What Can Higher Education Learn from Outcomes-Based Health Care Funding?

Higher education has long struggled to solve an important problem: how to fairly recognize and reward institutions for serving their students well when they enroll drastically different populations.¹ To date, the federal government’s response to this question has been to ignore it altogether while states have attempted to introduce outcomes-based or performance-based funding models to various levels of success.² As a result, more often than not, these systems have left colleges with little financial incentive to improve student outcomes or move toward more equitable practices designed to serve all students well.³

Massive new investments into the higher education system that are now being proposed create new openings for federal policymakers to revisit how we could consider student outcomes as part of federal higher education policy. Outcomes-based funding has often gotten a bad rap in the higher ed space because some initial designs at the state level tended to favor wealthy, predominantly white institutions and did little to improve student success or distribute resources more equitably to the schools and students who need them the most.⁴ However, a 2020 report from the Education Trust found that intentionally designing outcomes-based funding models to use metrics that are inclusive of race and income could help achieve greater equity and make these funding models more successful. Luckily, higher education doesn’t have to reinvent the wheel as it can, and should, look to other sectors for lessons on how to implement policies that recognize and reward institutions that are doing their part to get good outcomes while serving traditionally underserved populations.

A prime example is the health care sector, which has faced similar challenges in trying to find ways to incentivize more equitable access to quality care. One method, known as risk adjustment, attempts to create more funding parity across the health care system by rewarding insurers that are doing their fair share to provide quality care to harder-to-serve patients. This memo explains how health care has successfully integrated risk adjustment into its funding system to improve patient care and outcomes, and it highlights four key lessons for policymakers and advocates looking to better hold higher education institutions accountable for student outcomes moving forward.

A Quick Primer on Risk Adjustment as an Outcomes-Based Funding Model in Health Care

Lowering the cost of health insurance and providing better medical care have long been the two staples of health care policy conversations. There has been widespread debate about how to achieve these twin goals and determine the right approach when health care can be an unpredictable market and unexpected costs can easily arise.⁵ Fortunately, new funding models are being used to control these costs and help patients get the care they need by capturing more accurate and complete data on patient conditions and incentivizing insurers to better care for them.

One model, recently implemented as part of the Affordable Care Act (ACA), is designed to offset the cost of providing health insurance to higher-risk individuals and proactively account for patient condition and care. It is a process known as “risk adjustment,” and it built on similar policies enacted previously in the Medicare program. In the health care sector, insurance covers unexpected, bad events that result in risk for individuals and communities. But some conditions like chronic diseases produce predictable expenses and can require a lifetime of care, resulting in higher costs. Risk adjustment selects many costly conditions, predicts future expenditures for those conditions, and pays insurers extra for covering patients who have them. It uses objective metrics based on patient demographics and health conditions to distribute funds, which will be discussed in more detail below. And it was a key component of protecting people with pre-existing conditions because insurers that on average enroll patients with more costly conditions now receive additional funds distributed by the Center for Medicare & Medicaid Services (CMS) to pay for the extra care they must provide. Previously, insurers could charge higher premiums to people with pre-existing conditions—or even refuse to cover them altogether.

When done properly, risk adjustment incentivizes insurers to enroll individuals with varying health statuses and provide every patient quality care.⁶ And while risk adjustment considers patient characteristics on the front-end, it also risk-adjusts how outcome measures are reported on the back end, so it doesn’t punish those who see “high-risk” patient populations and undermine the incentives. This allows risk adjustment to capture both inputs and outputs when it comes to patient care. The federal government has no actual role paying for this process, as insurers who serve low-risk healthy populations are required to provide the additional funds to subsidize those who cover patients who need a greater level of care. In 2019, 561 insurers participated in this program and 554 received a transfer.⁷ In the same year, risk adjustment transfers across all risk pools totaled to $10.8 billion ($5.4 billion in payments and the same amount in charges).⁸ Of the participating insurers, most will receive a transfer. It’s common that some insurers insure only risky patients, but it’s less likely that they’ll insure no risky patients. So, they still get a small payment for the few risky patients they do insure but the dollar amount is significantly lower than what other insurers receive. Distributing funding in this way can ultimately help lower the cost of coverage, provide more accurate reimbursement to insurers for care, and improve patient outcomes because both insurers and providers have better data on their conditions.⁹ 

What Higher Education Can Learn from Risk Adjustment in Health Care

Like in the health care sector, conversations around creating a more consumer-centric and equitable system have peaked in higher ed in recent years, even more so in the wake of the COVID-19 pandemic. Most institutions ultimately want to get the best outcomes possible for their students, as do doctors for their patients. Equipping schools with the resources they need to serve students well is a big part of helping them reach that goal.¹⁰ But we have a lot of work left to do. As recently reported by the National Student Clearinghouse, around 40% of all college students still haven’t completed a degree after six years.¹¹ And 9.7% of all students in repayment for certain direct federal student loans currently default within six years, with that number spiking to 32% for Black borrowers.¹² Further, a college degree can often be a ladder out of poverty, but the system consistently fails students who need that economic mobility the most, as less than 15% of low-income students earn a four-year degree compared to the more than 60% of wealthy students who do.

A funding method that aims to improve these outcomes, among others, could help target funds to institutions that are educating their fair share (or more) of historically underserved students. It’s important to note that we are not equating certain types of students to patients with chronic illnesses, but to acknowledge that not all institutions are playing the same role in our system. We should find ways to create funding parity in higher education and recognize and reward institutions that are admitting and achieving good outcomes with traditionally marginalized populations, and we can apply four key lessons from similar efforts in the health care sector to make this happen.

Lesson 1: Use Multiple Performance Metrics.

What Happened in Health Care: For any funding policy incorporating outcomes to work, it needs to use objective and agreed-upon metrics that aren’t arbitrary. Risk adjustment in health care took this concept and ran with it. Policymakers recognized that a variety of metrics were needed to fully capture an insurer’s patient population and the care those patients receive.¹³ To do this, policymakers created a hierarchy of over 100 different medical conditions, with each enrollee anonymously evaluated based on how many of those conditions they meet.¹⁴ Additionally, other factors are considered like whether the individual has a combination of conditions or the condition’s severity. Demographic data are also included like age, gender, marriage status, where a patient lives, prescription drugs, disability status, and employment status.¹⁵ The formula used in risk adjustment is extremely complex involving many data points that took years to determine. This effort was necessary to ensure the data was perceived as unbiased, accurate, and high quality.¹⁶ Part of risk adjustment’s success is also a result of the health care sector working together to agree upon which metrics to use, creating buy-in that has helped with implementation.

Lesson for Higher Ed: The lesson that higher ed can draw from health care is clear: use multiple performance metrics when incorporating outcomes into new formulas and models. It’s hard to evaluate outcomes when looking at only one data point, as that one metric may not show the whole picture of how a college or university is serving its students. Federal graduation rates are an excellent example of this problem, as they only account for first-time, full-time students and leave out the 53% of students who are part-time or have transferred schools. They can also easily be gamed by institutions who could hand out diplomas without actually equipping students with the skills to get quality jobs.¹⁷ And this is not the only higher ed metric that has its drawbacks when examined in isolation. Workforce-related data, like a student’s earnings a certain number of years post-enrollment, are only reported to the federal government for students receiving federal financial aid, which accounts for only 71% of the current college-going population (and much less at certain institutions).¹⁸ And for many years, much of this data was only offered at the institution level, so we had no idea how students fared within specific programs. Using multiple performance metrics can also make it harder for higher education institutions to game the system, because it’s not a make-or-break situation that’s based on one factor alone.

Various data points tracked by the federal government could help allocate funds, but they need to work together to give a full picture of how institutions are serving their students in all stages of the college-going process. Lifting a 2008 ban on student-level data would improve metrics around retention, completion, post-enrollment earnings, and loan repayment rates, and others that could be used in combination to assess which schools are truly successful at serving traditionally underserved students. And like with risk adjustment, it will be easier to implement if there is buy-in from across the system on what metrics are used.

Lesson 2: Make Thresholds Secret.

What Happened in Health Care: Determining which insurers serve low-risk or high-risk populations requires the use of an actuary, who labels an insurer a “high-risk insurer” or “low-risk insurer” based on the risk adjustment formula, which considers its enrollees’ anonymized individual risk scores. Once these scores are determined, CMS instructs insurers with lower actuarial risk to make payments to higher-risk insurers. To prevent any manipulation of the system by insurers or hospitals, a handful of the objective metrics used to calculate a risk score annually are kept secret from insurers for much of the year.¹⁹ This ensures that no company can game the system by maximizing their payout at a level just above the thresholds. Keeping the process fair and accurate is crucial to making sure funds go to insurers who need the additional dollars to care for their patient populations and prevent others from qualifying for additional funding without serving the patients that need it most.

Lesson for Higher Ed: A question that often comes up in higher education accountability and student outcomes conversations is whether the metrics being used can be gamed or manipulated by colleges and universities to improve their standings or avoid sanctions. There are already examples of colleges and universities taking advantage of existing loopholes in federal law to do just that.²⁰ For example, an important outcome metric currently tracked by the federal government is the Cohort Default Rate (CDR). CDR measures the share of students who default on their federal student loans within three years of entering repayment.²¹ If a college or university’s CDR is at 30% for three years in a row or 40% for a single year, it can lose access to federal financial aid. Knowing that the consequence is the complete loss of federal aid, some predatory institutions have gamed the metric by pushing students into forbearance to keep them out of default just until the three years have passed, even when this is not in students’ best interests.²² Making thresholds a secret could help prevent any gaming by predatory colleges and universities. If institutions don’t know what the thresholds are, it will take away an easy incentive to manipulate the system by skating by just above them.

Lesson 3: Counter-Incentivize “Skimming” Among Institutions.

What Happened in Health Care: A key consideration in outcomes-based funding is preventing predatory behavior by those being regulated before it starts. In health care, this meant finding ways to counter-incentivize skimming (when insurers sell policies only to those who are healthy or avoid insuring them when they get sick).²³ The risk adjustment program helps do this on the front end by preventing skimming while distributing funding more equitably. This also allowed policymakers to stop insurers from charging more to sicker people who have a greater need for coverage (known as community rating) and risk-skimming, where insurers avoid covering those who are likely to cost more (also known as risk selection).²⁴ In addition to these funding mechanisms, the health care sector also risk-adjusts how it reports out outcome measures, like cancer hospitals’ quality rankings and mortality rates, to avoid penalizing providers who exclusively or primarily see “high-risk” patient populations. Risk adjustment was designed to help prevent these consequences in both individual and small group markets inside and outside the health care exchanges by spreading financial risk across the markets.²⁵ Separate risk adjustment systems operate in individual and small group markets (unless these markets are combined) and these separate systems all work toward the same goal of preventing skimming and risk selection in the insurance markets.

Lesson for Higher Ed: Like in health care, policymakers should try to mitigate “creaming,” which is the practice of having schools avoid enrolling “riskier” populations for fear of being penalized by an outcomes-based funding system. This could be accomplished by allowing schools to earn incentives for choosing to enroll and successfully matriculate traditionally underserved populations. Some policymakers have already suggested possible ways to do this through existing legislative language. Creating a “Title I” for higher ed, for example, would provide additional funding to institutions serving large numbers of low-income students.²⁶ Similar to Title I in the K-12 system, institutions that meet a minimum percentage of Pell Grant eligible students enrolled (for example) could get additional aid and the same could be done for institutions graduating high numbers of low-income students. Alternatively, institutions who choose not to enroll their fair share of lower-income students could be penalized through a reduction in funding, such as what’s laid out in the Senators Jeanne Shaheen (D-NH) and Todd Young’s (R-IN) Student Protection and Success Act.²⁷ The funds collected from this penalty could then be used to support institutions who serve their low-income students well, similar to how low-risk insurers subsidize higher-risk ones. Over the long term, this could encourage schools to prioritize students that have not been their focus in the past, increasing access for often underserved student populations to quality higher education.

Lesson 4: Distribute Funding Across Entities Equitably.

What Happened in Health Care: Risk adjustment is a system that recognizes that not all positive outcomes in health care are created equal. It is a zero-sum game, because the basic premise of risk adjustment is to balance out funding based on unequal populations. Insurers or hospitals that treat a large amount of chronically ill people receive additional funding, and those who have fewer than average chronically ill people pay into a fund that’s used to make the payments to the other insurers.²⁸ This is not to penalize insurers that are treating fewer chronically ill patients and are considered to have very good outcomes for their patient populations. These insurers will still receive funding for the few chronically ill patients they do treat, but most of the funds will go to those who are treating large numbers of chronically ill people based on the risk adjustment formula that accounts for these conditions. A major goal of risk adjustment is to incentivize value-based care and better outcomes among sick patients. So targeting funds to those who are chronically ill and may face worse outcomes and higher medical costs throughout their lifetimes makes sense to encourage insurers or hospitals to focus resources on these patients. Those in health care should strive for good outcomes for any patient being treated, but risk adjustment helps boost those efforts by getting additional funding into the hands of those who need it most and recognizing that keeping patients healthy is more time and resource intensive for some than others. 

Lesson for Higher Ed: Selective colleges like Harvard University or Princeton University have graduation rates around 97% for first-time, full-time students, but they also have large endowments and receive numerous donations that allow them to provide personalized support services to students to help get them through college. They’re also highly selective and choose only those students with the highest GPAs and test scores, who are already very likely to succeed in higher education.²⁹ But not all colleges and universities have these resources or populations. Small, liberal arts colleges, community colleges, and Historically Black Colleges and Universities (HBCUs) have much smaller endowments, if any, and are often hit hard with declining state appropriations during economic downturns.³⁰ Yet these are the colleges that take in and educate way more than their fair share of traditionally underserved students, who may need additional supports to complete higher education.

Some policymakers have attempted to mitigate these inequities through proposals that would direct more funding to schools with additional capacity needs. For example, the Access, Success, and Persistence in Reshaping Education Act (ASPIRE Act), co-sponsored by Senators Chris Coons (D-DE) and Jacky Rosen (D-NV), would attempt to correct for inequitable funding and distribute it more fairly to the schools and students who need it by directing dollars from institutions who aren’t admitting or serving “riskier” students well to institutions who are.³¹ It would put the emphasis on access and completion efforts at institutions serving a greater number of low-income students. The Biden administration has also demonstrated a commitment to closing equity gaps in the Build Back Better Act. Like the Title I for higher ed idea, the Build Back Better Act would provide $500 million to colleges and universities serving high numbers of low-income students to invest in completion and retention activities.³² These grants would go to underfunded and open access institutions who show a commitment to students who have been traditionally underserved. This investment could play a big role in increasing student retention and completion while distributing funds more equitably among institutions.

Let’s be clear, this is not to give a pass to colleges and universities who have terrible student outcomes. Any college with a 10% graduation rate, for example, is not fulfilling its mission to improve the lives of its students and create economic mobility. But, targeting funding to those who need the resources most to achieve positive outcomes is a concept we can draw from health care and apply to higher education. Boosting the resources that colleges and universities can use to graduate more students and prepare them for their careers will help improve outcomes at historically under-resourced colleges. We want to pair access to college with success in college and beyond, and getting money to the students and colleges that need the help the most will let us do that. 

Conclusion

Student outcomes in higher education matter. It’s time to consider new ways we can incentivize institutions to prioritize the success of their students and enroll those who need the mobility of higher education the most. Fortunately, the health care sector has grappled with similar questions over time and has created a system that more fairly and appropriately accounts for these inequities and incentivizes and rewards positive outcomes for all patients. Higher education can and should look to the health care sector to learn important lessons about how to accomplish these goals in a complex system.