Skip to content
Q&A

New Research Points to How to Reduce the $100 Billion Americans Spend Annually on Unnecessary Health Care

Measuring low-value care at the health system level sheds light on the drivers of low-value care and how to take action to address them.

Hospital room with robotic operating machine.
Getty Images

Each year, millions of patients in the United States receive care that is medically unnecessary and provides little or no clinical benefit to them. Such “low-value” services — which range from unnecessary CT scans and lab tests to overprescribing and avoidable surgeries — can harm patients physically and financially, including by leading to further unnecessary treatments.

In addition to exposing patients to potential harm, low-value care wastes crucial health care resources. The United States spends up to $100 billion annually on procedures and treatments that are low value, often because providers have a financial incentive to provide them. 

A new study supported by Arnold Ventures and published in JAMA Internal Medicine measures low-value care use across and within individual health systems, highlighting hotspots of low-value care and its drivers. The researchers examined 41 individual low-value services across 556 health systems spanning the country and identified specific health system characteristics associated with higher use of low-value services. This study builds on prior research measuring low-value care at the regional or national level by measuring low-value care at the more actionable level of individual health systems. Such groundwork can provide a starting point for policymakers, health plans, and employers looking to curb the use of low-value care.

The study’s lead author, Dr. Ishani Ganguli, an assistant professor at Harvard Medical School, sat down with Arnold Ventures to discuss the study’s novel approach and what it tells us about how to reduce the use of services that don’t benefit patients.

This interview has been edited for length and clarity.

Headshot of Arnold Ventures
Arnold Ventures

How common is low-value care? In your analysis, how much variation did you find in the amount of low-value care that health systems are providing?

Headshot of Dr. Ishani Ganguli
Dr. Ishani Ganguli

Low-value care remains common — by recent estimates, about a third of all traditional Medicare enrollees receive some low-value medical service in a given year. Patients continue to receive these services for a number of reasons. The U.S. health care system financially rewards ordering more services, even in nonprofit health systems. Clinicians order low-value services out of habit (with the desire to be thorough or to match their peers), in response to patient request (whether real or perceived), or due to concerns about malpractice lawsuits. In addition, primary care physicians who have easy access to specialists within their health system may be more likely to refer their patients to these specialists; these specialists, in turn, may order more low-value services.

In this study, we found a great deal of variation in low-value service use between health systems, especially for laboratory tests and drugs. Systems ranged from 0% to more than 50% of their eligible patients receiving low-value lab tests like prostate cancer screening tests for older men. Systems ranged from 0% to more than 60% of their eligible patients receiving low-value drugs, like antipsychotic medications for patients with dementia.

Headshot of Arnold Ventures
Arnold Ventures

As a physician, how did you become interested in the topic of low-value care? Have you seen firsthand an example of its impact on patients that you could share with us?

Headshot of Dr. Ishani Ganguli
Dr. Ishani Ganguli

I first became interested in this topic as a medical resident. I noticed, and was bothered by, tests and treatments my peers and I were ordering not for a compelling clinical reason but instead out of habit — like daily bloodwork for hospitalized patients. Such decisions not only harm patients (for example, by interrupting their sleep for an early morning blood draw and by causing anemia), but they also create more work for medical teams. Low-value services can also cascade into downstream tests, treatments, and new diagnoses that on average may cause more harm than good, which I have previously written about.

Headshot of Arnold Ventures
Arnold Ventures

Low-value care has primarily been reported at the national or regional level. Your study uses the first system-level measure of low-value care. How does measuring low-value care use within individual health systems help advance efforts to make sure patients receive high quality, affordable care?

Headshot of Dr. Ishani Ganguli
Dr. Ishani Ganguli

Health systems — as opposed to standalone physician practices — are increasingly where Americans get their care. These systems may also be key drivers of decisions to order low-value services. For example, health systems can affect low-value service use through the policies they set, their clinical workflows, the technologies they invest in, their workplace culture, whom they hire, and how they pay their clinicians. While our claims-based measures of low-value services are not perfect, we hope that they can help individual systems target and intervene on relevant low-value services using these same levers. Patients might use publicly reported measures to choose which doctor they see or to start a conversation with their doctor about whether they really need a certain test or treatment. Policymakers can use them to hold systems accountable for reducing low-value care.

The U.S. health care system financially rewards ordering more services, even in nonprofit health systems.
Dr. Ishani Ganguli assistant professor at Harvard Medical School
Headshot of Arnold Ventures
Arnold Ventures

You found health systems with certain characteristics were more likely to underperform. What do these findings tell us, and why do they matter?

Headshot of Dr. Ishani Ganguli
Dr. Ishani Ganguli

We found systems that used more low-value care had lower shares of primary care physicians — in line with numerous other studies linking primary care to lower spending and better health outcomes. These poor-performing systems also tended to serve more minority patients, which is unfortunately consistent with other studies showing lower quality care in hospitals and practices serving these populations. These systems tended to be headquartered in the South or West — geographic patterns that match with many prior studies about low-value care use. And they tended to serve areas that had higher health care spending overall — demonstrating that low-value health care spending doesn’t occur in a vacuum.

These associations suggest which systems to target when designing low-value care interventions. They also point to the types of policies that may help reduce low-value care. For example, reforming the way clinicians are reimbursed or using loan repayment programs to increase the share of clinicians working in primary care.

Headshot of Arnold Ventures
Arnold Ventures

Your study found that systems in Accountable Care Organizations (ACOs) had the same use of low-value care as systems not in ACOs. Why do you think this is, and what can we do to encourage ACOs to reduce low-value care?

Headshot of Dr. Ishani Ganguli
Dr. Ishani Ganguli

Exactly — systems with or without at least one ACO contract did not differ in their use of low-value care, even though this payment model incentivizes systems to lower spending for their attributed patient population. This aligns with a 2015 study showing only modest reductions in low-value care use in ACOs compared to non-ACOs and a survey by members of our team showing ACOs have not prioritized reducing low-value care. It’s not surprising when considering that ACOs may focus their efforts on reducing large sources of upfront spending that can be easily targeted, rather than on low-value services that often have low unit costs (though even these add up in aggregate and can cause cascades of further low-value care and other harms).

Low-value services are also trickier to eradicate and require thoughtful solutions built into clinical workflows. ACOs have not historically been held to quality standards on low-value care avoidance. Quality reporting typically focuses on measures of underuse, not overuse, but this may be a helpful next step that policymakers and others could take to strengthen incentives to focus on low-value care.

Headshot of Arnold Ventures
Arnold Ventures

What are promising steps you think purchasers, payers, and policymakers can take to reduce low-value care?

Headshot of Dr. Ishani Ganguli
Dr. Ishani Ganguli

Policymakers can track system performance over time and hold ACOs accountable for avoidance of low-value care as mentioned above. In addition, payers and employers can include low-value care avoidance in their contract negotiations with providers and experiment with high-value provider networks.

Grants

Arnold Ventures funds projects to understand problems and identify policy solutions.

Map of the U.S. made of Arnold Ventures icons