The Leonard Davis Institute of Health Economics’ (LDI) Sept. 11 virtual seminar, “Consumer-Directed Health Care: Promise or Pitfall,” was a debate about a deceptively simple question: Can patients be turned into effective health care shoppers, and would doing so meaningfully reduce U.S. health care costs?

Behind this question is a powerful idea from economics that has long shaped efforts to control health care costs: Turn patients into shoppers.

Called “patient-directed health care,” the theory sounds straightforward. Give patients the prices of medical treatments and services, tell them how providers compare on quality, and market forces should take over. Patients will shop around, weigh price against quality, and choose the best care for the money.

But patients have stubbornly refused to behave like textbook consumers.

Research has increasingly shown that simply giving people more price and quality information does not reliably send them shopping for the best health care bargain. Behavioral economics helps explain why. Medical decisions are made by people who may be sick, frightened, pressed for time, loyal to a doctor, confused by complicated choices, or simply looking for the easiest and fastest way to get care.

In those circumstances, emotion, habit, trust, urgency, and convenience can matter more than price — precisely the kinds of forces that traditional economic models of consumer choice tend to underestimate.

As he opened the discussion, moderator David Grande, MD, MPA, LDI Interim Executive Director, noted: “Since the 1990s, we’ve seen several experiments and evaluations of consumer-driven health care. Today we examine what the evidence has shown about the approach. Have these models reduced costs and improved efficiency as hoped? Do they shift financial risk onto patients and discourage necessary care? And what is the policy outlook for consumer-directed health care as a federal policy solution?”

Participating in the discussion were two top national authorities in the field. The first was Sherry Glied, PhD, Professor of Public Service and former Dean of NYU’s Robert F. Wagner Graduate School of Public Service, former Assistant Secretary for Planning and Evaluation at the U.S. Department of Health and Human Services, and former member of the President’s Council of Economic Advisers. The second was Stephen Parente, PhD, MPH, Professor and the Minnesota Insurance Industry Chair of Health Finance at the University of Minnesota and Associate Dean of the Carlson Global Institute, former President of the American Society of Health Economists, and a former Chief Economist for Health Policy on the President’s Council of Economic Advisers.

These were the main points of debate:

Do consumers really shop for health care?

Parente: Consumers can behave more like shoppers if prices are visible, incentives are meaningful, and the system gives them usable tools. He pointed to retail clinics, elective services, pharmacies, imaging, and laboratory tests as areas where more conventional consumer markets could develop. His larger argument was that today’s failure to shop may partly reflect the fact that the health care system makes shopping extraordinarily difficult.

Glied: Decades of evidence suggest that consumers generally do not want to shop for medical care in this way, even when given tools and financial incentives. Very few patients use price-comparison tools, and once patients enter care, they generally follow their physician’s recommendations rather than independently comparing prices. She argued that consumer-directed theory has run well ahead of the empirical evidence.

Is health care fundamentally “shoppable”?

Parente argued that at least some medical services are inherently shoppable — particularly discrete, scheduled services such as MRIs, laboratory tests, and certain consultations. He envisioned an Amazon-like system in which a patient could see a service, provider, price, and available appointment and book it immediately.

Glied countered that this misunderstands how most patients encounter medicine. Someone whose doctor orders an MRI usually goes where the doctor recommends, even if cheaper imaging centers are available. Patients have deliberately selected physicians as expert agents and generally trust their judgment. In her view, elective diagnostic testing purchased independently by healthy consumers may become highly shoppable, but that is different from the core medical care market.

The sharpest exchange involved Parente’s example of enormous price differences for the same MRI performed by the same physician. He cited transparency data showing prices ranging from roughly $220 to $1,800 and argued that a patient facing a large deductible should obviously care about getting the lower price. Glied’s response was essentially this: Evidence says they usually don’t care enough to do the work required to find it.

Who should be responsible for controlling prices?

Parente: Consumers should have a meaningful role. If patients spending their own deductible dollars can see that the same service costs dramatically different amounts, they have an incentive to choose the cheaper option. He also argued that insurers administering much of the employer-sponsored market may lack sufficiently strong incentives to minimize every medical price.

Glied: The consumer is the wrong actor to assign that job. Insurers and employers should be doing the sophisticated shopping and negotiating. Patients hire insurers precisely because individual consumers cannot realistically master complicated provider networks, contracts, quality differences, and medical prices.

Grande distilled the disputed point: Whose problem are wildly varying medical prices to solve — the patient with the high deductible, the insurer negotiating provider prices, or the employer paying for coverage?

Price transparency: Solution or Disappointment?

Parente sees enormous price variation as both evidence of a dysfunctional market and an opportunity. Consumers often have little idea how large the differences are, he argued, and better transparency and technology could change their behavior.

Glied rebutted with the example of a heavily promoted New York transparency tool that allowed consumers to compare provider prices. Few patients used it, and researchers found no evidence that consumers shifted toward cheaper providers. Ironically, some physicians apparently found the information more useful: Lower-priced doctors saw competitors’ prices and raised their own prices.

Parente responded that consumers don’t understand how much prices vary because the information has been hidden and difficult to use.

Would easier access to cheaper care actually reduce spending?

Parente said allowing patients to directly locate and purchase inexpensive imaging could make that market more efficient.

Glied challenged the premise: If you make MRIs cheaper and easier to obtain, won’t people simply get more MRIs? Prices might fall while utilization rises, meaning total spending might not decline.

High-deductible plans and Health Savings Accounts

Parente described high-deductible plans and Health Savings Accounts (HSAs) as mechanisms that give consumers a financial stake in medical spending, with HSAs allowing tax-advantaged funds to accumulate for future care. He argued that properly designed high-deductible options can work and cited the University of Minnesota’s experience as evidence that such plans can remain attractive when their premium savings are preserved.

Glied was much more skeptical. She said evidence suggests high deductibles can reduce utilization, but the savings become much less convincing when plans are paired with HSAs and exemptions for preventive care. She also argued that HSAs have evolved partly into tax shelters disproportionately benefiting higher-income people, rather than functioning primarily as mechanisms for disciplined health care purchasing.

Her policy prescription was therefore almost the reverse of Parente’s direction: Eliminate FSAs and substantially narrow HSAs. Parente, meanwhile, suggested expanding experimentation with catastrophic coverage — low-premium, high-deductible health insurance designed primarily to protect people from extremely large medical bills caused by serious illnesses, injuries, or other major health emergencies.

Can AI be used to finally make consumer-directed care work?

Here the two moved somewhat closer together while retaining their basic disagreement.

Parente sees AI-enabled apps, transparent price data, and eventually access to patients’ electronic medical records as potentially creating the consumer interface that health care has long lacked. Younger, digitally native consumers may be far more comfortable using such tools to compare and obtain services.

Glied agreed that AI could transform health care pricing — but predicted that its greatest impact would occur behind the scenes. Insurers and employers could use AI to analyze enormous transparency datasets, identify anomalous prices, and negotiate more effectively with providers. In other words, Parente sees AI potentially creating smarter patients, while Glied sees it creating smarter institutional purchasers.

Their final AI recommendations reflected that difference. Parente emphasized giving consumer applications broad access to longitudinal electronic medical records so AI can help patients and providers navigate care. Glied emphasized AI-assisted triage — getting patients to the appropriate level of care rather than expecting them to become sophisticated medical shoppers.

Is the real problem that the market itself is broken?

Despite Glied and Parente’s differences, they were in agreement that extreme provider price variation and market concentration cannot simply be ignored.

Glied said rate setting can make sense in local markets dominated by one or a few hospitals because normal competition cannot function without competing suppliers.

More strikingly, the strongly market-oriented Parente acknowledged that the price-transparency data had “stretched my tolerance” for relying entirely on market forces. The enormous price differences suggested to him that parts of health care behave more like regulated monopolies than competitive markets. He therefore entertained limits or price “bands” rather than unrestricted pricing.

That may have been the seminar’s most important area of consensus: The debate is no longer simply government regulation versus consumer choice. Both speakers acknowledged that the underlying structure of U.S. health care markets — concentrated providers, opaque contracting, weak purchasing incentives, and extraordinary price variation — may prevent either conventional competition or consumer shopping from working as economic theory predicts.

Parente’s case was essentially that consumer-directed health care has never been given the transparent, technologically sophisticated marketplace necessary to prove what it can do; Glied’s was that more than 20 years of experience already provide considerable evidence that most patients neither want nor are well positioned to serve as the primary price setters of the health care system.

As the session approached its end, moderator Grande asked the two what single change in policy or technology would most improve how the market works and move the country toward controlling health care spending.

Parente: “In the future, if we’re going to have consumer apps for shopping, those apps are going to have to access people’s medical records, and they can’t be just the clinic you saw in the last couple weeks. They have to `be ubiquitous across all the other aspects of a person’s care because that is what will allow the AI piece to work better, benefiting both providers and patients.”

Glied: “What I hope will happen with AI is that we will get a lot better at doing triage. One of the things that keeps us from having a competitive system here is that it is very hard to know when you should be eating at McDonald’s and when you should be eating at a five-star restaurant. We are currently not using our resources efficiently, and a lot of the waste in care comes from using the wrong level of care for a given patient.”

---