McKinsey & Company LogoMcKinsey on Healthcare
Sign in
Stock hero image. Not critical for content.
Payer insights

Transitions in coverage type are the norm for most consumers over time

Filed under: Consumer engagement, Individual insurance, Medicaid, Medicare

Transitions in health insurance coverage appear to be the norm for most Americans over time. Our analysis of point-in-time coverage data reveals striking age- and income-related patterns that imply change over time in both the types of coverage people are most likely to have and their probability of being uninsured. As the quality and availability of data needed to track individuals across coverage programs improve, a more complete picture of these lifetime dynamics is emerging. Analysis of the data by age and income, as well as recent studies on lifetime income dynamics, suggest that coverage transitions over the course of an individual’s lifetime are common. As a result, payors may want to view short-term fluctuations in coverage as well as long-term transitions of their members as interrelated phenomena, with implications for how their think about their business.

Quantifying transitions

We undertook our exploration initially because of questions about the rate of coverage transitions between Medicaid and the individual market. Given Medicaid expansion and the introduction of income-related subsidies in the individual market, many experts1 had predicted that income volatility—associated with common life events such as marriages, divorces, family formation, income growth, and job loss—would cause a high percentage of low-income adults to move between Medicaid and individual market coverage each year.2

We first reviewed external literature. For example, researchers from UC Berkeley, using a simulation model,3 studied expected coverage transitions in light of policy changes under the Affordable Care Act. They estimated that, of the non-elderly enrolled in Medi-Cal (California Medicaid) at the beginning of a year, 25% would no longer be enrolled at the end of that year.4 Our analysis of CMS enrollment reports indicates that that the nationwide annual Medicaid disenrollment rate may be even higher—as much as 37% per year.5 Given that many eligibility categories (e.g., aged, blind, or disabled) are not based solely on income, the populations susceptible to income-driven coverage transitions (e.g., parents and single adults) would experience even higher exit rates than this average rate would imply. A recent study of income dynamics using data from the Social Security Administration confirmed that those with low incomes experience not only overall income growth trends but also greater income volatility relative to higher income cohorts.  Both persistent and short-term increases in income would contribute to a loss of Medicaid coverage for individuals, should their incomes rise above eligibility thresholds.6

At present, little is known about what happens to individuals who exit the Medicaid program. The UC Berkeley researchers estimated that more than one-third of these individuals in their state would switch to employer-sponsored insurance (ESI); the remainder would become eligible for a qualified health plan (QHP). However, the researchers did not determine the percentage of individuals who actually enrolled in a QHP once eligible. An analysis of Washington State data suggests that about 1% of all Medicaid enrollees in that state transitioned to a QHP between October 2014 and September 2015.7 This point estimate of movement between Medicaid and QHPs is smaller than one may have expected, but it also does not capture transitions out of Medicaid to other coverage types—whether ESI, a temporary gap in coverage, or a long-term lack of insurance. A broader picture of the complete set of coverage transitions Medicaid individuals undergo is needed to fully understand the long-term movements of these individuals.

To further understand this broader set of coverage transitions, we used the McKinsey Predictive Agent-based Coverage Tool (MPACT) to examine the overall 2014 U.S. health insurance market. Results reveal that coverage patterns vary within each age cohort and income level (Exhibit 1).8 Nevertheless, certain trends based on age and income begin to emerge.

Exhibit 1

The patterns in coverage by age cohort are highlighted in Exhibit 2. Before age 18, Medicaid and ESI are the dominant coverage types. Between the ages of 18 and 34, the rate of ESI coverage increases, but the proportion of people without health insurance spikes. Between 35 and 65 (the peak earning years for most people9), the uninsured rate decreases and ESI predominates. After the age of 65, most people have Medicare coverage, but only a minority rely on that program alone for coverage.

This point-in-time analysis of insurance coverage is consistent with the longitudinal income dynamics study referenced above, which confirms that short-term income volatility occurs within the context of long-term upward trends in income. Although volatility declines as individuals’ ages and incomes increase, large negative income shocks remain a real possibility.10 As more data become available to characterize the longitudinal coverage movements of individuals, we may likewise be able to link the point-in-time insurance snapshots outlined above more directly to coverage transition patterns.


Exhibit 2

In short, the patterns underscore that coverage transitions are common: both short-term fluctuations and long-term trends. For example, many middle-aged, commercially insured individuals may have been on Medicaid at one time, and most of them will be on Medicare in the future. Similarly, an individual may move between Medicaid coverage, a QHP, and uninsured status over the course of a couple of years. Importantly, these are not isolated phenomena. In other words, individuals may experience short-term coverage transitions in the context of a longer-term path through different types of coverage.

Implications for payors

Our findings suggest that payors might benefit from better understanding how consumers transition among coverage types over time (in both the short and long term), since these transitions have important business implications. In general, coverage transitions represent opportunities for payors to either retain, gain, or lose members as consumers switch from one form of coverage to another (e.g., they become ineligible for Medicaid or lose ESI but become eligible for individual insurance or Medicare Advantage). Engagement strategies could have the potential to increase retention across coverage types in these scenarios. For instance, the newly issued Medicaid managed care regulations from the Centers for Medicare & Medicaid Services allow exchange carriers to reach out to Medicaid beneficiaries, potentially building relationships before a coverage transition becomes necessary.

At present, most ESI enrollees have only a limited choice of insurance carriers. However, if an increasing number of those with job-based insurance obtain coverage through the public or private exchanges (as some predict will happen, to varying degrees), their carrier choice will likely expand. Should this occur, carriers will increasingly have the opportunity to serve members in multiple phases of their lives. 

More immediately, understanding coverage transitions can help payors identify commonalities that transcend coverage type—commonalities that could help them realize efficiencies within their organizations. Our research shows, for example, that the behaviors and preferences of consumers covered by Medicaid, individual insurance, and ESI are notably similar in certain ways. In our surveys, 19% of Medicaid enrollees and 22% of those with ESI coverage said they exchanged text messages with a healthcare provider.12 Use of a mobile app to enhance prescription drug compliance was reported by 16% and 14%, respectively. These similarities (often defined around a consumer’s health, income, education, or other demographics, independent of coverage type) suggest that the patient engagement apps and other mobile tools payors develop may appeal to consumers in multiple lines of business. Other strategies to engage consumers could also appeal to consumers across coverage types.

Another example: Medicaid beneficiaries and consumers with individual coverage have similar preferences when their primary care physician (PCP) no longer accepts their insurance. When asked about this scenario in our surveys, 56% of Medicaid beneficiaries and 57% of consumers with individual coverage said they would choose a new PCP; only 10% of both groups said they would go out of network to see their original PCP. This finding suggests that lines of business might be able to take similar approaches to network and plan design—though further research is needed to better understand the underlying drivers of individuals’ provider preferences when covered under different insurance types.

Of course, coverage type will remain a critical characteristic for payors that want to attract, retain, and manage members, and thus many elements of the enrollment and renewal processes will remain within separate lines of business. And, certainly, differences in the health risk profiles of different coverage types cannot be ignored. For example, our research indicates that adult Medicaid enrollees have a higher rate of chronic illness than do those with individual or ESI coverage and may be helped by care management programs tailored to their needs.

Nevertheless, by understanding coverage transitions, payors may be better able to serve the needs of their members efficiently and effectively. They could also help reduce the number of people who fall into coverage gaps and become uninsured. More broadly, there are opportunities to improve data availability and further analyze coverage transitions to better understand their causes and impacts.

An abbreviated version of this blog post appears on the NEJM Catalyst website.

  1. Examples include:
    (1) Sommers and Rosenbaum. Issues in health reform: How changes in eligibility may move millions back and forth between Medicaid and Individual Exchanges. Health Affairs. 2011.
    (2) Buettgens et al. Churning under the ACA and state policy options for mitigation. Urban Institute: Timely analysis of immediate health policy issues. 2012.
    (3) Sommers et al. Medicaid and Marketplace eligibility changes will occur often in all states; Policy options can ease impact. Health Affairs. 2014. 
    (4) For potential transitions under a Medicaid/Basic Health Program arrangement: Curtis and Neuschler. Income volatility creates uncertainty about the state fiscal impact of a Basic Health Program (BHP) in California. Institute for Health Policy Solutions. 2011.
  2. It is important to note that issues related to coverage transitions and continuity are most relevant where coverage rates among younger and lower-income adults are high. Some populations (e.g., low-income individuals in states that opted not to expand Medicaid) may be more likely to lose coverage than to transition between coverage types when common life events occur.
  3. A combination of the California Simulation of Insurance Markets (CalSIM) model version 1.7 and data from the Survey of Income and Program Participation were used to estimate steady-state coverage transitions (e.g., in 2019).
  4. Dietz et al. The ongoing importance of enrollment: Churn in Covered California and Medi-Cal. UC Berkeley Labor Center.  April 1, 2014.
  5. This 37% annual disenrollment is likely an upper bound as disruptions during the redetermination process may force individuals to disenroll for a short period of time.
  6. Guvenen et al. What do data on millions of U.S. workers reveal about life-cycle earnings risk? Federal Reserve Bank of New York Staff Reports. February 2015.
  7. Washington Health Benefit Exchange. September enrollment data. Enrollment Reports & Data. Washington Healthplanfinder, 2015.
  8. The McKinsey Predictive Agent-based Coverage Tool (MPACT) is a behavioral micro-simulation model that projects the post-reform health insurance coverage landscape under different scenarios from 2013-2022. The model is built upon a granular demographic database built by bringing together multiple public data sources such as the Census, American Community Survey, Small Area Health Insurance Estimates, and many others.
  9. U.S. Census Bureau, Current Population Survey.
  10. Guvenen, Fatih et al. What do data on millions of U.S. workers reveal about life-cycle earnings risk? Federal Reserve Bank of New York Staff Reports. February 2015.
  11. Congressional Budget Office. The effects of the Affordable Care Act on employment-based health insurance, March 2012. Congressional Budget Office. Updated estimates of the insurance coverage provisions of the Affordable Care Act, April 2014.
  12. McKinsey Consumer Health Insights survey, 2015; McKinsey Consumer Health Insights Medicaid survey, 2015.

Also of Interest