Pattern of Functional Change During Rehabilitation of Patients With Hip Fracture

Transcription

1 111 ORIGINAL ARTICLE Pattern of Functional Change During Rehabilitation of Patients With Hip Fracture Nancy K. Latham, PhD, PT, Diane U. Jette, DSc, PT, Reg L. Warren, PhD, Christopher Wirtalla, BA ABSTRACT. Latham NK, Jette DU, Warren RL, Wirtalla C. Pattern of functional change during rehabilitation of patients with hip fracture. Arch Phys Med Rehabil 2006;87: Objective: To examine the rate of functional change in 2 domains, activities of daily living (ADLs) and mobility, over 2 time periods during hip fracture rehabilitation. Design: Retrospective analysis of data contained in an administrative dataset. Setting: Seventy skilled nursing facilities (SNFs). Participants: People (N 351) receiving rehabilitation in SNFs from March 1998 to February 2003 after hip fractures. Interventions: Not applicable. Main Outcome Measure: Rate of change in scores in the ADL and mobility domains of the FIM instrument during 2 time intervals of rehabilitation. Results: The rate of functional change across 2 time intervals was constant for mobility (mean change in FIM points per day,.46 vs.49), but declined in the second time period for ADLs (mean change in FIM points per day,.55 vs.41). Executive function, length of stay (LOS), and medical complexity were related to rate of change in mobility, and baseline ADLs, executive function, living setting, and LOS were related to rate of change in ADLs. There was an interaction between rehabilitation phase and baseline mobility. People with lower baseline mobility had an increased rate of change during the second interval (mean change in FIM points per day,.41 vs.55), whereas those with higher baseline mobility had a decreased rate of change (mean change in FIM points per day,.50 vs.43). Conclusions: The pattern of functional change over time differed for ADL and mobility domains, and for specific groups of patients. The results have implications for goal setting and discharge planning. Key Words: Activities of daily living; Hip fractures; Rehabilitation by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation HIP FRACTURE IS A COMMON health problem with serious consequences. One quarter of older people who suffer a hip fracture will be dead or will need total assistance to walk 6 months later. 1 It is estimated that 350,000 people in the From the Health and Disability Research Institute, Boston University, Boston, MA (Latham); Physical Therapy, Simmons College, Boston, MA (Jette); and Senior- Metrix Inc, Nashville, TN (Warren, Wirtalla). Supported by the Dudley Allen Sargent Fund, Boston University. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Nancy K. Latham, PhD, PT Health and Disability Research Institute, Boston University, 53 Bay State Rd, Boston, MA 02215, nlatham@bu.edu /06/ $32.00/0 doi: /j.apmr United States experience a hip fracture each year, resulting in hospital costs alone of $6 billion. 2 Moreover, the number of hip fractures is expected to rise substantially in the coming decades. 3 However, despite the frequency of this condition and the serious consequences associated with it, little is known about the progress and pattern of functional changes that might be expected during rehabilitation of patients with hip fracture. 4,5 Most studies that have examined change in function during rehabilitation have reported the total change from a baseline measurement at admission to a discharge measurement, with some studies using the overall change in function to calculate the change in function per day or week. 6-8 Some studies have also investigated the patient characteristics that affect the overall amount of change during rehabilitation. Although studies that examined changes from baseline to discharge provide an overall indication of the rate of change, they do not provide information about how function changes during rehabilitation. The pattern of functional improvement using measurements taken at intervals between admission and discharge has been reported to a limited extent for rehabilitation conditions other than hip fracture. Those studies included patients with spinal cord injury (SCI), traumatic brain injury (TBI), and stroke. 9,10 However, to our knowledge, there have been no analyses of the pattern of change, or any exploration of how patient characteristics influence interim change, in patients with hip fracture. A further limitation of previous studies of functional change during rehabilitation is that the studies examined change in the total FIM score, or in the motor and cognitive subscales of FIM. 6-8 The total FIM score distinguishes major gradients in functioning and is therefore valid and appropriate for use in estimating the total amount of care required. 11,12 The internal validity of the motor and cognitive subscales has also been supported in 20 different clinical conditions, including lowerlimb fractures. 11 The motor FIM, however, has items that assess independence in a wide range of activities, from eating and grooming, to sphincter control, to walking and stair climbing. When the functional changes in people with hip fracture are examined using only the motor subscale level, one might miss important but subtle changes in specific domains of physical functioning. Stineman et al 13 conducted factor analyses that found that FIM can be divided into 4 domains for people with lowerextremity fracture: activities of daily living (ADLs), sphincter management, mobility, and executive control. For people with hip fracture, functional independence in ADLs (eating, grooming, bathing, dressing, toileting) and mobility (bed to chair or wheelchair transfer, toilet transfer, tub or shower transfer, walking or wheelchair mobility, stairs) are likely to be most affected because the condition results in pain and loss of joint range of motion and strength in a lower limb. Although one might suspect that the rates and patterns of change would differ between ADLs and mobility, and that patient characteristics would be associated with rate of improvement, to date these factors have not been explored. Understanding the rate and pattern of change in different domains of function during

2 112 HIP FRACTURE FUNCTIONAL CHANGE PATTERNS, Latham rehabilitation may possibly enhance goal setting and discharge decisions, thus leading to improved outcomes. Our purpose in this study, therefore, was to examine the patterns of change in functional independence in patients with hip fracture across an episode of rehabilitative care in a skilled nursing facility (SNF). Specifically we compared the daily rate of change in FIM mobility and ADL domains during 2 distinct periods of care. We also explored whether patient characteristics such as age, medical complexity, length of stay (LOS), and baseline functional level contributed to different rates of functional change. METHODS Design This study was a secondary data analysis using data from an administrative dataset compiled and owned by SeniorMetrix Inc. The privately held company has contracts with health plans to assist them improve the quality and efficiency of rehabilitation services in SNFs for patients covered under Medicare Advantage reimbursement plans. Medicare Advantage allows Medicare beneficiaries to choose among health maintenance organizations (HMOs), provider-sponsored organizations, and other fee-for-service plans that provide care under contract to Medicare. Under the program, HMOs contract with Medicare to provide benefits in return for monthly per-person payments. The study was approved by the Institutional Review Board of Boston University. Participants The sample was derived from patients with hip fracture who were treated for the first time in 1 of 70 SNFs that provided care for health plans with contracts with SeniorMetrix from March 1998 to February Ninety-two percent of the sample was admitted from an acute care facility. All patients had a diagnosis of hip fracture as indicated by their International Classification of Diseases, 9th Revision, coding and had at least 2 FIM measurements taken after the baseline measurement. Patients who died, or who were admitted to an acute care facility from the SNF, were not included. Additionally, patients whose FIM measurements were taken less than 4 or more than 8 days between the first and second or second and third FIM measurements, were not included. We limited the data set in these ways to create a somewhat homogeneous sample and to obtain FIM data that were collected from the patients within similar time frames. The sample, therefore, consisted of 351 patients; their characteristics are provided in table 1. Procedure Data for all patients were collected regarding demographics, medical complexity at admission, and functional status at baseline (within 72h of admission), 4 to 8 days after baseline (mean standard deviation, ), and 4 to 8 days after the second measurement (mean, ). The medical complexity score was based on clinicians evaluations of patients medical conditions other than the primary diagnosis, and whether the conditions were relevant, active, and/or likely to influence functional change during the SNF episode of care. Physicians, nurses, and therapists were trained and credentialed in the use of the medical complexity scale and the score was based on team consensus. The scale is scored from 0 to 5, with 0 indicating no disease other than the primary condition, and 5 indicating a moribund condition. Concurrent validity of the scale is suggested by the fact that in patients with stroke, lower Table 1: Patient Characteristics Characteristic Values Sex Male 101 (28.8) Female 250 (71.2) Age (y) (36.8) (63.2) Prior living setting Not at home 69 (19.7) At home 282 (80.3) Medical complexity (23.6) (53.3) (23.1) Baseline ADL stage Stage (49.3) Stage (50.7) Baseline executive stage Stage (21.1) Stage (78.9) Baseline mobility stage Stage (64.1) Stage (35.9) LOS (wk) (30.8) (38.7) (30.5) Mean FIM interval 1 (d) Mean FIM interval 2 (d) Mean mobility domain FIM Baseline measurement Second measurement Third measurement Mean ADL domain FIM Baseline measurement Second measurement Third measurement NOTE. Values are n (%) or mean standard deviation. function, as determined by patients function-related group, 14 was related to greater medical complexity (Spearman.23, P.001). The scale s reliability has not been formally tested; however, every clinician who collected data was trained and credentialed (by SeniorMetrix) in the use of the scale and had passed a written examination based on 10 case studies. Functional status data included FIM scores. Measures FIM instrument. The FIM is a measure of function in 13 motor areas and 5 cognitive areas. There are 18 items, each rated on a scale of 1 to 7, with 7 equating to total independence. Various subscales may be derived based on the scores for each item. Stineman et al 15,16 have described 4 domains for FIM (ADLs, sphincter management, mobility, executive control) and within each domain, 7 stages of functional independence. Each stage is based on FIM item scores and approximates the average amount of assistance the patient needs and the amount of effort required by the patient. The validity and reliability of the FIM instrument have been documented in previous studies. 11,17,18

3 HIP FRACTURE FUNCTIONAL CHANGE PATTERNS, Latham 113 Data Analysis We selected independent variables that previous studies have suggested might be related to functional recovery after hip fracture. Age was categorized into 2 groups ( 80y or 80y). Baseline function was categorized based on functional independence stages for the mobility, ADLs, and executive function domains described by Stineman et al. 15 Categories for ADL and executive function variables included stage less than 3 and stage 3. Because few patients were classified at higher mobility stages at baseline, the categories for mobility were stage 1 and stage 1. Medical complexity was categorized into 3 categories using the medical complexity scale ( 3, 3, 3; range, 0 5). LOS was categorized into 3 groups ( 2wk, 2 3wk, 3wk). We categorized quantitative independent variables to enhance interpretation and clinical usefulness. Categories were based on both the distributions of our data and our assumptions about the clinical relevance of certain cut-points. The dependent variable was determined by first adding the individual FIM scores for the items in each domain at each measurement. With 5 items in the mobility domain, scores could range from 5 to 35. Scores for the ADL domain, with 6 items, could range from 6 to 42. Each domain s baseline score was subtracted from the score obtained at the second measurement. The remainder was then divided by the number of days within the first time interval to provide a measurement of change per day, or rate of change. Similarly, the score obtained for each domain at the second measurement was subtracted from the score obtained at the third measurement, and the remainder divided by the number of days in the second time interval. To compare the rate of change in FIM across the 2 time periods, and to examine the factors related to any differences in the rate of change, we conducted 2 repeated-measures analyses of variance, with either mobility or ADL FIM score change per day as the dependent variable. Measurement interval was included in the models as a 2-level within-subjects variable (interval 1, baseline to second measurement; interval 2, second to third measurement). Age, baseline executive function stage, mobility, or ADL stage (depending on the dependent variable), LOS, medical complexity, and whether the patient was living at home prior to the hip fracture were added to the model as between-subject variables. We examined main effects for each variable and interactions between the within-subject and between-subject variables. For independent variables with more than 2 levels, we conducted post hoc analyses (.05 for each analysis). For each analysis, we estimated adjusted means for rates of change in FIM domain scores for the 2 measurement intervals and for the levels of each of the between subjects variables. These data are displayed in figures 1 and 2. RESULTS Rate of FIM Mobility Domain Change per Day The rate of change per day for the FIM mobility domain was fairly constant in the first and the second phases of rehabilitation (mean change in FIM points per day,.46 vs.49) with no difference found between the 2 time periods. There was an interaction (P.014) between rehabilitation interval and the participants baseline mobility stage (see fig 1). People with worse mobility at baseline had a greater rate of change during the second interval of rehabilitation compared with the first, whereas those with better mobility at baseline had the opposite pattern (ie, a slower rate of change during the second interval). There was no interaction between the rehabilitation interval and the other variables (table 2). However, main effects were found for Day per FIM Change Mobility Mean Adjusted FIM Measurement Interval 3 variables: baseline executive function stage (P.001), LOS (P.001), and medical complexity (P.028). People who had a higher executive function stage at baseline (ie, better cognitive and communication ability) had overall a greater rate of change in mobility per day (.57 FIM points per day) compared with people with worse cognition and communication at baseline (.38 points per day). The overall mobility domain change per day was different across all levels of LOS, with those with longer LOSs having smaller changes per day. Those with less medical complexity (0 2) displayed greater overall mobility domain change per day (.68 FIM points per day) than those with both greater levels of complexity (.47 domain points per day for those scoring 3,.27 domain points per day for those scoring 4 or 5). Differences in the patients ages and residential settings prior to admission were not significantly associated with a greater rate of change in mobility (see table 2). Rate of FIM ADL Domain Change per Day The daily rate of change in the FIM ADL domain was not constant between the 2 rehabilitation time intervals; there was modest difference (P.047) between the daily rate of change in the ADL domain in the first and second rehabilitation intervals. A greater rate of change occurred during the first interval compared with the second interval (mean change in FIM points per day,.55 vs.41). There were significant overall effects on daily rate of change in the FIM ADL domain for 4 other variables: baseline ADL function (P.017), baseline executive function (P.001), living setting prior to admission (P.032), and LOS (P.001) (table 3). People with better ADL function at baseline (ie, baseline ADL stages of 3 ) displayed less overall rate of change in ADL domain scores (.43 domain points per day) than those with lower baseline ADL stage (.54 domain points per day, see fig 2). In contrast to the pattern of change associated with baseline ADL stage, people with higher stages of baseline executive function (3 ) had a greater overall rate of change in ADL domain scores (.63 FIM points per day) than those with lower- 2 Baseline Mobility Stage 1 Stage >1 Fig 1. Adjusted mean mobility domain change per day in 2 time periods, by baseline mobility (adjusted for age, baseline executive function and mobility stages, LOS, medical complexity, and whether the patient was living at home prior to the hip fracture).

4 114 HIP FRACTURE FUNCTIONAL CHANGE PATTERNS, Latham Variable Table 2: Mobility FIM Domain Change per Day Mean 95% CI 95% CI Mean 95% CI 95% CI Interval 1 Lower Boundary Upper Boundary Interval 2 Lower Boundary Upper Boundary Measurement interval Age (y) Baseline mobility stage Baseline executive function stage* Medical complexity* LOS (wk)* Prior living setting Not at home At home Abbreviation: CI, confidence interval. *Significant main effects (P.05). Significant interaction (P.05). executive function stages (.33 FIM points per day). Those who were living at home prior to the fracture had a higher overall rate of change in ADL FIM scores (.54 domain points per day) than those who were not (.42 domain points per day). People who had a longer LOS (ie, 3wk) had less overall change in ADL domain score per day (.34) than those with less than 2 weeks or 2 to 3 weeks LOS (.50 and.60 domain points per day, respectively). There was no significant association between medical complexity or age and the rate of change in ADLs and, overall, there was no significant interaction between the time interval and any of these variables (see table 3). Variable Table 3: ADL FIM Domain Change per Day Mean 95% CI 95% CI Mean 95% CI 95% CI Interval 1 Lower Bound Upper Bound Interval 2 Lower Bound Upper Bound Measurement interval* Age (y) Baseline ADL stage* 1 or Baseline executive function stage* 1 or Medical complexity LOS (wk) Prior living setting* Not at home At home *Significant main effects (P.05).

5 HIP FRACTURE FUNCTIONAL CHANGE PATTERNS, Latham 115 Day per FIM Change ADL Mean Adjusted FIM Measurement Interval 2 Baseline ADL Stage 1-2 Stage 3+ Fig 2. Adjusted mean FIM ADL domain change per day in 2 time periods, by baseline ADLs (adjusted for age, baseline executive function and ADL stages, LOS, medical complexity, and whether the patient was living at home prior to the hip fracture). DISCUSSION We found that ADL and mobility functional independence domains had different patterns of change and patterns of characteristics associated with these changes. While the daily rate of change in the mobility domain was fairly constant over the 2 time intervals, the ADL domain had a modest decline in the rate of change in the second time interval. Given that the ADL and mobility FIM items are often combined within 1 overall motor FIM scale (includes mobility, ADL, and sphincter management items), the different patterns of recovery in the 2 domains are potentially masked when patients recovery is examined. For example, graphic representation of data by Bode and Heinemann 9 suggested a fairly linear increase in overall motor FIM score over an episode of care for patients with stroke, SCI, or TBI. The rate of change in ADL function appears to decline somewhat over time when controlled for potentially confounding factors, whereas the rate of change in mobility function appears to be similar during the 2 time intervals. These differences were not the result of a ceiling effect for either domain because no subjects in the final time interval reached the highest category for the mobility domain, and only 0.3% reached the highest interval in the ADL domain. The difference in the rate of change between the 2 domains may be due to the relative level of difficulty in performing the items within each domain, particularly for people recovering from hip fracture. Many of the ADL tasks require lower levels of motor function than do the mobility tasks. Also, the ADL tasks are also largely accomplished with the upper limbs and may not, therefore, be as greatly affected by the impairments associated with a hip fracture as are mobility tasks. Additionally, the rate of change in ADL scores was greater during both measurement intervals for subjects with lower levels of ADL function at baseline than for those with higher levels. Together, these findings may reflect an early emphasis in rehabilitation on the lower level activities such as dressing that comprise the ADL domain, particularly for those with a low level of function. Improvement in performing ADL tasks can be accomplished through therapy that instructs and assists patients in adapting to tasks, rather than addressing physiologic impairments such as strength and flexibility that may take considerable time to improve. The mobility tasks measured by the FIM, such as transfers and walking, not only require strength, balance, and flexibility, but also are directly affected by hip fracture, including pain and loss of joint motion. No other studies, to our knowledge, have examined the rate of change in function using the mobility and ADL domain scores that we examined. Further research, therefore, is needed to substantiate our findings. The interaction of the measurement interval with mobility stage at baseline suggests that the rate of change in mobility accelerates over time in patients with low mobility function at baseline, and decelerates in those with higher levels. The finding implies that improvement can be expected even for patients with hip fracture who have a low level of mobility independence at admission and initially show slow improvements. This result seems inconsistent with that of Adunsky et al, 19 who found that patients with hip fracture who started with a low total FIM score, showed a daily gain in that score that was less than that for patients with higher baseline FIM scores. In that study, however, only the mean daily rate of change during the first 3 weeks was reported, thereby potentially masking the increased rates of improvement we found. In our study, rates of change over time were related not only to baseline functional independence, but also to LOS, medical complexity, and independence in executive function. Similarly, Bode and Heinemann 9 and Ween et al 10 showed that the rate of improvement in function was related to LOS. Subjects with greater rates of improvement demonstrated shorter LOS. As noted by Bode and Heinemann, the analysis of data from patients without regard to LOS may lead to the inaccurate impression that rate of improvement plateaus from admission to discharge. Our findings are also, in part, consistent with those of Patrick et al, 6 who reported that comorbidity was related to total FIM change per day. The patients in their sample were, on average, aged 82 years and had a variety of conditions. In a sample of relatively younger patients with lower-limb joint arthroplasty, Lew et al 8 found that the existence of comorbidities was not related to the change in FIM score per day. These findings suggest that an interaction between age and comorbidity may affect the rate of recovery. We did not explore this interaction because the majority of the patients were more than 80 years old. There are some important considerations in the evaluation of our findings. First, the intervals during which the FIM measurements were taken varied across and within patients. We attempted to address this problem by limiting the dataset used for analyses to include only patients with 4 to 8 days between measurements. To some extent, the effects of the differences in the duration of the intervals were controlled because the change in FIM occurring during the intervals was divided by the number of days in the interval. This approach to normalizing scores, however, assumes that the change in FIM is linear, an assumption that may not be true. A second issue is that the FIM, and particularly the domain scores, may not be sensitive to changes over short periods of time. Bode and Heinemann, 9 however, addressed each of these issues. Controlling for LOS, they used weekly measures and demonstrated a fairly linear increase in FIM motor scores in most groups of patients. A second consideration is that the differences in FIM domain rates of change across the various levels of the between subjects factors, although statistically significant, were small. It is unclear whether those small differences are clinically meaningful. We believe, however, that recognizing the differences in rates of improvement in ADL and mobility function, and differences in rates of improvement based on patient characteristics, is important to allow clinicians to improve the focus of their patient management, including realistic discharge planning and goal setting over time. For example, the information could influence clinicians in

6 116 HIP FRACTURE FUNCTIONAL CHANGE PATTERNS, Latham their goal setting for patients who have poor mobility at admission and whose rate of improvement seems slow the first week; the clinician may anticipate greater changes during the second week. This also suggests that the use of predictive tools that assume linear change in the calculation of FIM scores are useful, 20 at least for predicting mobility function at discharge. Clinicians could use this knowledge to ensure that they carefully monitor and rapidly progress the patient s goals so that clinicians continue to provide an adequate challenge throughout the rehabilitation episode. The difference in slopes of recovery between initial and later stages of the skilled episode may also be helpful to therapy providers for caseload management. The limitations of this study also include the use of secondary data. These data were generated for purpose of clinical and utilization management decision making, and using them for research purposes requires cautious interpretation. The limits of external validity must also be considered. The study sample consisted only of patients covered under Medicare Advantage for their SNF stay. The patients appeared to have basic demographic characteristics similar to a large national sample of patients treated in SNFs in There are reports, however, that patients covered in Medicare managed care organizations have less therapy, 22,23 smaller changes in function, 22 and shorter LOS, 22,23 than patients reimbursed under fee-for-service. These differences may affect the relations we describe. Additionally, the SNFs represented in the study sample are from only 4 regions of the United States and LOS and patient mix may reflect local market tendencies. Future research is necessary to replicate our analyses in patients with other types of conditions, in other rehabilitation settings, and with a broader array of third-party payers. We believe that use of the mobility and ADL domains of the FIM as separate subscores may be helpful to clinicians and encourage analyses using those subscores to enhance decision making. CONCLUSIONS The pattern of functional change over time differed for ADL and mobility domains, and for specific groups of patients. Rate of change in mobility is similar during the first and second approximately 6 days of rehabilitation in an SNF, whereas the rate of change in ADLs appears to be greater during the first approximately 6 days than during the second 6 days. We found that people with severe baseline mobility impairments experience a slower rate of mobility improvement in the initial stage of rehabilitation, but the rate of change increases in the second phase. This pattern is reversed for patients with higher levels of mobility at baseline. For both domains, executive function and LOS were associated with the overall rate of change. The use of domainspecific FIM scores allows a more precise understanding of the functional changes taking place over time during hip fracture rehabilitation. Acknowledgment: We acknowledge the contribution of Marie Jette, who provided assistance with the database used for the analyses in this study. References 1. Hannan EL, Magaziner J, Wang J, et al. Mortality and locomotion 6 months after hospitalization for hip fracture. JAMA 2001;285: Sloan FA, Taylor DH, Picone G. Costs and outcomes of hip fracture and stroke, Am J Public Health 1999;89: Cumming RG, Nevitt MC, Cummings SR. Epidemiology of hip fractures. Epidemiol Rev 1997;19: Parker MJ, Handoll HH, Dynan Y. Mobilisation strategies after hip fracture surgery in adults. Cochrane Database Syst Rev 2002; (2):CD Cameron ID, Handoll HH, Finnegan TP, Madhok R, Langhorne P. Co-ordinated multidisciplinary approaches for inpatient rehabilitation of older patients with proximal femur fracture. Cochrane Database Syst Rev 2001;(3):CD Patrick L, Knoefel F, Gaskowski P, Rexroth D. Medical complexity and rehabilitation efficiency in geriatric inpatients. J Am Geriatr Soc 2001;49: Adunsky A, Lusky A, Arad M, Heruti RJ. A comparative study of rehabilitation outcomes of elderly hip fracture patients: the advantages of a comprehensive orthogeriatric approach. J Gerontol A Biol Sci Med Sci 2003;58: Lew HL, Lee E, Date ES, Zeiner H. Influence of medical comorbidities and complications on FIM change and length of stay during inpatient rehabilitation. Am J Phys Med Rehabil 2002;81: Bode RK, Heinemann AW. Course of functional improvement after stroke, spinal cord injury and traumatic brain injury. Arch Phys Med Rehabil 2002;83: Ween JE, Mernoff ST, Alexander MP. Recovery rates after stroke and their impact on outcome prediction. Neurorehabil Neural Repair 2000;14: Stineman MG, Shea JA, Jette A, et al. The Functional Independence Measure: tests of scaling assumptions, structure and reliability across 20 diverse impairment categories. Arch Phys Med Rehabil 1996;77: Stineman MG. Measuring casemix, severity, and complexity in geriatric patients undergoing rehabilitation. Med Care 1997;35: JS Stineman MG, Jette A, Fiedler R, Granger C. Impairment-specific dimensions within the Functional Independence Measure. Arch Phys Med Rehabil 1997;78: Jette DU, Warren RL, Wirtalla C. Rehabilitation in skilled nursing facilities: effect of nursing staff level and therapy intensity on outcomes. Am J Phys Med Rehabil 2004;83: Stineman MG, Ross RN, Fiedler R, Granger CV, Maislin G. Staging functional independence validity and applications. Arch Phys Med Rehabil 2003;84: Stineman MG, Ross RN, Fiedler R, Granger CV, Maislin G. Functional independence staging: conceptual foundation, face validity, and empirical derivation. Arch Phys Med Rehabil 2003;84: Ottenbacher KJ, Hsu Y, Granger CV, Fiedler R. The reliability of the functional independence measure: a quantitative review. Arch Phys Med Rehabil 1996;77: Dodds TA, Martin DP, Stolov WC, Deyo RA. A validation of the functional independence measurement and its performance among rehabilitation inpatients. Arch Phys Med Rehabil 1993;74: Adunsky A, Levenkrohn S, Fleissig Y, Arad M, Heruti RJ. Rehabilitation outcomes in patients with full weight-bearing hip fractures. Arch Gerontol Geriatr 2001;33: Warren RL, Wirtalla C, Leibensberger A. Preliminary observations on reduced utilization in skilled nursing facility rehabilitation. Am J Phys Med Rehabil 2001;80: Iwanenko W, Fiedler R, Granger CV, Lee M. The Uniform Data System for Medical Rehabilitation: report of first admissions to subacute rehabilitation for Am J Phys Med Rehabil 2001; 80: Kramer AM, Kowalsky JC, Lin M, Grigsby J, Hughes R, Steiner JF. Outcome and utilization differences for older persons with stroke in HMO and fee-for-service systems. J Am Geriatr Soc 2000;48: Angelelli JJ, Wiber KH, Myrtle R. A comparison of skilled nursing facility rehabilitation treatment and outcomes under Medicare managed care and Medicare fee-for-service reimbursement. Gerontologist 2000;40: