The Effect of the Introduction of a Case-Mix−Based Funding Model of Rehabilitation for Severe Stroke: An Australian Experience

Article Outline

• Abstract
• Methods
  • Participants
  • Measures
  • Statistical Analysis
• Results
  • Length of Stay
  • Discharge Destination
  • Functional Status at Admission and Discharge
  • Change in FIM Motor Score and FIM Motor Gain per Day
  • Relationship Between LOS and Improvement in FIM Motor Score
  • Comparative Analysis of the Participating Units
  • Comparison With Previously Published National and International Studies
• Discussion
  • Study Limitations
• Conclusions
• Acknowledgment
• References
• Copyright

Abstract 

Brock KA, Vale SJ, Cotton SM. The effect of the introduction of a case-mix–based funding model of rehabilitation for severe stroke: an Australian experience.

Objective

To compare resource use of, and outcomes for, rehabilitation for severe stroke before and after the implementation of the Casemix and Rehabilitation Funding Tree case-mix−based funding model.

Design

Prospective, observational cohort study.

Setting

Eight inpatient rehabilitation centers in Australia.

Participants

Consecutive sample of 609 patients with severe stroke.

Interventions

Not applicable.

Main Outcome Measures

Rehabilitation length of stay (LOS), discharge destination, and FIM instrument motor score at discharge.

Results

The average rehabilitation LOS changed significantly between the preimplementation year and the implementation year (Mann-Whitney U, P=.001). There were no significant differences in discharge destination. FIM motor score at discharge showed significant reduction in improvement (Mann-Whitney U, P=.001) between the preimplementation year and the implementation year. There were no significant correlations between LOS in rehabilitation and gain in function for either the preimplementation year (Spearman ρ, P=.07) or the implementation year (P=.15).

Conclusions

The change in funding model was associated with a decrease in inpatient costs and with an associated increase in disability at discharge. Our results suggest that the rate of improvement in severe stroke is variable; also, they support the use of funding models for stroke rehabilitation that allow flexibility in resource allocation.

Key Words: Rehabilitation, Stroke, Treatment outcome

THE INTRODUCTION OF case-mix–based funding to health care in Australia has been a major focus for the last 15 years. The Australian health care system provides free hospital care for all Australians. The purpose in introducing case-mix−based funding was to reimburse hospitals more fairly through a system that allocates higher levels of payment for the more complex and difficult cases.¹ In rehabilitation, case-mix−based funding has been viewed as a way to achieve greater efficiency and equity in allocation of resources.²

Rehabilitation presents specific challenges to the development of case-mix−based funding models. Compared with the acute care sector, in rehabilitation there tends to be a smaller number of cases accumulating a relatively larger number of bed days and the degree of severity and resource use within any one diagnostic group may be greater.³ Case-mix−based funding models also have the potential to affect access to rehabilitation by providing financial incentives for admitting patients who are more likely to be profitable, rather than those who may receive significant benefits from the rehabilitation process.⁴, ⁵

In 2001, the Casemix and Rehabilitation Funding Tree (CRAFT) model was implemented in Victoria. CRAFT has 2 classifications for stroke, split on the Modified Barthel Index score (15-item version).⁶ Payment is based on the average length of stay (LOS) for the class, with short stay cases having a higher per diem rate and longer stay cases having a lower per diem rate. There is currently no cap on the per diem payments. The funding model is designed to provide an incentive to enhance efficiency while allowing some variability in resource allocation so that rehabilitation units can meet the requirements of patients without incurring unacceptable levels of financial risk. Before CRAFT was introduced, shadow funding was implemented for 1 year, during which units were funded according to CRAFT, with top-up payments to match historical funding, if required.

The introduction of CRAFT caused a level of concern among clinicians. The patients believed to be most at risk of adverse outcomes were those with severe stroke. There is a potential for these patients to have less access to rehabilitation and/or have less successful outcomes because of pressures for shorter LOS. Therefore, our purpose in this study was to compare resource use and outcomes for rehabilitation of severe stroke in the year before CRAFT was implemented, the year of shadow funding, and the year of implementation. We also examined the variability in resource use in those 3 years and the relationship between resource use and functional gain. We then compared the results of this data set with previously published national and international studies.

Back to Article Outline

Methods 

Participants 

Eight of the 25 designated rehabilitation services in Victoria provided prospective data for the 3 years. Seven centers were located in Melbourne, Australia, with representation from all the metropolitan health regions. One regional center participated. The participating units were a representative sample of rehabilitation units in Victoria, inclusive of rehabilitation delivered in an aged-care environment and in rehabilitation centers with younger patients. Participating units were asked to forward data on all patients who met the criteria of recent severe stroke, even if data were incomplete, to ensure having data for every admission. We defined severe stroke as a FIM instrument⁷ motor subscale score of 46 or below (range, 13−91). This level has been identified in 2 major studies⁸, ⁹ as the level of function associated with an increase in resource use in rehabilitation. All participating units obtained approval from their relevant ethics committee for submission of deidentified data without a requirement for informed consent.

Measures 

We used LOS both in rehabilitation and in the acute hospital as indicators of resource use. The FIM⁷ was used to determine severity of stroke on admission to rehabilitation and as an outcome measure. The FIM has been extensively investigated as to its reliability¹⁰ and validity,¹¹, ¹², ¹³, ¹⁴ and is used to benchmark rehabilitation services in the United States.¹⁵ The FIM has motor and cognitive components that measure separate domains of function.¹³ We used the motor subscale (FIM motor), a measure of independence in activities of daily living, mobility, and continence, to measure admission and discharge status. In Victoria, some rehabilitation units collect Modified Barthel Index data, rather than the FIM. In this study, Barthel Index scores were mapped to FIM scores using the conversion chart developed by the Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP), utilizing a deterministic mapping approach derived from a comparative study of FIM motor and Barthel Index scores.⁹

To facilitate analysis of outcomes, FIM motor scores on discharge were categorized into 3 levels: good, fair, and poor outcomes. A “good” outcome was defined as a patient achieving a FIM motor score of 65 or above (maximum, 91). With a score of 65, patients usually require either supervision or minimal assistance with mobility and self-care,¹⁶ indicating that the patient’ physical care requirements for daily activities are minimal. Scores above 46 indicate some improvement (fair outcome) and scores under 46 indicate a large physical burden of care at discharge (poor outcome).

Change in FIM motor score was calculated as a measure of improvement during rehabilitation. The calculation of FIM motor change scores has been validated by extensive research into the scaling properties of the FIM.¹³ FIM motor gain per day (FIM change divided by LOS) was also calculated as a measure of efficiency. Data regarding discharge destination were collected.

Statistical Analysis 

Data were screened and examined for assumptions of normality, using a variety of techniques including visual inspection of histograms and scatterplots, examination of skewness, kurtosis, and the Kolmogorov-Smirnov statistic. Positive skewness was demonstrated for data regarding LOS in rehabilitation, LOS in the acute hospital, FIM motor score at admission, FIM motor score at discharge, and change in FIM motor score. Negative skewness was demonstrated for FIM motor gain per day. The assumption of normality was rejected for all variables. Because data did not conform to normal Gaussian distributions, the Mann-Whitney U test (reported as z) was used for examination of continuous and scale-dependent variables and chi-square was used for categorical dependent variables. Significant chi-square test statistics were followed up with examination of the chi-square table cell’s standardized residuals (z).

Descriptive data of mean, median, standard deviation (SD), 25th and 75th interquartiles, and minimum and maximum were prepared for each year of the study. Descriptive data for the year of shadow funding were included to enable assessment of the consistency of the data over time. Data from the preimplementation and implementation years were compared, using the Mann-Whitney U test, to investigate whether the change in funding model had influenced LOS in rehabilitation, LOS in the acute hospital, FIM motor score at admission, FIM motor score at discharge, change in FIM motor score, or FIM motor gain per day. The categorized function at discharge and discharge destination data were investigated using chi-square analysis, followed by analysis of standardized residuals. We examined the relation between change in FIM motor scores and LOS using scattergrams and descriptive data. The correlation between change in FIM motor and LOS was tested using Spearman ρ. Alpha was set at the .05 level for all analyses.

Back to Article Outline

Results 

Six hundred nine patients with severe stroke were admitted to the rehabilitation units in the 3-year period. Fifty-four percent were men. Table 1 shows demographic and medical data for each of the 3 years.

Table 1. Demographic and Medical Data for Participants in Each of the 3 Years of the Study

NOTE. Values are mean ± SD or as otherwise indicated.

Length of Stay 

Table 2 shows the distribution of scores for LOS in rehabilitation in each of the 3 years. There was a significant difference between the preimplementation year and the implementation year (z=−3.23, P=.001), with longer LOS in the preimplementation stage. There was no significant difference in average LOS in the acute hospital between the preimplementation year and the implementation year (z=−.61, P=.54) (see table 1).

Table 2. LOS in Rehabilitation (in days)

Value	Preimplementation	Shadow Funding	Implementation
Mean ± SD	64.6±41.2	67.9±41.7	53.9±34.7
Median (range)	57.5(4–381)	60.0(6–278)	45.5(5–259)
25th percentile	37	40	29
75th percentile	79	89	68

Discharge Destination 

Table 3 shows the discharge destination for each year. To compare discharge destination, we combined the 2 categories of “home alone” and “home with family.” The “other” category, inclusive of discharges to acute hospitals, was excluded. There was no significant relationship between discharge destination and year of the study ( test=8.81, P=.07).

Table 3. Discharge Destination

Destination	Preimplementation	Shadow Funding	Implementation	Total
Home alone	26(12.1)	16(7.9)	11(6.1)	53(8.9)
Home with family	117(54.4)	134(66.3)	105(58.0)	356(59.5)
Low-level care⁎	20(9.30)	24(11.9)	14(7.7)	58(9.7)
High-level care†	40(18.6)	20(9.9)	32(17.7)	92(15.4)
Other	12(5.6)	8(4.0)	19(10.5)	39(6.6)
Total	215(100.0)	202(100.0)	181(100.0)	598(100.0)‡

NOTE. Values are n (%).

Functional Status at Admission and Discharge 

The average FIM motor score at admission for the total sample was 35.8±7.6. There was no significant difference between the preimplementation year and the implementation year with respect to admission FIM motor scores (z=1.19, P=.24). Table 4 shows the distribution of FIM motor scores at discharge for each of the 3 years. There was a significant difference in FIM motor scores at discharge between the preimplementation and the implementation year (z=−3.19, P=.001).

Table 4. Distribution of Scores for FIM Motor

We further examined functional status at discharge by categorizing outcomes as a good outcome (score of ≥65), moderate (47−64), or poor (≤46). Chi-square analysis indicated that there was a relationship between the year of study and outcomes ( test=14.9, P=.005). Examination of standardized residuals showed that the significant differences pertained to the implementation year data, with a lower than expected number of patients achieving a score of 65 and a higher than expected number scoring between 47 and 64 (table 5).

Table 5. Categorized FIM Motor at Discharge

FIM Motor Score	Year 1⁎	Year 2†	Year 3‡	Total
≤46	47(22)	35(17.5)	43(23.5)	125(20.9)
47to64	42(19.6)	45(22.5)	60(32.8)§	147(24.6)
≥65	125(58.4)	120(60.0)	80(43.7)§	325(54.4)
Total	214(100.0)	200(100.0)	183(100.0)	597(100.0)

NOTE. Values are n (%).

Change in FIM Motor Score and FIM Motor Gain per Day 

Table 4 shows the distribution of change in FIM motor scores and FIM gain per day for each of the 3 years. The data for change in FIM motor score revealed a significant difference between the preimplementation year and the implementation year (z=−3.54, P<.001), with patients in the implementation year improving less. There was no significant difference in FIM gain per day between the preimplementation year and the implementation year (z=−.88, P=.38).

Relationship Between LOS and Improvement in FIM Motor Score 

There was no significant correlation between LOS and improvement in FIM motor score for either year (preimplementation year: Spearman ρ=.12, P=.07; implementation year: ρ=.11, P=.15). The scattergrams showed considerable improvement in function for both short- and long-stay patients and similarly poor gains in function for both short and long stay patients. It is possible that the relation between LOS and improvement in function may have been skewed by the patients who remained in the rehabilitation units for a long period awaiting placement for high level residential care. We repeated the correlation analysis for the cohort, excluding patients discharged to high level care, leaving 291 cases. With the high level cases removed, there was a weak yet significant correlation between the 2 variables for the preimplementation year (ρ=.19, P=.02) and no significant correlation for the implementation year (ρ=−.12, P=.17).

Comparative Analysis of the Participating Units 

We examined the data from each participating center to determine whether the results of the analysis reflected changes across most units, or a subset of units. Six of the 8 units showed a 10% or greater decrease in mean LOS, with half of these units having a decrease of more than 20%. One unit had a decrease of less than 10% and another had an increase of 10%. To determine the variability of LOS, we examined the range of SDs as a proportion of the mean for each unit. In the preimplementation year, the SD ranged between 34% and 66% of the mean. In the implementation year, it ranged between 50% and 61% of the mean. We also examined the correlation between LOS and improvement in FIM motor score for each of the participating units (combined data from the 3 years). None of the units showed significant correlations between LOS and improvement in FIM motor score (P range, .08−.72). It is notable that the total number of patients dropped from 218 in the first year to 188 in the third year. Two units showed a marked reduction in the number of patients admitted with severe strokes, 1 unit having half the number of patients admitted in the first year with severe stroke and 1 unit having one third the number.

Comparison With Previously Published National and International Studies 

To compare this data set with previous studies, we reanalyzed the data according to the classification models used in 3 major studies: the AN-SNAP,⁹ the Uniform Data System for Medical Rehabilitation’s FIM-FRG,¹⁷ and the Copenhagen stroke study.¹⁸ Comparison with these classification models required restratifying the Victorian sample according to age and/or severity of disability to match the criteria in each of the 3 studies. The classification model for AN-SNAP utilizes 2 classes for stroke, split according to age.⁹ For the FIM-FRG’s classification, sufficient data were available from the Victorian sample for comparison with band STG-1 only (the lower band of stroke; FIM motor <38 and age <74y).¹⁷ The other FIM-FRG bands require FIM cognition scores (not available in this data set) or include higher functioning cases.

The Victorian data were also analyzed by severity of disability, as utilized in the Copenhagen stroke study.¹⁸ Within the parameters of our study, the Copenhagen study had 2 stratifications: “severe” (Barthel Index score between 25 and 45) and “very severe” (Barthel Index score <25).¹⁸ The “very severe” classification was not considered here because the Copenhagen study was a population-based methodology, whereas cases in this study were selected as having potential for rehabilitation. For “severe” stroke, our data were converted from FIM motor data to Barthel Index data, using the AN-SNAP method discussed previously.⁹ Table 6 shows the results of the Victorian sample when stratified according to the classifications in the original studies.

Table 6. Comparative Results of the Victorian Sample With 3 Previously Published Classification Systems

Back to Article Outline

Discussion 

This study demonstrated that the introduction of the CRAFT case-mix−based funding model in Victoria was associated with a decrease in resource use as measured by LOS in rehabilitation. This represents a significant reduction in the costs of rehabilitation for severe stroke. The reduction in costs, however, was associated with a reduction in outcomes. Patients discharged from rehabilitation after CRAFT was introduced required more assistance with everyday tasks than in the year before implementation. There was no improvement in rehabilitation efficiency as demonstrated by FIM motor gain per day.

Overall, the results indicate that there were no significant changes in access to rehabilitation for patients with severe stroke. Analysis of data from each participating unit, however, suggests that changes in the selection of patients for rehabilitation may have occurred in 2 of the 8 units, resulting in decreased access to rehabilitation for patients with severe stroke in those geographic areas.

In this study, we found higher levels of disability at discharge from inpatient care after the CRAFT model was introduced. After inpatient rehabilitation, all participating units provide ambulatory services at varying levels. It is unknown whether patients in the implementation year continued to improve after discharge to achieve the higher levels of outcome seen in the preimplementation year. Follow-up studies for severe stroke in the context of the new funding model are indicated.

In this study, there was no significant relationship between LOS and improvement in FIM motor score. Short and long stay cases attained similar discharge functional status and discharge disposition. This may be indicative of differing rates of recovery, which have an impact on LOS. Bode and Heinemann¹⁹ found different rates of recovery in cases with similar FIM scores on admission, with slower recovery linked to longer LOS. Similarly, the neurologic impairments underlying functional status can affect outcomes and the rate of recovery. Han et al⁵ found that rate of recovery of subjects in specific FIM-FRG classifications was significantly different in relation to the presence of impairments in motor, somatosensory, and/or visual systems. The family support available to the patient, the home environment, and access to outpatient rehabilitation services are also major considerations in the timing of discharge.¹⁹ In this study, the variability in LOS remained high after introduction of the CRAFT model, which indicates that the practice of matching resources to clinical indications continued postimplementation.

To place the Victorian model in context with national and international resource use and outcomes, the Victorian data were compared with previously published data from the AN-SNAP study, an Australia wide review of rehabilitation data,²⁰ from the FIM-FRG in the United States,¹⁷ and the Copenhagen stroke study.¹⁸

Compared with the AN-SNAP²⁰ data, for both younger and older stroke cases, the median FIM motor discharge score was higher, with a longer LOS in rehabilitation and a higher rate of discharge home (see table 6). There was a similar pattern when we compared the younger, more severe cases in the Victorian sample with the data from the FIM-FRG classification.¹⁷ Comparison of the Victorian study with the Copenhagen stroke study¹⁸ for “severe” stroke (Barthel Index score between 25 and 45) yielded very similar data between the 2 studies with regard to outcomes and discharge destination. The Copenhagen study does not provide data regarding LOS according to level of initial disability; however, subjects with severe stroke, as measured by the Scandinavian Stroke Scale, had mean hospital LOSs (acute and rehabilitation combined) of stroke survivors ranging between 60 and 100 days.²¹ Mean LOS, acute and rehabilitation combined, for the Victorian sample was 76 days. While it is difficult to draw conclusions from these comparisons, the review suggests that access to longer periods of rehabilitation for some subjects with severe stroke is associated with better outcomes overall.

Study Limitations 

Limitations of this study relate to both deficiencies in the data acquired and in lack of information about additional variables. In Victoria, rehabilitation units use both the FIM and the Modified Barthel Index as primary outcome measures, necessitating a recoding of data by an inexact method. There is inherent error in this process, but we believe that this is unlikely to have affected the main conclusions of the study. More detailed data collection regarding discharge destination for cases that were rated in the “other” category would clarify any effects on this outcome measure. In future studies, it would be useful to capture data about the areas of impairment resulting from the stroke and the social supports available to the patient, to map the rate of recovery of functional ability over time, and to follow cases for 12 months poststroke.

The results of this study suggest that the length of rehabilitation required to achieve good outcomes is highly variable. An optimal funding model would have incentives to contain costs and, at the same time, enable providers to vary the resources used to meet the needs of individual patients. Extended rehabilitation stays carry a significant cost burden, however; discharge at an earlier stage of rehabilitation may increase the proportion of patients who need to be discharged to high-level care, or increase the level of carer burden for patients who are discharged home. Because so many stroke patients are cared for by their elderly spouses, this increased burden of care may pass to a very vulnerable group in the community. We acknowledge the necessity of rationing expenditure on rehabilitation. At the same time, care must be taken to ensure that measures introduced to reduce costs do not result in increased costs in residential or community care.

This study has shown that among severe stroke survivors, there is considerable variability in the time required to achieve functional gains. These findings support the approach taken by the CRAFT funding model to maximize the flexibility of its clinical application while still having a positive effect on costs. This is in contrast to funding models that provide fixed payments based on diagnosis and initial disability, such as the United States inpatient rehabilitation prospective payment system,²² or models that base payment on rate of functional gain, as proposed by Saitto et al.²³

The primary goal of rehabilitation is to restore function. Complete recovery from severe stroke is rare and treatment is aimed at both achieving independence and decreasing carer burden. If changes are made in the provision of rehabilitation services, this big picture must be kept in focus to ensure that the effectiveness of the changes is evaluated by long-term outcomes for patients and their families so that the most effective means of decreasing disability in the community can be clarified.

Back to Article Outline

Conclusions 

The CRAFT case-mix−based funding model introduced into Victoria was associated with a decrease in the use of resources and in a small reduction in patient outcomes at discharge. There was a high degree of variability in LOS before the funding model was implemented and this variability was maintained postimplementation. This suggests that the rehabilitation service providers varied resource allocation according to the individual requirements of their patients. We found a low association between hospital LOS and functional gain in severe stroke, which suggests that rates of recovery were variable. The ability to allot available resources according to individual patient requirements is an important consideration in developing appropriate funding models for rehabilitation in severe stroke.

Back to Article Outline

Acknowledgment 

We thank the Leadership and Management Group and the Neurology Group of the Australian Physiotherapy Association (Victorian Branch) for their support.

Back to Article Outline

References 

1. Department of Health, Housing, Local Government and Community Service. Casemix: a new direction in health care management (The National Casemix Education Series, Part A). Canberra: Australian Institute for Health and Welfare; 1993;
   • View In Article
2. Duckett S, Gray L, Howe A. Designing a funding system for rehabilitation services (Part 1: Rationale and recent developments). Aust Health Rev. 1995;18:30–44
   • View In Article
   • MEDLINE
3. Duckett S, Gray L, Howe A. Designing a funding system for rehabilitation services (Part 2: Policy objectives and options for achieving efficiency and quality of care). Aust Health Rev. 1995;18:62–77
   • View In Article
   • MEDLINE
   • CrossRef
4. Gagnon D, Nadeau S, Tam V. Clinical and administrative outcomes during publicly-funded inpatient stroke rehabilitation based on a casemix group classification model. J Rehabil Med. 2005;37:45–52
   • View In Article
   • MEDLINE
   • CrossRef
5. Han L, Law-Gibson E, Reding M. Key neurological impairments influence function-related group outcomes after stroke. Stroke. 2002;33:1920–1924
   • View In Article
   • CrossRef
6. Granger CV, Albrecht GL, Hamilton BB. Outcomes of comprehensive medical rehabilitation: measurement by PULSES Profile and the Barthel Index. Arch Phys Med Rehabil. 1979;60:145–154
   • View In Article
   • MEDLINE
7. Uniform Data System for Medical Rehabilitation. Guide for use of the Uniform Data Set for Medical Rehabilitation including the Functional Independence Measure. Buffalo: Research Foundation, State University of New York; 1995;Version 4
   • View In Article
8. Stineman MG, Escarce JJ, Goin JE, Hamilton BB, Granger CV, Williams SV. A case-mix classification system for medical rehabilitation. Med Care. 1994;32:366–379
   • View In Article
   • MEDLINE
   • CrossRef
9. Eager K, Gordon R, Hodkinson A, et al. The Australian National Sub-Acute and Non-Acute Patient Classification (AN-SNAP): report of the National Subacute and Nonacute Casemix Classification Study. Wollongong: Centre for Health Service Development, University of Wollongong; 1997;
   • View In Article
10. Ottenbacher KJ, Hsu Y, Granger CV, Fiedler RC. The reliability of the Functional Independence Measure: a quantitative review. Arch Phys Med Rehabil. 1996;77:1226–1232
    • View In Article
    • Abstract
    • Full-Text PDF (987 KB)
    • CrossRef
11. Granger CV, Cotter AC, Hamilton BB, Fiedler RC. Functional Assessment Scales: a study of persons after stroke. Arch Phys Med Rehabil. 1993;74:133–138
    • View In Article
    • MEDLINE
12. Heinemann AW, Linacre JM, Wright BD, Hamilton BB, Granger CV. Prediction of rehabilitation outcomes with disability measures. Arch Phys Med Rehabil. 1994;75:133–143
    • View In Article
    • MEDLINE
13. Linacre JM, Heinemann AW, Wright BD, Granger CV, Hamilton BB. The structure and stability of the Functional Independence Measure. Arch Phys Med Rehabil. 1994;75:127–132
    • View In Article
    • MEDLINE
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;77:1101–1108
    • View In Article
    • Abstract
    • Full-Text PDF (948 KB)
    • CrossRef
15. Granger CV, Ottenbacher KJ, Fiedler RC. The Uniform Data System for Medical Rehabilitation: report of first admissions for 1993. Am J Phys Med Rehabil. 1995;74:62–66
    • View In Article
    • MEDLINE
    • CrossRef
16. Fiedler RC. The Rasch measurement model. UDS Update. 1993;7:1–10
    • View In Article
17. Stineman MG, Fiedler RC, Granger CV, Maislin G. Functional task benchmarks for stroke rehabilitation. Arch Phys Med Rehabil. 1998;79:497–504
    • View In Article
    • Abstract
    • Full-Text PDF (4790 KB)
    • CrossRef
18. Jorgensen HS, Nakayama H, Raaschou HO, Vive-Larsen J, Stoier M, Olsen TS. Outcome and time course of recovery in stroke (Part 1: Outcome. The Copenhagen study). Arch Phys Med Rehabil. 1995;76:399–405
    • View In Article
    • Abstract
    • Full-Text PDF (817 KB)
    • CrossRef
19. Bode R, Heinemann A. Course of functional improvement after stroke, spinal cord injury, and brain trauma. Arch Phys Med Rehabil. 2002;83:100–106
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (508 KB)
    • CrossRef
20. Lowthian P, Disler P, Ma S. The Australian National Sub-acute and Non-acute Patient Casemix Classification (AN-SNAP): its application and value in a stroke rehabilitation programme. Clin Rehabil. 2000;14:532–537
    • View In Article
    • MEDLINE
    • CrossRef
21. Jorgensen H, Nakayama H, Raaschou H, Olsen T. Acute stroke care and rehabilitation: an analysis of the direct cost and its clinical and social determinants. Stroke. 1997;28:1138–1141
    • View In Article
    • MEDLINE
    • CrossRef
22. Stineman M. Prospective payment, prospective challenge. Arch Phys Med Rehabil. 2002;83:1802–1805
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (58 KB)
    • CrossRef
23. Saitto C, De Bernardini L, Di Benedetto P, et al. A new system for the prospective payment of inpatient rehabilitation: a summary of the research project: “Development of an iso-resource classification for the prospective payment of rehabilitation care.”. Eur Medicophysica. 2003;39:205–209
    • View In Article