Genetic Discrimination and the Future of Health Insurance

Have health insurance companies outlived their useful life?

The Genetic Information Nondiscrimination Act (GINA) is taking effect.  The employment provisions are effective on November 21.  The health coverage provisions began to apply on May 21, 2009, with group coverage required to comply with on the start of the plan or policy year following that date.  As a Monday NYT story describes,  the purpose of GINA is to prohibit discrimination in employment and health coverage on the basis of genetic condition.  The Times cites examples of denial of coverage on genetic grounds;  some additional “horror stories” are told of people denied coverage due to a genetic marker for a medical condition.

Genetic discrimination is widely regarded as an “unfair” basis on which to deny health coverage.   As then-President George W. Bush explained in 2001:

Genetic discrimination is . . . unjustified — among other reasons, because it involves little more than medical speculation. A genetic predisposition toward cancer or heart disease does not mean the condition will develop. To deny employment or insurance to a healthy person based only on a predisposition violates our country’s belief in equal treatment and individual merit.

This assessment seems to match our general sense of fairness.  But why is a genetic predisposition different from other predispositions for insurance underwriting purposes?  Does hypertension mean that stroke or heart attack will develop?  Does the occurrence of a heart attack mean that another will occur?  Not necessarily, but both are statistically valid indicators of future health costs, as are some genetic predispositions.  One person’s medical speculation is another’s actuarial probability.  So why ban the use of genetic information, but not the use of preexisting illness?  Or, for that matter, age and gender?

GINA passed overwhelmingly, but Congress has otherwise been slow to limit health-based underwriting.  Part of the explanation for carving out genetic information is its novelty, and the fact that one’s DNA is never one’s “fault.”  But manifestation of genetic conditions is also not one’s fault, and GINA has nothing to say about underwriting based on symptomatic genetic disease.  Maybe a more powerful force behind GINA’s passage was researchers’ fear that genetic discrimination would drive test subjects away from genetic research.  Dr. Francis Collins, then-Director of NIH’s National Human Research Genome Institute, testified to this concern:

[T]he science of genomic medicine is rocketing forward. But fear of genetic discrimination threatens to slow both the advance of such groundbreaking biomedical research and the integration of the fruits of that research into our nation’s health care.

So, is GINA’s incursion into health insurers’ common freedom to charge more for those likely to be sick an anomaly?  It does seem to run afoul of what Donald Light, in 1992, wryly called the “inverse coverage rule”:

the inverse coverage law: the more people need coverage, the less coverage they are likely to get, or the more they are likely to pay for what they get.

Increasingly, Americans seem to have had it with business models premised on denying care to those most in need.   A recent WSJ/NBC poll found that the most popular currently proposed reform provision “by a mile” was that forbidding exclusions for pre-existing medical conditions.

Here’s a modest proposal.  It is time to end insurers’ use of risk selection.  Insurers face steady erosion in their ability to risk-select as that practice increasingly rubs against our basic moral principles.  In addition, health coverage is less and less like “insurance,” as routine care is covered, and as it is easier and easier for insurers and consumers alike to discern who is likely to need to draw down on the “insurance” pool.   Insurers already do a great deal of their business, without bearing risk or using risk selection to generate income, administering self-funded plans for large employers (as ASOs or TPAs).  It is time for government to bear the risk (which it increasingly does anyway), and for insurers to be set to tasks that add social value.  They add value in their ability to form, maintain, and administer provider networks.  Why should their profit or loss depend on their ability to identify and exclude the needy?   Clearly health reform in 2009 won’t get us this far, but let’s start admitting it: exclusion from coverage and care on the basis of need for health care should not be a business model supported in America.

Making the Case for the Public Plan, Part I: The Difficulty of Private Health Insurance Regulation

As health reform moves to the top of the Congressional agenda, we will be hearing a lot about a possible “public option” in the plan. Earlier this Spring I began thinking about whether a public option was absolutely necessary to a successful reform. I started out hoping that it wasn’t, because Republican leaders despise it, and Democrats have sometimes let the “perfect be the enemy of the good” in health reform. But I’m now convinced that a public option is necessary, and I hope to spend a few posts explaining why.

To begin with, we should get clear on exactly what insurers do. I have tried to summarize it in a one page chart, which appears here. The right column focuses on the purely positive role of insurers–how they add value to the health care system. With massive amounts of data at their disposal, they can identify best and worst providers, good and bad treatments, and even spot dangerous side effects in drugs and devices. They can invest in new technology to better process claims. To the extent that they retain long-term relationships with customers, they have an incentive to reduce costs by keeping those patients healthy.

But the structure of the US health insurance market makes it difficult for most private insurers to respond to such incentives. About 21% of insurance policyholders cancel their plans in any given year, meaning that the average customer’s commitment to a plan lasts for about three years. That’s just not enough time for an insurer to gain much investing in the health of its members.* There are many more profitable strategies–which lead me to the left side of the column, bad insurer practices.

Health care costs are highly concentrated among a small portion of the population. As AHRQ notes, “Half of the population spends little or nothing on health care, while 5 percent of the population spends almost half of the total amount.” (The famed 80/20 rule also applies in health care expenditures.) This creates almost irresistible pressures for private insurers to “risk select;” i.e., to avoid covering those who need care most. While “pre-existing conditions” exclusions and recissions are most common in the individual insurance market, they and other tactics can undermine the idea of risk-pooling at the core of any feasible insurance scheme. Given that many private insurers began thriving by cherry picking (and lemon dropping) the healthiest (and sickest) customers, they have long resisted regulation of risk selection.

But now, as the chances for reform increase, leading private insurers are beginning to soften their approach in order to argue that a public plan is not necessary. They are promising to accept “guaranteed issue” coverage, “with no pre-existing condition exclusions.” They have even promoted plans for “risk adjustment,” which “spreads costs for the highest-risk individuals.” Would regulation like that preclude the need for a public option?

I don’t think so, because there are so many other ways for insurance companies to drive away the sickest customers. As noted in the chart, subtler selection can include refusal to respond to needs of high cost patients in order to drive them away, and attracting a disproportionate share of low‐risk individuals. For example, a plan might decide to increase coverage of gyms and cosmetic procedures (to attract fit customers) and devise complex forms to be filled out monthly in order for a patient to get oxygen or insulin (to repel customers with congestive heart failure or diabetes). These are not merely hypothetical concerns. The Netherlands is often held up as a model for US reform because of recent moves there to make their traditionally solidaristic system more market-oriented. But risk selection threatens to unravel the Dutch “middle ground:”

[After the Dutch moved in a more American direction, insurers] have more tools for managing care, which can also be used to select risks. . . . Insurers have more room to define the precise entitlements of their insured groups, which can be used to select favorable risks. Third, insurers are allowed to sell mandatory health insurance together with any other type of non–life insurance (such as supplementary health insurance, sick leave insurance, and car insurance), which prior to 2006 was not allowed.

In particular, supplementary health insurance can be an effective tool for risk selection, because insurers are allowed to reject applicants based on their health status. Fourth, insurers are free to give premium rebates to groups for the mandatory basic insurance, which prior to 2006 was not allowed. A group can have any risk composition, and the “organizer” of the group can selectively enroll preferred members only. Although the rebate for the basic insurance is at most 10 percent, insurers can give these groups any rebate on supplementary health insurance or other insurance products. . . . Given the increasing incentives and expanding tools for risk selection, further improvements of the risk-equalization method are necessary to prevent insurers from engaging in risk selection, which occurs, for example, in Switzerland.

US insurers are sure to import methods like that, and to continue along current lines of risk selection. As health policy expert Karen Pollitz has noted, all of the following tactics can be used to risk select:

–“Street” underwriting
–Selective marketing (including in competing markets)
–Renewal rating
–Closed blocks
–Benefit designs
–Payment practices
–Provider network design

Congress or HHS or state insurance commissioners could try to outlaw or restrict risk selection practices one by one. But as Pollitz has noted, as of 1997, the “US Department of Labor had resources to review each employer-sponsored group health plan under its jurisdiction once every 300 years.” The Bush years probably did not significantly address that shortage. Moreover, “state insurance department staff levels declined 11% in 2007 while premium volume increased 12%.” The personnel simply aren’t there, and when they are, they are as likely as not to be outgunned by private sector attorneys, lobbyists, and experts-for-hire. The right way to discipline private insurers is to have competition from a public option–not to allow them to continue a risk-selection race-to-the-bottom by deflecting regulation.

I have taught health care regulation at both Seton Hall and Yale Law Schools, and my students have always been dismayed by the cat-and-mouse games that regulators and insurers play to control (and evade control of) risk selection. I have very little faith that DOL, HHS, or their state equivalents (who are also often tasked with regulating life and auto insurance and banks) can really make private insurers accountable, no matter how ingeniously the insurance exchanges are designed.

So that’s a case for the public plan largely based on the problems with private insurance regulation. For a positive case, which I’ll develop in my next post, I’ll focus on the middle column of the chart–eternally contested insurer actions designed to ration access to providers.

*For recognition of this problem in the context of bariatric surgery, and a creative plan for solving it, see Ronen Avraham and K.A.D. Camara, The Tragedy of the Human Commons, 29 Cardozo Law Review 479 (”bariatric surgery is just one example of insurers’ failure to cover prospectively efficient treatments. A similar confluence of insureds switching insurers frequently, high transaction costs of individualized contracts, and medical-industry lobbying explain insurers’ failure to cover other prospectively efficient treatments.”).

X-Posted: Concurring Opinions.