Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Health Aff (Millwood). Author manuscript; available in PMC 2014 July 26.
Published in final edited form as:
PMCID: PMC4111896

Transitioning to Medicare before Age 65

Pamela Farley Short, Ph.D.* and France Weaver, Ph.D.**


Disabled workers who start receiving Social Security Disability Income (SSDI) must wait 24 months to qualify for Medicare. Legislation introduced in Congress would eliminate this waiting period, to guarantee that people with disabilities severe enough to qualify for SSDI will not be uninsured. We provide a longitudinal view of Medicare enrollment before age 65 by following a national sample from ages 55 to 64. One person in six was covered by Medicare before turning 65. A quarter of new enrollees were uninsured during the waiting period. There were great disparities in reliance on Medicare and coverage in the waiting period.

Medicare, which provides nearly universal coverage for Americans after age 65, also covers nearly 7 million American under 65.1 Most early Medicare enrollees receive Social Security Disability Insurance (SSDI) payments, after meeting eligibility tests based on the severity and likely duration of disabilities that prevent them from working. Those who qualify must wait 24 months from the start of SSDI payments before enrolling in Medicare. Because there is a 5-month waiting period from the onset of disability to the start of SSDI disability payments, people who can no longer work have to wait a total of 29 months for Medicare. At any given time, an estimated 1.3 to 1.5 million newly eligible SSDI recipients are waiting to qualify for Medicare.2,3

Because most health insurance in the United States is obtained at work, and SSDI recipients are unable to work, the Medicare waiting period is a significant hole in the safety net protecting American workers against the risk of disability. To help fill this hole for people insured by employers when they become disabled, the Consolidated Omnibus Budget Reform Act of 1986 (COBRA) allows disabled employees to extend enrollment in an employer’s plan for an additional 11 months beyond the standard 18 months available to all former employees. Thus, COBRA provides an extension of 29 months that matches the SSDI/Medicare waiting period. However, employers can require former employees to pay the entire premium plus a 2% administrative fee during the first 18 months and then up to 150% of the premium during the additional 11 months allowed to disabled workers.4 As the average total premium for a single worker was about $4,200 in 2006,5 COBRA premiums can be expensive in relation to SSDI income payments, which averaged $978 per month ($11,700 per year) in December 2006.6

Persons with disabilities may also qualify for Medicaid during the Medicare waiting period, if their incomes and assets are low enough and they meet state tests for disability that are sometimes more stringent than SSDI’s.7 When Riley linked household-reported health insurance from the National Health Interview Survey to Social Security and Medicare administrative records for 1994 to 1996, he found that 21% of SSDI recipients were covered by Medicaid during the waiting period.8 From administrative data for selected states, Dale and Verdier estimated that Medicaid covered approximately 40% of people in the Medicare waiting period in early 2002.9 Furthermore, Dale and Verdier estimated that up to one-third of those waiting for Medicare lacked any coverage at all, while Riley estimated that 25.8% were uninsured.

Stories describing how uninsured people with severe disabilities struggle to get through the waiting period without health insurance are heartbreaking.10 These stories also suggest that the inadequacy of their care damages their health and adds to Medicare costs later. In recognition of these concerns, legislation was introduced in the last two Congresses to eliminate the Medicare waiting period. The current version, “Ending the Medicare Disability Waiting Period Act of 2007,” would phase out the waiting period over 10 years and eliminate it immediately for individuals with life-threatening conditions. Companion bills were introduced by Rep. Gene Green in January 2007 (H.R.154) and Senator Jeff Bingaman in September 2007 (S.2102) and are in committee.11

In this article, we provide new data related to enrollment in Medicare before age 65 and coverage during the Medicare waiting period. In comparison to previous descriptions of these coverage patterns, which have offered a cross-sectional view for a calendar year, we provide more of a lifetime perspective on early transitions to Medicare. Using longitudinal data from the Health and Retirement Study, we follow the health insurance status of a nationally representative cohort of Americans (born from 1936 to 1941) from age 55 to age 64. This birth cohort turned age 55 during calendar years 1991 to 1996 and reached age 65 from 2001 to 2006. In addition to estimating how many people were covered by Medicare before they reached age 65, we examine coverage during and prior to the waiting period, along with demographic and socioeconomic factors associated with coverage differences at this stage of life.

Study Data and Methods

The Health and Retirement Study (HRS) is an ongoing longitudinal survey conducted by the University of Michigan since 1992.12 Participants enter the survey at different ages, from 51 to 61, and are interviewed every two years until death. HRS asks detailed questions about the health insurance of survey participants at the time of each interview. Different birth cohorts are part of the survey, but here we focus on individuals born from 1936 through 1941. Furthermore, we concentrate on the ten years before each person turned 65, the usual age for enrolling in Medicare. For the cohort we selected, using data from the first six waves of the HRS from 1992 through 2004, we can observe each person’s health insurance status every two years from age 55 or 56 to age 63 or 64.

Because of missing data and the difficulty of linking information on SSDI and Medicare, we could not determine coverage during the Medicare waiting period from the HRS questions on SSDI.13 Instead, given the two-year interval between HRS interviews and the two-year Medicare waiting period, we consider that a person reporting Medicare for the first time in Wave t was in the Medicare waiting period at Wave t-1. We also observe coverage prior to the waiting period at Wave t-2. This approach assumes that all new enrollees are subject to the 24-month waiting period, ignoring the special rules applying to end stage renal disease (3-month wait) and amyotrophic lateral sclerosis (no wait). Because HRS does not enroll anyone younger than age 51 (except to collect spousal data), we begin our analyses at age 55 to be able to “look back” to Wave t-2 to determine coverage prior to the waiting period. Even so, because subjects enter HRS at different ages, we cannot observe t-1 for some subjects who entered Medicare after age 55; for others, we observe t-1, but not t-2. We drop subjects lacking data for waves prior to Medicare enrollment from analyses of coverage during or prior to the waiting period.

To get estimates of the U.S. population in the birth-year cohort selected for analysis, we use the 1992 HRS survey weights. To adjust for attrition not attributable to deaths, we post-stratify the weights of respondents with complete health insurance data from age 55/56 to age 63/64 to the initial weighted distribution of the cohort at age 55/56 (excluding decedents) by gender, marital status, race and ethnicity, and health insurance in 1992. Finally, we take into account the complex survey design of HRS in estimating confidence intervals.

At age 55/56, our analytic sample includes 4,044 individuals, corresponding to an estimated population cohort of 13.9 million Americans born from 1936 through 1941 who were alive at age 55/56. By age 63/64, the estimated size of the cohort is 12.7 million, with cumulative deaths reaching 8.5% of the initial sample.

Study Results

Health insurance status by age

Early enrollment in Medicare plays an important role in offsetting declines in employer-sponsored insurance as people approach age 65 (Exhibit 1). Among Americans born from 1936 to 1941, Medicare enrollment more than doubled from 667,000 at age 55/56 to 1.5 million by age 63/64. In percentage terms, Medicare coverage grew from 4.8% of the cohort at age 55/56 to 12% at age 63/64. Medicaid coverage also increased, but to a lesser extent. In contrast, the proportion covered by employers (including Tricare for military dependents and retirees) decreased from 67.6% to 60.4%. The percentage uninsured declined, from 12.6% to 9.5%, as the reduction in employer insurance was more than offset by increases in other types of insurance (mainly Medicare).

Exhibit 1
Type of health insurance by age (U.S. population born from 1936 to 1941)

Medicare at any time from ages 55 to 64

All together, 15.2% of the cohort (representing 2.1 million persons) was covered by Medicare at some time from age 55 to 64 (Exhibit 2). In other words, 1 person out of 6 alive at age 55 was on Medicare before age 65. Of the total, 13% entered Medicare between ages 55 and 64, and 2.2% were already enrolled at age 55 (data not shown).

Exhibit 2
Types of health insurance ever held from ages 55 to 64 (U.S. population born from 1936 to 1941)

There was considerable variation in the likelihood of early Medicare enrollment by race, marital status, and socioeconomic characteristics (Exhibit 3, first column). One in three African Americans was covered by Medicare before turning 65, compared to 12.9% of Whites. Persons who were married at age 55 were much less likely to be on Medicare before 65 than persons who were not married, with the highest rate for widowed persons (26.2%). Finally, we observe a strong gradient in early Medicare enrollment by socioeconomic status, as measured by education and poverty level at age 55. For example, 38 percent of persons below the poverty line at age 55 were on Medicare before age 65, compared to 8.4% with incomes above 400% of the poverty line.

Exhibit 3
Percent ever on Medicare from ages 55 to 64 and type of health insurance during the Medicare waiting period, by socio-demographic characteristics (U.S. population born from 1936 to 1941)

Coverage during Medicare waiting period

We estimate that 24% of new Medicare enrollees were uninsured during the waiting period (Exhibit 3). About half were covered by their own (presumably former) employers (33.6%) or their spouse’s employer (16.7%) during the waiting period. Seventeen percent were covered by Medicaid.

Here, too, there were significant differences by race, marital status, education, and income. Similar proportions of Whites and African Americans were uninsured during the waiting period, but the two racial groups drew their coverage from different sources. A third of African Americans were on Medicaid, compared to 10.5% of Whites. The rate of spousal coverage for Whites was three times the rate for African Americans. Married Medicare enrollees were far more likely to have employer coverage in the waiting period than divorced, separated or never married individuals, who were the most often uninsured. More than a third of individuals who were poor at age 55/56 were uninsured, nearly three times the rate for the top income category. The higher rate of Medicaid coverage for the poor (39.8%) did not offset much higher rates of employer insurance at higher incomes. More than half of new Medicare enrollees in the top income category were covered by their own employment-related insurance in the waiting period. The differences by education were similar to, but less extreme than the differences by income.

Loss of employer insurance

Sixty-two percent of new Medicare enrollees were covered by employers prior to the waiting period (data not shown). Of the group covered by employers prior to the waiting period, more than three-quarters were also covered by employers during the waiting period (Exhibit 4). More than half retained coverage through their own employers and 24% were covered by their spouses’ employers. Approximately 12% of persons with employer coverage prior to the waiting period became uninsured during the waiting period, while 4% went on Medicaid.

Exhibit 4
Percent distribution of insurance during the Medicare waiting period for persons covered by employers before the waiting period.

Coverage prior to the waiting period for those uninsured during the waiting period

Here, we try to better understand the circumstances that led people to be uninsured during the waiting period (Exhibit 5). About half of those uninsured during the waiting period were also uninsured prior to the waiting period. The loss of employer insurance accounted for 36% of those uninsured during the waiting period, including 27% previously covered by their own employers and 9% by a spouse’s employer.

Exhibit 5
Percent distribution of health insurance before the Medicare waiting period for persons uninsured during the waiting period.


Our estimates of coverage during the Medicare waiting period, for a cohort in the Health and Retirement Study that turned 55 in the early to mid 1990’s, are similar to Riley’s 1994–1996 estimates from the National Health Interview Survey. Given the limitations of our estimates —which rely on self-reports of health insurance status, miss SSDI recipients who died before qualifying for Medicare, and identify the waiting period from survey questions about Medicare enrollment instead of SSDI receipt—this consistency is reassuring. Like Riley, we find that a quarter of those enrolling in Medicare before age 65 were uninsured in the waiting period. Our estimate of Medicaid enrollment in the waiting period (17%) is also similar to Riley’s (21%). These survey-based estimates suggest less Medicaid coverage, fewer uninsured, and implicitly more private health insurance during the waiting period than Dale and Verdier estimated from administrative data.

The discrepancy in Medicaid estimates is especially concerning. These estimates enter into the estimation of new government costs to eliminate the waiting period, because Medicaid coverage during the waiting period is already financed from Federal and state funds. Our Medicaid estimates from HRS could be influenced by studying a single cohort over time, instead of a cross-section, and the focus on Medicare enrollment after age 55.

Consequently, as an additional check, we combined four years of data from the Medical Expenditure Panel Survey (2001–2004) to identify 115 people who had not reached age 65 by the end of the MEPS survey year, but transitioned to Medicare during the year. When we examined their coverage in the month before they started Medicare, the weighted percentages were 24% Medicaid, 6% other public (including Tricare), 24% uninsured, and 46% privately insured.

Although the MEPS estimates corroborate Medicaid estimates from the other two surveys, all three surveys exclude people in institutions, who are especially likely to qualify for Medicaid.14 To the extent that people with severe physical or mental disabilities are less likely to participate in surveys, surveys may further underestimate Medicaid during the waiting period. Finally, we note an anomaly in the reported timing of Medicare enrollment in MEPS that may undermine the validity of that survey in studying the waiting period: although we limited attention to calendar years before subjects turned 65, to avoid problems with the exact timing of the enrollment of 65 year olds, a large proportion (about a third) of new enrollment before age 65 is reported at age 64.

When Medicare was extended to persons with disabilities in 1972, the waiting period was included as a way of reducing costs, preserving existing private insurance, and keeping people with less severe or more temporary disabilities from seeking health insurance through the Social Security system.15 The waiting period still serves those functions,16 but advocates for its elimination focus particularly on people who are uninsured during those two years and argue that no one with health problems severe enough to qualify for SSDI should be left without health insurance.17

Eliminating the waiting period does indeed seem to be the best way to help new SSDI recipients who are currently uninsured. They are a relatively disadvantaged group in economic terms, with little access to other sources of health insurance. According to our estimates of coverage prior to the waiting period, only a third have employer insurance that they could extend under COBRA (even with subsidies to help pay the premiums). Putting them on Medicaid would not only impose additional costs on states, but would waste administrative expense and effort enrolling people in Medicaid for two years before they move to Medicare.

Given the arguments for eliminating the waiting period for some new SSDI recipients, the simplicity of eliminating it entirely--for all new SSDI recipients--is appealing. Enrolling everyone immediately in Medicare would provide more equal access to the health services needed by persons with severe disabilities, reducing current disparities in coverage during the waiting period. It would also shift much of the financial burden to the Federal government from new SSI recipients and their families, employers and active workers, and states. However, the relative numbers of SSDI recipients uninsured in the waiting period (24%), enrolled in Medicaid (17%), and privately insured (59%) imply that the Federal costs would be considerably more than the cost of insuring new SSDI recipients who are currently uninsured (although there would be some offsetting Federal savings involving Medicaid).

In particular, as we estimate that 62% and 48% of Medicare enrollees were covered by employers before and during the waiting period, respectively, preserving a role for employers in insuring new SSDI recipients might yield enough Federal savings to be worth the added complication (and likely political resistance from employers). From the government’s perspective, even with generous subsidies to new SSDI recipients to defray the cost of employer premiums, premium subsidies would likely cost less than enrolling persons with severe disabilities in Medicare. Because premiums are based on claims averaged over all enrollees, and the average claims of persons with disabilities are likely to be well above the employer average, the government could save money by paying premiums to employers instead of claims under Medicare. In this way, employers and active workers would continue to share the cost of health care for disabled workers for two years. Additionally, because of the coverage gaps in Medicare, some SSDI recipients would get more comprehensive coverage through employers than through Medicare. For example, those with prescription benefits through an employer would avoid the “doughnut hole” in Medicare.

Thus, the Federal government could guarantee health insurance for all new SSDI recipients by enrolling some immediately in Medicare, but heavily subsidizing private health insurance for others. By eliminating the waiting period only for people without access to employer insurance, the government could strongly encourage enrollment in employer plans for two years. Eliminating the Medicare waiting period for everyone, with premium subsidies in place, would allow many persons with disabilities to choose between private insurance and Medicare. Unfortunately, while the Federal government might save money by encouraging more employer coverage during the waiting period, shifting costs to employers in this way might also accelerate the ongoing erosion of employer-sponsored health insurance.

In closing, we note that American workers have a bigger stake in policy decisions involving the Medicare waiting period than annual estimates of Medicare enrollment might suggest. In any given year, Medicare covers only 2.5% of the non-institutionalized population under age 65.18 However, we found that 1 out of 6 people alive at age 55 were on Medicare before age 65 (and, therefore, had to negotiate the waiting period). If they were aware of these odds, workers might be more interested in changing the Social Security disability program to guarantee immediate access to affordable health care.


This project was funded by the Penn State Center on Population Health and Aging under NIA grant P30 AG024395 and Penn State’s College of Health and Human Development. The authors are grateful to Cindy Mitchell from Penn State’s Population Research Institute for programming assistance.


1. Centers for Medicare and Medicaid Services. [accessed on 29 May 2007];Medicare Enrollment: National Trends 1966–2005. 2007
2. Riley G. Health Insurance and Access to Care among Social Security Disability Insurance Beneficiaries during the Medicare Waiting Period. Inquiry. 2006;43:222–230. [PubMed]
3. Dale SB, Verdier JM. Elimination of Medicare’s Waiting Period for Seriously Disabled Adults: Impact on Coverage and Costs. New York: The Commonwealth Fund; 2003. [PubMed]
4. Flynn P. COBRA Qualifying Events and Elections, 1987–1991. Inquiry. 1994;31:215–220. [PubMed]
5. Kaiser Family Foundation and Health Research and Educational Trust. [accessed on 30 May 2007];Employer Health Benefits 2006.
6. Social Security Administration. [accessed on 1 June 1 2007];OASDHI Monthly Statistics, April 2007.
7. Dale and Verdier. Elimination of Medicare’s Waiting Period. [PubMed]
8. Riley. Health Insurance and Access to Care among SSDI Beneficiaries. [PubMed]
9. Dale and Verdier. Elimination of Medicare’s Waiting Period. [PubMed]
10. Hayes RM, Beebe D, Kreamer H. Too Sick to Work, Too Soon for Medicare: The Human Cost of the Two-Year Medicare Waiting Period for Americans with Disabilities. New York: The Commonwealth Fund; 2007. [accessed on 30 May 2007]. [PubMed]
11. Library of Congress (Thomas) [accessed on 13 November 2007];Ending the Medicare Disability Waiting Period Act of 2007.
12. [accessed on 22 June 2007];The Health and Retirement Study.
13. Before 2000, the HRS questionnaire did not reliably distinguish between receipt of disability income from SSDI and SSI.
14. HRS follows subjects who enter institutions, but the initial sample is limited to the non-institutionalized population.
15. See the text of the committee reports quoted in Dale and Verdier.
16. Based on empirical analyses in an unpublished working paper written in 2002, Jonathan Gruber and Jeffrey Kubik concluded that eliminating the waiting period would encourage a significant increase in SSDI applications. Gruber J, Kubik J. NBER Working Paper No. 9148. Health Insurance Coverage and the Disability Insurance Application Decision.
17. Bingaman Senator Jeff. [accessed on 12 June 2007];Statement introducing “Ending the Medicare Disability Waiting Period Act of 2005” on June 9. 2005
18. Fronstin P. Sources of Health Insurance and Characteristics of the Uninsured: Analysis of the March 2006 Current Population Survey. Washington, DC: Employee Benefits Research Institute; 2006. Issue Brief 298. [PubMed]