Sommerfeld, Disa Kathryn PT, PhD; Johansson, Helena PT; Jönsson, Anna-Lena OT; Murray, Veronica; Wessari, Tuula; Holmqvist, Lotta Widén PT, PhD; von Arbin, Magnus MD, PhD
The process of recovery after acute stroke is most rapid during the first days and weeks following the stroke event and continues for a long period of time.1 Because of the ever-increasing cost of hospital care, there is pressure from various administrative bodies to reduce duration of hospital stay in an effort to control costs. To select the optimal level of care for elderly persons with stroke-related disability, it is important to be aware of adequate discharge destinations and to have reliable predictors for the length of institutional stay (LOS) (ie, in hospital or nursing home). Several methods have been studied for predicting recovery and LOS after acute stroke.2,3 Common predictors of outcome in the acute stage are consciousness-,4 continence-,5 and movement-related functions.6 In a previous study,7 we looked for positive predictors of LOS (ie, predictors of short LOS) at 10 days after stroke onset, for persons 65 years and older. At 10 days after acute stroke, a score of 4 or more points on the Rivermead Mobility Index (RMI)8 (especially the ability to rise from a chair in less than 15 seconds and maintain standing for 15 seconds) and a score of 35 or more points on the Barthel Index (BI)9 had the greatest positive impact on the chance of being returning to home within 3 months.
A review of the Danderyd Hospital admission and discharge data revealed that LOS in the stroke unit after admission for acute stroke has been reduced from a mean of 10 days to 5 days. Ninety percent of all persons with stroke admitted to the hospital are treated in the stroke unit, which provides comprehensive multidisciplinary examination and intervention. About half the persons in our stroke unit are discharged home within 5 days. There are 4 possible discharge destinations from our stroke unit for persons 65 years and older: (1) home with home care services as necessary; (2) transfer to a geriatric rehabilitation ward for those able to participate in rehabilitation; (3) transfer to sheltered accommodation in a homelike environment with or without rehabilitation service; or (4) transfer to a long-term care facility, where minimal or no rehabilitation service may be provided.
In the present prospective study, we used a different study population to find the most feasible and optimal prognostic instruments among those commonly used in the clinic, collecting data on the 5th day following acute stroke. We were primarily interested in determining predictors for LOS as well as discharge destination for persons 65 years and older. The potential predictors included were defined according to the international classification of functioning.10 Information about stroke characteristics included cause (ie, cerebral infarction or intracerebral hemorrhage), whether the lesion was visible on computed tomographic (CT) scan and side of the body with hemiparesis or hemiplegia. We also gathered information on environmental factors, that is, previous living arrangements and availability of family support (ie, living with another person or not) and personal factors (age and gender).
The study was reviewed and approved by the ethics committee at the Karolinska Hospital. Informed consent was obtained from all participants or their next of kin.
Persons who were 65 years or older remaining hospitalized on the 5th day following acute stroke were eligible for enrollment in the study. Potential subjects also must have demonstrated movement-related impairments (≤17 points on the Scandinavian Stroke Scale)11 or limitations in mobility or self-care (≤11 on the RMI8 or ≤65 on the BI9).
Exclusion criteria included a diagnosis of subarachnoid hemorrhage and cerebellar lesions, a priori dependence in personal activities of daily living (ADLs), diagnosed dementia, severe confusion that prevented the person from living alone, and a variety of non–stroke-related disorders that impacted on the person's activity level. Any individual readmitted to the hospital could not reenroll in the study.
Eligible participants were consecutively recruited from the stroke unit at Danderyd Hospital in Stockholm, from September 1999 to May 2002, with recurrent breaks for administrative reasons.
Minimum requirements for discharge home to unassisted living included the following: the ability to eat without assistance, normal urination and bowel function according to the BI9 subtests 6 and 7, the ability to transfer between bed and wheelchair independently, and sufficient understanding to be able to call for help. Persons who did not meet these minimum requirements were discharged to a nursing home or other setting with access to personal assistance around the clock. These persons are further referred to in the text as being institutionalized, that is, not discharged home.
The final study sample consisted of 115 persons (Table 1). All persons underwent functional assessment 4 to 6 days after stroke onset. All test instruments were in routine use in the stroke unit. Physical therapists were responsible for examination of consciousness, somatosensory function, and mobility. Speech and language pathologists assessed language function. Occupational therapist assessed other body functions and ability to perform ADLs. The examiners were co-trained to minimize risk of interrater error.
In the text, we will use “day 5” to represent the 5th day following acute stroke. Participants would be considered “discharged home” if they returned to their home setting within 3 months of their acute stroke.
Measures of Body Function
Consciousness was assessed with the Glasgow Coma Scale (GCS) (possible range, 3–15 points).12 The GCS scores of 13 to 15 indicate mild brain injury; 9 to 12, moderate brain injury; and 8 or less, severe brain injury. For persons with stroke, the GCS is considered valid and reliable and a low score is an important predictor of negative outcome.4,13
Language function was dichotomized as either intact (no aphasia) or impaired (any aphasia). No distinction was made between different subtypes of aphasia. There is a high long-term mortality rate among persons with aphasia following stroke, such that aphasia may therefore be considered a negative predictor of stroke outcome.14
Perceptual space function was assessed using the line15 and letter cancellation tasks.16 The score represents the number of targets marked by the person on standardized forms. In the line cancellation task, 4 lines at the center of the page are used for demonstration, leaving 36 lines for the person to mark as demonstrated. The suggested range of normal performance is 32 to 36 marked lines.17 In the letter cancellation task, the person seeks to mark 30 target “A's” among 120 distracting letters, displayed at random on the test form. One “A” at the center of the page is used to initially demonstrate the task. Because there is no report of normal values for this test, we opted to a similar strategy to that of the line cancellation task, using 26 to 29 marked targets as indicator of intact perception. The persons were allowed unlimited time to complete the tests. These cancellation tasks are widely used and are considered valid tests of perceptual space function. At least 2 different cancellation tasks are recommended to identify all persons with impaired attention.18 Perceptual space function was considered impaired if the person marked fewer than 32 lines or 26 letters.
Perceptual function with respect to color sense and constructional practice was evaluated using the Block Design Test (the Block Test).19 There are 16 cubes with different colors on each surface (red, white, blue, yellow, diagonally red/white, and diagonally blue/yellow). A standard protocol is followed during testing, with increasing complexity using more cubes across 7 testing tasks. The tester demonstrates how to build 2 structures using designs presented on pictures and then asks the person to continue on their own. A single point is given for each correctly completed construction. Normal values for persons 65 years and older are not available, and there is no defined time limit for testing of elderly persons. Participants were allowed unlimited time to complete the test. The block test is considered reliable20 and valid21 for evaluating everyday spatial functioning. Perceptual functions assessed within 3 weeks after stroke have been found to be important predictors of long-term functional outcome.22
Emotional function (reported sadness) was assessed according to the Montgomery-Åsberg Depression Rating Scale (MADRS)23 item 2. Energy and drive (lassitude) was assessed according to the MADRS23 item 7. The test is considered reliable and valid for individual items.23 Elderly persons with poststroke depression have significantly more periods of hospitalization than those who have not.24
Mental function was assessed using the Mini-Mental State Examination (MMSE).25 A normal range of values has been established at 24 to 30 points.26 The reliability and validity of the MMSE is well supported26 and it has been recommended for use in the initial assessment of persons with stroke.27 Mini-Mental State Examination predicts discharge home in elderly persons with stroke.28 Because of difficulty in completing the MMSE, persons with aphasia or dysarthria were not assessed by the MMSE.
Touch function was determined by testing the ability to perceive pinprick (metal pin) and light touch (cotton wool) on the upper arm, forearm, dorsal hand, thigh, calf, and dorsal foot. The person was classified as having impaired touch function if they were unable to perceive pinprick and/or light touch on 1 or more locations on the affected side of the body. These tests are frequently used and are considered to have satisfactory reliability with a rough grading into normal or impaired touch function.26
Proprioceptive function, awareness of the relative position of body parts,10 of the affected upper limb was tested with the Thumb Localizing Test.29 The upper limb of the affected side is positioned passively and the person is asked to pinch the thumb of that limb with the opposite thumb and index finger. The test is repeated 4 times, the limb being positioned in different positions on each trial. Proprioceptive function is considered normal if the person is able to locate the thumb on the affected side with the eyes closed in 3 of 4 trials. The test is considered valid29 but has not been tested for reliability.
If either touch or proprioceptive function was classified as impaired, the person was considered to have impaired somatosensory function. Normal somatosensory function, 10 days poststroke, predicts early discharge home.30
Urination function was graded as continent or incontinent, according to the BI subitem 6.9 The person was considered incontinent if unable to control the bladder or if he or she was catheterized. If the person was unable to walk to the toilet, had only occasional accidents, or needed assistance with toileting, they were not considered incontinent. Urinary incontinence is considered to be an extremely negative predictor, with less recovery among those with incontinence.5
Mobility was assessed with the RMI8 (possible range, 0–15 points). The RMI is a hierarchical instrument that begins with the person's ability to turn over in bed in item 1 and ends with the person's ability to run in item 15. The RMI is considered reliable,31 valid,32 and sensitive to change.31,32 Persons with 4 or more points, at 10 days after stroke, are considered to have good prognoses.7
ADL was assessed with the BI9 (possible range, 0–100 points). The BI assesses personal ADL (feeding, bathing, grooming, dressing, bowel control, bladder control, toilet use, transfer, mobility, and the ability to manage on stairs). The BI is widely used and is considered reliable,33 valid,34 and sensitive to change.33 A BI score of less than 35 of 100 points in the early poststroke period is thought to indicate poor prognosis.7,35 Urinary function and mobility items, although included in the BI, were also evaluated and analyzed separately in the present study.
LOS was recorded during the period 5 to 90 days after stroke onset.
Median values and interquartile ranges, on day 5, were calculated for LOS, GCS, line and letter cancellation task, Block Test, MADRS item 2 and 7, MMSE, RMI, and BI. Univariate survival analyses within 3 months poststroke were estimated using Cox proportional hazards regression, assessing the unadjusted hazard ratios (HR) and 95% confidence intervals (CI). Considering the multiple structures of the data, Cox proportional hazards regression was performed, including all variables from the univariate analyses. Both stepwise and best selection procedures were undertaken. With the best selection procedure, we tried to find a specified number of best models containing 1 initial variable and adding in variables until the model contained all potential predictors. The best procedure is based on a global score chi-square statistic. For 2 models having the same number of predicting variables, the model with the highest score statistic is considered the best. Models with interaction between the predicting variables were also evaluated, that is, to determine whether the effect of one variable is significantly dependent on the effect of another variable. In the survival analyses, the variables were dichotomized according to cutoff values known to predict short LOS, that is, mild versus moderate or severe brain injury for GCS; intact versus impaired language function, MADRS, MMSE, somatosensory function, and urination function; less than 4 points versus 4 or more points on the RMI;7 and less than 35 points versus 35 or more points on the BI.7,35 When no predictive values for the LOS were available, as for the line and letter cancellation tasks, the variables were dichotomized according to normal versus impaired function. Stroke characteristics were dichotomized as cerebral infarction or intracerebral hemorrhage; lesion, visible on CT scan or not; impaired on left or right side of the body; and living alone or with others. For the Block Test, the median score was used and, for age, the mean value was used as cutoff. Length of institutional stay is presented with Kaplan-Meier curves. The positive prediction value of being discharged home at 1, 2, and 3 months, respectively, was estimated for variables reaching statistical significance in the multivariate analyses of LOS. It was only possible to include participants assessable according to each of the tests on day 5 in the analyses of LOS. Those who died between inclusion and 3 months were retained in the analyses of LOS and were regarded as not being discharged home. P < .05 was considered statistically significant. Data were analyzed using SAS System 8.2 (SAS Institute Inc, Cary, North Carolina) and Statistica 5.1 (StatSoft, Inc, Tulsa, Oklahoma) for Windows.
As can be seen in Table 2, it was possible to assess only consciousness and urination function, mobility, and ADL in all persons (n = 115) on day 5. Table 2 also shows median scores, interquartile ranges, and distribution of persons, dichotomized as normal/disabled, according to each test instrument assessed on day 5. Median LOS for all 115 persons was 80 days (27–90).
Three months after stroke, 60 persons (52%) were living at home, 12 (10.5%) had died, and 43 (37.5%) were still institutionalized. Among the institutionalized persons, all but 8 had received rehabilitation services in the geriatric ward. Seven of these 8 persons scored 0 on both the RMI and BI on day 5, poststroke, and 1 person (96 years old) scored 1 on the RMI and 15 on the BI on day 5, poststroke.
Univariate Survival Analyses
The LOS up to 3 months was significantly shorter for persons with mild brain injury (GCS ≥ 13), normal emotional function (MADRS item 2 = 0, no apparent sadness), normal somatosensory function, urinary continence, RMI ≥ 4 points, BI ≥ 35 points assessed on day 5; no previous stroke; or the present lesion not visible on CT scan (Table 3).
There were no statistically significant differences in LOS up to 3 months regarding language, perceptivity, energy and drive, and mental functions; social situation; side of present symptom; type of lesion; age; and gender.
Multiple Survival Analyses
Best selection regression analyses for LOS up to 3 months including 1, 2, 3, and 4 variables, respectively, assessed on day 5, gave the following chi-square (χ2) scores:
RMI score ≥ 4 (χ2 = 51.4) (P < .001)
RMI score ≥ 4 and BI score ≥35 (χ2 = 56.3) (P < .001)
RMI score ≥ 4, BI score ≥35 and GCS ≥13 (χ2 = 59.8) (P < .001)
RMI score ≥ 4, BI score ≥35, GCS ≥13 and no previous stroke (χ2 = 63.2) (P < .001)
The HR and CI for the final model containing the variables RMI, GCS, and previous stroke are presented in Table 4. The BI was not a significant variable in the multiple analyses, but if the RMI was removed from the multiple analyses, the BI had a statistically significant positive impact on LOS. All additional variables lost their statistical significance in the multiple analyses when estimated together with the significant variables presented previously. There was no statistically significant interaction effect between the variables.
The positive prediction value, that is, the proportion of persons discharged home on day 30, 60, and 90, respectively, categorized according to the respective cutoff on the GCS, RMI, and BI assessed on day 5, and previous stroke, is presented in Table 5.
The proportion of persons not discharged home, categorized according to the respective cutoff assessed on day 5, is shown in Figure 1 (GCS), Figure 2 (RMI), Figure 3 (BI), and Figure 4 (previous stroke).
More than half the participants in our stroke unit who are 65 years or older are discharged home at 5 days after stroke. We looked for predictors for LOS for those not discharged by day 5 for persons 65 years and older who were still in hospital 5 days after acute stroke, as these persons constituted a large and important subgroup among all stroke victims. To our knowledge, this is the first study presenting a clinically useful predictor possible to use on every person 65 years and older as early as 5 days after acute stroke. Multiple analyses for LOS, including functioning assessed on day 5, stroke characteristics, social living, and personal factors, showed that (1) mild brain injury, (2) the ability to rise from a chair in less than 15 seconds and stand there for 15 seconds with or without an aid (RMI ≥ 4 points), and (3) no previous stroke had the greatest positive impact on LOS. Of these variables, the ability to rise from a chair in less than 15 seconds at day 5 was the strongest predictor of short LOS. Additional variables had no significant influence on the results of the multiple analyses.
One limitation of the study is that the data used in analysis were collected almost 10 years ago. Although LOS may be even shorter at this point in time and rehabilitation typically takes place in other settings, our results illustrate which participants will recover enough to be living outside an institution after 3 months and what factors predict a short LOS. We did not follow the participants after discharge and thus we do not know how well they fared or whether they were readmitted to hospital after discharge. However, we were aware of whether they had a new stroke and were readmitted to the stroke unit. These circumstances constitute a limitation of the study; however, the purpose of the study was not to evaluate the discharge routines of the hospital but to find predictors matching the actual clinical situation.
One cannot expect to find an instrument to predict outcome or LOS after stroke that will be feasible in every possible situation, given variation in functioning and disabilities of participants and different contextual factors. Our selection of participants was made according to local prerequisites, including possible access to different levels of care. Furthermore, the intention was that the persons included should be at least 65 years old. Our participants were quite disabled, as was indicated by the fact that more than 40% were incontinent at day 5.
Mobility and ADL, assessed by the RMI and the BI in the present study, are activities that are to be considered after stroke.36 Both RMI and BI were very strong positive predictors of LOS, but, in combination, the RMI was stronger. Both the RMI and the BI are easily and quickly performed on any person with stroke.
History of previous stroke is thought to be a negative predictor of outcome after stroke2 in some studies, whereas others report that previous stroke has little effect on outcome.3 In our study population, “no previous stroke” was one of the statistically significant positive predictors of short LOS in the multiple analyses at day 5.
Impaired consciousness immediately after stroke is a strong predictor of negative outcome.4,13 This was reflected in both the univariate and the multiple analyses of the present study. Impaired consciousness on day 5 was a reliable negative predictor, whereas being conscious on day 5 was an insufficient positive predictor of LOS (Figure 1).
Although consciousness was a competitive contributor in the multiple analyses, no other body functions, personal or contextual factors, or stroke characteristics reached statistical significance when estimated together with the current activity tests. The isolated importance of perceptual, somatosensory, and urination function is obvious for the individual person. Although these body functions were single predictors of short LOS, they lost their significance when calculated with strong predictors such as RMI and BI. None of the body functions, except urinary function, were assessable in all persons on day 5 after stroke, thus limiting the usefulness of these test instruments at this time point after stroke.
RMI ≥ 4 had the highest chi-square score in the best selection model and was a statistically significant predictor in the multiple analyses of LOS at day 5. Persons with stroke, scoring 4 or more on the RMI on day 5, are 4.8 times as likely to be discharged home early and within 3 months, compared with those scoring less than 4 on the RMI (Table 4). Furthermore, RMI ≥ 4 assessed on day 5 had the highest positive prediction value of being discharged home within 90 days after stroke, and also within 30 and 60 days after stroke (Table 5).
The RMI and the BI measure whether an activity can be performed but not how it is accomplished. These instruments are comprehensive because many body functions contribute to the ability to perform the tested activities.27,37 All possible impairments may not be reflected in the activity tests, however. Thus, relying on these instruments only may miss key global and/or specific mental impairments that, if addressed, might impact on function. Regardless of activity scores, mental function must be assessed as part of the discharge planning process.
In addition to the variables identified in our previous study,7 we evaluated whether consciousness, perceptual, emotional, and mental functions might be predictors of LOS, 5 days after stroke. However, only consciousness function had any significant influence on LOS in the multivariate analyses.
A decision about discharge destination and subsequent need for rehabilitative or instrumental care after acute stroke is based on factors related to medical appraisal and functional status, as well as environmental and personal factors, and may vary by region or country. In this study, we found that LOS can be predicted using data collected on the 5th day following acute stroke, much sooner than the 10-day model of our previous work. The ideal situation is to be able to predict discharge destination immediately after stroke onset. However, reliable predictors will likely be more difficult to identify earlier, as there is often rapid recovery within 5 days.1
In the present study, we confirmed the negative prognostic value of unconsciousness after acute stroke. Having no previous stroke was confirmed as an important positive predictor. Among the test instruments used on day 5, the strongest predictor of short LOS was RMI ≥ 4, corresponding to the ability to rise from a chair in less than 15 seconds and remain standing for an additional 15 seconds. In the clinical setting, this test could serve as a feasible prognostic instrument to use as early as 5 days after stroke onset in persons 65 years and older.
The authors thank physical therapists Elsy Eek, Kerstin Ekdahl, and Helena Lindell and occupational therapist Sofia Biderholt, Department of Geriatric Medicine, Danderyd Hospital, for assessing the participants. They also thank Elisabeth Berg, Department of Learning, Informatics, Management and Ethics (LIME), Karolinska Institutet, for valuable statistical advice.
1. Sprigg N, Gray LJ, Bath PM, et al. Early recovery and functional outcome are related with causal stroke subtype: data from the tinzaparin in acute ischemic stroke trial. J Stroke Cerebrovasc Dis. 2007;16:180–184.
2. Kwakkel G, Wagenaar RC, Kollen BJ, Lankhorst GJ Predicting disability in stroke—a critical review of the literature. Age Ageing. 1996;25:479–489.
3. Counsell C, Dennis M Systematic review of prognostic models in patients with acute stroke. Cerebrovasc Dis. 2001;12:159–170.
4. Riachy M, Sfeir F, Sleilaty G, et al. Measuring neurological severity at hospital admission by GCS is the main determinant of future functional capacity in CVA patients at discharge and after one year. BMC Neurol. 2008;8:24.
5. Kolominsky-Rabas PL, Hilz MJ, Neundoerfer B, Heuschmann PU Impact of urinary incontinence after stroke: results from a prospective population-based stroke register. Neurourol Urodyn. 2003;22:322–327.
6. Hendricks HT, van Limbeek J, Geurts AC, Zwarts MJ Motor recovery after stroke: a systematic review of the literature. Arch Phys Med Rehabil. 2002;83:1629–1637.
7. Sommerfeld DK, von Arbin M Disability test 10 days after acute stroke to predict early discharge home in patients 65 years and older. Clin Rehabil. 2001;15:528–534.
8. Collen FM, Wade DT, Robb GF, Bradshaw CM The Rivermead Mobility Index: a further development of the Rivermead Motor Assessment. Int Disabil Stud. 1991;13:50–54.
9. Mahoney FI, Barthel DW Functional evaluation: the Barthel Index. Md State Med J. 1965;14:61–65.
11. Scandinavian Stroke Study Group. Multicenter trial of hemodilution in ischemic stroke—background and study protocol. Stroke. 1985;16:885–890.
12. Teasdale G, Jennett B. Assessment of coma and impaired consciousness: a practical scale. Lancet. 1974;2:81–84.
13. Appelros P, Nydevik I, Seiger A, Terent A Predictors of severe stroke: influence of preexisting dementia and cardiac disorders. Stroke. 2002;33:2357–2362.
14. Laska AC, Hellblom A, Murray V, Kahan T, Von Arbin M Aphasia in acute stroke and relation to outcome. J Intern Med. 2001;249:413–422.
15. Albert ML A simple test of visual neglect. Neurology. 1973;23:658–664.
16. Mesulam MM Attention, Confusial States and Neglect Principles of Behavioural Neurology. Philadelphia, PA: FA Davis Company; 1985.
17. Stone SP, Halligan PW, Wilson B, Greenwood RJ, Marshall JC Performance of age-matched controls on a battery of visuo-spatial neglect tests. J Neurol Neurosurg Psychiatry. 1991;54:341–344.
18. Hartman-Maeir A, Katz N Validity of the Behavioral Inattention Test (BIT): relationships with functional tasks. Am J Occup Ther. 1995;49:507–516.
19. Goldstein K, Scheerer M Abstract and concrete behavior: an experimental study with special tests. Psychol Monogr. 1941;239:1–151.
20. Dureman I, Kebbon L, Österberg E Manual till DS-Batteriet. Stockholm, Sweden: Psykologiförlaget AB; 1971.
21. Groth-Marnat G, Teal M Block design as a measure of everyday spatial ability: a study of ecological validity. Percept Mot Skills. 2000;90:522–526.
22. Nys GM, van Zandvoort MJ, de Kort PL, et al. The prognostic value of domain-specific cognitive abilities in acute first-ever stroke. Neurology. 2005;64:821–827.
23. Montgomery SA, Åsberg M A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134:382–389.
24. Jia H, Damush TM, Qin H, et al. The impact of post-stroke depression on healthcare use by veterans with acute stroke. Stroke. 2006;37:2796–2801.
25. Folstein MF, Folstein SE, McHugh PR “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–198.
26. Wade DT Measurement in Neurological Rehabilitation. Oxford, UK: Oxford medical publications; 1992.
27. Zwecker M, Levenkrohn S, Fleisig Y, Zeilig G, Ohry A, Adunsky A Mini-Mental State Examination, cognitive FIM instrument, and the Loewenstein Occupational Therapy Cognitive Assessment: relation to functional outcome of stroke patients. Arch Phys Med Rehabil. 2002;83:342–345.
28. 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.
29. Hirayama K, Fukutake T, Kawamura M “Thumb localizing test” for detecting a lesion in the posterior column-medial lemniscal system. J Neurol Sci. 1999;167:45–49.
30. Sommerfeld DK, von Arbin M The impact of somatosensory function on activity performance and length of hospital stay in geriatric patients with stroke. Clin Rehabil. 2004;18:131–137.
31. Chen HM, Hsieh CL, Lo SK, Liaw LJ, Chen SM, Lin JH The test-retest reliability of 2 mobility performance tests in patients with chronic stroke. Neurorehabil Neural Repair. 2007;21:347–352.
32. Hsieh CL, Hsueh IP, Mao HF Validity and responsiveness of the rivermead mobility index in stroke patients. Scand J Rehabil Med. 2000;32:140–142.
33. Hsieh YW, Wang CH, Wu SC, Chen PC, Sheu CF, Hsieh CL Establishing the minimal clinically important difference of the Barthel Index in stroke patients. Neurorehabil Neural Repair. 2007;21:233–238.
34. Collin C, Wade DT, Davies S, Horne V The Barthel ADL index: a reliability study. Int Disabil Stud. 1988;10:61–63.
35. Kalra L, Smith DH, Crome P Stroke in patients aged over 75 years: outcome and predictors. Postgrad Med J. 1993;69:33–36.
36. Duncan PW, Jorgensen HS, Wade DT Outcome measures in acute stroke trials: a systematic review and some recommendations to improve practice. Stroke. 2000;31:1429–1438.
37. Mercier L, Audet T, Hebert R, Rochette A, Dubois MF. Impact of motor, cognitive, and perceptual disorders on ability to perform activities of daily living after stroke. Stroke. 2001;32:2602–2608.
functional assessment; length of stay; prediction; stroke