STUDIES HAVE SHOWN that rehabilitation efforts for

Transcription

1 ORIGINAL ARTICLE Rehabilitation Outcomes in a Population of Nonagenarians and Younger Seniors With Hip Fracture, Heart Failure, or Cerebral Vascular Accident Douglas Conner, PhD, Carol Barnes, MS, Cynthia Harrison-Felix, PhD, Nora Reznickova, MD 1505 ABSTRACT. Conner D, Barnes C, Harrison-Felix C, Reznickova N. Rehabilitation outcomes in a population of nonagenarians and younger seniors with hip fracture, heart failure, or cerebral vascular accident. Arch Phys Med Rehabil 2010;91: Objectives: To compare rehabilitation characteristics and patient outcomes between nonagenarians and younger seniors with hip fracture (HFx), heart failure (HF), or cerebral vascular accident (CVA). Design: Data only, retrospective cohort. Setting: Seven skilled nursing facilities providing rehabilitation services to a managed care organization. Participants: Subjects (N 2563; age, 65y) with HFx, HF, or CVA receiving rehabilitation services. Interventions: Not applicable. Main Outcome Measures: Patient and rehabilitation characteristics influencing FIM score at discharge and the proportion of patients discharged to the community were compared between nonagenarians and younger seniors with HFx, HF, or CVA. Results: Patients with higher FIM scores were discharged with better function. Different patient characteristics were important for successful rehabilitation for different conditions and outcomes. Except for HFx, nonagenarians had and discharge characteristics similar to those of younger seniors, although fewer were discharged to the community. Nonagenarians and younger seniors with CVA were most similar for all measures. Conclusions: Fewer nonagenarians were admitted from the community and fewer were discharged to the community, even if admitted from the community. Nonagenarians with HFx differed most strikingly from their younger counterparts in and discharge measures, as well as total discharge FIM score and discharge to the community. Nonagenarians and younger seniors with CVA were most similar for all measures. Our results suggest that a large proportion of the nonagenarian population can benefit from rehabilitation efforts for these 3 conditions; however, more rehabilitation resources may be required for some conditions to achieve similar outcomes. Key Words: Nonagenarian; Recovery of function; Rehabilitation; Rehabilitation outcomes. From the Institute for Health Research, Kaiser Permanente Colorado, Denver, CO (Conner); Care Management Institute, Kaiser Permanente, Oakland, CA (Barnes); Research Department, Craig Hospital, Englewood (Harrison-Felix); and Continuing Care Department, Colorado Permanente Medical Group, Denver, CO (Reznickova). 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. Correspondence to Douglas Conner, PhD, Institute for Health Research, Kaiser Permanente Colorado, PO Box , Denver, CO , douglas.a.conner@kp.org. Reprints are not available from the authors /10/ $36.00/0 doi: /j.apmr by the American Congress of Rehabilitation Medicine STUDIES HAVE SHOWN that rehabilitation efforts for many diagnoses in older adults are successful, yet older adults with more frequent rehabilitation needs are less likely to be referred for rehabilitation than younger adults. 1-8 This occurs in part because much of our knowledge about older adults is based on inpatient and nursing home studies of patients with multiple chronic conditions and disabilities, 9,10 and, in many cases, rehabilitation research has not included older adults. 11 There is even less knowledge when we look at rehabilitation outcomes with the very old, particularly nonagenarians. Nonagenarians have become the most rapidly growing segment of the older population in the United States, 12,13 and, with the aging baby boomer generation, our need to examine the public health and scientific implications of this age group are even more pressing. 9 Some studies have looked at rehabilitation for nonagenarians with orthopedic diagnoses (primarily HFx), but reports for other diagnostic groups are not as common. Kevorkian et al 14 examined rehabilitation outcomes in a small population of nonagenarians compared with a sample younger than 90 years undergoing rehabilitation, most because of HFx. They found that nonagenarians were less likely to have been admitted from home, were women, and were less likely to be discharged to home. However, age was not a factor in total FIM score gain or FIM score gain per day. The goal of this study was to examine patient and rehabilitation characteristics influencing total discharge FIM score and discharge to the community in a large sample of nonagenarians compared with younger seniors with HFx, HF, and CVA. We hypothesized that nonagenarians can undergo successful rehabilitation; however, more effort usually is required for this age group. Information about the influences on rehabilitation outcomes may assist providers in developing appropriate treatment plans for patients undergoing rehabilitation. METHODS Design This was a retrospective analysis of the records of patients admitted to SNFs for rehabilitation after a hospitalization or CVA HF HFx MCO SNF List of Abbreviations cerebral vascular accident heart failure hip fracture managed care organization skilled nursing facility

2 1506 REHABILITATION OUTCOMES IN OLDER ADULTS, Conner decline in function. All patients were enrollees in a large MCO. The institutional review board of the MCO approved this study. Setting The setting included 7 SNFs in the Denver, CO metropolitan area. SeniorMetrix Inc, the contractor for the collection and database management of outcome information from participating facilities, provided rehabilitation records. 15 The MCO contracts for rehabilitation services with these SNFs for its members. Participants Participants were 65 years and older, admitted to an SNF for rehabilitation after a hospitalization or decline in function between May 1998 and September The initial sample consisted of 13,270 rehabilitation records. Patients in the sample represented a variety of diagnostic categories, including orthopedic conditions (29.2%), pulmonary conditions (9.0%), stroke (8.6%), cardiac conditions (7.7%), and other disabling conditions (27.3%; neoplasms, various diseases of the major organ systems, metabolic and immune disorders, injuries, and complications). We excluded all diagnoses except for HPx, HF, and CVA because of the small numbers of other diagnoses or, in the case of other disabling conditions, the diversity of conditions. The final sample size was 2563 records. Assessment of Function at Discharge Functional ability at and discharge was measured by using the FIM. 16 The FIM was used by providers at the time of to the SNF and at discharge. The FIM is a validated and reliable measure of a person s functional ability. The FIM consists of a cognitive subscale score (5 items), a motor subscale score (13 items), and an overall score computed as the sum of the 2 subscale scores. Scores for each item range from 1 (total dependence) to 7 (total independence), resulting in a cognitive FIM score ranging from 5 to 35 and a range of 13 to 91 for the motor FIM score. 16 The maximal FIM score therefore is 126 points. Discharge Destination The number of patients discharged to the community was measured as the number of patients discharged to home, board and care, or assisted living facility. Measurement of Other Characteristics Age, sex, ethnicity, days postonset, and medical complexity were abstracted from rehabilitation records. Days postonset is the number of days from the precipitating event to the initiation of rehabilitation treatment. Medical complexity is a clinicianbased measure of the effect of a patient s comorbid conditions on function used by SeniorMetrix Inc and is similar to the American Society of Anesthesiologist s rating scale. 17 It is measured on a scale of 0 to 5 (0 no systemic disease other than primary medical diagnosis; 1 premorbid, inactive, and/or irrelevant systemic disease; 2 active relevant systemic disease not limiting function; 3 active systemic disease limiting function; 4 active systemic disease severely limiting function; 5 moribund/terminal condition). Rehabilitation characteristics included number of therapy hours (physical, occupational, and speech therapies combined) and rehabilitation length of stay (number of days from the first to the last day of therapy). We did not use SNF length of stay because SNF discharge may be influenced by social and other factors in addition to clinical factors, particularly in older populations. 8 For example, a patient may have concluded their course of therapy but may not be discharged immediately after therapy because no discharge setting has been identified, transport is not available, or relatives are unprepared to receive the patient. Data Analysis Statistical analyses were performed by using SAS statistical software. a Total discharge FIM score and proportion of patients discharged to the community were the 2 outcomes of interest. Significant mean differences between nonagenarians and younger seniors were tested by using Wilcoxon 2-sample tests for continuous measures. Chi-square or Fisher exact test was used for categorical measures. Patient and rehabilitation characteristics influencing the 2 outcomes were identified by using multiple linear regression for total discharge FIM score and multiple logistic regression for discharge to the community. Separate models were developed for characteristics alone and for and rehabilitation characteristics. The predictive ability of each model was compared by using the proportion of explained variance (R 2 ) for regression models, and the discriminatory ability of the logistic models was indicated by using the C statistic. Diagnostics were run on both the linear and logistic regression models. RESULTS Significant Univariate Comparisons at SNF Admission Admission measures included days postonset; medical complexity; cognitive, motor, and total FIM scores; sex; ethnicity; and to an SNF from a community or other setting (table 1). Nonagenarians had significantly lower total, cognitive, and motor FIM scores, and were less likely to have lived in the community before SNF. There were no differences in days postonset, medical complexity, sex, or ethnicity. Nonagenarian patients with HF were admitted to SNFs sooner (fewer days postonset) and were less likely to be white, although most patients for all diagnoses were white. There were no differences in total, cognitive, or motor FIM scores or medical complexity. Nonagenarians were equally likely to have lived in the community before SNF. Nonagenarian patients with CVA had fewer days postonset and were less likely to be living in the community. No differences in medical complexity; total, cognitive, or motor FIM scores; or ethnicity were found. Significant Univariate Comparisons at SNF Discharge Measures during an SNF stay and at discharge included total hours of therapy received, number of days that the patient received therapy during the SNF stay, total discharge FIM score and FIM score gain, and the proportion of patients discharged to the community (overall and only for those admitted from the community) (table 2). Nonagenarian patients with HFx had a significantly lower discharge FIM score than younger seniors. Fewer nonagenarians with HFx or CVA were discharged to the community. This was true even for those admitted from the community. The proportion of nonagenarians and younger seniors with HF who were discharged to the community did not differ. Admission Influences on Total Discharge FIM Score and Discharge to the All measures (except sex) contributed to total discharge FIM score. Being a nonagenarian resulted in a lower

3 REHABILITATION OUTCOMES IN OLDER ADULTS, Conner 1507 Table 1: Admission Measures for Younger Seniors and Nonagenarians for HFx, HF, and CVA Younger Seniors Nonagenarians Measures Condition Mean SD Median N Mean SD Median N Continuous measures Age HFx NA HF NA CVA NA Days postonset HFx HF CVA Medical complexity HFx HF CVA Total FIM score at HFx HF CVA Cognitive FIM score at HFx HF CVA Motor FIM score at HFx HF CVA Number Discharged (%) Number Discharged (%) Categorical measures Sex (% women) HFx 898 (74.2) (79.4) HF 159 (62.6) (73.6) CVA 161 (59.4) (66.7) Living situation at HFx 1122 (92.8) (89.4) (% living in the community) HF 242 (95.3) (91.7) CVA 625 (96.0) (89.4) Ethnicity (% white) HFx 1203 (99.5) (99.7) HF 252 (99.2) (94.4) CVA 646 (99.2) (98.8) Abbreviations: NA, not applicable. *Wilcoxon 2-sample test for continuous measures, chi-square or Fisher exact test (for cells 5) for categorical measures. Home, assisted living, or board and care. P* total discharge FIM score than being a younger senior. Total explained variance was 72% (table 3). For patients with HF, higher motor and cognitive FIM scores, less medical complexity, and quicker SNF after the precipitating event contributed to increased total discharge FIM score. Medical complexity had the greatest influence. Sex, being admitted from the community, or being a nonagenarian did not influence total discharge FIM score. Total explained variance was 53% (see table 3). Patients with CVA with higher motor and cognitive FIM scores, less medical complexity, and fewer days postonset contributed to an increase in total discharge FIM score. Being a nonagenarian did not influence total discharge FIM score for patients with CVA. Total explained variance was 70% (see table 3). Adding the number of therapy hours to the regression models for total discharge FIM score had a significant, but small, effect on discharge FIM score. The increase in adjusted R 2 value for HFx was 1.6 points; for HF, 3.3 points; and for CVA, 4.4 points; small increases in the total explained variance for each model (results not shown). Patients with HFx who were discharged to the community had higher motor and cognitive FIM scores, were admitted from the community, were admitted sooner, and were more likely to be younger women. The discriminatory ability of the model or C statistic was.76 (table 4). For patients with HF, a higher motor FIM score was the only significant predictor of discharge to the community. The discriminatory ability of the model or C statistic was.64 (see table 4). For patients with CVA, higher motor and cognitive FIM scores and being younger, women, and admitted from the community predicted a community discharge. The discriminatory ability of the model or C statistic was.74 (see table 4). Number of therapy hours during an SNF stay did not contribute significantly to community discharge for patients with either HFx or CVA. Number of therapy hours was a significant predictor for patients with HF. Each hour of therapy increased the probability of community discharge by 5%, and the C statistic increased from.64 to.67 (results not shown). DISCUSSION This study showed the importance of a higher FIM score in contributing to both total discharge FIM score and discharge to the community for both age groups. Better functional status at SNF resulted in better function at discharge. It also showed that different patient characteristics were important for successful rehabilitation for different conditions and outcomes. For example, community was important for patients with HFx and CVA for both outcomes but not for patients with HF. Being a man resulted in higher total discharge FIM score, but being a woman increased the

4 1508 REHABILITATION OUTCOMES IN OLDER ADULTS, Conner Table 2: Discharge Measures for Younger Seniors and Nonagenarians by HFx, HF, and CVA Younger Seniors Nonagenarians Measures Condition Mean SD Median N Mean SD Median N Continuous measures Total FIM score at HFx discharge HF CVA Total FIM score gain at HFx discharge HF CVA Therapy hours HFx HF CVA Rehabilitation length of stay HFx HF CVA Number Discharged (%) Number Discharged (%) Categorical measures Discharge to the community HFx 749 (62.0) (41.4) HF 159 (62.6) (55.6) CVA 423 (65.0) (48.2) Discharge to the community for those admitted from the community HFx 731 (65.5) (44.4) HF 153 (63.2) (57.6) CVA 417 (66.7) (50.0) *Wilcoxon 2-sample test for continuous measures; chi-square or Fisher exact test (cells 5) for categorical measures. Total discharge FIM score range, 18 to 126. Total number of rehabilitation hours a patient received during his/her SNF stay. P* likelihood of community discharge (except for patients with HF). Fewer nonagenarians with HFx and CVA were admitted from a community setting than younger seniors. The proportion of nonagenarians for each diagnosis admitted from the community (89.4% 91.7%) in this study was higher than the distribution of all nonagenarians between institutional and community settings in the general population found in another study (73.1%; living at home plus service building [similar to an assisted living facility]), 9 but we did not find a larger proportion of female nonagenarians in our sample given the preponderance of women in the general population who were 90 years or older. 18 Hip Fracture Nonagenarians admitted with an HFx had poorer function, had greater medical complexity, and were less likely to be admitted from the community. The lack of difference in time to SNF may be caused by the emphasis on early surgical repair and rehabilitation for HFx. A number of studies have shown that earlier surgical repair was associated with more positive short- and longer-term outcomes 19,20 (but see Majumdar et al 21 ). At SNF discharge, nonagenarians with HFx also showed consistently poor functional outcomes and a trend for more therapy than younger seniors. They also were less likely to be discharged to the community. Why are differences in and discharge measures for patients with HFx between nonagenarians and younger seniors so large compared with those for patients with HF and CVA? Other studies have suggested that nonagenarians with HFx experience significantly greater loss of function, greater morbidity, and greater mortality than younger seniors, and this may account for some of the poorer outcomes in this study. 22 However, it is worth noting that 41% of nonagenarians with HFx were discharged to a community setting. Heart Failure Nonagenarian patients with HF were admitted to SNFs with similar functional status, and a similar proportion was admitted from a community setting, although some other studies have reported that older cardiac rehabilitation patients generally are less fit, with lower maximum oxygen consumption and aerobic capacity. 23 Nonagenarian patients with HF were admitted to SNF facilities more than 2 days sooner than younger seniors. There were no significant differences in rehabilitation outcomes for nonagenarians with HF compared with younger seniors. Other studies have shown that the amount of gain achieved from cardiac rehabilitation was similar in younger and older age groups. 23 A number of investigators have mentioned that the goals of rehabilitation therapy in older cardiac patients may differ from those of younger patients and that older patients often have increased exercise/rehabilitation limitations because of age-related changes in cardiac and lung function, muscle strength and endurance, cognitive status, and comorbid condition burden. 24,25 This was not reflected in the outcomes we measured. Cerebral Vascular Accident Patients with CVA had the fewest differences between the 2 age groups. Nonagenarian patients with stroke did not differ in functional status or medical complexity compared with younger seniors. Fewer nonagenarians were admitted from the community, and they were admitted earlier than younger seniors. One other study compared days to by age and found an increase of 3 to 9 days to SNF for younger ages (age groups, 65, 65 79, 79y), which

5 REHABILITATION OUTCOMES IN OLDER ADULTS, Conner 1509 Table 3: Multivariate Regressions of SNF Admission Measures on Discharge FIM for HFx, HF, and CVA Variable Measure Estimate SE P* HFx HF CVA Admission motor Admission cognitive Medical complexity Days postonset Female sex Nonagenarian Admission motor Admission cognitive Medical complexity Days postonset Female sex Nonagenarian Admission motor Admission cognitive Medical complexity Days postonset Female sex Nonagenarian *Multivariate linear regression (adjusted R 2 : HFx.72, HF.56, and CVA.70). Admission motor FIM score range, 13 to 91; cognitive FIM score range, 5 to 35. matches our trend. However, these were more than 2 times longer than the days postonset for our younger seniors. 26 Nonagenarian patients with stroke were less likely to be discharged to the community. Granger et al 27 also found a lower rate of community discharge with increasing age, but we did not find significantly lower FIM scores at and discharge, or increasing days postonset for older patients, as they did. Other studies have shown an association between age and functional ability or living setting that we did not find Some of these differences may be caused by restricting our subject sample to only those with a CVA diagnosis. Another possible explanation is that our patients with CVA may have had greater severity of stroke, regardless of age, because of their to an SNF rather than to home or other rehabilitation venues. Finally, Inouye et al 30 also have found an association between functional status and better postrehabilitation outcomes for patients with stroke. Discharge to the Fewer nonagenarians were discharged to a community setting for all 3 conditions than younger seniors, even if admitted from the community. This lower rate of community discharge for older patients also has been found in other studies. 27,32,33 Nonagenarians were less likely than younger seniors to be discharged to home and more likely to be discharged to either assisted living or board and care facilities, where presumably more supportive care was available (results not shown). Despite this, more than 50% of nonagenarians were still discharged to a community setting for HF and CVA diagnoses, and more than 45%, for HFx. Study Limitations There are a number of potential study limitations. This was a retrospective study based on the records of 2563 rehabilitation patients distributed among 3 diagnoses. Because of the large sample size, statistical significance may have been found without corresponding clinical significance; however, many of our P values were small, and we believe that our findings were clinically significant. Our source of administrative data did not contain information about the severity of a diagnosis, which may have a significant effect on outcomes. We also were unable to look at the potential contribution of comorbid conditions on rehabilitation outcomes except for their possible indirect influence on medical complexity. This study did not look at other variables that may influence a patient s ability to return to the community, such as amount of family support or availability of other community support services. Some of these limitations may account for the lower R 2 and C statistic values for some of our models, suggesting that other unmeasured factors also are important. Finally, results of this study Table 4: Multivariate Logistic Regressions of SNF Admission Measures on Discharge to the for HFx, HF, and CVA Impairment Group HFx HF CVA Measure Adjusted Odds Ratio 95% CI P* Admission motor Admission cognitive Medical complexity Days postonset Female sex Nonagenarian Admission motor Admission cognitive Medical complexity Days postonset Female sex Nonagenarian Admission motor Admission cognitive Medical complexity Days postonset Female sex Nonagenarian Abbreviation: CI, confidence interval. *Multivariate logistic regression (C statistic: HFx.76; HF.64; CVA.74). Admission motor FIM score range, 13 to 91; cognitive FIM score range, 5 to 35.

6 1510 REHABILITATION OUTCOMES IN OLDER ADULTS, Conner may not be applicable to all SNF rehabilitation programs because patients were members of a group model MCO that uses a geriatric rehabilitation model with their contracted facilities providing on-site physician services and care coordination. CONCLUSIONS This study found a number of differences between nonagenarians and younger seniors with HFx and HF, but not CVA. Fewer nonagenarians were admitted from the community and fewer were discharged to the community even if admitted from the community. Nonagenarians with HFx differed most strikingly from their younger counterparts in and discharge measures, as well as total discharge FIM score and discharge to the community. Nonagenarians and younger seniors with CVA were most similar for all measures. Our results suggest that a large proportion of the nonagenarian population can benefit from rehabilitation efforts for these 3 conditions; however, more rehabilitation resources may be required for some conditions to achieve similar outcomes. Acknowledgments: Chris Wirtalla, MS, SeniorMetrix Inc, Nashville, TN, provided the rehabilitation database and assisted the investigators with the measures collected. Don A. Backstrom, PT, MBA, GCS, Kaiser Permanente Colorado, provided valuable comments on the manuscript. References 1. Hardy SE, Gill TM. Recovery from disability among communitydwelling older persons. JAMA 2004;291: Lieberman D, Lieberman D. Rehabilitation following stroke in patients aged 85 and above. J Rehab Res 2005;42: Bagg S, Paris Pumbo A, Hopman W. Effect of age on functional outcomes after stroke rehabilitation. Stroke 2002;33: Couser JI Jr, Githmann R, Abdugany HM, Kane CS. Pulmonary rehabilitation improves exercise capacity in older elderly patients with COPD. Chest 1995;107: Austin J, Williams R, Ross L, Moseley L, Hutchison S. Randomized controlled trial of cardiac rehabilitation in elderly patients with heart failure. Eur J Heart Fail 2005;7: Penrod JD, Litke A, Hawkes WG, et al. Heterogeneity in hip fracture patients: age, functional status, and comorbidity. J Am Geriatr Soc 2007;55: Cruise CM, Sasson N, Lee MHM. Rehabilitation outcomes in the older adult. Clin Geriatr Med 2006;22: Nakayama H, Jørgensen HS, Raaschou HO, Olsen TS. The influence of age on stroke outcome: the Copenhagen Stroke Study. Stroke 1994;25: Von Strauss E, Fratiglioni L, Viitanen M, Forsell Y, Winblad B. Morbidity and comorbidity in relation to functional status: a community-based study of the oldest old (90 years). J Am Geriatr Soc 2000;48: Lindgren A-M, Svärdsudd K, Tibblin G. Factors related to perceived health among elderly people: the Albertina Project. Age Ageing 1994;23: Ferrara N, Corbi G, Bosimini E, et al. Cardiac rehabilitation in the elderly: patient selection and outcomes. Am J Geriatr Cardiol 2006;15: U.S. Bureau of the Census. Current population reports: special studies, 65 plus in the United States. Washington (DC): Government Printing Office; U.S. Bureau of the Census. Population projections of the United States by age, sex, race, and Hispanic origin: 1995 to Washington (DC): Government Printing Office; Kevorkian CG, Ergeletzis D, Rintala D. Nonagenarians on a rehabilitation unit: characteristic, progress and outcomes. Am J Phys Med Rehabil 2004;83: Senior Metrix Inc. Postacute rehabilitation database. Nashville, TN: Senior Metrix Inc; Dodds TA, Martin DP, Stolow WC, et al. A validation of the functional independence measurement and its performance among rehabilitation patients. Arch Phys Med Rehabil 1993;74: Owens WD, Felts JA, Spitznagel EL Jr. ASA physical status classifications: a study of consistency ratings. Anesthesiology 1978;49: Han MK, Postma D, Mannjino DM, et al. Gender and chronic obstructive pulmonary disease: why it matters. Am J Crit Care Med 2007;176: Sircar P, Godkar D, Mahgerefteh S, Chambers K, Niranjan S, Cucco R. Morbidity and mortality among patients with hip fractures surgically repaired within and after 48 hours. Am J Ther 2007;14: Honig H, Rubenstein LV, Sloane R, Horver R, Kahn K. What is the role of timing in the surgical and rehabilitation care of community-dwelling older persons with acute hip fracture. Arch Intern Med 1997;157: Majumdar SR, Beaupre LA, Johnston DW, Dick DA, Cinats JG, Jiang HX. Lack of association between mortality and timing of surgical fixation in elderly patients with hip fracture: results of a retrospective population-based cohort study. Med Care 2006;44: Formiga F, Lopez-Soto A, Sacanella E, Coscojuela A, Suso S, Pujol R. Mortality and morbidity in nonagenarian patients following hip fracture surgery. Gerontology 2003;49: Fattirolli F, Burgisser C, Guarducci, Rinaldi LA, Mosotti G, Marchionni N. Cardiac rehabilitation in the elderly. Ital Heart J Suppl 2005;6: Pasquali SK, Alexander KP, Peterson ED. Cardiac rehabilitation in the elderly. Am Heart J 2001;142: Vigorito C, Incalzi RA, Acanfora D, Marchionni N, Fattirolli F. Recommendations for cardiovascular rehabilitation in the very elderly. Monaldi Arch Chest Dis 2003;60: Cruise CM, Sasson N, Lee MHM. Rehabilitation outcomes in the older adult. Clin Geriatr Med 2006;22: Granger CV, Hamilton BB, Fiedler RC. Discharge outcome after stroke rehabilitation. Stroke 1992;23: Ween JE, Alexander MP, D Esposito M, Roberts M. Factors predictive of stroke outcome in a rehabilitation setting. Neurology 1996;47: Osberg JS, DeJong G, Haley SM, Seward ML, McGinnis GE, Germaine J. Predicting long-term outcomes among post-rehabilitation stroke patients. Am J Phys Med Rehabil 1988;67: Inouye M, Kishi K, Takada M, et al. Prediction of functional outcomes after stroke rehabilitation. Am J Phys Med Rehabil 2000;79: Di Carlo A, Lamassa M, Pracucci G, et al. Stroke in the very old: clinical presentation and determinants of a 3-month functional outcome: a European perspective. Stroke 1999;30: Smith PM, Ottenbacher KJ, Cranely M, et al. Predicting follow-up living setting in patients with stroke. Arch Phys Med Rehabil 2002;83: Tesio L, Franchignoni FP, Perucca L, Porta GL. The influence of age on length of stay, functional independence and discharge destination of rehabilitation inpatients in Italy. Disabil Rehabil 1996;18: Supplier a. SAS Institute Inc, 100 SAS Campus Dr, Cary, NC