Health Affairs

Insurance market reforms under the Affordable Care Act are designed to increase the number of Americans with insurance—and to shed the current system in which health plans have an incentive to enroll healthier people while avoiding the sick. One of the arrangements that will make the new system workable is risk adjustment—a process by which health insurance plans will be compensated based on the underlying health status of the people they enroll, and therefore protected against losing money by covering people with high-cost conditions.

But implementing risk adjustment could prove challenging. The statistical methods used in risk adjustment are technically complex. There are questions about the ability of the states, which have to carry out the risk adjustment, to collect accurate data and implement methodologies that result in fair payments to plans. This policy brief explains what risk adjustment is and how it works, and it examines policy issues involved in implementation.

Much of the practical knowledge that exists about implementing risk adjustment comes from experience with the Medicare program. For example, about one in four Medicare beneficiaries purchase a Medicare Advantage plan, which is a private health insurance plan that offers Medicare benefits. Payments to such private plans have always been adjusted to reflect differences in the health risks of their enrollees, initially by adjusting payments by demographic characteristics, including age, sex, and Medicaid eligibility.

Since 2000, risk-adjusted payments to Medicare Advantage plans have used data on patient diagnoses obtained from hospital admissions. Medicare’s risk-adjustment techniques have also been refined by incorporating diagnostic information from beneficiaries’ use of outpatient care and prescription drugs. Risk adjustment is also being used by many state Medicaid programs and by the Massachusetts health insurance “Connector,” a type of insurance exchange that distributes both publicly subsidized and private health coverage.

RISK ASSESSMENT AND RISK SCORE: In risk adjustment, a third party, such as the federal government or a state, collects and organizes data from insurance claims and clinical diagnoses for all enrollees in every participating health plan or provider organization in a particular market. Using what’s known as a risk-assessment tool or methodology, this entity then converts the data into a risk score for each person. Individual risk scores are then aggregated into an overall score for each insurance plan.

Here’s how the system works: If the average risk score for the overall population is defined as 1.0, a healthy young man might receive a score of 0.4 based on historical claims data, while a young woman with asthma might be scored at 1.5, and an older person with diabetes might be scored at 2.3. A plan having an aggregate score of 1.2 for its enrollees would receive a 20 percent add-on to its average per person payments, while a plan with an aggregate score of 0.8 would experience a 20 percent reduction in payments.

In practice, individual risk scores, built from data on patient demographics, disability, institutional status, and diagnoses, are used to help determine monthly payments made to plans for each person enrolled in Medicare Advantage, Medicare Part D prescription drug benefits, and many state Medicaid managed care programs.

The private health insurance market prior to passage of the Affordable Care Act was organized very differently than it will be once the law is fully implemented in 2014. Overall, insurers operating in the individual or small-group insurance markets had an incentive to enroll healthy, younger people and a corresponding disincentive to enroll older, less healthy people. These factors contributed to rising levels of uninsurance. They may also have resulted in higher premiums in the private insurance market, as health care providers raised their charges to private payers to help cover the costs of uncompensated care.

Traditionally, private health insurers operating in the individual and small-group markets have based their premiums on health history, or what is known as “experience rating.” A person with diabetes or a heart condition would be charged a higher premium than one whose health records showed only trouble-free annual checkups. Likewise, a small company with above-average care needs and costs paid more to cover its employees than it would pay if its employees were healthier.

In addition, insurers selling individual health insurance policies also have been able to deny coverage to people because of past or current health conditions. The Department of Health and Human Services (HHS) estimated that without the Affordable Care Act, up to 129 million nonelderly Americans with preexisting conditions would be at risk of losing their insurance coverage when they needed it most or might not be able to purchase affordable individual insurance coverage.

But as noted above, insurance reforms contained in the Affordable Care Act made major changes to the insurance market and insurance regulation. As of September 23, 2010, health plans were barred from excluding children from insurance coverage because of preexisting health conditions. As of January 1, 2014, plans will also be barred from using preexisting condition restrictions to prevent adults from receiving coverage. Plans will be barred as well from charging premiums based on health history.

After 2014, however, health plans will still be able to vary premium levels for individuals and small businesses based on certain factors, including age, family size, geographic region, and tobacco use. The law will allow no more than a 3:1 difference in price across age groups, meaning that older people can be charged three times as much as younger people, and no more than a 1.5:1 difference in price for individuals who use tobacco, meaning they can be charged up to 50 percent more than nonusers.

HEALTH INSURANCE EXCHANGES: In conjunction with these new insurance rules, the Affordable Care Act requires the creation of health insurance exchanges in each state, with open enrollment in the health coverage they provide scheduled to begin in October 2013. Through the exchanges, individuals and companies with no more than 100 employees will be able to shop for health insurance policies, known as “qualified health plans,” that will offer a package of “essential health benefits” and meet other standards. States have the option of establishing and operating exchanges on their own or having the federal government run one for them. (See the Health Policy Brief published on February 9, 2012, for more information on exchanges, and the Health Policy Brief published on April 25, 2012, for more information on essential health benefits.)

Because the exchanges will feature standardized benefit options and restrict insurers’ ability to base premiums on their enrollees’ health status, plans that enroll a sicker-than-average enrollee population will be in danger of losing money, while plans that enroll relatively healthier enrollees will probably be overpaid. Ultimately, if too many plans lose money, some could go out of business, and the overall system could be seriously destabilized.

To prevent this from happening, the law requires the use of risk adjustment to reallocate premium income among plans to account for differences in their enrollees’ aggregate health conditions, and therefore the likely cost of paying for their care. Risk-adjustment methods similar to those used in Medicare Advantage or the Medicare Part D prescription drug benefit may be used, or states can propose an alternative methodology subject to approval by HHS.

If a state does not establish a system of risk adjustment, HHS will establish and operate one for that state. On March 23, 2012, HHS published a final rule on risk adjustment along with two related programs—a transitional “reinsurance program” and so-called risk corridors, both of which are discussed below. HHS plans to publish its proposed risk-adjustment methodology in the fall of 2012.

Risk adjustment will be required of all qualified health insurance plans sold to individuals and small groups both within and outside of the exchanges. “Grandfathered” health insurance plans—those in existence at the time the Affordable Care Act was signed into law—are exempt from risk adjustment as well as many other provisions of the health care reform law. However, it is widely expected that many of these plans will disappear over time.

MITIGATING RISK: The Affordable Care Act creates two additional mechanisms to carry out risk adjustment and help mitigate financial risk for insurers during 2014–2016—the first three startup years of the risk-adjustment program (Exhibit 1). The first, a “transitional reinsurance program,” will be run by each state or, if a state elects not to have one, by HHS. The second, “risk corridors,” will be run by HHS.

Exhibit 1
Download Powerpoint Slide

• Transitional reinsurance program. All nongrandfathered health insurance issuers and self-insured group health plans in a given state will be assessed contributions based on their relative market share to a temporary reinsurance program for that state. The state will then channel funds to any plans in the individual market that end up covering people who have extraordinarily high medical costs. Although the details remain to be worked out, the reinsurance payments will cover a percentage of claims paid above a certain level—referred to as an “attachment point”—and up to a specified cap or maximum amount.
  
  
• Risk corridors. Financial arrangements known as risk corridors will be structured to protect insurers from the consequences of having to pay for care for high-cost individuals. These will be particularly useful in a period of transition, such as is likely to be the case in 2014 when many sicker people and those with preexisting health conditions will be buying coverage through insurance exchanges for the first time. Under this arrangement, insurance plans whose costs turn out to be at least 3 percent less than their target cost projections—presumably because their enrollees experienced fewer health problems than expected—will pay a percentage of the money they saved to HHS. The agency will then use that money to compensate insurers whose actual costs turned out to be more than 3 percent higher than projected—presumably because their enrollees had more health expenses than initially projected. The payments in effect will cover a portion of any losses that the plans incurred on high-cost individuals.

The reform law’s individual insurance requirement will bring many previously uninsured people into the exchanges. Plans will have little or no data with which to predict future service needs of these enrollees and estimate their premiums, but they will need to do so nonetheless. Some people will have pent-up needs that will lead to high service use in the short run but then taper off; others may have chronic illnesses that result in high costs over a long period.

Risk adjustment involves technically complex and data-intensive methods that will challenge insurers’ ability to predict costs and the ability of federal and state agencies charged with implementing the new arrangements. Here are some of the issues that health insurers, states, and the federal government will be facing.

• Data systems and methodology. Risk adjustment can be undertaken in three different time periods: on a prospective basis, that is, at the start of a given plan year; concurrently, or while a year is underway; or retrospectively, after a plan year is over. In prospective adjustment, data from a past year can be used to project current year payments. Or data from the current year can be used, subject to being reviewed and possibly changed at the end of the year—for example, if some enrollees turn out not to be as sick as expected. Collecting and analyzing all of these data requires that insurance plans and states alike have sophisticated electronic capabilities to collect and analyze information. Some states may decide to contract with an outside third party or private service provider to set up and manage their risk-adjustment systems, especially if they lack the experience or necessary systems infrastructure.
  
  Moreover, although much has been learned from Medicare’s experience with risk adjustment under Medicare Advantage, it may not translate well to the populations covered and the plan offerings in the individual and small-group markets once the Affordable Care Act is fully implemented. For example, a simulation by the RAND Corporation found that risk adjustment had varied success in adjusting payments across the different types of plans—bronze, silver, gold, and platinum—to be offered in the exchanges. The availability of diagnosis-based information will also be critical to making risk adjustment more accurate. For example, a simulation involving nonelderly patients conducted by Johns Hopkins University researchers found that under- and overpayments were reduced fivefold when data on diagnoses were used compared to use of only demographic factors, such as age and sex.
  
  
• Centralized vs. distributed data systems. Regardless of who operates a data system for a state, each state will have to choose between running what is termed a centralized data collection system or a distributed data collection system. Centralized systems are generally run by government agencies or contractors and work by collecting data supplied by insurers. It is easier with a centralized system to ensure the integrity and consistency of data and thus to produce accurate assessments of the impact on insurers. However, centralized data collection systems may also elicit concerns about medical privacy because the data come from private health plans, not from public programs like Medicare or Medicaid.
  
  Some states may opt instead for using a decentralized data collection system, in which insurers calculate their own scores based on centrally determined formulas and data elements, and then forward this information to the states. Although this approach eliminates the need to centrally report, collect, and store private health information, it may also produce less-accurate results while increasing the administrative burden on the states, which still have to assemble and collate the submitted results. As mentioned, states may also adopt federal methodologies, or even decline to operate exchanges at all, in which case HHS will run both the exchanges and the risk-adjustment system.
  
  
• Addressing “upcoding.” The success of risk adjustment will also depend upon the accuracy of the data that come from health plan payments to providers. This is because the costs that insurers will have to pay for care won’t just be driven by those individuals’ underlying health status. The costs will also be driven by how doctors, hospitals, and other providers “code” the cases for payment by insurers. Providers may code patients’ conditions at the highest level of severity consistent with the treatment they provide, a practice known as upcoding. They may also treat and seek payment for additional medical conditions, such as arthritis or diabetes in a heart patient, even when these conditions might not be relevant to the particular clinical event.

Upcoding can undermine risk adjustment if it distorts the actual health-risk profile of a plan, for example, by suggesting that the people that the plan has enrolled are actually sicker than they really are. Thus, those running risk-adjustment systems will have to audit plans to enforce coding “integrity”—that is, consistent use of diagnosis and procedure codes to negate any effect of upcoding.

As noted, HHS plans to publish its proposed risk-adjustment methodology in the fall of 2012. More broadly, federal and state regulators as well as members of Congress will be likely to watch closely as exchanges go live and insurance reforms kick in during 2013 and 2014, and risk adjustment is fully implemented in 2014 and beyond.

There has been agreement among many policy makers since the early 1990s that structured insurance market competition is an important path to cost containment and improved health system performance. At the same time, the scale of insurance reforms and the exchanges created by the Affordable Care Act are largely new and untried. Moreover, the success of these innovations hinges on methodologies that are challenging to understand even by experts, let alone the lay public. The outcome of this massive experiment will not be known until it is tested in practice.

Yet the ultimate success of health reform depends critically on this esoteric science of risk adjustment. Appropriate payment is the key to creating an environment where health plans will compete on the basis of quality and efficiency, rather than avoidance of high-cost patients. This is the kind of competition that many believe will be necessary to curb unsustainable growth in the cost of care, and risk adjustment is central to it.

American Academy of Actuaries, “Risk Assessment and Risk Adjustment,” May 2010.

Brown, Jason, Mark Duggan, Ilyana Kuziemko, and William Woolston, “Will Risk-Adjustment Decrease Health Care Costs? New Evidence from the Medicare Advantage Program,” National Bureau of Economic Research, August 21, 2011.

Department of Health and Human Services, “Bulletin on the Risk Adjustment Program: Proposed Operations by the Department of Health and Human Services,” May 1, 2012.

Hall, Mark, “Risk Adjustment under the Affordable Care Act: A Guide for Federal and State Regulators,” Commonwealth Fund, Issue Brief No. 1501, Vol. 7, May 2011.

HealthCare.gov, “Standards Related to Reinsurance, Risk Corridors, and Risk Adjustment Final Rule,” March 16, 2012.

Park, Edwin, “Ensuring Effective Risk Adjustment an Essential Step for the Success of the Health Insurance Exchanges and Market Reforms under the Affordable Care Act,” Center on Budget and Policy Priorities, May 18, 2011.

Pope, Gregory, John Kautter, Melvin J. Ingber, Sara Freeman, Rishi Sekar, and Cordon Newhart, “Evaluation of the CMS-HCC Risk Adjustment Model,” RTI International, March 2011.

Weiner, Jonathan P., Erin Trish, Chad Abrams, and Klaus Lemke, “Adjusting for Risk Selection in State Health Insurance Exchanges Will Be Critically Important and Feasible, but Not Easy,” Health Affairs 31, no. 2 (2012): 306–15.

Winkelman, Ross, Julie Pepper, Patrick Holland, Syed Mehmud, and James Woolman, “Analysis of HHS Final Rules on Reinsurance, Risk Corridors, and Risk Adjustment,” State Health Reform Assistance Network, April 2012.

Written by
Rob Cunningham
Health Policy Consultant
(Cunningham was previously a deputy editor and contributing writer at Health Affairs.)

Editorial review by
Katie Merrell
Senior Research Scientist
Social and Scientific Systems

Jonathan P. Weiner
Professor of Health Policy and Management
Bloomberg School of Public Health
Johns Hopkins University

Ted Agres
Senior Editor
Health Affairs

Anne Schwartz
Deputy Editor
Health Affairs

Susan Dentzer
Editor-in-Chief
Health Affairs

Health Policy Briefs are produced under a partnership of Health Affairs and the Robert Wood Johnson Foundation.

Cite as:
"Health Policy Brief: Risk Adjustment in Health Insurance," Health Affairs, August 30, 2011.

Sign up for free policy briefs at:
www.healthaffairs.org/healthpolicybriefs