|
The cost effectiveness of post-acute care for elderly Medicare beneficiaries |
Chen Q, Kane R L, Finch M D |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology Post-acute care (PAC) was studied under four different settings. These were home without formal care, home care, a nursing home (skilled nursing facility, SNF) and a rehabilitation facility.
Type of intervention Rehabilitation and treatment.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised patients participating in a study of PAC, which was conducted by the University of Minnesota.
Setting The setting was tertiary care and the community. The economic study used data on PAC services in the USA.
Dates to which data relate The effectiveness data were obtained for the period 1988 to 1989. Since expenditures on PAC were used, it was not possible to break down the cost data into resource utilisation and unit costs. The PAC costs related to the period 1987 to 1990. The price year was not reported.
Source of effectiveness data The effectiveness data were obtained from a single study of the PAC conducted by the University of Minnesota.
Link between effectiveness and cost data The costing study was conducted retrospectively on the same patient sample as that used in the effectiveness analysis.
Study sample The sample of patients used in this study was drawn from an analysis of PAC conducted by the University of Minnesota. The patients were enrolled during 1988 and the follow-up took place during 1989. From the University of Minnesota's study, the authors of the current study included Medicare beneficiaries, aged 65 years and over, who were suffering from one of the 5 DRGs listed in the 'Hypothesis/Study Question' field. The study focused on patients from 51 hospitals in 3 cities (Twin Cities, Pittsburgh, and Houston), chosen for the differences in hospital and PAC settings. One hundred and eleven patients were excluded from the analysis. These patients represented small sub-groups who experienced CHF or hip procedures, and who were sent to SNFs or rehabilitation facilities. The reason for their omission was the impact of the small sample size on the stability of the overall results. The overall sample size was 2,137.
The numbers in each group were reported according to DRG and discharge location. The sample included 2,137 patients. Of the 487 patients who had a stroke, 160 were discharged home, 125 to home care, 123 to a nursing home and 70 to rehabilitation facilities. Of the 336 patients suffering from COPD, 209 were discharged home and 127 received home care. Of the 444 patients with CHF, over 62% (276) were discharged home and 168 received home care. Of the 264 patients who underwent a hip procedure, 183 patients were discharged home and 81 to home care. Of the 606 patients treated for hip fracture, 120 were discharged home, 115 to home care, 282 to a nursing home, and 89 to rehabilitation facilities. The authors also analysed a homogeneous sub-group of 557 patients, of which 86 were discharged home, 130 to home care, 240 to a nursing home, and 101 to rehabilitation facilities.
Study design The study was a multi-centred, retrospective non-randomised study. Patients from three different cities in the USA participated in the study. The three cities (Twin Cities, Pittsburgh and Houston) were chosen for their different hospital and PAC structures. The four different types of PAC were available in each of the three cities. The patients were interviewed at hospital discharge, with follow-up interviews conducted at 6 weeks, 6 months and one year after discharge for patients and their carers.
Analysis of effectiveness The analysis on which the clinical study was based (intention to treat or treatment completers only) was not explicitly reported. However, the retrospective nature of the study implies that data were collected for all of the patients in the study. The health outcome used in the analysis was functional status. The change of functional status was measured by the activities of daily living (ADL). In particular, the ADL score was created using the weighted magnitude estimation technique. This converted ADL scores to a scale between 0 and 100. A score of 0 indicated complete dependence or death, while a score of 100 indicated no disability. As the patients were not randomised to the treatment arms, significant differences were evident in patients undergoing different types of PAC. The authors corrected for this selection bias using the method of instrumental variables.
Effectiveness results In general, the analysis indicated that, at 6 weeks after discharge, patients sent home with formal home care experienced a larger improvement in ADL (taking account of selection bias), irrespective of DRG.
The improvement in ADL resulting from home care ranged from 11.2 to 29.1% at 6 weeks, from -2.3 to 32.3% at 6 months, and from -21.3 to 32.7% at one year after discharge.
The greater functional improvement in patients receiving home care was also evident at the 6-month and 1-year follow-ups.
Compared with the other forms of PAC, stroke patients discharged to an SNF experienced the lowest improvement in ADL at all three follow-up time periods. The improvements were 0.86% at 6 weeks, -5.9% at 6 months and -6.6% at one year.
Compared with home care, home without formal care resulted in a lower improvement in ADL for patients with COPD, CHF, or hip procedures. At 6 weeks, the improvements were 3.1% (COPD), -0.6% (CHF) and 20.7% (hip procedures), respectively. At 6 months, the improvements were -5.2% (COPD), -14.7% (CHF) and 29.3% (hip procedures). Finally, at one year, the improvements were -11.3% (COPD), -27.9% (CHF) and 30.8% (hip procedures).
Interestingly, patients recovering from a hip fracture experienced the lowest improvement in ADL under home care for the first two follow-up periods (18.3% at 6 weeks and 21.7% at 6 months). However, a year after discharge, the hip-fracture patients sent to SNFs had the lowest ADL improvement (17.2%).
The authors did not test the statistical significance of the differences across types of PACs or DRGs.
Clinical conclusions According to the analysis of the effectiveness data, at 6 months and one year following discharge, home care resulted in higher functional improvement for all five DRGs, compared with the other discharge locations.
Modelling Due to the open, non-random nature of this study, the authors found significant differences between the characteristics of the patients. For example, at the time of hospital discharge, the patients discharged to SNFs or rehabilitation facilities were more dependent than those sent home with or without formal care. The method of instrumental variables was used to correct for this selection bias. The first step of this procedure involved the estimation of the PAC discharge location using a multinomial logit model. This estimated variable, together with all but one of the independent variables used in the multinomial logit equation, was then included in ordinary least-squares regressions to estimate the costs and the outcomes of the PAC.
Measure of benefits used in the economic analysis The measure of benefits used was the improvement in functional outcome, as measured by the ADL score. The total ADL score was calculated using the weighted magnitude estimation technique. This had the advantage of assigning different valuations to different disabilities.
Direct costs Medicare expenditures were obtained from the Medicare Automated Data Retrieval System (MADRS) for the period 1987 to 1990. This period covered one year prior to admission to hospital and one year following discharge. Where data were missing from MADRS, the authors estimated Medicare payments by multiplying the resource use from MADRS by the costs obtained from the literature. However, despite using this method, the authors did not analyse the resource utilisation and the costs separately. The total costs of PAC were calculated as the sum of Medicare and Medicaid payments, payments by other insurance, and out-of-pocket payments. The costs were not discounted.
Statistical analysis of costs No statistical analysis of the costs was included.
Indirect Costs The indirect costs were not reported.
Sensitivity analysis Due to the correlation between functional outcomes and costs, the authors conducted a sensitivity analysis of the incremental cost-effectiveness ratio (CER). Taylor's approximation method was used to calculate the 95% confidence interval of the CER.
Estimated benefits used in the economic analysis The incremental benefits were measured in terms of the change in the improvement in ADL in one PAC setting relative to another. At 6 weeks after discharge, home care resulted in an additional 24.74 percentage point improvement in ADL for stroke patients relative to SNF. The incremental benefit of home care was lower relative to home without formal care (12 percentage points) and rehabilitation (1.8 percentage points). In general, although the absolute improvement in ADL for stroke patients in each of the PAC settings declined at the 6-month and 1-year follow-ups, the incremental benefit of home care was largest when compared with SNF. The incremental benefit of home care relative to home without formal treatment ranged from 8.2 percentage points for COPD patients, to 11.8 percentage points for CHF patients, to 6.8 percentage points for hip procedure patients. Furthermore, the incremental benefit to home care generally remained positive for these groups of patients at the 6-month and one-year interviews. The exception was COPD patients discharged to home with no formal care at the one-year follow-up, who experienced a relatively smaller fall in ADL. At 6 weeks after discharge, the incremental benefit to hip fracture patients of home care was largest when compared to home without formal care (10.8 percentage points). The incremental benefit of home care was lower relative to discharge to a SNF (8.6 percentage points) and a rehabilitation facility (4.8 percentage points). The pattern of incremental benefits was similar at 6 months and one year following discharge, although the incremental benefit of home care was largest when compared with SNF.
Cost results The Medicare and total costs of the different PAC regimes were calculated for the 5 DRGs. The total costs, which included Medicare and Medicaid expenditures, other insurance payments and out-of-pocket expenses, were greater than the Medicare payments. The costs were reported for the three follow-up time periods (6 weeks, 6 months and one year after discharge).
At 6 weeks after discharge, home without formal care resulted in lower total costs for three of the five DRGs. These were $3,103 for stroke patients, $1,504 for COPD patients, and $447 for patients that underwent a hip procedure.
The least costly option in terms of Medicare costs for CHF and hip fracture groups was home with formal care. The costs were $1,989 for CHF patients and $2,283 for hip fracture patients.
The most expensive PAC type for stroke and hip fracture patients was rehabilitation, with Medicare costs amounting to $12,210 in the stroke group and $11,256 in the hip fracture group.
At 6 months following discharge, home care resulted in cost-savings for stroke and hip fracture patients. These amounted to savings of $1,077 in Medicare payments for the stroke group and $815 in total expenditures. The corresponding savings for the hip fracture group were $1,701 (Medicare payments) and $1,300 (total expenditures). However, home care for COPD, CHF and hip procedure patients resulted in an increase in the range of $74 to $1,865 in Medicare costs, and in $435 to $2,177 in total costs.
Rehabilitation was still the most costly form of PAC for stroke and hip fracture patients at 6 months and one year after discharge, although the costs were higher than those at 6 weeks.
Cost-savings from home care were still evident for stroke and hip fracture patients one year after discharge, amounting to $73 and $3,772 for Medicare payments.
Home care for hip fracture patients still resulted in the lowest total costs ($8,967). However, the cheapest PAC regime in terms of the total costs for stroke patients was home without formal care ($12,850).
For the COPD, CHF and hip procedure patients, home care was more costly than home without formal care. The difference ranged from $1,222 to $4,312 for Medicare payments and from $1,814 to $5,703 for the total payments.
Synthesis of costs and benefits The incremental CER was used to compare the costs and the benefits of each PAC type. In general, the incremental CER was higher when the costs included Medicaid, other insurance and private expenditures, compared with Medicare payments alone. This ratio was calculated for each DRG at the three follow-up time periods for the Medicare costs and the total costs. For each DRG, discharge home with formal care resulted in a greater improvement in functional outcome, sometimes at lower costs. At 6 weeks, the incremental CER for home care (relative to the cheaper alternative, home with no formal care) ranged from -$42 to $115. At 6 months, the incremental CER for home care ranged from -$211 to $245. At one year, the incremental CER for home care ranged from -$832 to $955. For stroke patients, discharge to an SNF resulted in a lower improvement in ADL, at higher costs, and so was dominated. The sensitivity analysis was undertaken using Taylor's method of approximation to calculate the 95% confidence interval of the incremental CER. The results were robust to this sensitivity analysis.
Authors' conclusions Compared with post-acute care (PAC) in skilled nursing facilities (SNFs) or rehabilitation facilities, home care was more effective and less costly for stroke and hip fracture patients. Further, the results indicated that formal home care was more cost-effective than discharge to home without formal care. The cost-effectiveness ratios (CERs) showed mixed support for PAC in rehabilitation facilities. Discharge to an SNF was dominated by the strategy of discharge home with no formal home health care. Overall, the authors deduced that higher costs do not necessarily lead to the greatest improvement in functional outcome.
CRD COMMENTARY - Selection of comparators Instead of explicitly specifying a particular comparator, the authors compared the four main forms of PAC that represented standard practice in the USA at the time of the study. Although, as the authors comment, the data analysed in the study do not capture recent developments in PAC provision. The authors did not use recent data and, therefore, acknowledged that they may not have identified new developments in the types of PAC.
Validity of estimate of measure of effectiveness The use of the improvement in ADL as the measure of benefit and effectiveness was justified, given the insensitivity of quality-adjusted life-years as a measure of effectiveness in an elderly sample and also the inability to adjust the latter for selection bias. The ADL score was calculated using a weighted magnitude estimation technique, which implies that different values were assigned to different disabilities. Given the retrospective nature of the study, the authors' analysis of the characteristics of the sample indicated that the patients were not identical. In light of this selection bias, the measure of effectiveness was adjusted for selection bias using the instrumental variable method. Analysis of a sample of homogeneous patients yielded results similar to those obtained from the larger sample. This suggests that the authors have successfully corrected for selection bias.
Validity of estimate of measure of benefit No summary benefit measure was used.
Validity of estimate of costs In their analysis of the costs, the authors examined Medicare payments and also total payments, which included Medicare, Medicaid, other insurance and price expenses. All the categories of cost relevant to the perspective adopted were included in the analysis. The costs and the quantities were not reported separately. Therefore, the authors did not examine the implications for resource use for each DRG, under each of the four PAC settings. The indirect costs were not considered. These may be significant, particularly if the patients are receiving informal care when discharged to home without formal home health care.
Other issues This study was novel in its examination of the costs and cost-effectiveness of different types of PAC. Thus, comparisons with similar studies were not possible. In terms of generalisability, the authors expressed concern that, due to the restriction of the sample to three cities, the results of the study may not be applicable to the entire US population of elderly Medicare beneficiaries. In providing the results of a combined analysis of the sample from across the three cities, the authors assumed that there were no differences between the PAC settings in the three cities. However, it may have been interesting to examine any differences in the costs and the effectiveness of PAC between the three cities. The authors considered the initial PAC location as the final discharge location for the patient. In effect, some patients moved several times between PAC settings, which may have had implications on the costs and the effectiveness. While the actual number of patients moving to different PAC settings may have been small (a precise number was not reported), the authors examined the consequences of changing this assertion and concluded that similar results were obtained. The statistical significance of the difference in the effectiveness or costs was not tested.
Implications of the study On the basis of the results of their study, the authors propose that an integrated single PPS, linking payments for both acute and PAC for a specific DRG, be implemented. This is instead of the current system of separate PPSs for PAC.
Source of funding Supported by the Health Care Financing Administration (Cooperative Agreement No 17-C-98891) and the Office of the Assistant Secretary for Planning and Evaluation, US Department of Health and Human Services.
Bibliographic details Chen Q, Kane R L, Finch M D. The cost effectiveness of post-acute care for elderly Medicare beneficiaries. Inquiry: a Journal of Medical Care Organization Provision and Financing 2000; 37(4): 359-375 Indexing Status Subject indexing assigned by NLM MeSH Aftercare /economics; Aged; Arthroplasty, Replacement, Hip /economics /rehabilitation; Cost-Benefit Analysis; Diagnosis-Related Groups /economics; Heart Failure /economics /rehabilitation; Hip Fractures /economics /rehabilitation; Humans; Lung Diseases, Obstructive /economics /rehabilitation; Medicare /economics; Outcome Assessment (Health Care); Prospective Payment System /economics; Research Support, U.S. Gov't, Non-P.H.S.; Research Support, U.S. Gov't, P.H.S.; Stroke /economics /rehabilitation; Subacute Care /economics; United States AccessionNumber 22001000671 Date bibliographic record published 31/12/2002 Date abstract record published 31/12/2002 |
|
|
|