Can we reduce healthcare waste while making Americans healthier? A new Medicare pilot program that uses AI is trying to do just that, but as usual, the devil is in the details. Here’s what’s promising about the program, and which elements could use rethinking.
What is the WISeR pilot project?
Federal agencies have long recognized the widespread problem of overuse. In 2022, Medicare spent $5.8 billion on just 31 low-value services that are more likely to harm than help patients. These low-value services were delivered at a whopping rate of 71 per 100 Medicare beneficiaries.
Now, the Centers for Medicare and Medicaid Services (CMS) is looking to reduce overuse for Medicare beneficiaries, starting with a new six-year program model in six states. The recently-announced Wasteful and Inappropriate Service Reduction model (WISeR) will require prior authorization from providers for 17 kinds of services that are commonly overused or historically have had a higher risk of waste, fraud, and abuse.
Services requiring prior authorization in WISeR program (Source: Federal Register)
- Electrical Nerve Stimulators
- Sacral Nerve Stimulation for Urinary Incontinence
- Phrenic Nerve Stimulator
- Deep Brain Stimulation for Essential Tremor and Parkinson’s Disease
- Vagus Nerve Stimulation
- Induced Lesions of Nerve Tracts
- Hypoglossal Nerve Stimulation for Obstructive Sleep Apnea
- Epidural Steroid Injections for Pain Management excluding facet joint injections
- Percutaneous Vertebral Augmentation for Vertebral Compression Fracture
- Cervical Fusion
- Arthroscopic Lavage and Arthroscopic Debridement for the Osteoarthritic Knee
- Incontinence Control Devices
- Diagnosis and Treatment of Impotence
- Percutaneous Image-Guided Lumbar Decompression for Spinal Stenosis
- Skin and Tissue Substitutes
- Application of Bioengineered Skin Substitutes to Lower Extremity Chronic Non-Healing Wounds
- Wound Application of Cellular and/or Tissue Based Products, Lower Extremities
Among these are services that have been on lists of low-value care to avoid for over a decade, such as injecting cement into the spine for a spinal fracture caused by osteoporosis (known as vertebroplasty) and surgery to remove damaged cartilage or bone in the knee using an arthroscope for patients with osteoarthritis (called knee arthroscopy). From 2020–2023, U.S. hospitals performed 100,000 vertebroplasties and 3,500 knee arthroscopies that met overuse criteria, per Lown Institute data.
| Number of procedures meeting overuse criteria, 2020-2023 | Hospitals with at least one procedure meeting overuse criteria, 2020-2023 | |
|---|---|---|
| Knee arthroscopy | 3,433 | 1,303 |
| Vertebroplasty | 98,429 | 2,073 |
Other services are ones that government agencies or journalists have previously flagged for abuse or fraud. For example, a 2023 investigation by the Office of the Inspector General identified 80,000 steroid injection sessions that exceeded Medicare’s coverage limitations, costing Medicare $13 million. A 2021 OIG investigation found that 40% of nerve stimulation implantations did not comply with Medicare requirements, costing Medicare over $600 million.
And the New York Times reported on an explosion in Medicare spending on skin substitutes made from placenta and other medical waste, as some providers allegedly exploited reimbursement loopholes.
The “wise” in WISeR
For those of us who have been advocating for high-value care, it’s encouraging to see CMS use their authority to target overuse and fraud. CMS has also added some guardrails to ensure the program doesn’t impede access to needed care:
- None of the services subject to prior authorization are “inpatient-only, emergency services, or services that would pose a substantial risk to patients if substantially delayed.”
- Any recommendations for Medicare to not pay a provider for these services have to be reviewed by an appropriately licensed clinician using “standardized, transparent, and evidence-based procedures.”
- And CMS is considering implementing a “gold card” program to exempt certain providers from prior authorizations if 90% of their requests get approved in a provisional period.
Words of caution
However, some elements of the WISeR program have set off alarm bells for health policy experts. CMS already requires prior authorization for a small number of outpatient services, some durable medical devices, and repeated non-emergency ambulance transport. These prior authorizations are managed by Medicare Administrative Contractors (MACs). Yet, in the WISeR program, CMS is contracting this work to outside private organizations. These private agencies are incentivized to deny more claims, as they get paid based on a “share of averted expenditures.”
The reliance on artificial intelligence to identify these overused services could also turbocharge the number of prior authorization requests unnecessarily, if these tools are not properly vetted or well-implemented. Given that Medicare Advantage plans have used AI to systematically deny claims for beneficiaries, the reluctance to use AI for prior authorization in traditional Medicare is understandable.
Targeted vs systemic approach
The WISeR program takes a systemic approach to reducing overuse and fraud, by applying prior authorization requirements for these services to all providers in the pilot states. Some policy experts argue that we should be taking a more targeted approach, going harder after the most obvious fraudsters rather than making all doctors go through the prior authorization process.
There are benefits to targeting particular providers with high rates of overuse. Data from the Lown Institute Hospitals Index shows how overuse in hospitals can be concentrated in just a few hospitals or even providers. For example, some hospitals in the Lown Index data set had rates of unnecessary vertebroplasty four times higher than the national average rate, while others did no vertebroplasties during those years. In one hospital with especially high rates of vertebroplasty overuse, a single physician was conducting 71% of those procedures. (For more on back surgery overuse, see our 2024 launch page and presentation.)
However, we can’t point to just a few bad apples as the source of all overuse. Despite high levels of variation across the country, overuse is also systemic. A survey of hundreds of practicing internal medicine doctors found that 94% said they had experienced “care cascades,” follow-up testing and treatments due to incidental findings that had no clinical benefit for patients.
The discourse around WISeR brings up an important question we should be asking around targeted versus systemic approaches to overuse. If overuse and fraud are events that we want to avoid as much as possible, it makes sense to take a systemic approach and apply the same criteria to all providers. The benefit to using prior authorization is that it prevents overuse before it starts, as opposed to catching it later. However, there’s no reason why we can’t do both, and harness the power of AI to identify serial overusers for enforcement.
Will WISeR work?
If we want all Americans to have the opportunity to be healthy, we need to provide access to needed care, while reducing the harm from overuse. With estimates of spending on low-value care ranging from $100 billion to $700 billion, reducing overuse is also an imperative for lowering the cost of care.
Ideally, WISeR will prevent certain types of fraud and overuse, reducing patient harm and saving Medicare millions. However, the success of this program depends on implementation. Are the AI tools being used accurate enough to catch the right services? Are CMS’ guardrails strong enough to ensure that people don’t lose access to medically necessary care? Lessons learned from the program in these first states will show whether this prior authorization model has real potential to reduce overuse.