Pre-ACA Market Practices Provide Lessons for ACA Replacement Approaches

Significant changes to the Affordable Care Act (ACA) are being considered by lawmakers who have been critical of its general approach to providing coverage and to some of its key provisions. An important area where changes will be considered has to do with how people with health problems would be able to gain and keep access to coverage and how much they may have to pay for it.  People’s health is dynamic. At any given time, an estimated 27% of non-elderly adults have health conditions that would make them ineligible for coverage under traditional non-group underwriting standards that existed prior to the ACA. Over their lifetimes, everyone is at risk of having these periods, some short and some that last for the rest of their lives.

One of the biggest changes that the ACA made to the non-group insurance market was to eliminate consideration by insurers of a person’s health or health history in enrollment and rating decisions.  This assured that people who had or who developed health problems would have the same plan choices and pay the same premiums as others, essentially pooling their expected costs together to determine the premiums that all would pay.

Proposals for replacing the ACA such as Rep. Tom Price’s Empowering Patients First Act and Speaker Paul Ryan’s “A Better Way” policy paper would repeal these insurance market rules, moving back towards pre-ACA standards where insurers generally had more leeway to use individual health in enrollment and rating for non-group coverage.1  Under these proposals, people without pre-existing conditions would generally be able to purchase coverage anytime from private insurers.  For people with health problems, several approaches have been proposed: (1) requiring insurers to accept people transitioning from previous coverage without a gap (“continuously covered”); (2) allowing insurers to charge higher premiums (within limits) to people with pre-existing conditions who have had a gap in coverage; and (3) establishing high-risk pools, which are public programs that provide coverage to people declined by private insurers.

The idea of assuring access to coverage for people with health problems is a popular one, but doing so is a challenge within a market framework where insurers have considerable flexibility over enrollment, rating and benefits.  People with health conditions have much higher expected health costs than people without them (Table 1 illustrates average costs of individuals with and without “deniable” health conditions). Insurers naturally will decline applicants with health issues and will adjust rates for new and existing enrollees to reflect their health when they can.  Assuring access for people with pre-existing conditions with limits on their premiums means that someone has to pay the difference between their premiums and their costs.  For people enrolling in high-risk pools, some ACA replacement proposals provide for federal grants to states, though the amounts may not be sufficient.  For people gaining access through continuous coverage provisions, these costs would likely be paid by pooling their costs with (i.e., charging more to) other enrollees.  Maintaining this pooling is difficult, however, when insurers have significant flexibility over rates and benefits.  Experience from the pre-ACA market shows how insurers were able to use a variety of strategies to charge higher premiums to people with health problems, even when those problems began after the person enrolled in their plan.  These practices can make getting or keeping coverage unaffordable.

Table 1: Average Health Costs for Non-Elderly Adults With and Without Deniable Health Conditions, by Age Range, 2014
Age Range Average Costs
With Deniable Condition Without Deniable Condition
18-34 $5,190 $1,809
35-44 $6,371 $2,279
45-54 $10,195 $2,657
55-64 $11,537 $4,641
18-64 $8,853 $2,527
SOURCE: Kaiser Family Foundation analysis of data from the Medical Expenditure Panel Survey.

The discussion below focuses on some of the issues faced by people with health issues in the pre-ACA non-group insurance market.  These pre-ACA insurance practices highlight some of the challenges in providing access and stable coverage for people and some of the issues that any ACA replacement plan will need to address. Many ACA replacement proposals have not yet been developed in sufficient detail to fully deal with these questions, or in some cases may defer them to the states.

We start by briefly summarizing key differences between the ACA and pre-ACA insurance market rules for non-group coverage that affect access and continuity of coverage.  We then focus on pre-ACA access and continuity issues for three different groups: (1) people transitioning from employer coverage or Medicaid to the non-group market; (2) people with non-group coverage who develop a health problem; and (3) people who are uninsured (are not considered to have continuous coverage) who want to buy non-group coverage.  After that, we discuss how medical underwriting and rating practices can segment a risk pool, initially and over time, and challenges that this poses for assuring continuous coverage.  We end by reviewing some of the policy choices for addressing the challenges that have been raised.

Non-Group Insurance Market Practices Before the ACA

The ACA significantly simplified the rules for health insurance enrollment, rating and benefits in the non-group market.  Generally, benefits are the same for all policies offered in a state, with four levels of cost sharing (bronze, silver, gold, and platinum).  Insurers cannot consider a person’s health at enrollment or in determining their premium.  People can enroll in any plan during an annual open enrollment period or other times under special circumstances (called special enrollment periods), such as the loss of prior coverage.

The ACA was a substantial departure from prior insurance practices in most states, where insurers had far more flexibility over enrollment, rating and benefits.  State laws and practice varied — for example, a few states required insurers to accept all applicants and prohibited rating variation based on health, similar to the ACA — but this was not the norm.  In most states, insurers were permitted to consider health in their enrollment and rating decisions.  Some of the more important differences between ACA and pre-ACA market rules are described here.  Their implications for providing access to coverage and assuring continuous and stable coverage are discussed in the next sections.

  1. Medical Screening of Applicants. The first and most obvious difference is that insurers could ask applicants about their health and generally could deny coverage to people with health problems.  They also could choose to accept the applicant at a higher premium, and, in many states, could accept the applicant but limit the terms of the coverage to exclude benefits related to a specified health condition (for example, an insurer could exclude benefits related to asthma).  Underwriting decisions could vary with the type and level of coverage sought: an insurer could deny enrollment in a policy with a lower deductible to an applicant with a relatively minor condition, such as acne, but might accept them in a higher deductible plan or in a plan without drug coverage.
    As will be discussed in the next section, the Health Insurance Portability and Accountability Act (HIPAA) provided access to coverage for people with at least 18 months of prior coverage, if the most immediate prior coverage was in a group health plan (generally a plan offered pursuant to employment by a public or private employer, but not Medicaid or Medicare).  Insurers were required to accept these applicants (called “HIPAA-eligible” individuals) without a pre-existing condition exclusion, but generally could charge them much higher rates than other applicants.  States could specify an alternative coverage mechanism for HIPAA-eligible applicants; 38 states specified an alternative, with most specifying a state high-risk pool.  HIPAA-eligible individuals without health problems could choose to apply for medically-underwritten non-group policies, but doing so made them subject to preexisting condition exclusion provisions (see Medical Underwriting and Pre-existing Condition Exclusion Provisions box below).
  2. Multiple Rating Classes for Similar People in the Same Policy. Another difference is that premiums for people of the same age from the same place could be quite different for the same policy.  Except for a differential for smoking, people of the same age from the same place face the same (unsubsidized) premium for the same plan under the ACA.  Prior to the ACA, there were many rate classifications. For example, there could be a rate for new applicants who have no health conditions, there could be several “substandard” rate tiers for people with health problems, there could be different rates for people based on how long they have had the policy (durational rating, described more below, which means that a newly issued 40 year old would pay a lower rate than a 40 year old who bought the same policy two years prior), there could be different rates based on how the policy was purchased (through an agent, directly from the insurer, through a trade group), the person’s occupation, and others.  Also, from year to year, the rates in each class could change by different percentages, increasing the differences for similar people in different rating classes.
  3. Rating by Policy or Block. A third difference relates to how premiums are established for different policies offered by an insurer in a state.  Under the ACA, where all policies cover the same essential health benefits, an average expected cost is estimated for all projected enrollees across all of an insurer’s non-group products in a state, and premiums for particular policies are determined on the objective differences (i.e., cost sharing and provider network) from the average cost. In contrast, prior to the ACA, premiums were established for each policy (or a group of policies, sometimes called a block) based on the expected claims costs for the people expected to be enrolled in that policy or block, projected over current and future years.  Importantly, the experience of each policy or block is developed independent of the costs or results expected in other policies or blocks, which means that two policies that are almost the same could have very different premiums associated with them based on the anticipated costs of who is projected to be enrolled (and who has actually enrolled).  As discussed more below, a policy or block of policies no longer for sale to new people (called a closed policy or block) would likely have much higher premiums for the same benefits than a policy currently available to new enrollees.2
  4. Broad Variation in Benefits Across Policies. Another difference is that there was significant variation in the benefits covered by pre-ACA policies, including options that excluded entire classes of benefits such as prescription drugs or mental health.  Under the ACA, all policies cover the same essential health benefits, with variations largely relating to cost sharing and network.  Pre-ACA policies sometimes had annual or lifetime limits on specific or total benefits: for example, a policy might limit prescription benefits to $500 per year.  Most states specified some benefits that needed to be covered or at least offered to applicants by insurers.
  5. Limited Ability to Switch Among Non-Group Plans. A fifth difference relates the ability of a person with non-group coverage to switch policies without re-submitting to medical underwriting.  Before the ACA, people who were accepted into a non-group policy were not necessarily able to switch into new non-group policies, at renewal or otherwise, either from their current insurer or from others, without passing medical underwriting.  Insurers sometimes offered people the ability to elect different policies at renewal (usually the ability to take a policy with higher cost sharing to moderate a rate increase), but they were not required to do so and did not have to allow current policyholders to move to different policies.

Medical Underwriting and Pre-Existing Condition Exclusion Provisions

Prior to the ACA, insurers often used the health of individual enrollees in making decisions about their coverage.  Medical underwriting is the process by which an insurer acquires information about the health of applicants for coverage and uses the information to make decisions about whether to offer coverage, what coverage to offer, and what premium to charge.  Applicants for non-group coverage generally were required to answer a long series of questions about their health and health history, and often were required to provide authorization for the insurer to obtain their medical records.  In the non-group market, insurers generally were permitted to use the information to decline the application, accept the applicant for a reduced scope of coverage, or accept the applicant at a higher premium.

A pre-existing condition exclusion provision is a contract term that permits an insurer to exclude coverage for benefits sought by an enrollee during a defined period after the coverage begins (for example, twelve months) if the insurer can show that the claim relates to a condition that existed before the policy was issued.  State laws varied in defining pre-existing conditions for non-group coverage; for example, in how far an insurer could look back to detect the condition, or in whether the condition must have been actually treated or whether a reasonable person would have sought treatment.  This exclusion allowed insurers to exclude benefits for pre-existing conditions that were not necessarily detected during the medical underwriting process.

While there are many other differences between ACA and pre-ACA non-group market rules (e.g., permitted cost sharing, limits on age rating), these have the most implications for providing access to and continuous coverage for people with health problems.  Most fundamentally, medical screening divides people by health at initial enrollment, and the inability to switch policies can trap people who develop health problems into much more expensive coverage.  The potential implications of this are discussed below.

Issues Raised by Pre-ACA Non-Group Market Rules for Access to Coverage and Continuous Coverage

To examine the issues raised by these pre-ACA market rules, we look at three different groups of people:

  1. People transitioning from existing coverage and applying for non-group coverage
  2. People with non-group coverage who develop health problems
  3. People without recent prior coverage applying for non-group coverage

People Transitioning From Existing Coverage to Non-group Coverage

Assuring access to non-group coverage for people who maintain continuous coverage has been a priority for proponents of changing the ACA.  One of the attributes of the ACA is that people who lose eligibility for coverage can obtain replacement coverage in the non-group market on the same terms as others covered in the market, without consideration of their health.

A large number of people who lose their coverage might want or need access to non-group coverage.  Looking at the 2012 through 2013 period (the 24 months immediately prior to the ACA coverage expansion), about 32 million people with coverage lost it and became uninsured for some period. People without health problems leaving previous coverage generally could purchase underwritten policies in the market.  Some, but not all, people with health problems who had previous coverage could qualify for designated non-group policies without regard to their health.

As noted above, prior to the ACA, federal law provided guaranteed access to non-group coverage for people with at least 18 prior months of continuous coverage if their most recent prior coverage was an employer plan and if they did not have a gap in coverage of more than 63 days.  These HIPAA-eligible individuals qualified for specified policies (most often, coverage in a state high-risk pool, but sometimes designated plans offered by non-group insurers), with no pre-existing condition exclusion.  Their premiums were almost always much higher than the rates charged to applicants who could pass medical underwriting.

The HIPAA non-group market provisions were perceived generally to have fairly limited effect, primarily because the coverage made available could be expensive.  Several factors limited HIPAA’s effectiveness in assuring access to non-group coverage:

  1. Eligibility. The guaranteed access and waiver of pre-existing condition exclusion provisions were available only to a limited group of people: those whose most recent previous coverage was involuntarily terminated and employment-based.  Limiting the option to people leaving employer group plans left out people coming from public coverage such as Medicaid or who lost a prior non-group plan because they moved out of area served by their insurer.  A few states expanded the requirement to include other types of coverage, but it was not the general rule.  The provisions also left out people who wanted to switch plans within the non-group market, for example, because of network changes in their existing plan or if it had become unaffordable (discussed below).
  2. Cost. Federal HIPAA portability provisions also did not limit the premiums that could be charged for the specified plans available to HIPAA-eligible people.  Most states used a high-risk pool to serve HIPAA-eligible people, where premiums typically ranged from 125% to 200% of the estimated standard premiums for non-group coverage.  With a couple exceptions, income-based subsidies were not available in high-risk pools, making it quite difficult for people with modest incomes.  In states where insurers were required to make private policies available to HIPAA-eligible individuals, insurers often were able to charge much higher premiums to HIPAA-eligible individuals with health problems; for example, insurers could develop separate rating classes for HIPPA-eligible individuals who could meeting medical underwriting standards and those would could not.  A few states limited the additional premium that could be charged to HIPAA-eligible individuals who could not pass medical underwriting.
    In addition, HIPAA only extended guaranteed availability to people after they had exhausted their eligibility for continuation coverage under COBRA or under state continuation laws.  Continuation coverage can be expensive: COBRA premiums are 102% of the full cost of the employer plan for at least 18 months.  Affording COBRA can be difficult for people who lost their job and may not have new work.
    The requirement to exhaust continuation coverage and the relatively high premiums served to limit the number of people who could afford to take advantage of the guaranteed availability opportunity under HIPAA.  As discussed below, people who could pass medical underwriting could save these expenses and enroll in lower-cost plans, but they would not get the full benefits of having continuous coverage.
  3. Combining Guaranteed Access and Waiver of Pre-Existing Condition Exclusion in the Same Provision. The law provided for guaranteed access to coverage and the waiver of pre-existing condition exclusion provisions only in specified policies, which tended to be quite expensive.  HIPAA-eligible individuals who were healthy and could pass medical underwriting could get a non-group policy for much less than the policies offered generally to HIPAA-eligible people, but in choosing the cheaper policy they sometimes exposed themselves to a new pre-existing condition exclusion period, despite the fact that they had at least 18 months of continuous coverage.  Many people may not even have understood that they were making this tradeoff.

A different kind of issue facing people leaving employer group coverage or Medicaid who wanted to maintain continuous coverage were the limits on benefits in many non-group policies.  One of the significant changes in non-group coverage under the ACA was the establishment of a fairly comprehensive essential health benefit package.  In particular, pre-ACA non-group policies had significant limits on mental health benefits (mental health parity requirements, which applied to employer-group plans for employers with more than 50 employees, did not apply to non-group coverage), and, unless required by states, typically excluded coverage for many policies, and also did not cover costs associated with pregnancy or routine delivery.  Some state high-risk pools, which were the only options for HIPAA-eligible individuals with health problems, had tight limits on coverage for prescriptions.3,4

Prior to the ACA, non-group coverage was decidedly less comprehensive than employer group coverage.  Substantial shares of non-group enrollees did not have coverage for routine maternity, substance abuse or mental health services, and it was not uncommon for policies to have relatively low annual benefit limits for prescription drugs or mental health services.  Even though insurers were able to medically screen applicants in most instances, they still imposed significant limits on benefits where there is a greater chance of purchasers selecting coverage based on the need for particular services.  Unlike the group market, where employers select levels of benefits for all their employees, insurers are wary of non-group purchasers who are willing to pay the relatively high cost for more comprehensive benefits.  These benefit limits, along with the rating issues discussed in the next section, meant that the non-group market was not a good long-term coverage option for many people, including those who wanted to start a family or who developed mental health problems.

People enrolled in non-group coverage who develop health problems

Another aspect of maintaining continuous coverage is being able to keep the coverage you obtain on a reasonable basis.  Prior to the ACA, non-group coverage generally was guaranteed renewable, which meant that enrollees had the right to renew their coverage (with certain limited exceptions) by paying their premiums.  Insurers also generally were not permitted to vary renewal premiums based on an enrollee’s individual health or claims.  Insurers, however, through selectively closing policies or blocks of business to new enrollees and through certain rating approaches, were able to access higher premiums than enrollees who developed health problems after they enrolled.  As discussed above, people with non-group coverage generally were not able to switch carriers or move to a new policy (in an open block of business) unless they could pass medical screening.  As a result, they could find themselves essentially locked into policies with escalating premiums that could be difficult to afford.

This can happen several ways.  The medical underwriting process allows insurers to protect themselves from adverse selection (see The Issue of Adverse Selection box below), but it also produces complicated dynamics that can segment risk by health even after people in good health have been accepted into coverage.  Medically screening new applicants, and declining applicants who are unhealthy, produces a group of healthy new enrollees whose expected claims costs over the short term could be meaningfully below the costs for an average mix of people.  Prior to the ACA, the expected low costs for these enrollees would be reinforced because the group also would generally be subject to a pre-existing condition exclusion provision for the first year that eliminated coverage for claims for pre-existing health conditions not uncovered during the medical underwriting process.  Over time, however, some of the group of enrollees would develop health problems, and the average costs of the group would grow each year; by year three or four after their enrollment the expected costs for the group would roughly equal the expected costs for an average mix of people.  This is sometimes referred to as “underwriting wearing off.” An insurer, at any given time, will have a group of recently underwritten enrollees, with relatively low expected costs, and other groups enrolled for varying lengths of time, with the tendency for those enrolled longer to have worse average health. If an insurer closed these older products to new enrollees – and allowed healthy enrollees in them to sign up for new, medically-underwritten products – premiums for existing enrollees would escalate over time, and those with medical conditions would essentially be trapped into paying those higher premiums because they could not switch to other coverage.

The Issue of Adverse Selection

Prior to the ACA, insurers used medical underwriting in the non-group market to protect themselves and their policyholders from adverse selection.  Unlike coverage offered to large employer groups, where insurers anticipate getting a mix of better and worse health risks when they accept a new group, non-group coverage is sold person-by-person.  While virtually everyone wants to have health insurance, people with high or ongoing health needs are more likely to sign up at any given price, a tendency referred to as adverse selection.  Adverse selection occurs not only in the decision of whether or not to purchase coverage, but also in decisions about how much coverage to get (people in poorer health tend to want more comprehensive benefits and less cost sharing) and in decisions about whether or not to keep coverage (people in better health are more likely to drop coverage or move to less coverage in the face of premium increases).  The relatively high cost of health insurance makes adverse selection more acute (premiums can be a large portion of a monthly budget, so there is a tendency for healthier people to forgo coverage if they do not think they will need it).  This is particularly an issue in the non-group market where enrollees pay the full premiums.

There are several ways insurers can reflect these differences in their rating and enrollment practice rates, and this a place where problems can occur for people who develop health problems after enrollment.  One option is for insurers to combine the new and existing enrollees in blocks of business that are being actively marketed (called “open” here), so that low expected costs of new enrollees can help offset the higher costs of enrollees who have been covered longer.  As long as there is a reasonable stream of people entering and leaving the block, premiums can remain reasonably spread over the entire group.  Insurers also can pool the expected total claims of each durational group of enrollees over their average expected length of enrollment.  This requires charging new and early-duration enrollees for more than their expected costs during their early years, setting aside a portion of the premium (i.e., creating a reserve) that can be used to offset the higher costs for those who keep their policies for longer periods.

Some insurers, however, may not want to pool the lower costs of new entrants with the higher costs of longer-term enrollees.  For example, insurers with larger and older blocks of business may find that they cannot compete well for new enrollees against insurers without as much existing business, because those insurers would have a higher proportion of new healthy enrollees and could have lower rates for new business, particularly if the new carrier is not reserving for the effects of underwriting wearing off.  An insurer also might develop a new group of policies based on a new approach (for example, a policy where it shares risk with an Accountable Care Organization (ACO) network) where it does not want to pool experience with its existing policies in determining rates.  An insurer also may want to increase its market share by being more competitive for new enrollees, which it might do by setting the premiums for new enrollees closer to their expected first year costs.

Insurers that want to reduce the pooling of newer and longer-term enrollees have several ways to do so.  One is to use the duration of enrollment as an explicit rating factor.  Insurers using durational rating can set initial rates relatively low for new enrollees, but will need to raise them relatively rapidly each year (on top of increases for rising health costs generally) for these enrollees to reflect their higher expected claims at later durations.  Another option is for an insurer to stop selling policies in blocks of business to new enrollees, directing them to new policies in a new block of business without any existing enrollees.  Because premiums are set based on the expected costs for specific policies or blocks of business, premiums for the new policies do not need to reflect the costs of the existing enrollees in the closed block, and future premiums for the closed block will reflect only relatively higher average costs of the existing enrollees.

Both of these practices end up harming enrollees who develop health problems.  Enrollees facing the relatively higher premiums under durational rating or in a closed block will look for lower cost alternatives.  Healthier enrollees who can pass medical screening will move to lower cost policies (essentially starting over as new entrants), while people with health problems who cannot move will have to stay and pay the higher premiums being charged.  The new round of higher premiums will cause more of the healthier enrollees to leave, resulting in higher expected costs for those remaining and higher premiums, a cycle that will continue until most enrollees have left the block.

People without recent prior coverage applying for non-group coverage

There was a substantial number of people without health insurance prior to the ACA, many of whom had been without coverage for long periods of time.5 The primary reason people went without coverage was its cost, although in some cases people were unable to qualify for coverage due to their health.6  The two factors sometimes worked together; many states had high-risk pools or similar options for people with health problems who were denied non-group coverage, but the high premiums and other limitations could make these options difficult for people to afford and the pools had fairly low enrollment.

High-risk pools are being discussed as an important part of ACA replacement proposals.  About 227,000 people were enrolled in 35 state high-risk pools at the end of 2011, including HIPAA-eligible individuals, which was equal to just over 2% of non-group market enrollment nationally.7  A few states with relatively lower premiums, such as Maryland, Wisconsin, Minnesota, and Oregon, covered somewhat higher shares of their people.8  Enrollment in state high-risk pools tapered off with the opening of the federal Pre-Existing Condition Insurance Pool, created and funded under the ACA, which served many of the people who previously would have been covered in the state pools.

State high-risk pools varied in terms of benefits, premiums, and funding.9  As noted above, in many states the high-risk pool served as the state-designated mechanism to cover HIPPA-eligible individuals.  There were a few common themes: premiums generally were calculated as a percentage of estimated standard premiums in the non-group market (typically 125% to 200% of standard premiums); coverage for pre-existing conditions was limited for a period after enrollment; pools generally offered several benefit options, most states had lifetime benefit limits and a few had annual limits; premiums did not cover the cost of benefits, with the difference subsidized by state and federal payments (a few states had dedicated revenue sources) or assessments on insurers.

A combination of factors limited the attractiveness of pre-ACA state high-risk pools.  The relatively high premiums made coverage difficult to afford for people with low or modest incomes, and only a couple of states had subsidies for lower-income enrollees.  In addition, pools generally had pre-existing condition exclusion periods for enrollees who were not HIPAA-eligible individuals, which means that people were required to pay for coverage that would not cover the illnesses that had made them eligible for the high-risk pool in the first place for six months to a year or more (depending on the state).  A few state pools also had annual limits on some or all benefits, and the majority had lifetime benefit limits.  Given the populations served, these limits could affect those with high cost chronic conditions, such as the ongoing need for expensive prescriptions.

A few states addressed access for people with health problems by requiring all insurers (or in some cases, one or more designated insurers) to accept applicants even if they were in poor health.  Premiums in these states tended to be much higher than premiums in states that permitted medical underwriting, which limited participation in non-group coverage significantly and made coverage even more difficult to afford for people with modest incomes.

Discussion

There were many aspects of the pre-ACA non-group market that made it difficult for people with health problems to get and keep non-group coverage.  Any proposal for replacing the ACA will have to determine which, if any, of these previous insurance practices will once again be permitted.  Medical screening was the most obvious barrier, combined with high premium costs for people who were HIPAA-eligible.  Even people who purchased coverage when they were healthy sometimes were unable to keep it because certain rating approaches could cause their premiums to spiral.  Returning to a less structured, less regulated non-group market raises questions about how people with health problems will be treated in terms of access to and cost of coverage.  Health insurance underwriting and rating is complex, and reviewing how the pre-ACA market operated provides information about the types of issues that people with health problems may confront if the ACA market structure is replaced.

KFF Headquarters: 185 Berry St., Suite 2000, San Francisco, CA 94107 | Phone 650-854-9400
Washington Offices and Barbara Jordan Conference Center: 1330 G Street, NW, Washington, DC 20005 | Phone 202-347-5270

www.kff.org | Email Alerts: kff.org/email | facebook.com/KFF | twitter.com/kff

The independent source for health policy research, polling, and news, KFF is a nonprofit organization based in San Francisco, California.