THE HONORABLE R. JAMES NICHOLSON
DEPARTMENT OF VETERANS AFFAIRS
HOUSE VETERANS AFFAIRS COMMITTEE
June 30, 2005
Mr. Chairman and Members of the Committee: Thank you for the opportunity to discuss the budget forecasting and finances of the Veterans Health Administration. Accompanying me this morning is our Under Secretary for Health, Dr. Jon Perlin and our General Counsel and Chief Management Officer, Mr. Tim McClain.
Mr. Chairman, in considering our budget planning and execution, I’d like to address three topics. First, how does VA rationally project resource requirements for the health care needs of Veterans? Second, why is there discrepancy from projections and what is the current status of resources? And, finally, what can we do to improve the budget formulation process and the current budget status?
Projecting Resource Requirements:
The Veteran’s Health Care Eligibility Reform Act of 1996 established a uniform package of health care services for enrollees. The legislation also established a priority-based enrollment system and required the VA Secretary to annually assess veteran demand for VA health care to determine which priority levels of veterans will be eligible to enroll for care based on the resources available to provide timely, quality care to all enrollees.
Eligibility reform contributed to the transformation of the Veterans Health Administration (VHA) from a health care system that provided episodic, inpatient care to a health care system that provides a full range of comprehensive health care services to enrollees. The focus on health promotion, disease prevention and chronic disease management has resulted in more effective and more efficient health care. As a result, the range of health care services utilized by VHA patients began to mirror that of other large health care plans. Therefore, VHA decided to follow private sector practice and use a health care actuary to predict future demand for VA health care services. Mr. Chairman, transforming from a hospital system to a health care system has facilitated VA’s ability to take a leadership position in health care quality in the United States. A recent Washington Monthly article stated the Veterans Health Administration gives the “best care anywhere.” Additionally, the results of a recent study conducted by the independent RAND Corporation revealed that based on 348 measures of performance, VA provides systematically better care in disease prevention and treatment.
In the past, VHA budgets (and most Federal budgets) were based on historical expenditures that were adjusted for inflation and then increased based on proposed new initiatives. However, rather than an arbitrary increase over prior budgets, with the implementation of eligibility reform and the shift to ambulatory care, VHA needed to more rationally budget for veteran requirements in a transformed health care system. It also needed to be able to continually adjust its budgetary projections for effects of shifting trends in the veteran population, increasing demand for services, and the escalating cost of health care, e.g., pharmaceuticals.
As a result, VA engaged Milliman, Inc., to produce actuarial projections of veteran enrollment, health care service utilization, and expenditures. Milliman consults to health insurers and as such, is the largest and most respected actuarial firm in the country in the area of providing actuarial health care modeling.
VHA Enrollee Health Care Demand Model
The VHA Enrollee Health Care Demand Model (model) develops estimates of future veteran enrollment, enrollees’ expected utilization for 55 health care services, and the costs associated with that utilization. These projections are available by fiscal year, enrollment priority, age, VISN, market, and facility and are provided for a 20-year period.
The model provides risk-adjustment and reflects enrollees’ morbidity, mortality, and their changing health care needs as they age. Because many enrollees have other health care options, the model reflects how much care enrollees receive from the VA health care system versus other health care providers. This is known as VA reliance. Enrollee reliance on VA is assessed using VA and Medicare data and a survey of VA enrollees. The VA/Medicare data match provides VA with enrollees’ actual use of VA and Medicare services, and the survey provides detailed responses from enrollees regarding any private health insurance and their use of VA and non-VA health care.
The model projects future utilization of numerous health care services based on private sector utilization benchmarks that are adjusted for the unique demographic and health characteristics of the veteran population and the VA health care system. The actuarial data on which the benchmarks are based represent the health care utilization of millions of Americans and include data from both commercial plans and Medicare, and are used extensively by other health plans to project future service utilization and cost.
The model produces projections for future years using health care utilization, cost, and intensity trends. These trends reflect the historical experience and expected changes in the entire health care industry and are adjusted to reflect the unique nature of the VA health care system. These trends account for changes in unit costs of supplies and services, wages, medical care practice patterns, regulatory changes, and medical technology.
Each year, the model is updated with the latest data on enrollment, health care service utilization, and service costs. The methodology and assumptions used in the model are also reviewed to ensure that the model is projecting veteran demand as accurately as possible. VHA and Milliman develop annual plans to improve the data inputs to the model and the modeling methodology. Notably, Mr. Chairman, perhaps going to a focus of the Committee today, on average for the past three years, patient projections have been within -0.6 percent of actual patients and enrollee projections have been within +1.9 percent of actual enrollees.
As required by eligibility reform legislation, VA annually reviews the actuarial projections and determines whether or not resources are available to meet the expected demand for VA health care and develops policies accordingly. For example, the model’s projection of continued significant growth in enrollment in Priority 8 formed the basis of VA’s decision to suspend Priority 8 enrollment in January of 2003, to ensure that resources were available to provide timely, quality health care to enrolled veterans.
Over the past six years, VHA has integrated the model projections into our financial and management processes. The VA health care budget is now formulated based on the model projections, as are the impact of most policies proposed in the budget.
Some services VA provides are not modeled by Milliman. These include readjustment counseling, dental services, the foreign medical program, CHAMPVA, spina bifida, and non-veteran medical care. Demand estimates and budgets for these programs are developed by their respective program managers.
Enrollee demand for long-term care services is modeled by VHA. The VHA long-term care model uses utilization rates from nationally recognized surveys adjusted for the unique characteristics of the enrollee population and known reliance factors to account for distance (access to VA facilities), multiple eligibilities, and case management to project demand for both nursing home care and community-based care.
Discrepancy from Projections and Status of Health Care Resources:
Actuarial modeling is the most rational way to project the resource needs of a health care system like the Veterans Health Administration. As noted, this is the approach utilized private sector. Unlike private sector, however, where projections are used to formulate budgets for the next year or even the next “open season,” the Federal budget cycle requires budget formulation using data two and one-half to three and one-half years ahead of budget execution.
For example, the data used to formulate the budget for 2005 derive from health care utilization in 2002, in this case, the last full year of data before the Department’s 2005 budget formulation began. While it is remarkable that the budget has been as accurate as it has, a lot can change in three years.
The actuarial projection model forecast numbers of enrollees. The number of patients from the enrollee pool is a derivative calculation based on what has been, to date, a fairly constant relationship. One factor that has compounded the projections is the increased utilization of health care services by enrolled Veterans in all priority levels and from all combat eras.
The actuarial model forecasted 2.3% annual growth in healthcare demand in FY2005. We discovered that growth has accelerated through April,2005 to 5.2% above FY2004, which is almost 3% above our annual projection. This constitutes a substantial increase in workload and resource requirements.
In 2002, we were not yet a nation with large numbers of service members deployed to combat zones. Appropriately, VA continued to use separation data from the Department of Defense to project potential rates of utilization separating service members. Our FY2005 budget assumed that 23,553 VA patients (at a cost of $81 million) would be veterans of the Global War on Terrorism. The number of these patients in 2005 is now estimated to be 103,000, so we are $273 million short. This additional cost is a substantial but not a predominant (or even the majority) component of the increased medical care cost in 2005.
Fortunately, many are seeking routine services. Some require dental care that was deferred as they deployed for combat. Others require more intensive care for both the physical and psychological consequences of combat. About 60 percent of the combat veterans who have come to VA are reservists or members of the National Guard. Veterans deployed to combat zones are entitled to two years of eligibility for VA health care services following their separation from active duty even if they are not immediately otherwise eligible to enroll at VA. Because of this, these combat veterans then come to VA in numbers much higher than if they were to separate from DoD without a combat history. The general DoD separation trends data available from the routine 2001 separation planning report could not anticipate the numbers of reserve service members who were subsequently activated and then separated from service.
In summary, the increased medical care cost in 2005 is nearly $1.0 billion of which $273 million (28%) is associated with veterans returning from the current combat theatres.
Questions have been raised about the timing of the information disclosed about VA’s 2005 budget situation. I want to be clear that we continue to feel that we can meet the needs of timely, high-quality health care for veterans. In fact, I indicated this in my letter of April 5 to Chairman Hutchison of the Senate Subcommittee on Military Construction and Veterans Affairs, in which I stated that, “whenever trends indicate the need for refocusing priorities, VA’s leaders ensure prudent use of reserve funding for these purposes. That is just simply part of good management.”
In a similar fashion, at his confirmation hearing on April 7, 2005, then Acting Under Secretary for Health Perlin, testified to the Senate Veterans Affairs Committee that reserve funds were being used to meet operational needs in 2005. This generated some subsequent questions from the Committee, and in a letter on April 12, Dr. Perlin wrote that the projected carryover might be diminished to address operational demands on our system, including the care of returning combat veterans of Operation Iraqi Freedom and Operation Enduring Freedom, noting that “we do feel confident that VHA has sufficient resources for the remainder of 2005.”
The following week, on April 19, VA staff met with Ranking Member and members of the minority and majority staff of the House Appropriations Subcommittee to discuss the Veterans Equitable Resource Allocation (VERA) model. During this meeting there was protracted discussion of the health system’s financial status in 2005, including the management decision to reallocate capital funds for direct patient care. During that same week, I met with the OMB Director to update him on the current status and to alert him to potential issues for Fiscal Year 2006 suggested by preliminary and incomplete data. We agreed to monitor the situation as more complete and actual data emerged.
In May, we performed our periodic actuarial model update for FY2006 with more current and accurate data. This further validated the emerging phenomenon of increasing workload. This was discussed internally as part of the Department’s mid-year financial review. In the first week of June, VA staff met with OMB staff for its annual mid-year management review where we discussed in general terms the implications of FY05 management decisions on the FY06 budget. Similarly, VA staff met on June 3 with majority staff members of the House and Senate Veterans Affairs Committee, where they had very candid dialog about the implications of the reallocation and use of funds projected for carryover into the base for the FY06 budget.
On June 23, the Under Secretary for Health offered testimony on the actuarial model and its limitations. Actuarial modeling for 2005 forecast a growth rate of 2.3 percent, and as of April 2005, VA was experiencing workload growth at the rate of 5.2 percent annually, explaining the need to reallocate funds and devote carryover funds for patient care. As discussed in the hearing, VA’s 2005 increased medical care cost is nearly $1.0 billion, which VA will manage by reducing the 2006 carryover balance by $375 million and deferring $600 million of non-critical capital expenses for a few months.
I think that the record shows that VA has been very forthcoming with information regarding both the status of our budget and the responsible management decisions we have made as 2005 unfolds. It is our first responsibility to provide the highest quality care to veterans. It is our next responsibility to be good stewards of the substantial resources entrusted to us for that care. While resources have been adequate to make reallocation decisions to meet the most essential needs in 2005, it is now clear that the budget picture for 2006 needs to be revisited. We are working with OMB to reach a satisfactory resolution for 2006 that assures VA is there for all eligible veterans.
After looking at what additional efficiencies may be possible in what is arguably the nation’s most efficient health system, I believe that the additional resources relative to the President’s Budget that are necessary to provide timely, high quality care to the Veterans in 2006 amount to approximately $1.5 billion. This includes $375 million to repay the carryover, nearly $700 million for increased workload, and $446 million for an error in estimating long-term care costs. The Administration will come forward to the Congress shortly with a proposal to provide VA the additional resources. This amount assumes enactment of the policies in the President’s Budget. If Congress does not accept any of the policies in the President’s Budget, additional resources will be needed.
In a sense, VA and other Federal agencies like DoD who use actuarial modeling to project resource requirements two and one-half to three years hence push the performance envelope compared to private sector, which uses these data at one year. In fact, the 2.9 percent margin of error we experienced is far better than the 11 percent error that occurred when budgets were projected by inflating an historical base. Mathematically, at three years, a 2.9 percent margin of error is pretty good. Still, we recognize that the consequences are not.
In order to improve the model and budget process going forward, additional model inputs are required. We must figure out how to better approximate changes needed to compensate for the lag in data in our estimates. In addition, we need to do a better job of linking DoD experience with our input.
The development of the actuarial model has been an evolutionary process. It is a prerequisite for the data necessary for the Secretary’s annual enrollment decision which matches enrollment levels to resource availability. Enhancements to the model include more detailed and robust adjustments for enrollee reliance, morbidity, and mortality, adding new data sources, and expanding the number of services modeled. Future planned improvements include access to data on enrollee’s use of Medicaid, Tricare, and military treatment facilities, the integration of the VHA long-term-care model into the actuarial model, and modeling additional services such as dental care.
Mr. Chairman, in closing, I believe that the VHA Enrollee Health Care Demand Model is a valuable budgeting and planning tool for projecting VA health care utilization. We look forward to working with you to ensure that we continue to provide timely and high-quality health care to our Nation’s Veterans.
U.S. Department of Veterans Affairs - 810 Vermont Avenue, NW - Washington, DC 20420
Reviewed/Updated Date: November 10, 2009