35; 95% confidence interval [CI], .012-.059; P , .005). In addition, multivariate logistic analysis identified clinical management by BCP as a significant determinant of home discharge (odds ratio, 1.24; 95% CI, 1.08-1.44; P , .005). Conclusions: Our retrospective cohort study demonstrated that clinical management provided by BCP in early rehabilitation can lead to functional recovery of acute stroke. Key Words: Board-certificated physiatrist—acute stroke—early rehabilitation—Functional Independence Measure—home discharge. Ó 2015 by National Stroke Association
From the Department of Rehabilitation Medicine, The Jikei University School of Medicine, Tokyo, Japan. Received November 17, 2014; revision received December 20, 2014; accepted December 22, 2014. Address correspondence to Masahiro Abo, MD, PhD, Department of Rehabilitation Medicine, The Jikei University School of Medicine, 3-25-8, Nishi-Shimbashi, Minato-Ku, Tokyo 105-8461, Japan. E-mail: [email protected]. 1052-3057/$ - see front matter Ó 2015 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2014.12.026
The clinical importance of early rehabilitation for acute stroke has been recently emphasized in the field of Physical Medicine and Rehabilitation (PM&R) in addition to other medical fields. Several clinical studies have already confirmed the beneficial effects of early stroke rehabilitation.1,2 In addition, the guidelines for acute ischemic stroke published by the American Heart Association and the American Stroke Association also highlight the importance of early stroke rehabilitation.3
Journal of Stroke and Cerebrovascular Diseases, Vol. -, No. - (---), 2015: pp 1-6
1
S. KINOSHITA ET AL.
2
Provision of appropriate early rehabilitation to acute stroke patients requires active participation of physiatrists. Physiatrists are usually involved in the clinical management of multidisciplinary rehabilitation team that consists of nurses, physical therapists, occupational therapists, speech therapists, and medical social workers.4 In some countries, a legal body of the PM&R oversees the certification of physiatrists after an appropriate residency training program.5,6 Therefore, it is desirable to have board-certificated physiatrists (BCP) with sufficient knowledge and experience about stroke rehabilitation as the primary care providers in the early rehabilitation of patients after stroke. In Japan, however, early stroke rehabilitation at acute hospitals is not always provided by BCP. At some hospitals, physicians with other specialties head the team that provides early stroke rehabilitation. This is in part due to the shortage in the number of certified BCP. So far, there are no clinical data regarding the impact of BCP involvement in early rehabilitation management for acute stroke patients in Japan. We hypothesized that the involvement of BCP in early poststroke rehabilitation correlates with good functional recovery compared with management provided by other specialties. To test the hypothesis, we performed retrospective cohort study to clarify the impact of participation of BCP in early stroke management on functional recovery after acute stroke using the Japan Rehabilitation Database containing the clinical data of a large number of acute hospitals throughout Japan.
Subjects and Methods Data Source The Japan Rehabilitation Database was established with financial support from the Ministry of Health, Labor, and Welfare of Japan in 2012.7 The database contains detailed clinical data collected on patients who were discharged from the participating hospitals during the period between January 2005 and December 2013. The database is divided into different sections based on the diagnosis, such as stroke, femoral neck fracture, spinal cord injury, and other disorders. The database for stroke patients includes mainly the identifiers of the following patient characteristics: age, sex, Functional Independence Measure (FIM, scores range from 18 for totally dependent to 126 for totally independent),8 length of stay at the acute hospital, days from stroke onset, type of stroke, discharge destination, amount of rehabilitative exercise per day, 5-grade modified Rankin Scale (mRS level 1, no significant disability; level 5, severe disability),9 National Institutes of Health Stroke Scale (NIHSS) score, self-exercise, regular conferencing, and involvement of a BCP as the responsible physician. The medical staff (physicians, therapists, and nurses) at each participating hospital recorded the data and submitted them electronically through the
Internet to the office of Japan Association of Rehabilitation Database. Because of the anonymous nature of the data, informed consent was waived.
Subjects Among the acute hospitals that provided clinical data for the database, we selected 19 acute hospitals where both BCP and other physicians (non-BCP) provided clinical management for acute stroke patients. For this study, the clinical data of patients who were admitted to these 19 hospitals with a diagnosis of stroke during the period between January 2005 and December 2013 were collected from the Japan Rehabilitation Database. We only included patients for whom the following data were recorded: diagnosis of cerebral infarction or cerebral hemorrhage; admission within 3 days of onset; length of stay of less than 6 months; available information on attending physician (BCP or non-BCP); and FIM score at admission/ discharge at acute hospitals. Because the distinction of functional recovery between subarachnoid hemorrhage and other subtypes of stroke has been previously reported,10 we excluded patients with the diagnosis of subarachnoid hemorrhage from this study.
Board-Certificated Physiatrist The PM&R Associations in some countries define the requirement for board certification, which includes residency trainings for 2-6 years followed by examination.5,6,11 The requirements for BCP, which were set up by Japanese Association of Rehabilitation Medicine (JARM) include12,13 a 3-year residency program that covers the entire field of PM&R (eg, brain injury, spinal cord injury, cerebral palsy, musculoskeletal diseases, neurologic diseases, and amputation) at the institutions that were certificated by JARM, followed by submission of 2 abstracts at scientific PM&R meetings, 30 case reports, and a list of 100 cases, before receiving the final written and oral examinations. Currently, 80% of the trainees pass the examinations and are certified every year. In addition, JARM requires renewing of certification every 5 years after the last registration. In 2014, there are 1959 certified BCP in Japan. Because Japanese physicians are allowed to acquire more than 1 board certification, almost 70% of BCP hold board certification in another specialty.
Outcome Measures For this study, 2 widely accepted measures were applied, FIM effectiveness and discharge destination. Evaluation of FIM was performed and recorded at admission to and discharge from the acute hospitals. The outcome measure of this study, FIM effectiveness, was calculated using the following equation: ([discharge FIM – admission FIM]/[maximum FIM – admission FIM]).14 In addition, data regarding discharge destination
ROLE OF BOARD-CERTIFICATED PHYSIATRISTS IN EARLY STROKE REHABILITATION
3
Table 1. Patients’ characteristics Characteristics Gender Female Male Age at admission (y) Time to admission after onset (d) Daily exercise time (min/d) Duration of hospitalization (d) NIHSS at admission FIM at admission Prestroke disability (mRS, 2-5) Death Cerebral infarction Comorbidities Hypertension Diabetes Atrial fibrillation Past history of stroke
Total (n 5 3838)
BCP (n 5 814)
Non-BCP (n 5 3024)
P
1670 (43.5) 2168 (56.5) 74.4 6 12.6 1.23 6 .52 76.6 6 49.6 29.5 6 21.3 8.12 6 8.25 52.5 6 31.4 1163 (30.3) 25 (.7) 2578 (67.2)
344 (42.3) 470 (57.7) 73.5 6 12.5 1.18 6 .46 123.4 6 42.3 27.9 6 16.5 7.56 6 7.33 48.7 6 27.3 250 (30.7) 4 (.5) 555 (66.9)
1326 (43.9) 1698 (56.1) 74.6 6 12.5 1.24 6 .54 64.1 6 43.6 30.0 6 22.4 8.29 6 8.50 53.2 6 32.3 913 (30.2) 21 (.7) 2023 (66.9)
.41 ,.05 ,.005 ,.001 ,.005 ,.05 ,.001 .77 .52 .49
2412 (62.8) 771 (20.1) 682 (17.8) 106 (2.8)
521 (64.0) 160 (19.7) 154 (18.9) 23 (2.8)
1891 (62.5) 611 (20.2) 528 (17.5) 83 (2.7)
.44 .73 .33 .90
Abbreviations: BCP, board-certificated physiatrist; FIM, Functional Independence Measure; mRS, modified Rankin Scale score; NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation. Values are expressed as mean 6 SD or number of patients (%).
(discharge to own home or other institutions) were collected.
Statistical Analysis The demographic and clinical characteristics of the patients were compared between patients managed by BCP and those managed by other specialists by the chi-square test for categorical variables and by unpaired t test for continuous variables. Multivariate linear regression analysis for FIM effectiveness as the dependent variables was performed to calculate the correlation coefficients of the following independent variables: responsible physician (BCP or nonBCP), sex, age, preonset disability (mRS of $2), subtype of stroke, time to admission, daily exercise time, length of stay, NIHSS score at admission, hypertension, diabetes, atrial fibrillation, and history of stroke. Multivariate linear regression analysis was also performed for both cognitive and motor FIM effectiveness. Furthermore, 2 subanalyses were performed after dividing the patients into 2 groups based on the calculated mean of age at admission and FIM score at admission. Multivariate logistic regression analysis for the discharge to patients’ own home as the dependent variable was also performed to calculate the odds ratios of the following independent variables: same clinical characteristics applied into multivariate linear regression analysis for FIM effectiveness and at least 1 caregiver because the presence of caregiver was previously found to correlate with discharge distinction.15 In this study, we used generalized estimation equations to account for clustering
of observations within hospitals and provide more accurate confidence intervals (CIs). Statistical analysis was performed using the SPSS 19.0 (IMB SPSS Inc., Armonk, NY). The level of significance was set at P less than .05.
Results In total, 6332 stroke patients were admitted to 19 participating acute hospitals in Japan during the study period. Of these, 5666 patients were admitted within 3 days of stroke onset and discharged within 180 days from their admission. Among them, information about responsible physicians (BCP or non-BCP) and FIM at admission/discharge was available in 3838 patients who were considered eligible for the present study. Table 1 shows the clinical characteristics of the study patients. BCP cared for 814 patients (21.2%). The age, time to admission, length of stays, and NIHSS score at admission were significantly lower for patients managed by the BCP compared with those managed by non-BCP. The FIM at admission and duration of daily exercise time were significantly higher in patients managed by the BCP than those managed by non-BCP. The rate of self-exercise, regular conferencing, and social worker intervention was also significantly higher in the patients managed by BCP than in those managed by non-BCP. There was no significant difference in the mortality rate between the 2 groups. Multivariate linear regression analysis identified the involvement of BCP as the clinical provider as significant factor in FIM effectiveness (coefficient, .35; 95% CI, .012-.059; P 5 .003; Table 2). In addition, BCP involvement
S. KINOSHITA ET AL.
4
Table 2. Multivariate linear regression analysis for FIM effectiveness Variable
Coefficients (b)
95% CI
P value
Board-certificated physiatrist Age at admission (y) Female gender NIHSS at admission Time to admission after onset (d) Daily exercise time (min/d) Duration of hospitalization (d) Prestroke disability (mRS, 2-5) Cerebral infarction Hypertension Diabetes Atrial fibrillation Past history of stroke
.035 2.008 2.017 2.015 2.007 .000 2.002 2.181 .060 2.009 2.006 2.021 2.089
.012-.059 2.010 to 2.006 2.025 to 2.010 2.017 to 2.014 2.024 to .010 .000-.000 2.004 to 2.001 2.208 to 2.155 .034-.085 2.028 to .011 2.032 to .020 2.052 to .010 2.144 to 2.034
,.05 ,.001 ,.001 ,.001 .44 .87 ,.001 ,.001 ,.001 .38 .66 .19 ,.005
Abbreviations: CI, confidence interval; FIM, Functional Independence Measure; mRS, modified Rankin Scale score; NIHSS National Institutes of Health Stroke Scale.
was a significant positive factor for motor FIM effectiveness (coefficient, .037; 95% CI, .014-.060; P 5 .002), although the involvement did not significantly affect cognitive FIM effectiveness (coefficient, .019; 95% CI, .032-.070; P 5 .47). We divided patients into 2 groups based on the calculated mean of age at admission (mean, 74.4 6 12.6 years, ,75 and $75 years of age) and FIM score at admission (mean: 52.5 6 31.4, ,52 and $53 points). The results of subgroup analysis showed that the involvement of BCP was a significant factor for FIM effectiveness in the younger patient group (age ,75 years) and in patients with mild-to-moderate disability (admission FIM $53, Table 3). Multivariate logistic analysis identified clinical care by BCP as a significant correlate with home discharge (odds ratio, 1.24; 95% CI, 1.08-1.44; P 5 .003; Table 4).
Discussion The present study showed that the involvement of BCP in early rehabilitation management of poststroke patients was associated with good functional recovery. FurtherTable 3. Subgroup analysis of board-certified physiatrist for FIM effectiveness Subgroup
Coefficient (b)
Age at admission, y $75 .015 ,75 .073 FIM at admission $53 .072 ,52 .021
95% CI
P value
.001-.028 .036-.110
,.05 ,.001
.034-.109 2.003 to .042
,.001 .090
Abbreviations: CI, confidence interval; FIM, Functional Independence Measure.
more, acute stroke patients managed by BCP were likely to be discharged home from acute hospitals. Because randomized controlled trials to identify the impact of management by BCP are not feasible in real clinical setting, we performed retrospective cohort analysis based on a large database for this purpose. The database used in the present study, the Japan Rehabilitation Database, included data of a large number of patients under multicenter registry of several hospitals across Japan. Therefore, the strong statistical power of this study allows us
Table 4. Multivariate logistic regression analysis for discharge to patients’ own home Variable
OR
95% CI
P value
Board-certificated physiatrist Age at admission (y) Female gender NIHSS at admission Time to admission after onset (d) Daily exercise time (min/d) Duration of hospitalization (d) Cerebral infarction Prestroke mRS (score, 2-5) Hypertension Diabetes Atrial fibrillation Past history of stroke At least 1 caregiver
1.24
1.08-1.44
,.005
.97 .96 .86 .90
.97-.98 .77-1.18 .84-.87 .83-.96
,.001 .68 ,.001 ,.005
.99 .96
.99-1.00 .94-.98
.11 ,.001
2.00 .57 .98 .93 .93 .95 2.31
1.75-2.24 .49-.67 .83-1.17 .81-1.07 .77-1.13 .55-1.65 1.80-2.96
,.001 ,.001 .84 .29 .46 .85 ,.001
Abbreviations: CI, confidence interval; mRS modified Rankin Scale score; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio.
ROLE OF BOARD-CERTIFICATED PHYSIATRISTS IN EARLY STROKE REHABILITATION
to draw definite conclusions. In addition, generalized estimating equations were used to account for the potential correlation of outcome within hospitals in this study. In other medical fields, previous clinical studies identified the important role of services provided by boardcertificated physicians in facilitating the introduction of appropriate and aggressive therapeutic interventions, which often lead to better functional recovery.16-18 For example, clinical care provided by midcareer anesthesiologists with board certification was associated with less mortality within 30 days of admission than that provided by physicians without such certification. Furthermore, the mortality rate of critically ill patients admitted to the intensive care units and managed clinically by board-certified intensivists was lower than that of patients managed by other specialties.19 In the field of PM&R, Jeong et al7 published a study that described the positive impact of clinical management by BCP. The group investigated the extent of BCP involvement in 12 rehabilitation hospitals and showed that clinical outcome after the completion of rehabilitative program was better in patients admitted to hospitals with sufficient number of BCP than in those admitted to hospitals without BCP. The study was based on data from rehabilitation hospitals and involved patients with subacute-to-chronic stroke. To our knowledge, however, there are no studies that examined the relation between the involvement of BCP in early rehabilitative management and functional recovery of acute stroke. Thus, our study is the first clinical study that examined the direct relationship between BCP clinical management in early rehabilitation and good functional recovery of patients with acute stroke. What is the main reason for the significant association between BCP involvement and good functional recovery? Although our study design does not provide a definite answer to this question, several reasons can be speculated. Comparison of clinical data of patients managed by BCP and other specialties (Table 1) indicated that the duration of daily exercise was longer and frequency of regular meeting was significantly higher for patients managed by BCP than those managed by non-BCP. We speculate that this is because of the confidence of BCP in the safety of the prescribed rehabilitative training program even in patients with acute stroke compared with the hesitation of non-BCP in initiating early poststroke rehabilitation for fear of complication development. Furthermore, the better results with BCP could be because of the ability of these specialties to coordinate a multidisciplinary rehabilitative team through regular meeting. In this regard, a meta-analysis by Lohse et al20 indicated that high-dose rehabilitative training during the acute phase had a positive impact on clinical outcome of stroke patients. Although the high-dose exercise was not associated with the functional recovery in the current multivariable linear analysis (Table 2), the correlation between the
5
duration of exercise time and BCP who had tendency to prescribe high-dose training was thought to cause the impact to disappear. Thus, it seems that the ability of BCP to introduce early stroke rehabilitation and regular conferencing contributed to the achievement of good functional recovery. The improvement of FIM effectiveness in patients managed by BCP would contribute to the high likelihood of home discharge because our analysis was made after adjustment for FIM at admission, existence of caregiver, and diagnosis of stroke subtype. This finding indicates that BCPs were superior in terms of management of home discharge program with multidisciplinary coordination. Two subanalyses in this study found positive impact for BCP involvement in young patients and those with higher score of FIM at admission. These findings indicate that BCP involvement in early rehabilitation does not seem to be beneficial to clinical outcome in older patients and severely disabled acute patients. We consider that the lack of significant association between BCP management and functional improvement in the latter 2 groups is because of the small window available in training of these patients. In this regard, Denti et al21 reported that elderly stroke patients are less responsive to rehabilitative program compared with younger patients. Furthermore, Inouye et al22 reported markedly poor clinical outcome of acute stroke patients with low FIM score at admission despite the introduction of early stroke rehabilitation. Interestingly, the recovery of cognitive function was not associated with BCP management. What is the reason for this finding? In Japan, cognitive rehabilitation is seldom initiated in the acute phase of stroke and significant improvement of cognitive function is usually found in the subacute chronic phase of stroke. In addition to BCP involvement, the results of multivariate linear regression analysis and logistic regression analysis identified other clinical factors, such as female gender, NIHSS at admission, duration of hospitalization, prestroke mRS, and subtype of stroke, as significant determinants of functional recovery after rehabilitation program for acute stroke. This finding was, however, not novel as the prognostic impact of the same factors has already been identified.23-25 It is noteworthy that various medical and surgical interventions cannot have an impact on most of these factors. Thus, based on the finding of the important role of physiatrists, we recommend further development of the training system for physiatrist-in-training, with special emphasis on education and clinical experience on stroke rehabilitation. Currently, the number of BCP practicing in clinical settings differs markedly among countries. Although some developed countries in Europe have more than 1 BCP per 100,000 populations, most Asian developing countries have less than 1 BCP per 100,000 populations.4,5 Thus, there is a need to increase the number of BCP not only in Japan but also in other countries, together with
S. KINOSHITA ET AL.
6
the introduction of more sophisticated education system for better patient care. This study has certain limitations. First, because the database used was constructed without random sampling, we cannot generalize the study results. Second, information about acute medical and surgical treatment, such as intravenous administration of tissue plasminogen activator in the database, was incomplete to allow its inclusion in the multivariate analysis. Therefore, we were not able to assess the effect of acute treatment or the association between acute treatment and the involvement of BCP in the management of patients with acute stroke. Third, because no data were available after hospital discharge, the impact of BCP care in early phase on long-term outcome of stroke remains unknown.
Conclusions In this retrospective cohort study, analysis of the Japan Rehabilitation Database identified clinical management provision by BCP in the form of early rehabilitation for acute stroke patients as a significant predictor of good functional prognosis. Based on the positive results on the role played by BCP in the clinical management of such patients and the small number of BCP per capita in Japan, we recommend increasing the number of physiatrist trainees in Japan to provide better rehabilitation management for acute stroke patients. Acknowledgments: We would like to thank Japan Association of Rehabilitation Database for database provision. Japan Association of Rehabilitation Database did not provide the perspective in this study.
References 1. Bernhardt J, Dewey H, Thrift A, et al. A very early rehabilitation trial for stroke (AVERT): phase II safety and feasibility. Stroke 2008;39:390-396. 2. Bernhardt J, Thuy MN, Collier JM, et al. Very early versus delayed mobilisation after stroke. Cochrane Database Syst Rev 2009;21. 3. Jauch EC, Saver JL, Adams HP Jr, et al. Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2013;44:870-947. 4. Franceschini M, Caso V, Zampolini M, et al. The role of the physiatrist in stroke rehabilitation: a European survey. Am J Phys Med Rehabil 2009;88:596-600. 5. Ward AB. Rehabilitation medicine: the European perspective. Am J Phys Med Rehabil 2005;84:233-237. 6. Han TR, Bang MS. Rehabilitation medicine: the Asian perspective. Am J Phys Med Rehabil 2007;86:335-338. 7. Jeong S, Kondo K, Shiraishi N, et al. An evaluation of the quality of post-stroke rehabilitation in Japan. Clinical Audit 2010;2:59-66.
8. Chumney D, Nollinger K, Shesko K, et al. Ability of Functional Independence Measure to accurately predict functional outcome of stroke-specific population: systematic review. J Rehabil Res Dev 2010; 47:17-29. 9. Sulter G, Steen C, De Keyser J, et al. Use of Barthel index and modified Rankin scale in acute stroke trials. Stroke 1999;3:1538-1541. 10. Westerkam WR, Cifu DX, Keyser L. Functional outcome after inpatient rehabilitation following aneurysmal subarachnoid hemorrhage: a prospective analysis. Top Stroke Rehabil 1997;4:29-37. 11. Brandstater ME. International survey of training and certification in physical medicine and rehabilitation. Arch Phys Med Rehabil 2000;81:1234-1235. 12. Chino N. Certification and measuring competency in Japan, South Korea, and the Philippines. Arch Phys Med Rehabil 2000;81:1248-1249. 13. Izumi S. Postgraduate training and certification of physiatrists (rehabilitation physicians) in Japan. Jpn J Rehabil Med 2005;42:199-204. 14. Shah S, Vanclay F, Cooper B. Efficiency, effectiveness, and duration of stroke rehabilitation. Stroke 1990;21: 241-246. 15. Tanwir S, Montgomery K, Chari V, et al. Stroke rehabilitation: availability of a family member as caregiver and discharge destination. Eur J Phys Rehabil Med 2014; 50:355-362. 16. Sharp LK, Bashook PG, Lipsky MS, et al. Specialty board certification and clinical outcomes: the missing link. Acad Med 2002;77:534-542. 17. Prystowsky JB, Bordage G, Feinglass JM. Patient outcomes for segmental colon resection according to surgeon’s training, certification, and experience. Surgery 2002;132:663-722. 18. Silber JH, Kennedy SK, Even-Shoshan O, et al. Anesthesiologist board certification and patient outcomes. Anesthesiology 2002;96:1044-1052. 19. UP Iyegha, Asghar JI, Habermann EB, et al. Intensivists improve outcomes and compliance with process measures in critically ill patients. J Am Coll Surg 2013;216: 363-372. 20. Lohse KR, Lang CE, Boyd LA. Is more better? Using metadata to explore dose-response relationships in stroke rehabilitation. Stroke 2014;45:2053-2058. 21. Denti L, Agosti M, Franceschini M. Outcome predictors of rehabilitation for first stroke in the elderly. Eur J Phys Rehabil Med 2008;44:3-11. 22. Inouye M, Kishi K, Ikeda Y, et al. Prediction of functional outcome after stroke rehabilitation. Am J Phys Med Rehabil 2000;79:513-518. 23. Di Carlo A, Lamassa M, Baldereschi M, et al. Sex differences in the clinical presentation, resource use, and 3-month outcome of acute stroke in Europe: data from a multicenter multinational hospital-based registry. Stroke 2003;34:1114-1119. 24. Adams HP Jr, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Neurology 1999;53:126-131. 25. Ween JE, Alexander MP, D’Esposito M, et al. Factors predictive of stroke outcome in a rehabilitation setting. Neurology 1996;47:388-392.
