Fourth, the best estimate of the national effect of the direct entry option on medicare part A expenditures is that SNF expenditures in 1980 would have increased by $46 million and hospital expenditures would have decreased by $49 million. The net change would have been a $3 million-0.014 percentreduction in medicare part A expenditures. A reasonable set of bounds on this estimate for the net effect for medicare would be between a $28 million savings and a $13 million cost. I would like to review briefly the evaluation methodology used in the expenditure analyses. The original goal of the evaluation was to assess separately the effects of demonstration programs operating in two different States. In conducting this type of evaluation, a number of methodological issues have to be resolved. A typical approach is to look at the immediate and primary consequences of the demonstration and was adopted here. The economic analyses gave primary consideration to two issues that have the most impact. Those were changes in hospital utilization and changes in SNF utilization. Estimates regarding these factors define the order of magnitude of the ultimate cost or savings. Thus, these two issues, hospital utilization and nursing home utilization, became the primary focus. Quantitative analyses of patient and facility level data were used to address medicare part A utilization and expenditure issues. The analyses considered the extent to which hospital days were avoided by use of the direct entry option, as well as the additional SNF days and expenditures involved. We estimated that approximately 41 percent of the direct entry patients in Massachusetts and 67 percent in Oregon would have been hospitalized in the absence of the direct entry option. The major cause for the differences between the Massachusetts and Oregon results involve the unusually high concentration of rehabilitation hospital/SNF's in Massachusetts and their disproportional involvement in the demonstration. Massachusetts also experienced a low number of SNF admissions directly from the patient's home-a likely consequence of relatively low levels of program awareness and physician involvement in nursing home care in general in Massachusetts. The national projections of medicare expenditure consequences of the waiver were directly based on the experience observed in the two States. As I indicated before, a range of estimates was prepared based on simulations. These simulations took into account the probability of hospitalization, the expected tradeoff between hospital and medicare SNF days, and average per diem medicare expenditures in 1980. Our view of the most likely effect of a nationwide direct entry option is a national saving of about $3 million under conditions that were observed in the demonstration. While there is potential for medicare savings by eliminating the prior hospital stay requirement, the ultimate effect is complicated and depends on a variety of additional factors. I will now review four of those. First, a fundamental point of my remarks is that each of the two demonstrations led to different results. While I feel confident that we have done a correct evaluation of the two demonstrations, a number of considerations make national projections very difficult. I cannot absolutely say whether the ultimate effect of the waiver involves a cost or a savings to medicare. There are a number of factors that could increase or decrease the national estimates presented in the final report. These include other legislation that you are now considering. The most critical issue is the extent to which the waiver diminishes days of care in hospitals, which are about four times as expensive as a day of care under SNF. The data used to estimate the number of avoided hospital stays were derived from discussions about the case histories of 276 randomly selected direct entry patients with the directors of nursing in 8 SNF's in Oregon and 10 in Massachusetts that were relatively high users of the direct entry option. Our national projections are based on a combination of the results observed in both States. Factors exist, however, which could cause either greater savings or potentially substantial greater costs as a result of a direct entry option. For example, increased physician awareness of the option or a change in reimbursement incentives could lead physicians to admit patients to SNF's rather than hospitals when appropriate. Conversely, less stringent interpretation or enforcement of the medicare SNF criteria could lead to a substantial increase in the number of medicare SNF stays without an offsetting decrease in hospital days. The important point is that savings can only be realized if approximately 50 percent or more of the direct entry patients would have been hospitalized without the waiver. Second, our national estimate did not reflect the additional costs which might accrue to medicare as a result of reductions in hospital days, lower overall occupancy rates, and consequently higher per diem costs. Since empty beds are the source of potential cost savings of a direct entry option, it is important to consider the cost of an empty bed to medicare. Such consideration may reduce the savings of the program below the amount reported in the evaluation. Our view was that the added cost could be ignored since the overall effect of the program on hospital occupancy would be quite small, and the medicare share of any average fixed costs would also be quite small. Medicare only pays a portion of the fixed cost of an empty bed. The portion that medicare pays can be approximated by medicare's share of total hospital reimbursement. Other complications involve variations in financing methods between the States. Medicare's financial burden for an empty bed will vary depending on medicare utilization, reimbursement practices of other payors, especially medicaid, the importance of fixed costs, and overall occupancy levels. Third, given the attractiveness of the medicare SNF benefit to the nursing facilities, it is reasonable to ask, “Why did the facilities not transfer more patients than they did onto the medicare SNF benefit using the direct entry option?" I believe that the major limiting factors are the stringency and enforcement of the level of care criteria. The level of care criteria were strictly enforced for the direct entry patients by the SNF staff, as well as the demonstration staff. If interpretation or enforcement of level of care criteria is loosened under a national program that allowed direct SNF entry, SNF's presumably would act on their preference for medicare over medicaid reimbursement and increase the number of internal transfers. Thus, if reductions in the funding of utilization review weaken enforcement of the level of care criteria, SNF utilization could increase beyond the level observed in the demonstrations. In addition, fiscally pressed medicaid programs may attempt to switch medicaid nursing home patients to medicare status. Fourth, physicians are also likely to affect the cost and utilization of the direct entry option. There is evidence that most physicians have little incentive to admit patients directly into an SNF when a hospital admission is possible. A hospital is generally a more convenient location for physicians to care for their patients. Moreover, medicare reimbursement for physicians hospital visits requires less paperwork than nursing home visits. The tradeoff between nursing home and hospital stays among waiver patients is only part of the potential effect of the waiver program. It was not possible to quantify all the benefits available from the direct entry option, particularly the benefits that patients and families gained from a more appropriate treatment pattern or the management flexibility available to SNF's. In summary, while elimination of the hospitalization requirement does have the potential for cost savings for medicare, these additional considerations lead to uncertainty about the ultimate national consequences. The most important qualification that I wish to place on the study findings involve extensions beyond the experience of Massachusetts and Oregon to the Nation at large. The future national long-term care environment is likely to differ considerably from that studied under the demonstrations. If the constraints on utilization in the demonstrations are eased, utilization of the direct entry option, as well as regular medicare SNF admissions, would increase. Thus, the waiver could cost substantial amounts of dollars. I would like to make a brief transition to talk about one of the other issues that have been brought up today. In concluding this testimony, I would like to comment briefly on the broader issues of nursing home reimbursement policies, which I have considered also in studies for HCFA about medicare nursing home reimbursement policies. I think the general conclusions are relevant for medicare as well. The general conclusion there that I reached was that appropriately designed prospective reimbursement systems can be developed to affect positively those dimensions of nursing home care that are central to policymaking. I discuss the various options in my book "Public Pricing of Nursing Homes" that is being published next month by Abt Books of Cambridge, Mass. Such a prospective reimbursement system based on cost analyses would foster groupings of patients according to their functional limitations by paying for the costs of their care accordingly.1 This view originates from the reluctance of nursing home operators to admit the more disabled and costly patients that account for a very large percentage of patients backed up in hospitals awaiting nursing home placement. If it is suitably designed, prospective reimbursement could reduce the costs of admission both to nursing home operators and to public agencies and lead to other desirable public goals. Thank you for allowing me this opportunity today. I will be happy to answer any questions. [The prepared statement and additional material follow:] STATEMENT OF HOWARD BIRNBAUM, PH. D., DIRECTOR OF AGING AND PUBLIC Mr. Chairman, I am Howard Birnbaum, the Director of Aging and Public Programs at Abt Associates Inc., in Cambridge, Massachusetts. The major issue about which I am pleased to provide testimony involves the cost implications of the elimination of the prior hospital stay requirement for Medicare reimbursement of skilled nursing care. At the end of my remarks, I also will comment briefly on some of the broader issues involving nursing home reimbursement. However, I first would like to introduce these comments. For the last seven years, I have been involved in evaluative studies and research for the Department of Health and Human Services (and its predecessor, the Department of Health, Education and Welfare), including several projects for the Health Care Financing Administration, HCFA. These studies have considered a variety of issues affecting both acute and long term care, as well as several associated income security programs. It is my involvement as Director of a recently completed study for HCFA on the prior hospital stay requirement that brings me here today. You are well aware of the growing concerns about cost, utilization, and quality issues surrounding nursing home care which affect so many beneficiaries of the Medicaid and Medicare programs and their families. The problems include high and continually rising expenditures, inappropriate placements, and inadequate types of care. Although expenditures for skilled nursing care are less than two percent of the total Medicare budget, Medicare patients may enter the hospital and incur unnecessarily expensive hospital stays because of entry requirements for the Medicare Part A skilled nursing benefit. One possible policy response that has been under congressional consideration over the past decade is the elimination of the three day prior hospital stay requirement for Medicare reimbursement of care in skilled nursing facilities (SNFs). In 1977, the Health Care Financing Administration established demonstration programs in Massachusetts and Oregon to investigate the consequences of eliminating the prior hospitalization requirement. In 1979, HCFA contracted with Abt Associates Inc. to conduct an independent evaluation of these direct entry SNF demonstrations. SUMMARY OF RESULTS The key results from the evaluation, completed in September 1981,1 are: 1. Direct entry SNF patients accounted for approximately 10 percent of the covered Medicare SNF admissions to participating demonstration SNFs. The dominant constraints on direct entry utilization were the effective strictness and enforcement of the Medicare criteria governing admissions to SNF care. 2. There were no differences in the equality of care provided to direct entry and regular SNF Medicare patients. Nor were there any patient outcome differences. 1 See Birnbaum, Howard et al. "Why Do Nursing Home Cost Vary?" Medical Care. November 1981, Vol. 19, No. 11. 2"Evaluation of the Three Day Hospital Stay Requirement for Medicare SNF Reimbursement." Abt Associates Inc., Cambridge, Mass., Sept. 30, 1981. AAI Report No. 81-76, HCFA Contract No. 500-79-0051. 91-156 0-82——5 3. The best estimate of the total net effect on Medicare Part A expenditures was a net saving of 0.1 percent of Medicare Part A expenditures ($182,000) in Oregon and a net cost of 0.02 percent ($122,000) in Massachusetts. 4. The best estimate of the national effect of the direct entry option on Medicare Part A expenditures is that SNF expenditures in 1980 would have increased by $46 million and hospital expenditures would have decreased by $49 million. The net change would have been a $3 million (0.014 percent) reduction in Medicare Part A expenditures. A reasonable set of bounds on this estimate for the net effect for Medicare would be between a $28 million savings and a $13 million cost. ESTIMATION ISSUES I would like to review briefly the evaluation methodology used in the expenditure analyses. The original goal of the evaluation was to assess, separately, the effects of demonstration programs operating in two different states. In conducting this type of evaluation, a number of methodological issues have to be resolved. A typical approach is look at the immediate and primary consequences of the demonstration and was adopted here. The economic analyses gave primary consideration to two issues: Estimates of the increased costs to Medicare due to an increase in Medicare SNF stays; and Estimates of potential savings due to avoided hospital utilization. Estimates regarding these factors define the order of magnitude of the ultimate cost (or savings). Thus, these two issues (hospital utlilization and nursing home costs) became the primary focus of the evaluation. Quantitative analyses of patient and facility level data were used to address Medicare Part A utilization and expenditure issues. The analyses considered the extent to which hospital days were avoided by use of the direct entry option as well as the additional SNF days and expenditures involved. We estimated that approximately 41 percent of the direct entry patients in Masssachusetts and 67 percent in Oregon would have been hospitalized in the absence of the direct entry option. The major cause for the differences between the Massachusetts and Oregon results involve the unusually high concentration of rehabilitation hospital/SNFs in Massachusetts and their disproportional involvement in the demonstration. Massachusetts also experienced a low number of SNF admissions directly from the patient's home-a likely consequence of relatively low levels of program awareness and physician involvement in nursing home care. The national projections of Medicare expenditure consequences of the waiver were directly based on the experience observed in the two states. As I indicated before, a range of estimates was prepared based on simulations. These simulations took into account the probability of hospitalization, the expected tradeoff between hospital and Medicare SNF days, and average per diem Medicare expenditures in 1980. Our view of the most likely effect of a nationwide direct entry option is a national saving of about $3 million. While there is potential for Medicare savings by eliminating the prior hospital stay requirement, the ultimate effect is complicated and depends on a variety of additional factors. I will now review four of the more important factors. 1. A fundamental point of my remarks is that each of the two demonstrations led to different results. While I feel confident that we have done a correct evaluation of the two demonstrations, a number of considerations make national projections very difficult. I cannot absolutely say whether the ultimate effect of the waiver involves a cost or a savings to Medicare. There are a number of factors that could increase of decrease the national estimates presented in the Final Report. The most critical issue is the extent to which the waiver diminishes days of care in hospitals, which are about four times as costly as SNF days. The data used to estimate the number of avoided hospital stays were derived from discussions about the case histories of 276 randomly selected direct entry patients with the directors of nursing in eight SNF's in Oregon and ten in Massachusetts that were relatively high users of the direct entry option. Our national projections are based on a combination of the results observed in both states. Factors exist, however, which could cause either greater savings or potentially substantially greater costs as a result of a direct entry option. For example, increased physician awareness of the option or a change in reimbursement incentives could lead physicians to admit patients to SNF's rather than hospitals when appropriate. Conversely, less stringent interpretation or enforcement of the Medicare SNF criteria could lead to a substantial increase in the number of Medicare SNF stays without an offsetting decrease in hospi |