SEARCH

SEARCH BY CITATION

Keywords:

  • care processes;
  • diagnosis-related groups;
  • cost per patient;
  • cost per bed-day

SUMMARY

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

Background and purpose

With the “graying” of the population, hip fractures place an increasing burden on health systems and call for efficient forms of care. The aim was to compare two models of organizing hip fracture care at one university hospital working at two sites. The differences in organization were coordinated care provided in one of the sites and traditional care, divided between different institutions, in the other.

Material and methods

The study was conducted at a Swedish university hospital and included all 503 hip fracture patients, admitted during the 1-year period of February 2009 through January 2010. Patient gender, age, type of fracture, admission and discharge dates were documented. The patients were surveyed of their health-related quality of life at the time of admission and at 4 and 12 months after discharge. The costs for the inpatient care episode were estimated using three costing methods.

Results

The coordinated care model resulted in a shorter hospital stay and consistently lower costs. There was no difference between patient-reported quality of life.

Interpretation

The care of hip fracture patients coordinated by a geriatric ward throughout the whole care episode is more cost-efficient than uncoordinated where patients are transferred to other institutions for rehabilitation. © 2014 The Authors. The International Journal of Health Planning and Management published by John Wiley & Sons Ltd.


INTRODUCTION

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

It is a widely held view that hip fracture patients, as a patient group, account for a disproportionately high share of healthcare expenditures and therefore pose a problem in the management of healthcare systems (Autier et al., 2000; Braithwaite et al., 2003; Lawrence et al., 2005; Borgstrom et al., 2006; Gehlbach et al., 2007; Azhar et al., 2008). Various programs and procedures in healthcare have arisen to try to manage this problem, for example, public healthcare programs, new prevention strategies such as osteoporosis treatment, use of hip protectors, behavioral and lifestyle changes and new surgical methods (Kannus and Parkkari, 2006; Gehlbach et al., 2007).

Not a great deal is known about the effects on patient outcomes and resource utilization of different models for organizing hip fracture care. For example, in some cases, care provided in acute care hospitals may be separated from rehabilitation, which is provided by other institutions. In other instances, acute care and post-operative care are provided in a coordinated way. In the latter case, several medical and surgical specialties are involved, including orthopedic surgery, general medicine, geriatrics and rehabilitation. These speciality services are usually provided in various departments of the hospital in various combinations (Adunsky et al., 2002; Gholve et al., 2005; Chong et al., 2009; Miura et al., 2009).

Several international studies on hip fracture care report on efforts to shorten the length of stay (LOS) (Huusko et al., 1999; Hommel et al., 2008; Lofgren et al., 2010; Lippuner et al., 2011), to introduce integrated pathways (Olsson et al., 2009) and to reduce costs (Miura et al., 2009). These efforts are all intended to improve care for the patients as well as decrease the costs of their care. Frequently used metrics for comparing models for hip fracture care are treatment outcomes and treatment costs.

Patient outcomes should preferably be assessed from the patient's perspective. Measuring patient outcome as an improvement in health-related quality of life (HRQoL) is one way to make that assessment. Several outcome studies have used the instrument EuroQol to measure HRQoL (Brooks, 1996; Tidermark et al., 2003; Borgstrom et al., 2006; Ekstrom et al., 2009). Other studies have collected information on costs and calculated cost–utility ratios (Borgstrom et al., 2006; Strom et al., 2008).

A simple proxy for resource utilization is LOS (Miura et al., 2004; Garcia et al., 2012). Compared with detailed cost calculations, LOS is less specific but can be retrieved from hospital information systems. The added value of cost calculations over LOS when comparing different models of care has not been systematically studied.

The comparison of hip fracture care costs is a challenging task. One reason is that accounting principles and procedures differ between hospitals as well as between regional and national healthcare systems. There are three main methods used to calculate such costs: cost per bed-day (CPB), cost per patient per episode (CPP) and cost of the medical “product” using diagnosis-related group (DRG) categories. To our knowledge, there are no studies that compare the cost of hip fracture care using all three methods.

The primary aim of this study was to compare two models for organizing hip fracture care by measuring their outcomes and costs using LOS, HRQoL and direct costs for an inpatient episode as metrics.

MATERIAL AND METHODS

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

Study location

This study was performed in a Swedish university hospital with two sites, 30 km apart. In this study, the sites are labeled site 1 and site 2. Site 1 organizes hip fracture care as a coordinated care episode that includes the acute admission, surgery, post-operative care and rehabilitation (Figure 1). After the hip fracture is diagnosed, the patient is admitted to the geriatric ward at the hospital, sent to the operation theater for treatment and returned to the geriatric ward for rehabilitation. The patient is thereafter discharged back home or if necessary to a new housing. A hip fracture patient at site 2 is admitted to the orthopedic ward, which organizes the operation and immediate post-operative care (Figure 2). From this ward, the patient is transferred to a rehabilitation unit or geriatric ward outside the acute care hospital, and those units discharge the patient to home or residential care.

image

Figure 1. The patient path in site 1 and number of living patients discharged. Deceased n = 8. A&E, accident and emergency department; GW, geriatric ward; RD, radiology department; OpT, operations theaters

Download figure to PowerPoint

image

Figure 2. The patient path at site 2 and number of living patients discharged. Deceased n = 1. A&E, accident and emergency department; RD, radiology department; OW, orthopedic ward; OpT, operations theaters; GW, geriatric ward; RI, rehabilitation unit

Download figure to PowerPoint

The figures give an overview of the different care paths for patients at sites 1 and 2, respectively.

Patients

This study reports on all hip fracture patients from the catchment area of a Swedish university hospital, diagnosed and treated during the 1-year period February 2009 through January 2010. The following hip fracture patients were excluded from the study: patients under 65 years of age, patients with pathological fractures, patients with comorbidities that prohibited surgical treatment and patients who died before surgery. Of the 503 patients in the study, 285 patients were admitted to the geriatric ward at site 1, and 218 patients were admitted to the orthopedic ward at site 2 (Table 1).

Table 1. Gender, age and fracture types of cervical hip fracture patients treated at sites 1 and 2 (n = 503) (descriptive and bivariate analysis)
VariablesSite 1 (n = 285)Site 2 (n = 218)χ2dfp-value
  • df, degree of freedom; SD, standard deviation.

  • *

    χ2 for trend.

Gender  0.110.76
Female19668.8%14867.6%   
Male8931.1%7032.1%   
Age, years  0.58* 0.45
Age, mean (SD) 83.4 (8.6) 83.2 (8.4)   
Age group 65–745117.9%3817.4%   
Age group 75–8410035.3%6128%   
Age group 85–9410536.8%10447.7%   
Age group ≥952910.1%156.9%   
Fracture type  7.6950.17
Cervical not dislocated (garden1–2)3411.9%3013.7%   
Cervical dislocated (garden 3–4)10637.1%8137.1%   
Basocervical62.1%10.5%   
Trochanteric two fragment4917.1%5123.3%   
Trochanteric multi-fragment6221.7%4118.7%   
Subtrochanteric289.8%146.4%   

Data on gender, age and fracture type were retrieved from the patients' medical records. There were no significant differences between these two groups of patients.

Outcome measures

The outcome measures used in this study were the patients' LOS, HRQoL and the costs measured by the three different methods. LOS was the number of days from enrolment to discharge for the whole care episode in the respective models. HRQoL was assessed using the first part of the EuroQol instrument, the health status profile, EQ-5D (Brooks, 1996), which is a standardized, non-disease-specific instrument. In a large British population, Doland et al (1996) used the Time Trade-Off method to rate the different states in the profile (UK EQ-5D index tariff). The EQ-5D has been used in several clinical studies for both hip fracture patients (Tidermark et al., 2002; Tidermark et al., 2003) and for patients with other diseases/illnesses (Borgstrom et al., 2006; Ekstrom et al., 2009), and demonstrating acceptable reliability and validity of the measure.

At admission to the hospital, the patients were asked to rate their HRQoL during the week preceding the hip fracture. At 4 and 12 months after discharge, the hospital sent the patients a questionnaire asking them to rate their HRQoL.

Cost calculations

The costs of hip fracture care at sites 1 and 2 were calculated by using the three different methods: as LOS times average ward-specific CPB, as cost for the corresponding DRG and as individual CPP (Sveriges kommuner och landsting (2009)). The cost of stay in geriatric hospitals for post-operative rehabilitation (patients from site 2) was calculated by DRGs and by their CPB. The cost of rehabilitation at out-of-hospital private rehabilitation units (for site 2 patients) was expressed as LOS times the per bed-day reimbursement stipulated in contracts with the purchaser. The university hospital provided the researchers with ward-specific CPB values. The CPB is the same for all patients in the same ward regardless of diagnosis.

In Sweden, the NordDRG version of the DRG patient classification system (Fetter, 1986; www.socialstyrelsen.se/EDP, 2009) is used to calculate the cost of in-hospital services. Cost weights are calculated for each NordDRG patient group based on a national sample of hospital cost data, which includes direct costs as well as internally allocated costs (Medin et al., 2011). The purchaser reimburses hospitals for discharged patients as the cost weight of the specific NordDRG group times the hospital's average cost per patient. Hip fracture patients are classified into four NordDRG groups: DRG 209, 210, 211 and 235. The reimbursement is the same for all patients in the same DRG group. The NordDRG data were retrieved from the Stockholm County Patient Care Register, which covers all care to all patients funded by the county. Acute care hospitals funded by Stockholm County report CPP data according to a uniform cost model. The hospitals use these data in budgeting; the county uses these data for benchmarking and pricing (Sveriges kommuner och landsting (2009)). CPP data were obtained from the Stockholm County Purchasing Office.

The Stockholm County Purchasing Office provided the data on reimbursements paid to the geriatric hospitals and independent rehabilitation units. Total costs for the whole care episode for patients treated at sites 1 and 2 were calculated using the three cost methods. The costs at site 1 were calculated for the whole patient episode at the hospital, whereas at site 2, the costs were calculated for the in-hospital stay and the care provided by the geriatric hospital and/or the rehabilitation unit.

It is essential when three methods of calculation are compared that all major cost items are included. Each cost model covers hotel, staff, laboratory, radiology, surgery, anesthesia, implant and drug costs.

Statistical methods

The statistical analysis was performed using SPSS 18.0 for Windows software (SPSS, Inc., Chicago, IL, USA). Nominal variables were tested by a chi-square test and ordinal variables by a chi-square for trend. An independent t-test was used to compare the LOS at the two sites. This test was conducted because the deviation from normality was slightly skewed. All tests were two-sided. Levene's test was used for unequal variance and Friedman's test when comparing the non-normally distributed dependent groups.

RESULTS

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

Length of stay

The observed LOS for the whole care episode was 3.9 days shorter at site 1, compared with site 2—a statistically significant difference. Patients typically had a longer in-acute-hospital stay at site 1 than at site 2 because at site 2, a high percentage of patients were discharged to the geriatric hospitals and rehabilitation units (Figure 2). Table 2 and Figure 3 illustrate the means and distributions of LOS at sites 1 and 2.

Table 2. LOS for the whole care episode at sites 1 and 2 and for the different parts of the care path at site 2
 MeanSDMedian25th percentile75th percentile
  1. Levene's test at sites 1 and 2 was significant, and the p-value for unequal variance was 0.04.

  2. LOS, length of stay; SD, standard deviation.

Site 1 (n = 285)13.06.4128.716.0

Site 2 (n = 218)

divided in the different parts (number of patients)

16.912.3147.022.0
Orthopedic ward (n = 218)7.24658
Geriatric ward (n = 64)20.3121713.526.5
Rehabilitation unit (n = 68)116.78714
image

Figure 3. The distribution of in-hospital days at sites 1 and 2

Download figure to PowerPoint

As shown in Figure 3, site 1 had a more compressed distribution of LOS than site 2.

Health-related quality of life

As expected, in a group with old hip fracture patients and with a follow-up time of 1 year, the response rates from the three patient HRQoL questionnaires were moderate (Table 3).

Table 3. Number of respondents to EQ-5D questionnaires and age distribution of the respondents who answered at the three measurements points
 Site 1 (n = 285)Site 2 (n = 218)
n%n%
Number of respondents to EQ-5D questionnaires at the three measurements points
At admission1425010950
At 4 months2137513461
At 12 months1956812557
On all three occasions115407836
Age distribution on all three occasions (years)
65–7429251925
75–8446402937
85–9431273038
≥95980 

An analysis of the two groups, answering and non-answering at the three measurement occasions, shows no significant difference between the two sites (Table 4).

Table 4. Analysis of observations in EQ-5D groups at the two sites, respondents (n = 115 + 78 = 193) and non-respondents (n = 170 + 138 = 308)
VariablesRespondentsNon-respondents
  1. Data include the descriptive and bivariate analysis of gender, age, mental status and mortality of the patients at the two sites.

  2. t-test was used for continuous variables, chi-square test for nominal and categorical variables and chi-square for trend for ordinal variables.

  3. df, degree of freedom; SD, standard deviation; HRQoL, health-related quality of life.

 Site 1 (n = 115)Site 2 (n = 78)t/χ2 *dfp-valueSite 1 (n = 170)Site 2 (n = 138)t/χ2 *dfp-value
Gender    0.0410.83    0.1610.69
Female7867.8%5469.2%   11768.8%9266.7%   
Male3732.2%2420.8%   5331.2%4633.3%   
Age, years              
Age, mean (SD)81 (8.7)80.1 (7.7)0.721910.4785.5 (8.2)85.2 (8.3)0.313060.75
Age group 65–742925.2%1924.4%0.07  2313.5%1813%0.49 0.48
Age group 75–844640%2937.1%   5130%3223.1%   
Age group 85–943130%3038.5%   7644.7%7352.9%   
Age group ≥9597.8%00   2011.8%1510.9%   
Mental status    3.2 0.07    3.6 0.06
Clear9582.6%5671.8%   8047.1%8058%   
Suspected of dementia/dementia2017.4%2228.2%   9052.9%5842%   
Mortality           1.2 0.27
During care episode       84.7%118%   
Within 120 days after discharge       3319.4%3223.2%   
Within 365 days after discharge       3017.6%2316.7%   
HRQoL (EQ-5D index score)    0.2610.60       
Baseline (at enrolment)0.664 0.648           
4 months after0.520 0.472           
12 months after0.563 0.528           

A Mann–Whitney test was performed to examine the differences in HRQoL measured by EQ-5D at sites 1 and at 2 at admission, and at 4 and 12 months after discharge We did not observe statistically significant differences between the two groups for any of the measurements with the exception of dimension 1 (mobility) at baseline (p = 0.03) where site 2 had better scores.

Costs

A detailed analysis of the cost structure at the unit level showed interesting differences. Care at rehabilitation units per day was significantly lower than at geriatric hospitals (SEK 2800 vs. SEK 4600). The DRG cost for the whole care episode at site 1 was 27% lower than at site 2. Using what is arguably the most detailed cost method (CPP) for the whole care episode, the difference in cost was 9%.

Care episode costs calculated by the three costing models at each site are presented in Table 5. The three methods were used as a sensitivity analysis in order to recognize the effect of using alternative counting procedures.

Table 5. Calculation of average costs (in Swedish crowns) per patient for the whole care episode at sites 1 and 2
Type of costSite 1 (n = 285)Site 2 (n = 218)
  1. Bank of Sweden 2009 average exchange rate: 1 euro = SEK 10.62.

Cost per bed-day76 288130 699
Diagnosis-related group92 028126 163
Cost per patient115 163124 879

A division of costs in the different parts as presented in Table 6 is only doable for CPP. Neither DRG nor CPB costs can be divided in its different part because they are based on the hospital's average costs.

Table 6. Costs divided in the different parts in cost per patient (CPP) (SEK)
ResourceMeanSDMd95%CI for MdQ 25Q 75
  1. Staff expenses are included for orthopedic surgery, anesthesia, laboratory and radiology and total cost per bed-day (CPB) for the different departments, hospitals and units. Ort, Orthopedic; Ger, Geriatric.

  2. Bank of Sweden 2009 average exchange rate: 1 euro = SEK 10.62.

Hotel20 00013 00016 20013 300– 20 60010 90027 800
Staff at ward (physicians, nurses, assistant nurses, occupational therapist and physiotherapists)37 20024 00030 00025 400–37 50019 80050 900
Orthopedic surgery22 30013 20019 60017 700–21 40014 40026 600
Implant4300540029002 300–290016005600
Anesthesia3900410031002700–350016005000
Laboratory15 600810012 80011 700–14 80010 30018 200
Radiology4000360031002900–350023003500
Drugs4400400033002900–390012005300
Total CPP111 700 91 000   
Total CPB Ort sites 1 and 211 800     
Total CPB Ger site 15800     
Total CPB Ger hospitals4600     
Total CPB rehabilitation units2800     

DISCUSSION

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

We compared two different models for organizing hip fracture care at two sites of the same university hospital. The study covered 503 hip fracture patients, 65 years of age and older and who were admitted to the hospital during a 1-year period. The study identified all stages inside and outside the hospital for the whole care episode. Comparisons were made of the LOS, the patients' assessment of HRQoL and the costs of the whole care episode using the three cost models.

The whole care episode was 4 days shorter at site 1 than at site 2. With a growing population segment of elderly people (Lofgren et al., 2010), it is important that the number of hospital beds is used efficiently. As shown in Figure 3, site 1 had a narrower distribution of bed-days in addition to shorter total LOS.

It is also interesting to make a comparison with historical data. In 2005, the average LOS for hip fracture patients at the seven emergency hospitals in the county was 17.6 days (Lofgren et al., 2010). Site 1 demonstrated a greater reduction than site 2. To move patients to other institutions during a care episode extends the care and thus makes it more expensive.

In this study, the difference in the EQ-5D scores showed no significant differences between the two sites at admission or at 4 and 12 months after discharge. Patients at both sites 1 and 2 felt that their HRQoL had significantly deteriorated 1 year after the fracture. This is an understandable result considering the age of the patients and may not necessarily have significance as far as the comparison of the two sites is concerned.

There were no significant differences in HRQoL measured at admission and at 4 and 12 months after discharge. The EQ-5D has previously appeared to be an easy-to-use instrument for elderly patients with hip fractures and can be a relevant clinical end point in outcome studies (Tidermark et al., 2002). At admission, 50% of the patients were reached for an EQ-5D assessment. At follow-up, response rates varied between 57% and 75%, with small differences between sites. Response rates in age groups declined with higher age, one possible explanation being the corresponding increase in comorbidities such as dementia.

The clinical outcome and the patients' experiences have to be balanced against cost-saving measures when seeking models for organizing hip fracture care efficiently. We found consistent differences in care episode cost as measured by the three cost models. The coordinated care model at site 1 costs less than the divided care model at site 2 in all comparisons. Average care episode costs were consistently lower at site 1 than at site 2, regardless of the costing method used. The cost difference ranged from 9% to 42%, depending on the cost method used. This difference is partly explained by a shorter care episode length at site 1 compared with site 2 (13.0 vs. 16.9 days).

In the coordinated care model at site 1, a geriatric ward at the university hospital took responsibility for the supervision of the whole care process, from admission through post-operative care and rehabilitation. This model contrasts with the divided care model at site 2 where patients were admitted to an orthopedic ward for surgery and, after a brief post-operative phase, were transferred to separate units for rehabilitation. Our conclusion is that when a single unit coordinates all care without any transfers to other units, a more cost-efficient care is achieved. Another reason for having a coordinated care at a geriatric ward may be that early discharge to a rehabilitation institution (i.e., the transfer itself) can be associated with an increase in hospital stay (Logters et al., 2008) and the need for patients to adjust to new conditions that may delay rehabilitation (Tha et al., 2009).

Several other studies have also shown positive clinical and economic outcomes where, as at site 1, the orthopedic surgeons and geriatricians work together (Miura et al., 2009; Gonzalez-Montalvo et al., 2010).

Costs of hip fracture care have previously been calculated in Sweden but not by the alternative costing methods as in this study (Stromberg et al., 1997; Borgstrom et al., 2006). Such costs have less frequently been calculated for the whole care episode. The three methods were applied in a sensitivity test to show the difference of a more precise costing approach, the micro-costing method (Drummond, 2005). Our calculations revealed that there is a broad range of costs per episode in hip fracture care, depending on the choice of costing method: from SEK 76 288 (EUR 7183) to SEK 130 699 (EUR 12 307). The largest cost difference (42%) was observed when we used CPB. However, CPB is a rough cost method and does not account for differences in patient mix, which can differ significantly between a geriatric department and an orthopedic department. In addition, these costs are strongly affected by LOS.

In Stockholm County, hospital providers are reimbursed for care provided based on the DRG payments, whereas the rehabilitation units are paid per bed-days. The hip fracture care cost for the whole care episode, during the study period, was 37% higher for the cases where the divided care model was used compared with the coordinated care model. By using the CCP method, the cost difference was far lower and amounted to 9%. Nevertheless, 9% is a comparatively large difference between two forms of organization and supports the conclusion that there are savings in the coordinated model. This result indicates that a more precise cost method reduces the cost differential, of interest to both providers and purchasers of care.

The most detailed cost method, aggregated costs per patients, was available only for hospital stays in the university hospital. CPP calculations, consequently, included the geriatric hospital and rehabilitation unit costs as DRG and CPP prices. Some of the cost variation observed indicates selective referral patterns. Patients with minor medical needs at site 2 were presumably sent to the rehabilitation units, and patients with greater medical needs were sent to the geriatric departments. That procedure may explain why LOS was shorter in the rehabilitation units than in the geriatric hospital wards.

We conclude that an organizational model with coordinated care for hip fracture patients has a shorter LOS over the care episode, has no negative effects on patient outcome as measured by HRQoL and is less costly than the divided care model. This reduction in LOS may have other positive effects on patients and healthcare than simply lower costs.

LIMITATIONS

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

This study has several limitations. One limitation is that the whole care episode does not include rehabilitation provided after discharge in long-term or residential care. In Sweden, utilization of those care arrangements and services, which are the responsibility of municipalities, is not included in the county databases. Still, the similar results in HRQoL between the two models indicate that there are similar needs for services and assistance for the elderly.

A second limitation is that the cost of medical treatment and care, like “product costing,” relies on cost estimations and cost allocation rules based on conventions, and do not express “true costs.” Nevertheless, the fact that our three cost methods showed consistency in the cost differences between the two care models suggests that our methodology was justified.

Finally, this study analyzes a comprehensive patient population. Although we found no differences in gender, age and type of fracture between patients at the two sites, a randomized controlled trial during a longer period is needed to establish whether our results on the benefit of the coordinated care model are valid more generally.

CONCLUSION

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

The study provides evidence that a coordinated care model covering the whole care episode from admission to rehabilitation is shorter and less costly, without compromising outcomes measured as HRQoL, compared with a divided care model.

ETHICS STATEMENT

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

This study was part of two larger studies that were approved by the Stockholm Regional Ethical Review Board (Decisions 2008/6 23-312006/13-31/1 and 2009/130-32). The study was conducted according to the Helsinki Declaration.

ACKNOWLEDGEMENTS

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

The authors gratefully acknowledge a grant from the Swedish Vinnvård research program and the assistance of the Stockholm County officials, especially Stig Hagström, who provided information on costs and prices; Thomas Emilsson for help with the EQ-5D calculations; and Mesfin Tessma for advice on the statistical analysis.

The authors report no conflicts of interest.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES

All authors contributed to the conception and planning of the study, the analysis of the data and the preparation of the manuscript. Two authors, SL and GL, collected the data.

REFERENCES

  1. Top of page
  2. SUMMARY
  3. INTRODUCTION
  4. MATERIAL AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. LIMITATIONS
  8. CONCLUSION
  9. ETHICS STATEMENT
  10. ACKNOWLEDGEMENTS
  11. AUTHOR CONTRIBUTIONS
  12. REFERENCES