New Marketplace
Vertical Integration and Bold Experimentation (11:08)
  1. There is significant opportunity to improve value in health care; recent research indicates both the availability of that value improvement and where it may reside.
  2. To create and capture that improved value, it’s important to influence what doctors recommend.
  3. To influence doctors’ recommendations, change the financial incentives and information that providers face. The best approach to doing that is vertical integration.
  4. It is a mistake to elevate the status quo in asking whether the alternative is better than today. Instead, ask, “Is the alternative statistically worse than today?” If it’s a draw, try something else — experiment aggressively.

Peter Orszag, Vice Chairman and Global Co-Head of Healthcare at Lazard, dives into detail on these four points.

Opportunities to Improve Value in Health Care

“What have we learned over the past decade on opportunities to improve value?” asks Orszag. For him, the most interesting dimension involves variation within U.S. health care. In employer-provided care, maybe half or more of the variation across the U.S. has to do with price charge per unit of care. But in Medicare, almost all the substantial variation is utilization itself. “More stuff is done to a patient in Florida than in Minnesota,” he says.

“Everyone accepts those cost differentials, but there’s been this long debate about whether they reflect inadequate adjustment for the health status of the patient,” says Orszag.

There’s been some interesting research from the past 5 or 6 years on people who move from one part of the U.S. to another. “If it were just the health status of the patient in Florida being worse than in Minnesota, then when someone moves, spending on them shouldn’t change because it’s all about the patient,” Orszag explains. “What instead you see is that when someone moves from Florida to Minnesota, spending on them plummets; when someone moves from Minnesota to Florida, spending on them jumps significantly.”

One might assume that some of the people who move from Minnesota to Florida feel sick and then decide to move. But a recent study of people moving between military bases, which has nothing to do with health status, shows similar variation based on where they move.

“There’s a lot of variation in health care utilization that has nothing to do with the health status of the patient and has a lot to do with norms and behaviors among providers in different areas of the country,” says Orszag. “And therein lies a massive opportunity.”

Moving Toward Improved Value

To get at that opportunity to improve health care value, there needs to be a focus on what the provider is recommending. Orszag points to the evidence on transparency, for example. A recent study found that, on average, patients in need of a lower-limb MRI scan passed six lower-cost providers before arriving at the MRI facility that scanned them.

“They would have saved somewhere between 20 and 40% even out-of-pocket by going to one of the cheaper places, with no obvious differences in quality,” he says.

“There is repeated evidence that, to a first approximation, (a) health care costs are very concentrated, so you have to focus on the high-cost cases; and (b), when there’s a lot of cost involved, what typically happens is what the doctor is recommending.”

Influencing Doctors’ Recommendations: Vertical Integration

Changing what the doctor recommends involves two pieces of work: information and financial incentives. “That brings us to these vertically integrated models,” says Orszag. “In order to change what doctors are recommending and how hospitals are behaving, you do need to change the financial incentives.” One way of doing that is through the health care payment model changes that are occurring, such as bundled payments and accountable care organizations.

“You can think of vertical integration in which an insurance company buys the delivery assets directly as an accountable care organization on steroids,” Orszag explains. “It’s no longer just weak incentives. You own all of the upside and downside risk, and it’s permanent as opposed to being renegotiated or the baseline changing each year.”

“We all agree that the payment models need to evolve. I’m in favor of just doubling down on that thought and going to vertically integrated delivery systems in which insurance companies own at least part of the delivery system, if not all of it,” he says.

Suggestive evidence in favor of this comes primarily from vertically integrated health care organizations that, on average, have outperformed other delivery systems and insurance plans. The best case of a system with vertical integration in the United States is Kaiser Permanente.

“Just pause and imagine this for a moment,” says Orszag. “Kaiser Permanente has 8% of Medicare Advantage beneficiaries nationwide. They have 80% of the five-star-ranked membership in Medicare Advantage — vastly disproportionate to their underlying share. In addition, their cost trends have been lower than competitors and they tend to offer their plans at lower premiums as a result of a managing care more tightly.”

Kaiser Permanente is not a perfect example — one could argue that California has a low Medicaid share — but the organization is one data point that vertical integration might work.

Another data point, though one Orszag is hesitant to rely on heavily, is that “all major payers are experimenting with exactly this thought process.”

“They are all going into different forms of the delivery system in different ways, trying to figure out where best to play,” Orszag says. He discloses that Lazard advised Aetna on merging with CVS and Express Scripts on merging with Cigna. “We believe that there are benefits not only for the shareholders, but also for the health care system overall from those sorts of combinations.”

There is also evidence that plans combining the pharmacy benefit manager and medical insurance functions have lower per member per month costs and avoid many incompatibilities or inconsistencies.

Aggressive Experimentation

We need to evaluate the various interventions and experimentation going on differently, and “we need to be boldly experimenting at this moment in time,” says Orszag.

“There is a lot of value improvement that the system could deliver. We don’t know exactly which models are going to work and so some humility is necessary,” he says, “but we should not take low-powered studies that have inadequate numbers of observations in order to really pick up an improvement and then find no improvement and conclude that nothing works.”

That’s a nihilistic approach that a lot of the policy world is stuck in, a feedback loop of trying something small, not seeing enough statistical power to pick up a significant effect, and therefore concluding that nothing is working. “Given how bad the status quo is, I don’t think that’s the right approach.”

One indication of this: We can’t quite figure out why health care costs have decelerated so much. Medicare spending per beneficiary has declined in inflation-adjusted terms over the past 8 years, according to Orszag. “If I had stood up when I was budget director and said, ‘I think real per-person spending in Medicare will fall,’ a group like this would have laughed at me — yet that’s what’s happened,” he says.

The evidence suggests it’s not the economy; it’s something else, yet we can’t, through statistically significant studies, pinpoint exactly what has caused that deceleration. “I would suggest to you what’s happening is business models are changing; people are anticipating that the payment models are not going to be fee-for-service in the future,” Orszag says. “Each little intervention does not have a statistically significant effect, but they’re making changes today that in aggregate are driving that deceleration.”

“Fundamentally, there’s a lot of room for improvement, and I am quite optimistic about the future,” Orszag concludes. “We are going to evolve toward better payment models, and that’s going to involve more vertical integration. We didn’t really talk about the information side of things, but I think the clinical decision support and information flow will improve dramatically, also.”

From the NEJM Catalyst event Disrupting the Health Care Landscape: New Roles for Familiar Players, held at NewYork-Presbyterian, October 25, 2018.

More From New Marketplace
A Look at the Four Pillars of Primary Care

Pay for Relationship: A Novel Solution to the Primary Care Crisis

What society should and can pay for is care that enables relationships between patients and providers.

Examples of Stages of AI Technology Development and Diffusion

How Artificial Intelligence Is Changing Health Care Delivery

The development of intelligent machines holds great promise for making health care delivery more accurate, efficient, and accessible, but challenges remain for incorporating AI technology into clinical and administrative settings.

Recommendations to Resolve Information Asymmetry at the Strategic Level

Information Asymmetry: The Untapped Value of the Patient

The knowledge and preferences that patients could — and should — share with clinicians would restore balance to point-of-care interactions, leading to better outcomes and enhanced value.

Key Components for Health Care Systems to Address Patient Affordability

The Next Frontier in Reducing Costs of Care: Patient Affordability

To create meaningful point-of-care guidance so that patients can make informed medical and financial decisions, health system leaders and policymakers can develop interventions to address four major components of a proposed patient affordability scale.

Direct-to-Consumer Telemedicine Is the Biggest Coming Threat to Traditional Health Care Organizations

Survey Snapshot: Mega-Mergers and Telemedicine Accelerate Convenient Care Growth

NEJM Catalyst Insights Council members detail how providers are looking to direct-to-consumer telemedicine and partnerships to meet the differing needs of their patient populations.

Opelka01_pullquote - ACS IPU team-based surgical care bundles playbook

Developing a Playbook for IPU-based Surgical Care and New Payment Models

The complexity associated with most surgery lends itself to the integrated practice unit structure, with its focus on the care team and value-based payment.

Convenient Care Has Been Good Overall for the Health Care Industry

New Marketplace Survey: Convenient Care — Opportunity, Threat, or Both?

A survey of the NEJM Catalyst Insights Council shows conflicting views about both the value of convenient care and what respondents’ organizations should do.

Payer-Provider Partnerships Produce Better Quality Outcomes 3 - community health plan - physician partnership

New Research Shows How Payer-Provider Partnerships Can Accelerate Adoption of Evidence-Based Care

Five best practices that are replicable and scalable are facilitating improved clinical and financial outcomes today.

30-Day Mortality Rates at Non-Teaching and Major Teaching Hospitals 2013-2014 - value-based care at academic medical centers

What Value-Based Payment Means for Academic Medical Centers

Academic medical centers must become as dedicated to advancing operational and clinical efficiency as they have been to advancing the science of medicine.

Medicare Compared to Private Spending Cumulative Growth 2009-2019 - traditional Medicare coverage

Redesigning Medicare to Work for Everyone

A proposal to improve the Medicare benefit package.


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »


Platforming Health Care to Transform Care…

Health care leaders need to focus less on ownership and control of the delivery process,…

Achieving Value in Highly Complex Acute…

To improve both the value and outcomes of ECLS, Cedars-Sinai Medical Center created guidelines for…

Build vs. Buy: What Should Health…

The consolidation craze continues, but vertical integration has yet to demonstrate real progress toward the…

Insights Council

Have a voice. Join other health care leaders effecting change, shaping tomorrow.

Apply Now