By JONATHAN HALVORSON, PhD
With each passing year, the Affordable Care Act becomes further entrenched in the American health care system. There are dreams on both the far left and far right to repeal and replace it with something they see as better, but the reality is that the ACA is a remarkable achievement which will likely outlast the political lifetimes of those opposing it. Future improvements are more likely to tweak the ACA than to start over from scratch.
A critical part of making the ACA work is for it to support healthy, competitive and fair health insurance markets, since it relies on them to provide health care benefits and improve access to care. This is particularly true for insurance purchased by individuals and small employers, where the ACA’s mandates on benefits, premiums and market structure have the most impact. One policy affecting this dynamic that deserves closer attention is risk adjustment, which made real improvements in the fairness of these markets, but has come in for accusations that it has undermined competition.
Risk adjustment in the ACA works by compensating plans with sicker than average members using payments from plans with healthier members. The goal is to remove an insurer’s ability to gain an unfair advantage by simply enrolling healthier people (who cost less). Risk adjustment leads insurers to focus on managing their members’ health and appropriate services, rather than on avoiding the unhealthy. The program has succeeded enormously in bringing insurers to embrace enrolling and retaining those with serious health conditions.
This is something to celebrate, and we should not go back to the old days in which individuals or small groups would be turned down for health insurance or charged much higher prices because they had a history of health issues. However, the program has also had an undesired effect in many states: it further tilted the playing field in favor of market dominant incumbents.
The national competitive picture has gotten worse since the ACA was passed. Today the top three insurers enroll at least 80% of the individual market in 37 states (up from 33 states in 2011) and the top three insurers enroll 80% of the small group market in 41 states (up from 37 states in 2011). In the small group market, the number of insurers enrolling over 1,000 lives declined nationally from 506 in 2012 to 409 in 2016. Even more starkly, 19 of the 23 co-op insurers created with funding from the ACA are now defunct. These co-ops failed for a number of reasons, one of which was the annual risk adjustment payments they had to make to insurers that had enrolled the majority in their local markets for decades, mostly BlueCross/BlueShield plans. Since there are many factors at work here, it is natural to ask how risk adjustment could be implicated.
There are at least three potential problems:
- Some plans have advantages in maximizing risk scores, which may not reflect differences in true underlying member population risk or lead to better care;
- The risk adjustment model makes no allowance for plan size when one or two insurers dominate a market; and
- The risk adjustment model does not account for the fact that plan designs with tighter cost management methods are often avoided by less healthy people, creating a moral hazard.
Regarding the first point, risk scoring has become its own cottage industry, to which plans devote substantial resources out of necessity (if you don’t find and record as many risks as your peers, you in effect pay your competitors for each risk you haven’t found). The effort must be repeated each year, or else the diagnoses cannot be counted towards the risk score. The risk coding must document a care plan, but the reward amount is independent of any additional services delivered. Plans with more sophisticated data mining and outreach operations to confirm diagnoses will receive higher risk scores than plans with less sophisticated operations, even if underlying health conditions are the same. In addition, larger plans with long-tenured membership have an advantage, since they have multiple years of diagnosis and claims data to analyze for eligible conditions and identify likely diagnoses to confirm. Also, when large plans receive new members, the individuals are more likely to already be in their databases from a previous enrollment than is the case for small plans.
Even if a plan can purchase outside services to mine the data and find these health conditions, having every insurer do it independently, over and over each year, does not appear to be the most efficient use of resources and it creates winners and losers based on access to data.
A very large gap in plan size can create still more issues. In Alabama, the local BlueCross/BlueShield plan has long been the dominant insurer. The ACA helped drive its share of the small group market from 90% up to 97%, while the number of competitors dropped from six to three. Responding to its crisis, Alabama requested and received a 50% reduction in the size of risk adjustment transfers.
Part of the problem in situations like this is simple math: risk adjustment transfers are based on each insurer’s deviation from the average statewide risk score. A larger insurer will always be more insulated from random variance in risk, and a very large insurer will necessarily drive the statewide average to be much closer to its risk level than to its competitors. Consider an Alabama-like example in which a large insurer has 90% market share and a 10% higher risk score than the average of its competitors. Because the number of enrollees in each plan matters when calculating the average, the statewide average risk score average will be only 1% below the large insurer’s risk level—but 9% above competitors’. To oversimplify a bit, the small insurers would be forced to pay 9% of their premium to the giant.
Of course, if the small insurers had members with 10% higher risk in this example, they would receive 9% in premium while the dominant insurer would pay 1% of its premium…though this rarely happens. This is for several reasons, such as the tendency of the dominant plan to have longer tenured members as the “blue chip” plan in an area, or as mentioned above the advantage of having a larger data repository of state residents. Both of these make it easier to capture every diagnosis. In addition, these plans often have older and sicker members than small plans, which tend to have smaller networks, engage in more active cost control measures, and be less familiar names, which older and sicker consumers select less often. There is a selection bias of higher risk members towards plan designs which do the least to control total costs—at least, with respect to measures that matter in the selection process. These cost-inflating plans include more of out-of-network benefits and more high-cost providers in their networks, and have lower utilization management.
This point applies regardless of the size of the plan, and occurs even across a single insurer’s products. For example, in Pennsylvania Independence Blue Cross operates under two ACA plan IDs, one of which is for its Keystone HMO (think smaller network, no out-of-network benefit, referrals to see a specialist) and the other for its PPO (think no referrals and extensive out-of-network benefits). The HMO owed $74M, while the PPO received $82M in statewide risk adjustment payments in 2018. Essentially, risk adjustment in its current form undermines important cost reduction strategies. Since rates are generally required to be actuarially justified after rate adjustment is taken into account, this forces the HMO to have higher rates than it otherwise would have, suppresses enrollment in the plan trying to reduce costs, and subsidizes those who enroll in the more inflationary plan.
There are also difficulties faced by smaller plans when competing in an oligopolistic market: established giants have market power, deep relationships with providers, employers and brokers, and high brand recognition and familiarity. Going back to the case of an insurer with 90% of the market and 10% higher risk score, a non-dominant insurer could not be expected to price its products 9% higher to cover the ACA risk transfer cost. These plans generally have to price low to grow, and forcing the premium higher to reflect the statewide average health care cost may cause a plan to lose what little business it has. Despite widespread belief to the contrary, health insurance is a low margin business, with profits typically in the range of 3-5%. A consistent transfer amount anywhere near 9% can wreak havoc, and under the ACA risk adjustment program the transfers are sometimes much higher. Many of these points apply not only to very small insurers, but to larger insurers that have traditionally had a small or no presence in a given market (such as an insurer that has had a presence in Medicare Advantage but seeks to expand in the ACA individual or small group market).
New York State couldn’t be more different than Alabama in many ways, but it is undergoing a similar dynamic. In the small group market, one company, UnitedHealth, has long been by far the largest player, with roughly half of the statewide small group enrollment and over 70% of the greater NYC market. Inspired by the ACA, two new insurers (Health Republic and CareConnect) initially made a splash and were able to grab market share. They enrolled everyone they could…which ended up being disproportionately younger, healthier people willing to switch plans for something new and unfamiliar in order to save some money. Other plans in the market have attempted to grow by curating the network to a select group of providers to reduce premium, and in the process have also been left with a healthier population.
To be clear, risk adjustment is critically important to balance out differences that arise from some consumer preferences. For example, high deductible plans tend to attract healthier enrollees who don’t expect to use their insurance. Without risk adjustment, these plans would become even cheaper than they already are, while more comprehensive plans that attract sicker members would get disproportionately more expensive, setting off a downward spiral that pushes more and more people into plans that have the least benefits. Risk adjustment is a fantastic way to prevent this sort of self-destruction of insurance markets. However, in other cases where risk levels differ due to preferences, such as an aversion among older and sicker people to unfamiliar plans or narrow networks, forcing insurers to pay a penalty that completely compensates for preferences can be harmful to innovation and the public interest. It is still important to allow for some risk adjustment in these cases, but fully compensating for the selection bias can create a perpetual penalty for types of plans that are actually helpful to coordinate care and control costs.
Risk adjustment is by no means the only systemic issue that has caused companies to exit the market in almost every state, but it has compounded other problems and persists year after year. For the smaller plans still competing in Alabama and downstate New York, nearly all are facing large annual risk transfers to the dominant plans. Some insurers are still losing 20% or more of their total premium in 2019, well beyond what they could make up for with premium hikes without losing membership. Meanwhile, United alone received about one billion dollars in risk adjustment transfers in New York from 2014 to 2018. Less extreme versions of these imbalances occur in state after state, from New Mexico to Illinois to Vermont.
The bottom line is that risk adjustment is a crucial tool to direct the focus of health plans and improve fairness and stability, but rewarding insurers that are better at identifying health conditions independent of overall quality and health outcomes, and correcting for all of the differences in risk scores in a state, can misdirect focus and undermine the ability of innovative insurers and progressive products to compete in the marketplace. More on this, and on what a solution could look like, in a following post.
Jonathan Halvorson is a Senior Healthcare Consultant at Sachs Policy Group and has a long-term interest in the transformative potential of technology on the health care system