The Department of Veterans Affairs (VA) Pharmacy Benefits Management Services and the Institute for Clinical and Economic Review (ICER) recently announced a collaboration in which the VA will use the ICER’s drug assessments as part of its formulary development and price negotiations. This type of relationship might be normal outside of the United States (for example, in the United Kingdom, Germany, and Australia), where input from governmental health technology assessment organizations is used in determining health care coverage decisions. However, in the United States—with our multipayer health care system covering different populations (for better or worse)—this represents the first instance in which the federal government is incorporating nonpopulation-specific assessments and cost-effectiveness-based price benchmarks into drug coverage and purchasing decisions. The ICER assessments are neither specific to veterans nor the particular characteristics of the VA health care system. As the VA considers how to use this information, caution is warranted.

The VA already has a very restrictive formulary, covering only 59 percent of the top 200 prescription drugs in the United States compared to Medicare prescription drug plans that cover about 85 percent. Furthermore, the VA is statutorily guaranteed substantial price discounts on prescription drugs, on top of which it negotiates further discounts as part of creating its formulary. The end result is that the VA receives the steepest discounts in the country.

The current and prior administrations have placed a high priority on improving VA health care. While substantial progress toward better access has been made, more work could be done to make the VA’s formulary decision making more transparent. There are, however, important factors to consider and questions to be asked about the use of the ICER assessments in VA decisions: Will these assessments lead to improvement in our veterans’ access to needed care? Are more rational drug decisions likely to result? And, will veterans’ health improve because of this approach?

The answer to all these questions is the same: Perhaps, but only if the following considerations are addressed.

Consideration 1: Broaden The Clinical Evidence Beyond Randomized Controlled Trials

The decisions regarding which drugs should be included on a formulary need a strong clinical evidentiary backbone. The ICER assessments summarize the available clinical information, and surveys indicate that payers value these clinical summaries. However, the ICER reviews are typically performed just before or as drugs gain Food and Drug Administration approval based only on the results of randomized controlled trials. Payers such as the VA need to make coverage decisions at the time of or soon after drug approval, and reliance upon clinical trials is appropriate for these initial reviews.

Once a drug comes to market, a greater understanding of its effectiveness and safety and how it works in “everyday” situations become available, as does—importantly—a greater understanding of how it works in the VA population. The need to consider real-world performance is even more critical in the VA, where there are populations of patients with complex clinical problems, such as younger veterans with post-traumatic stress disorder (PTSD) and older veterans (covered by Medicare) with multiple chronic conditions. Veterans may see private providers and clinicians, as well as clinicians at VA facilities, where they may also come for their prescription medicines.

Clinical trials typically enroll younger patients without additional illnesses. In contrast, the population served by the VA has more comorbidities, may be older, and has different constellations of clinical situations, including homelessness, which is a categorical eligibility for VA health benefits.

Because the ICER assessments are not routinely updated, they don’t incorporate real-world experience with how therapies work in the VA environment. As this evidence accumulates, the value associated with each treatment will naturally evolve. To better reflect the full benefits and harms of therapies to veterans, the VA should conduct periodic reassessments and incorporate information specific to the population of individuals served. The VA has invested heavily in electronic health records, and those databases could be leveraged for this purpose.

Consideration 2: Consider The Perspective Of Veterans And Their Families As Well As The Health Care System

The ICER assessments are based on a health system or payer’s perspective and incorporate the costs that are most relevant to these entities (for example, impact on hospitals/emergency departments, use of medications, diagnostics, or surgery). A broader perspective would include factors such as the economic consequences of a therapy’s impact on a veteran’s ability to work/work more effectively, or reducing the care burden on family members or other caregivers. It should also include additional benefits and expenses from the Department of Veterans Affairs, as well as other federal and state programs.

For the population of veterans and the conditions they face, this broader societal perspective is both critical and likely of high value. If a therapy allows a disabled veteran to reenter the workforce or to become more self-sufficient, its societal and economic benefits are likely to outweigh its costs. Using a hypothetical salary of $40,000 per year, enabling veterans with PTSD to return to work would more than pay for all but the most costly medications. Similarly, if a disabled veteran’s health can be improved so that a caregiver need not be present throughout the day, the economic benefits likely outweigh the drug costs. The VA should follow the recommendations of the Second Panel on Cost-Effectiveness and perform its own societal assessment and not just depend upon a narrow payer or health system perspective.

Consideration 3: One Size Can’t Fit All—Or The Need For A Broader Set Of Value Inputs

Each ICER assessment provides the VA with a single score or price benchmark for a particular drug, based upon the clinical evidence and modeled cost-effectiveness. Value, however, is broader than cost and improvements in discrete clinical endpoints. It also includes quality of life, complexity of the regimen, and caregiver burden, attributes that are highly personal, vary widely, and are not amenable to a singular assessment. Moreover, assessments of value and of what is most important to patients differ by disease, by the stage of that disease, and by the personal determinants of those patients. For example, health care needs, preferences, and value of a therapy are likely to be very different for older Korean War veterans than for those from the Vietnam era, or those who fought more recently in the Middle East.

A singular assessment cannot reflect the differing needs among veterans and VA centers in disparate parts of the country. The ICER assessments provide a starting point, but the VA should incorporate additional information from other, more patient-centric approaches to determining value. These include efforts from the National Health Council (Patient-Centered Value Model Rubric), Avalere/FasterCures (Patient Perspective Value Framework), and the Innovation and Value Initiative, Memorial Sloan Kettering (DrugAbacus, which enables individualized assessments based upon user preferences), as well as value assessments by the American Society of Clinical Oncology and the National Comprehensive Cancer Center.

Consideration 4: Arbitrary Spending Limits On Individual Drugs Is An Incorrect Policy

The ICER approach uses a $915 million budget impact threshold for an individual drug, regardless of whether it addresses the needs of a small population (for example, chronic renal failure) or a large one (for example, diabetes, hypertension, depression). Arbitrary adherence to a spending threshold, even one specific to the VA, could lead to undesirable consequences. A recent study found that dividing up drug spending equally among newly released drugs has no clinical basis, may constrain the use of highly beneficial drugs (for example, hepatitis C therapies), and would not appropriately address aggregate health care spending.

Budget impact concerns may be valid. However, to allow dollars to flow to the most effective treatments, even if the standard of care evolves over time, assessments and policies should be at a disease level instead of at the sector level (hospital care, physician services, drugs, devices, diagnostics). For example, care for individuals with AIDS in the 1980s was hospital-centric and not very effective. Today, spending on this condition has shifted to medications that enable people living with HIV infection to treat it as a chronic condition instead of a death sentence from various opportunistic infections or cancers. By contrast, for patients with severe osteoarthritis of the hip, a reverse resource allocation is appropriate: surgical care in the hospital (joint replacement), instead of spending on chronic opioid pain medications. Arbitrary limits on sector spending or a priori division of the health care spending pie (for example, among hospital care, physician services, imaging, drugs) would neither be economically efficient nor in patients’ best interest.

Looking Forward

Incorporating a variety of drug assessments and information into the VA’s decision process might bring improved care and better outcomes for veterans. However, is the incorporation of the ICER assessments on top of the VA’s long-standing process for drug reviews a means to justify an even more limited and cost-constrained formulary? Or, more positively, might it broaden access to medications deemed valuable? Hopefully, the answer to the former is “no” and the latter, “yes.”

Addressing the considerations in this blog would make the positive outcomes more likely. The VA should proceed with caution.