By Michael Geruso, Timothy J. Layton, Jacob Wallace
National Bureau of Economic Research, August 2020
Abstract
Exploiting random assignment of Medicaid beneficiaries to managed care plans, we identify plan-specific effects on healthcare utilization. Auto-assignment to the lowest-spending plan generates 30% lower spending than if the same enrollee were assigned to the highest-spending plan, despite identical cost-sharing. Effects via quantities, rather than differences in negotiated prices, explain these patterns. Rather than reducing “wasteful” spending, low-spending plans cause broad reductions in the use of medical services—including low-cost, high-value care—and worsen beneficiary satisfaction and health. Supply side tools circumvent the classic trade-off between financial risk protection and moral hazard, but give rise instead to a cost/quality trade-off.
From the Introduction
Regulated competition between private health plans is becoming the dominant form of social health insurance in the United States. In 2017, 54 million Medicaid beneficiaries (69%) and 19 million Medicare beneficiaries (33%) were enrolled in a private managed care plan. In the same year, almost $500 billion of the $1.3 trillion spent on public health insurance programs went to private managed care plans.
In this paper, we identify the causal effects of the health plan in which a beneficiary enrolls on her healthcare utilization, the quality of care received, and proxies for satisfaction and health. The context of our analysis is Medicaid Managed Care (MMC), the privatized system through which most Medicaid beneficiaries receive benefits today.
In our setting, all plans are required to provide care at zero marginal cost to beneficiaries. It is therefore an ideal context for studying whether various non-cost-sharing plan features (e.g., networks, negotiated provider rates, patient follow-up and medication adherence programs, etc.) can constrain healthcare spending. In contrast, nearly all of the prior econometric literature studying how health plans affect utilization and health outcomes has focused on consumer cost-sharing provisions like copays, coinsurance, and deductibles. But a modern health plan is more than a set of consumer-facing prices, and our analysis sheds new light on the range of impacts generated by supply-side (non-cost-sharing) plan features. To facilitate a transparent comparison between our results and results from cost-sharing studies including the RAND Health Insurance Experiment (Manning et al., 1987) and more recent quasi-experimental work (Brot-Goldberg et al., 2017), we focus our analysis on the types of outcomes that have been the focus of this prior literature. These include overall service utilization and spending, utilization of high- and low-value care, conventional measures of healthcare quality, and surrogate health outcomes like avoidable hospitalizations.
As our first main result, we document statistically and economically significant causal variation in spending across plans. If an individual enrolls in the lowest-spending plan in the market she will generate about 30% less in healthcare spending than if the same individual enrolled in the highest-spending plan in the market. We show that risk-adjusted observational measures and causal estimates of plan spending effects are correlated, but find that the risk-adjusted measures tend to overstate causal differences in spending across plans. Plans that attract healthier patients thus do more to constrain spending—i.e., provide less care—consistent with a classic adverse selection model, where sicker individuals select plans providing more care. This fact has important implications for the use of observational measures of spending and quality as a basis for regulatory rewards or penalties.
After establishing important differences between risk-adjusted (OLS) plan spending effects and causal (IV) estimates, we investigate which factors drive the bottom-line causal differences. First, we find that almost all services are marginal. That is, lower spending plans tend to provide less of nearly everything. This includes inpatient and outpatient visits, primary care physician office visits, and high-value/cost effective drugs. Second, unlike in other markets, differences in provider prices do not explain the differences in healthcare spending across plans in our setting. In a decomposition, prices account for very little of the cross-plan spending differences.
Instead, spending differs because enrollees in low-spending plans use less care, with much of the utilization gap driven by the extensive margin. Importantly (and similar to the effects of deductibles in Brot-Goldberg et al., 2017), utilization reductions do not seem to focus on “low-value” care or “waste”: We estimate that low-spending plans reduce utilization of high-value drugs used to treat diabetes, asthma, and severe mental illnesses, as well as high-value screenings for diabetes, cancer, and sexually transmitted infections.
Finally, we show that the low-spending plans also increase avoidable hospitalizations and decrease consumer satisfaction, as measured by the propensity of auto-assigned enrollees to switch out of their plan post-assignment. These results suggest a clear trade-off between spending and beneficiary satisfaction and health.
We show that there is substantial causal heterogeneity across plans in spending and utilization that arises without any differences in consumer cost-sharing exposure.
Our findings complement a large literature extending back to the RAND health insurance experiment (Manning et al., 1987) that documents how consumer prices impact healthcare utilization. In RAND, and the studies that have followed, patient cost-sharing has proven to be a blunt instrument, affecting the use of low- and high-value services alike (Brot-Goldberg et al., 2017). These findings sparked interest in whether managed care tools offer a scalpel that can target inefficient spending and better manage the high-cost patients responsible for the majority of spending. But our results, along with prior work studying managed care in Medicare (Curto et al., 2017), indicate that supply-side tools exhibit many of the same features and limitations as demand-side tools. Their impacts on healthcare spending are blunt. They indiscriminately reduce utilization, limiting both high- and low-value care rather than targeting “waste.” In another similarity to the effects of consumer cost sharing (as found in Brot-Goldberg et al., 2017), lower-spending managed care plans in our setting do not appear to generate savings by steering patients to lower-cost providers or lowering negotiated prices.
Lastly, our work highlights how supply side tools can achieve spending reductions while circumventing the classic trade-off between financial risk protection and moral hazard noted by Zeckhauser (1970) and Pauly (1974). The spread of plan effects we estimate are similar to the utilization difference between the 0% and 95% coinsurance rate treatment arms in the RAND HIE. Thus, significantly constraining healthcare spending need not require exposing consumers to out of pocket spending. But there is no “free lunch” here, as we also document that these spending reductions come at the cost of beneficiary satisfaction and, ultimately, health.
Conclusion
Our results are important for understanding the potential for managed care to constrain healthcare spending growth. We show that the baskets of rationing devices implicit in managed care can have spending and utilization impacts significantly larger than what could be accomplished by exposing consumers to high deductibles and reasonable coinsurance and copays.
Importantly, rationing via managed care reduces spending without exposing consumers to financial risk, circumventing the classic trade-off between financial risk protection and moral hazard noted by Zeckhauser (1970) and Pauly (1974). These findings are particularly relevant for public insurance programs—including the low-income segments of HIX Marketplaces and Medicare—where policymakers have been reluctant to expose low-income consumers to financial risk.
However, these spending reductions appear to come with a utility cost. Willingness to remain enrolled in a plan is negatively related to that plan’s cost savings. And cost reductions are blunt—reducing utilization of all types of care, lowering traditional measures of healthcare quality, and increasing the likelihood of adverse health events.
Comment:
By Don McCanne, M.D.
Medicaid managed care plans have become very popular in states for the obvious reason that they reduce spending. This paper is important because it reveals how the cost savings are achieved.
Most Medicaid programs do not use cost sharing and thus they do not reduce spending that way, but, even if they did, these authors state that “managed care can have spending and utilization impacts significantly larger than what could be accomplished by exposing consumers to high deductibles and reasonable coinsurance and copays.” They also find that differences in provider prices do not explain the differences in health care spending across plans in this setting.
So how does managed care control spending? The plans consider all services to be marginal and thus they reduce utilization of all services regardless of the value of those services. “Lower spending plans tend to provide less of nearly everything. This includes inpatient and outpatient visits, primary care physician office visits, and high-value/cost effective drugs.”
This gives rise to cost/quality trade-offs. Reduction in beneficial health care services obviously reduces the quality of care. Low-spending plans increase avoidable hospitalizations. This cavalier attitude does not go without notice; beneficiary satisfaction is diminished.
How many more studies do we need? The private insurance industry has to go. And, yes, we have a replacement: single payer improved Medicare for All.
Stay informed! Visit www.pnhp.org/qotd to sign up for daily email updates.
