The recent release of risk corridor results by the Centers for Medicare and Medicaid Services (CMS) has highlighted a national issue for health insurers offering coverage in the individual market under the Affordable Care Act (ACA). In many states, some insurers did not project premium rates at sufficient levels to account for the risk of the newly insured population. Thus, there were insufficient risk corridor “payables” available from conservatively-priced plans to cover all risk corridor “receivables” for underpriced plans to compensate for risk corridor-eligible losses.

However, this phenomena is highly state-specific. In contrast to most of the country, Covered California, California’s health insurance exchange, had much greater success enrolling a diverse population and did not have a market that allowed the continuation of low-risk, remaining “transitional” plans once the ACA was implemented. As a result, the health risk scores of enrollees for individual plans in the exchange have stabilized in the second year (regressed to the mean), with risk being more evenly distributed across all of the plans.

At the same time, the overall risk score statewide for enrollees with chronic conditions has decreased, indicating that the 2015 cohort of enrollees is slightly healthier than the 2014 cohort. Understanding a population’s health risk, which is a proxy for anticipated health care costs, is critical to accurately setting the next year’s premiums. As an active purchaser, Covered California used this information to better negotiate 2016 premiums on behalf of Californians.

2016 Premium Growth Is Lower Than 2015

Each of the actuaries from the ten major insurance carriers participating in Covered California in 2015 used the risk score analysis along with claims data to set 2016 premium rates lower than anticipated and lower than the individual market’s historical average. (The exchange added two more health insurers—United Healthcare Group and Oscar—for 2016.) Partly as a result of sharing these analyses with carriers, Covered California’s premiums in 2016 are only increasing by a weighted average of 4.0 percent, which is less than the 2015 weighted premium growth of 4.2 percent. This also benefits the entire individual insurance market, since Covered California’s negotiated rates carry over into off-exchange premiums also.

The health risk findings disprove what many ACA detractors once opined: that health insurance exchanges will not be able to attract healthy enrollees, and the resulting death spiral will increase premiums to unsustainable levels. The ACA is working in California. Other health insurance exchanges can also use a comparative risk score analysis in conjunction with active purchasing tactics to reduce premium increases.

Developing The Risk Scores

An initial study analyzed the risk scores of Covered California’s 2014 enrollee population, finding that risk scores varied considerably by plan but that the overall enrollment group had a relatively low number of people with chronic conditions. Using the same approach, we updated the analysis for the 2015 enrollee cohort with more recent hospital discharge and emergency department data.

Using individual-level unique identifiers, we combined two datasets: data from Covered California’s enrollment system as of May 2014 and April 2015; and, the 2013 data files (the most recent available) from the California Office of Statewide Health Planning and Development (OSHPD ) emergency department and inpatient discharge datasets (a statewide all-payer claims database).

Next, we derived risk scores for each plan based on the individual-level diagnoses contained in the OSHPD datasets, when present, and the age and sex of each enrollee derived from Covered California eligibility data. In calculating these the scores, we applied risk weights developed using the Chronic Illness and Disability Payment System (CDPS) for the Temporary Assistance for Needy Families (TANF) population under California’s Medicaid program (MediCal). This allowed us to better align the derived risk scores with the Covered California population, the majority of whom earn less than 250 percent of the Federal Poverty Level and are non-disabled (similar to a TANF population). For the vast majority of individuals who did not have either a hospitalization or an emergency department visit in 2013, we derived the risk weights based on enrollees’ age and sex only.

As a side note, we used the already-available risk weights from MediCal in our 2014 analysis because of time constraints. To have risk scores available in time for the 2015 premium negotiating sessions, one of the authors had his team from the MediCal program use the available TANF risk adjustment weights, rather than use a longer process to input the more complex Hierarchical Condition Categories risk adjustment system used by the U.S. Department of Health and Human Services. In updating the analysis for 2016 premium negotiations, we continued to use the CDPS/TANF weights to provide consistent results and better year-over-year comparisons.

We then standardized the risk scores by dividing the unadjusted score for each health insurance carrier by the average risk score for Covered California’s entire population. The resulting adjusted risk score normalizes the distribution to 1.00 for the Covered California enrollment (but doesn’t include any off-exchange enrollment), such that scores below 1.00 are relatively lower risk than scores above 1.00. Comparing the normalized risk score of enrollees in each of the ten health insurance carriers allowed us to assess any differences in average risk scores across the carriers. Lastly, we compared the two-year historical trend to evaluate how risk scores changed.

Risk Scores Are Stabilizing And Improving

Figure 1 shows the 2014 and 2015 mean risk scores by carrier. Without exception, every carrier’s population mean risk score moved closer to 1.00 in Covered California’s second year, which is good news for the stability of premiums. It means that the enrollment of each specific carrier is moving towards the statewide average risk, so no carrier appears to be attracting mainly “good risks” or mainly “high risk” individuals. This enrollment pattern is important for Covered California’s long-term stability.


Source: Covered California

Additionally, overall hospitalization and emergency department data showed that enrollees had less severe conditions in 2015, compared with 2014. The raw (non-normalized) risk scores for those who were hospitalized decreased by over 9 percent from 2014 to 2015, even though the percent of enrollees who were hospitalized or had an emergency department visit remained stable. (In 2014, 9.2 percent of enrollees had a hospitalization or emergency department visit, compared to 9.3 percent of the 2015 enrollee cohort.)

The hospitalization and emergency department visits and the derived risk scores provide only proxies for an individual’s or a population’s risk, and do not perfectly reflect the actual health risk for each individual enrollee. The risk score analysis is not a crystal ball and will not predict the future, but the approach is rooted in actuarial science and accepted as an important gauge of future health costs.

The leveling off of risk scores across carriers and the overall reduction in hospitalizations and emergency department severity mean that Covered California has enrolled enough people, and enough healthy people, to stabilize premiums and keep premium growth within reasonable limits, which is great news for Californians.

Working By Design

Covered California is working by design. The Affordable Care Act was designed to level the playing field for all health insurers, regardless of the relative health of the enrollees they might attract, in comparison to their competitors. Health insurance carriers in a state must use a single risk pool with both exchange and non-exchange enrollees to avoid adverse risk selection. The guaranteed issue requirement allows individuals to buy much-needed health insurance even if they have pre-existing conditions. Subsidies in the form of premium tax credits provide financial assistance to those who do not have health insurance through employer sponsored coverage.

In Covered California, there is no death spiral. Premiums are not skyrocketing. In fact, premium growth is lower in 2016 than it was last year, and the number of uninsured Californians is decreasing. Also by design, Covered California serves as a sort of “way station” for those who are in between other sources of health coverage. These shorter-term enrollees are a critical part of Covered California’s stability and success.

California May Be First To Use All the Tools Available; Other States Can Follow

If the risk score analysis provided the justification to lower 2016 premiums, it was Covered California’s role as an active purchaser that served as the mechanism to do so. Covered California opted early on to use all of the tools available under the ACA, including the ability to select the health plans with the highest value to be sold in the exchange; to standardize all plans’ health benefits so that health insurance companies had to compete on price, provider network, and quality; and to actively negotiate premiums with carriers to get the best value for Californians.

Covered California also opted to require health plans contracting with the exchange in 2014 to immediately transition all of their policies to ACA-compliant products (with no “transitional grandfathered” products remaining), thereby making a short-term political sacrifice to gain long-term market stability. Health insurance exchanges in other states do not employ all the tools available in the ACA, and many states prolonged the transition of non-ACA compliant health plans.

California, like a few other states, also benefits from a state-wide all-payer claims database, which is needed for a population-level risk score analysis. But many other states have access to the federal government’s Healthcare Cost and Utilization Project data, which can be used in place of a state-specific claims database.

Covered California is an “early adopter” of the new tools, but only because the size of the state and its exchange allow it to leverage all of the tools available. Other states, too, can use similar methods to calculate the enrollee population’s risk and the relative health risk among insurance carriers. The critical step, though, is to use those data to ensure any premium increases are proportional to the anticipated cost of delivering health care. Other states’ health insurance exchanges can also take more actions to achieve low premium increases; exchanges can be sources for data-driven and actuarially sound analysis that keeps health insurance premiums affordable.