separator Health Economist header tab round table

Dissecting The Health Insurance Exchanges

A Penn Research Group Burrows Into the Complexity

Despite the potentially large impact on family finances and health, the average buyer on the Massachusetts health insurance exchange during 2007 to 2009 tended to spend less than 20 minutes selecting and purchasing a policy, according to researcher Amanda Starc. That evidence of short attention spans is one more piece of the health insurance exchange puzzle that Starc and her University of Pennsylvania colleagues are beginning to ferret out and piece together.
A Wharton School Assistant Professor of Health Care Management, Starc is a member of the Leonard Davis Institute
panel headerAmanda Starc

Alternative content

Amanda Starc, PhD, University of Pennsylvania Wharton School Assistant Professor of Health Care Management and LDI Senior Fellow
Bob Town

Alternative content

Robert Town, PhD, University of Pennsylvania Wharton School Associate Professor of Health Care Management and LDI Senior Fellow
Tom Baker

Alternative content

Tom Baker, JD, a Penn Law Professor, expert in insurance law and LDI Senior Fellow
Mark Duggan

Alternative content

Mark Duggan, PhD, University of Pennsylvania Wharton School Professor of Business Economics and Public Policy, former Senior Economist for Health Care Policy on the White House Council of Economic Advisers, and LDI Senior Fellow
of Health Economics' (LDI) Health Insurance Exchange (HIX) research group and has extensively studied the Massachusetts Connector, the pioneering exchange that is a model for those now required in every state by the Affordable Care Act.

She was one of four Penn HIX group members who took part in a recent roundtable discussion about some of the thorniest issues facing state and federal officials designing the nation's new network of health insurance exchanges. The other participants were Mark Duggan, a Wharton School Professor of Business Economics and Public Policy, and former Senior Economist for Health Care Policy on the White House Council of Economic Advisers; Tom Baker, a Penn Law Professor and expert in insurance law; and Robert Town, Wharton School Associate Professor of Health Care Management.

'Choice architecture'
Launched late last year, the 12-member group of health policy researchers is collecting data on federal and state government health insurance exchange structures. By analyzing the mechanics of these digital systems' "choice architecture," the Penn scholars hope to understand and identify strategies and elements that optimize the functions of the exchanges on which tens of millions of Americans in the individual and small-group market will soon be purchasing their health insurance.

Starc's study found that after a brief perusal on the Massachusetts exchange website, sixty percent of shoppers picked an insurance plan from the "Bronze" or lowest-priced policy tier. Using a heuristic or rule-of-thumb approach, twenty percent simply selected the cheapest and least generous plan. But was this behavior ultimately good or bad for the consumer?

"It depends on whether folks are enrolling in plans that make sense for them after they take enough time to figure out which plan that is," said Robert Town, who co-directs the research group with Tom Baker. "But the evidence suggests that people aren't very good at selecting appropriate health plans."

Long-term implications
Town notes bad decisions have long-term implications. "For instance, people with chronic conditions who need to take maintenance drugs and select a Bronze plan have to pay more out of pocket, which may lead them to be less diligent about purchasing their drugs, so their health deteriorates in that context. We see that in other places, like Medicare Part D where, when they go into the 'doughnut hole' and have to pay more for drugs, many seniors change their behavior and start pill splitting and the like."

Tom Baker pointed out other potential problems: "Someone may choose the cheap Bronze plan in the exchange but then that individual doesn't qualify for the program's cost-sharing," he said. "That person might actually be better off by selecting the more expensive Silver plan and getting the subsidy that makes it less expensive -- but he or she may not understand that."

Complicated for typical consumer
Mark Duggan said available evidence makes it clear the onus "is really on the states to make sure their exchanges maximize residents' possible benefit. So if you're a family of two adults and two kids, you log onto an Amazon-like exchange site on October 13, 2013 and try to figure out what plan of available options is going to work for your family, given your resources, given the multiple providers, given the multiple plans, given the cost parameters, given the doctors and hospitals you want, and it's complicated. I think there's going to be a ton of variation across the states with respect to how well they distill that complexity for a typical person without leaving out critical information. For example, it's going to be difficult for these exchanges to offer information about what providers are included in the various networks. Take Los Angeles County, where there are 130 hospitals and thousands of physicians. How do you know who's in what plans?"

And, noted Baker, once you do know, how do you keep that information current for customers making real-time purchase decisions? "A doctor may be in a network this week but not next week," said Baker, "or he or she could have stopped accepting new patients. It's a dynamic situation."

"There's some pretty good buzz among the consultants who are advising exchanges that the exchanges ought to be focusing on picking defaults for people that are simple and good enough," Baker continued. "So long as at least a couple of states actually do that, we could have an incredible opportunity to compare what works and what doesn't."

~ ~ ~

So far, the general public debate about health exchanges has been focused on the narrow political question of which states will or will not build exchanges. A good many pundits have suggested that the enormity of the task and fast-approaching deadlines make it all but impossible for a national system of exchanges to be functional enough to begin enrolling millions of customers in just 13 months.

But the Penn HIX group members disagree, and note that while individual states may face various obstacles, it is very likely the federal government will be setting up exchanges in as many as 30 states as scheduled -- although the shroud of secrecy in which the federal exchange structure is being developed has kept its true progress out of the news.

Feds will make deadline
"My sense is there will be some kind of exchange up everywhere on October 13, 2013," said Duggan. "The feds are bringing a lot of resources to this and they have experience in this kind of thing. They learned a lot from their work on Medicare Part D," he said, referring to the national online exchange-like system providing insurance coverage for prescription drugs for the elderly. It launched in 2006 after a frenzied two-year build, and during its first year of operation signed up 24 million people.

"That was a very complicated product aimed at millions and it went relatively smoothly. CCIIO has the manpower and the money to do the insurance exchange even better," said Duggan referring to the Center for Consumer Information and Insurance Oversight, an office within the Centers for Medicare & Medicaid Services (CMS) that is responsible for the overall exchange program.

Baker, who has been in contact with the CCIIO exchange office since it began work, agrees. "They're really all over this," he said. "And in the history of government programs, there's plenty of time to finish something like this on time."

The Department of Health and Human Services Department has budgeted more than $860 million for the overall exchange project. The website structure, a template that can be customized for each state, is being developed by CGI Federal Inc. The federal data server hub, which will pull from various federal and state databases to verify the identity, income, citizenship and legal address of exchange users in real time, is being built by Quality Software Services Inc. (QSSI) Both companies operate under tight security and secrecy.

Complaint from the states
"That's one of the complaints we're hearing from the states," said Town. "There's been no glimpse of what the federal exchange actually looks like or the rules underlying its structure and function. It's a serious issue because many states are still asking themselves 'should we do our own exchange or go with the federal model?'"

In a July briefing with insurance commissioners, CCIIO Interim Director Mike Hash said the exchange structure was complete enough to have begun system-to-system interface tests with some states and insurance companies. He said the project was on track to be fully operational by October of next year when the first exchange enrollments are scheduled to begin.

But, HIX group members point out, the ultimate success of the exchange concept depends on getting large quantities of small details right within some of the most complex data systems the health care industry has ever seen -- the most complex of which is the exchanges' risk adjustment mechanism.

~ ~ ~

As it greatly expands the national population of people covered by health insurance, the Affordable Care Act creates potential pitfalls for insurers in the new exchange market. For one thing, insurance companies can't refuse to insure anyone, no matter how sick the individual is likely to be. For another, insurers must cover new sorts of "bundled" or "episodic" treatments whose actual profit and loss potential is currently unknowable.

In less than two years, millions of new customers will stream into this system, potentially making the exchanges a perverse lottery for insurers -- randomly hobbling some with customer bases heavily weighted with relatively sicker people and rewarding others with customer bases equally oversubscribed with relatively healthier people.

Offsetting lopsided risk levels
To offset such perilously lopsided risk levels, the ACA requires each state exchange to implement an ongoing financial adjustment system that can quantify risk variation and transfer compensating payments from insurers with lower-risk customer pools to those with higher-risk pools.

Although the federal government is creating its own risk adjustment methodology, it hasn't publicly defined exactly how such a system will work. Its final risk adjustment rules are scheduled for publication this January.

The ACA authorizes the CCIIO to operate its own risk adjustment system in the exchanges it sets up in states that decline to create their own; states that do establish exchanges can elect to either use the federal risk adjustment system or create an alternative that must be approved by federal authorities.

Risk adjustment systems, which involve the ongoing analysis of all insurance transactions, customers and claims in each state, are massively complicated algorithmic structures; that complexity poses potential difficulties that haven't been widely discussed in the public debate about insurance exchanges.

For instance, in some ways, the risk adjustment process could serve as a tool for insurers seeking an unfair advantage, according to Town.

'Gaming' the exchange system
"The 'gaming' potential is a result of the risk adjustment mechanisms being so imperfect," he said. "If an insurer is able to work that in combination with subsidies, which are also complex, then that carrier may be able to enroll a lot of people who kind of 'look' sick and are subsidized and also get bonus risk-adjustment payments on top of that. An insurer may be able to make a killing by working both sides."

For this reason, Starc predicted that, "the entire country is going to get a lot sicker on paper."

Duggan agreed. "Imagine I'm an insurer and all the people with cancer sign up for my plan because my network has the state's best oncologists. I'm going to lose a lot of money if the state doesn't risk adjust for that. So, if you're a plan that happens to get relatively many cancer patients or heart attack patients, you can get higher payments, he said.

"But," he continued, "depending on the risk adjustment formula that's used, an insurer will have an incentive to give people the absolutely most thorough physical of their lives when they join because if there is even a trace of conditions like cancer or diabetes or whatever, the insurer may be able to get more risk adjustment money."

"The thing I worry about," Duggan said, "is the complexity -- we're talking about 50 different exchanges and risk adjustment systems. They're going to price some things wrong. They're going to pay too much on this condition and not enough for that condition. The devil is in the details and it's going to be unbelievably hard. Medicare's been trying to do this sort of risk adjustment for 30 years and still hasn't figured out how to do it."

~ ~ ~

Another issue with big implications that's not being widely discussed involves whether or not insurers might be allowed to "slice" up a state, particularly in relation to rural populations.

Duggan explains: "Think of Texas -- there are only five big markets there: Houston, Dallas, San Antonio, Austin and El Paso. But most of the state is outside of those. Typically, insurers don't work hard to get rural market share because the economies of scale are so spread out it's not worth a lot of money. So, how is a state going to handle the rural areas? If an insurer plans to offer a product in the Houston area, does that insurer have to offer that product statewide? It's a big decision facing states because it can have a huge effect on what kind of options are available on the exchanges, especially for people in rural and lower-income places."

Provider network adequacy
In a similar way, Baker said the exchanges' "network adequacy" will be an important issue. "The states have to decide how broad and generous the networks will have to be, and this will affect the power of providers who are negotiating with insurers. So, if you HAVE to have this particular hospital in this particular town, you're going to give that hospital a lot of bargaining leverage that's going to affect cost."

"I'm thinking of this," said Baker, "because I was just visiting my dad, who is a doctor in Genesee County in upstate New York where there is exactly one hospital. So if there is a rule that says Genesee Memorial HAS to be in a network, that hospital has a lot of bargaining power in that tiny little county."

"You also have the issue that some insurers could decide not to bother going into an area," said Duggan. "Then you end up with one or two insurers in an area and one of them might think 'why not price a family plan at $25,000?' because the consumers are only going to be bearing 4% or 6% of their income and the federal government will subsidize the rest. So, if states don't carefully think through both the market power on the provider side and the insurer side, there can be some completely unexpected things happening in terms of how expensive plans can be."

The bottom line, said Baker, is that "there are going to be some weird places in the country" unless states establish statewide policies that don't allow insurers to cherry pick by region.

~ ~ ~

Hoag Levins is a journalist and managing editor of The LDI Health Economist.

blog comments powered by Disqus

Share This Page

share icons



The LDI Blog


LDI Twitter

Data Briefs

LDI/RWJF Joint Reports

Main LDI Site

Health Economics Center


Business News Journal

~ ~ ~


~ ~ ~