Leveraging The Affordable Care Act To Enroll Justice-Involved Populations In Medicaid: State And Local Efforts

Abstract

The Affordable Care Act provides an unprecedented opportunity to enroll criminal justice–involved populations in health insurance, particularly Medicaid. As a result, many state and county corrections departments have launched programs that incorporate Medicaid enrollment in discharge planning. Our study characterizes the national landscape of programs enrolling criminal justice–involved populations in Medicaid as of January 2015. We provide an overview of sixty-four programs operating in jails, prisons, or community probation and parole systems that enroll individuals during detention, incarceration, and the release process. We describe the variation among the programs in terms of settings, personnel, timing of eligibility screening, and target populations. Seventy-seven percent of the programs are located in jails, and 56 percent use personnel from public health or social service agencies. We describe four practices that have facilitated the Medicaid enrollment process: suspending instead of terminating Medicaid benefits upon incarceration, presuming that an individual is eligible for Medicaid before the process is completed, allowing enrollment during incarceration, and accepting alternative forms of identification for enrollment. The criminal justice system is a complex one that requires a variety of approaches to enroll individuals in Medicaid. Future research should examine how these approaches influence health and criminal justice outcomes.

The United States has one of the highest incarceration rates in the world, with 6.9 million people in jail or prison or on probation or parole in 2013.^1,2 National surveys have shown that incarcerated people are at higher risk than the general population for having chronic and infectious diseases such as hypertension, asthma, HIV, and hepatitis.^3,4 The most recent national data available revealed that 24 percent of state prisoners reported substance dependence, 15 percent reported a mental health condition, and 42 percent reported both mental health and substance use disorders.⁵ Social and economic barriers such as low levels of literacy and high rates of poverty, homelessness, and unemployment also contribute to poor health in criminal justice–involved populations.⁶

Correctional facilities include prisons, which are operated by the state or federal government and typically incarcerate people who have a sentence of over a year; jails, which are typically operated at the county level and incarcerate people who have not yet gone to trial and those with sentences of a year or less; and community corrections settings such as probation and parole systems.⁷ In 2012 approximately 637,000 prisoners were released from federal and state prisons.⁸ Local jails experienced higher turnover rates, with a daily population of 731,000 but a yearly population of 11.7 million in 2013.⁹

Previous evidence suggests that the period directly following release from prison or jail is a time when former inmates are particularly likely to experience negative health outcomes. Ingrid Binswanger and coauthors found that in the two weeks following release, the mortality rate among released prisoners was 12.7 times the rate among the general population. The leading cause of death during this period was drug overdose.¹⁰

Discharge planning varies widely by jurisdiction, but incarcerated individuals are typically released with no more than a two-week supply of medication, which makes it difficult for them to manage their health problems.¹¹ Another likely contributing factor to diminished health status among justice-involved populations is lack of health insurance. An estimated 80 percent of incarcerated people lacked health insurance coverage or the financial resources to pay for medical care before 2014.¹²

The Affordable Care Act (ACA) has provided unprecedented new opportunities to address low rates of insurance coverage among people returning to the community after incarceration. Under the ACA, thirty states and the District of Columbia have opted to expand Medicaid to allow enrollment of nonelderly people whose income is less than 138 percent of the federal poverty level, including single or childless men. Before the ACA, a large majority of adult Medicaid enrollees were women, since enrollment was limited to specified categories of low-income individuals, including children and their caregivers, pregnant women, and individuals with disabilities. The expansion of Medicaid in these states to most low-income adults meaningfully affects justice-involved populations, which disproportionately consist of low-income men.¹³ Nathan Birnbaum and coauthors report that nearly all criminal justice–involved individuals are now eligible for Medicaid upon release in Medicaid expansion states.¹⁴

Before the ACA, Medicaid enrollment among justice-involved individuals with disabling mental illnesses was associated with increased use of health care services and reduced recidivism.^15,16 These findings suggest that enrolling people by the time of their release could enhance access to health care for a group with high health risk, while also reducing crime and incarceration. By federal regulation, Medicaid funds cannot be used to cover health services provided while a person is incarcerated.^17,18 As a result, an individual’s Medicaid enrollment must be either suspended or terminated during the period of incarceration. However, incarcerated people are eligible for Medicaid upon release.

Given the changing landscape of insurance eligibility under the ACA and the overlap between people newly eligible for Medicaid and the justice-involved population, a relatively small group of state and local jurisdictions have begun innovative programs to have incarcerated individuals enrolled in Medicaid by the time they are released. To date, no study has attempted to capture a national picture of the efforts that are taking place. We filled this gap by creating an inventory of efforts to enroll justice-involved individuals in Medicaid across the nation.

Study Data And Methods

PROGRAMS ENROLLING JUSTICE-INVOLVED PEOPLE IN MEDICAID

Between June and December 2014 we used two strategies to identify programs enrolling criminal justice–involved populations in Medicaid. In the first strategy we asked policy experts around the country to identify such programs. Experts in the field included leaders of programs that were then enrolling justice-involved populations in Medicaid as well as representatives of national health and criminal justice advocacy organizations and foundations. Experts were identified with the assistance of the study funder, the Laura and John Arnold Foundation, and of the Open Society Foundations’ Medicaid Jail Enrollment Initiative and Community Oriented Correctional Health Services. In the case of California, we initially identified programs via a survey conducted by Californians for Safety and Justice, and we then contacted these programs to gather information for the inventory.

In the second strategy we conducted an online search to identify additional programs enrolling individuals in Medicaid that were established before or in the context of ACA Medicaid expansions. Using key search terms, we identified and reviewed multiple information sources, including peer-reviewed journal articles, news media, and publicly available reports from both governmental and private organizations and agencies. For peer-reviewed journal articles, we searched PubMed and Google Scholar. For news media, we searched LexisNexis and Google News. For all searches, we used the following key search terms: prisoners, prison, jail, health insurance, Medicaid, and Affordable Care Act. Searches were conducted in combination with the names of the fifty states and the District of Columbia.

Once we had identified counties and states with programs that enrolled justice-involved individuals, we developed a standardized protocol to collect information on the characteristics of each program. We identified contacts to interview, and we reached out to them between August and December 2014 by e-mail, telephone, or both to collect summary information.

Some counties and states were operating programs with multiple components (for example, a component that enrolled people while they were in jail and another component that enrolled people who were in the probation system) that had overlapping staff and resources. As a result, we categorized program information at the jurisdiction level—that is, the county or state in which the program operated—instead of categorizing each individual component of the program.

We developed a coding instrument to categorize the key components of each program. They included program setting, enrollment personnel, point of screening for Medicaid eligibility, and target enrollment population. Options for program setting were local jails, state prisons, the probation or parole system, or other. Options for enrollment personnel were corrections staff; staff of a community-based or nonprofit organization; staff of a public health department, social service agency, or other health agency; or other. Options for point of screening for Medicaid eligibility were during booking, during incarceration, during release, after release, or other. Options for target enrollment population were people with serious mental illnesses, with substance abuse treatment needs, or with chronic somatic illnesses; other special populations; or no special population.

INVENTORY OF PROGRAMS

Our inventory included programs in place as of January 2015. We contacted leaders of each program in early 2015 to validate program information as of that January. We simultaneously administered a six-item survey via e-mail to collect information on whether the programs’ jurisdictions used any of four specific policy approaches to facilitate Medicaid enrollment. These were suspension instead of termination of Medicaid benefits for people who were incarcerated, use of presumptive eligibility for Medicaid (that is, presuming an inmate is eligible before the application has been processed, so that as soon as the inmate is released, Medicaid can be billed for his or her health care), allowing enrollment during incarceration, and accepting alternative forms of identification during the enrollment process.

All of the sixty-four programs in the final inventory validated the summary and coding of their jurisdictions, and sixty-one of them completed the survey about adoption of the four policy approaches. We report the status of programs in the inventory as of January 2015; our study did not capture program evolution after this date.

The inventory includes programs directly affiliated with county and state corrections departments that initiated Medicaid enrollment of justice-involved individuals during detention, incarceration, or the release process. To maintain a feasible scope for the study, we excluded community-based organizations, such as the Transitions Clinics in San Francisco, that initiated Medicaid enrollment of clients primarily in the community or in clinical settings once clients had left the justice system.¹⁹

LIMITATIONS

This study had several limitations. First, while the inventory is as comprehensive as was feasible, we may have missed some programsthatwereinoperationinJanuary2015. Our Internet search terms included all states. However, it was not feasible to search for every county or to systematically survey every county and state corrections department. County-based enrollment programs in California were identified via a recently conducted state survey. Thus, we were able to be somewhat more comprehensive in capturing all programs in California than was feasible in other states. Nonetheless, our use of a rigorous data collection process allowed us to identify all or nearly all programs across the nation during our study period.

Second, only forty-two of the sixty-four programs in our inventory were able to provide information on the number of individuals they had enrolled in Medicaid. Thus, the information on how many justice-involved individuals have been enrolled constitutes an underestimate.

Finally, we did not systematically collect information on sources of funding to establish and run these programs. However, the ability to sustain the programs will depend on a long-term viable funding source—in particular, to pay for screening and enrollment personnel. More information on how these programs have funded their activities would be helpful for other jurisdictions considering the feasibility of launching justice-based Medicaid enrollment efforts.

Study Results

We identified sixty-four programs that enrolled justice-involved individuals in Medicaid during detention, incarceration, or the release process. Fifty-seven of these were in states that had chosen to expand Medicaid, and seven were programs that targeted disabled populations and operated in states that had not expanded Medicaid as of January 2015.²⁰

Seventeen programs operated at state-level departments of corrections, and forty-seven operated at the county level. Collectively, the programs represented twenty-one states. The location of programs varied across census regions, with forty-four programs located the West and no more than nine programs in any of the other regions (Exhibit 1). Thirty-three of the programs in the West were located in counties in California (data not shown).

EXHIBIT 1.

Characteristics Of Sixty-Four Programs Enrolling Criminal Justice–Involved Populations In Medicaid

Characteristic	All programs		State-level programs		County-level programs
	No.	%	No.	%	No.	%
SETTING
Jail	49	77	4	24	45	96
Probation or parole system	27	42	3	18	24	51
Prison	17	27	15	88	2	4
Other	5	8	1	6	4	9
START YEAR
2014 or later	38	59	8	47	30	64
2012–13	14	22	3	18	11	23
Before 2012	12	19	6	35	6	13
CENSUS REGION
West	44	69	4	24	40	85
Northeast	9	14	6	35	3	5
Midwest	7	11	4	24	3	6
South	4	6	3	18	1	2
ENROLLMENT PERSONNEL
Staff of public health department, social service agency, or other health agency	36	56	4	54	32	68
Corrections staff	23	36	13	76	10	21
Staff of community-based or nonprofit organization	16	25	4	24	32	68
Other	9	14	3	18	6	13
POINT OF SCREENING FOR MEDICAID ELIGIBILITY
During incarceration	35	55	13	76	22	47
After release	30	47	3	18	27	57
Within 90 days before release	28	44	10	59	18	38
During booking	9	14	0	0	9	19
Other	1	2	0	0	1	2
TARGET ENROLLMENT POPULATION
Individuals with serious mental illnesses	11	17	2	12	9	19
Individuals with substance abuse treatment needs	4	6	0	0	4	9
Individuals with chronic somatic illnesses	3	5	1	6	2	4
Other special population	19	30	2	12	17	36
No special population	53	83	15	88	38	81

SOURCE Authors’ analysis of survey response data provided by programs in the inventory. NOTE Programs can belong to more than one category in some sections of the exhibit.

Thirty-eight (59 percent) of the programs started in January 2014 or later (Exhibit 1). Cumulative enrollment statistics were available for forty-two programs, and those forty-two programs reported having enrolled 112,520 justice-involved individuals in Medicaid as of the end of our study period (data not shown).

There was significant variation in how the sixty-four programs approached enrollment of criminal justice populations. Overall, the programs were more likely to be set up at the county than at the state level (as noted above); to be implemented in jails instead of other criminal justice settings; to employ staff of public health or other social service agencies to conduct Medicaid enrollment instead of, for example, corrections personnel; and to target a general population of justice-involved people instead of special populations such as people with serious mental illnesses (Exhibit 1).

Some programs operated multiple components in different settings and were therefore coded into more than one category for program characteristics. Forty-nine programs operated in jails, seventeen in the prison system, and twenty-seven in the probation or parole system. Other settings included work-release programs and community-based reentry centers.

As noted above, a majority of programs used personnel from public health or social service agencies to conduct enrollment. County-level programs were more likely than state-level programs to do so (68 percent versus 54 percent). The presence of non–criminal justice personnel in justice settings has necessitated the formation of partnerships across public agencies.

Enrollment personnel also included corrections staff members and staff members of community-based or nonprofit organizations. State-level programs were more likely than county-level programs to use corrections staff (76 percent versus 21 percent).

Overall, programs most often screened for Medicaid eligibility during in carceration (55percent) or after release from prison or jail (47 percent). Smaller percentages of programs enrolled individuals within ninety days before release or during booking. Compared to county-level programs, state-level programs were more likely to screen for eligibility during incarceration (76 percent versus 47 percent) and less likely to screen during the booking process (0 percent versus 19 percent). The lack of enrollment during booking at the state level compared to the county level may be due to the shorter sentences of inmates in jails—typically operated at the county level—which results in a shorter time between booking and release.

Eighty-three percent of the programs did not target a specific population in the correctional setting. However, some programs targeted people with serious mental illnesses, substance abuse treatment needs, or chronic somatic illnesses.

Several programs that did not target a specific population prioritized enrolling populations with high medical and behavioral health needs or had pilot-tested programs focused on a specific population. Programs that started before the Medicaid expansions or those that were located in nonexpansion states focused on people with categorical eligibility for Medicaid, such as those with disabling mental illnesses. Some programs replaced their targeted enrollment screening strategy after Medicaid expansion with population-based screening once nearly all inmates became eligible for Medicaid. Although programs varied in the strategies they used to enroll incarcerated individuals, several patterns of innovative practices emerged.

SUSPENSION VERSUS TERMINATION

Some states had policies that allowed people arriving at a correctional facility with Medicaid benefits to suspend those benefits, instead of terminating them, upon incarceration. For example, before 2007, inmates’ benefits in New York were terminated, and inmates had to wait months after being released for their reapplication to Medicaid to be processed. But in 2007 the state passed a law that made it possible to suspend an inmate’s Medicaid benefits and immediately reinstate the benefits upon the inmate’s release.²¹ Forty (66 percent) of the programs were in states with such policies and applied them (Exhibit 2).

EXHIBIT 2.

Innovative Practices Adopted By Sixty-One Programs Enrolling Criminal Justice–Involved Individuals In Medicaid

Practice	Description	All programs		State-level programs		County-level programs
		No.	%	No.	%	No.	%
Suspension of benefits	Program temporarily suspends, rather than terminates, existing Medicaid benefits upon incarceration	40	66	8	50	32	71
Presumptive eligibility	Program assumes Medicaid eligibility upon release and bills Medicaid for recently released inmates’ health services or prescriptions before application is completely processed	12	20	5	31	7	16
Authorizing enrollment during incarceration	Medicaid applications are submitted and processed before the release of an inmate	49	80	14	88	39	87
Alternative forms of identification	Alternative forms of identification and proof of income (such as corrections identification and jail release letter) are accepted	44	72	11	69	33	73

SOURCE Authors’ analysis of survey response data provided by programs in the inventory.

Despite the large number of programs that implemented suspension policies, many faced barriers in rendering them functional. For example, Colorado passed a law that called for incarcerated individuals with Medicaid benefits to have the benefits suspended and to maintain eligibility. But as of September 2015 the state had not been able to alter the Colorado Benefits Management System, the statewide system used to process benefits applications and determine eligibility, to accommodate the suspension of benefits (Terri Hurst, Colorado Criminal Justice Reform Coalition, personal communication, September 2, 2015).

PRESUMPTIVE ELIGIBILITY

Presumptive eligibility refers to the practice of assuming that a person is eligible for Medicaid before he or she has completed the enrollment process, which allows providers to deliver services without waiting until the Medicaid application is processed. This practice is typically used in states to allow qualified entities such as hospitals to provide temporary Medicaid eligibility and immediate medical services to specific populations, such as pregnant women and children. Twelve programs used presumptive eligibility to cover immediate health services and medications for recently released inmates (Exhibit 2).

For example, Connecticut applied the concept of presumptive eligibility to individuals released from the state prison system. Because the state begins the Medicaid enrollment process for all eligible inmates sixty to ninety days before release, presumptive eligibility largely applies to those who are released unexpectedly. Following an analysis showing that virtually all released inmates were eligible for Medicaid, the Connecticut Department of Social Services began piloting a program in 2010 to issue vouchers to released inmates that can be used at community-based pharmacies to receive needed medications. Inmates fill out condensed applications for the insurance exchange, which are processed by Department of Social Services specialists, and they can immediately receive needed prescriptions even if their Medicaid application is pending. This program was expanded to additional correctional facilities and courts in November 2012 (Colleen Gallagher, Connecticut Department of Corrections, personal communication, September 25, 2014).

AUTHORIZING ENROLLMENT DURING INCARCERATION

Forty-nine programs give people the ability to begin the enrollment process during incarceration when release is imminent, which allows inmates to leave a corrections facility with their Medicaid enrollment approved (Exhibit 2). For example, the Washington State Health Care Authority (HCA), the state’s Medicaid agency, released regulatory guidance to allow for enrollment up to thirty days before release if a correctional facility or county office has a memorandum of understanding in place with the HCA.^22,23 This has resulted in several counties’ launching enrollment programs. And in 2013 California passed Bill AB 720, which allowed each county board of supervisors and sheriff’s office to designate an entity to begin enrolling inmates in Medicaid thirty days before release.²⁴ This resulted in a rapid growth in such programs throughout the state.

Some states have faced infrastructure barriers to implementing such laws. For example, although Wisconsin allows individuals leaving prisons and jails to begin and complete the enrollment process before release via a telephonic application, as of September 2015 the state’s online Medicaid application had not been updated to allow a person to complete an application if he or she is incarcerated (Mike Bare, Community Advocates Public Policy Institute, personal communication, September 3, 2015).

ALTERNATIVE FORMS OF IDENTIFICATION

One barrier for justice-involved populations seeking Medicaid benefits is the lack of specific documents, such as proof of identity (for example, a driver’s license or birth certificate) and proof of income, that are typically required to complete a Medicaid application.²⁵ Forty-four programs have developed systems that allow people to use alternative documents as proof of income and identification (Exhibit 2). For example, Rhode Island Medicaid allows inmates to use their corrections identification to verify name, date of birth, and residency.²⁶

Discussion

Criminal justice–involved populations constitute a sizable component of the population eligible for Medicaid following state expansions under the ACA. As a result, new and expanded justice-based enrollment programs have emerged in states and counties around the country with the goals of improving population health and reducing recidivism. Many programs have opted to screen incarcerated people for eligibility once their release date has been set and to make use of enrollment personnel from health and social service agencies. Overall, however, programs are using a wide variety of methods depending on what best fits their criminal justice and Medicaid systems. We identified four promising innovative practices that could be used by states or counties considering launching justice-based Medicaid enrollment efforts: the suspension, instead of the termination, of Medicaid benefits when a person is incarcerated; presumptive eligibility; authorizing enrollment during incarceration; and allowing the use of alternative forms of identification for enrollment.

While the literature generally supports policies to suspend instead of terminate Medicaid benefits upon incarceration, there is variation in how well suspension policies work in practice.^27,28 David Rosen and coauthors found that most states with suspension policies reported being able to reactivate benefits within a month of release but that some states with the policies reported that reenrollment was not an automatic process.²⁸ Furthermore, Joseph Morrissey and coauthors concluded that some local jurisdictions suspend benefits, despite state policies that require termination of benefits.¹⁶ Future research should characterize states’ suspension activities to understand the effects on enrollment after release and identify best practices for the implementation of suspension policies.

Previous research on the effect of presumptive Medicaid eligibility among pregnant women indicates that the policy is associated with an increase in completing Medicaid enrollment and initiating prenatal care following the initial services provided under presumptive eligibility.^29,30 Future research should assess whether authorization of enrollment during incarceration and use of modified presumptive eligibility decreases the probability of gaps in medication or care immediately after a person is released from incarceration. It will be important to better understand the contribution of presumptive eligibility to insuring this group.

Previous research has also identified the lack of proper documentation as a key barrier to Medicaid enrollment.^17,25,31,32 Medicaid agencies often require numerous forms of identification for enrollment, which may have been confiscated or absent at booking or not be available following release. Former prisoners face cost and time barriers since obtaining some documents, such as birth certificates, may take several months.²⁵ These logistical barriers to acquiring necessary documentation are likely to be harder to surmount for programs initiating the enrollment process while a person is still incarcerated, since such people have limited access to required documentation. As a result, several programs allow the use of alternative documentations, such as a corrections identification card. Future research should examine the variation in these policies and identify mechanisms that facilitate the use of alternative forms of documentation, focusing particularly on how state information technology infrastructure (for example, databases that determine benefit eligibility and online systems that process applications) should be updated to accommodate them.

While our inventory focused on programs that were directly associated with state and local correctional agencies in enrolling individuals during detainment, incarceration, or the release process, several community-based programs also exist. For example, Camden Churches Organized for People, in New Jersey, enrolls released inmates as part of a larger program to reintegrate people back into the community. The efforts and effects of these initiatives should be examined to gain a richer understanding of insurance enrollment of justice-involved individuals.

Policies such as presumptive eligibility and suspension of benefits are an important first step in continuity of care. However, the provision of health insurance and medication alone cannot ensure that former inmates seek and receive needed health care after incarceration.¹²

Several programs in this inventory are engaged in other approaches to increase the likelihood that Medicaid enrollment will result in engagement in care. For example, state prison facilities in Rhode Island conduct educational courses to enhance prisoners’ health literacy. In Multnomah County, Oregon, the corrections and health departments partner with local community-based providers to schedule postrelease medical and behavioral health appointments. And in Shasta County, California, inmates in the county work-release program receive a two-day sentence reduction if they submit a Medicaid application. Moving forward, other programs may consider using these or other incentives for enrollment to increase engagement in medical care services.

Conclusion

The Affordable Care Act has provided an unprecedented opportunity to expand criminal justice–involved individuals’ access to health insurance. Because justice-involved populations have comparatively high health needs, many jurisdictions have used the ACA as an opportunity to initiate Medicaid enrollment for these populations. This first national review of programs providing that enrollment following passage of the ACA highlights certain facilitators of and barriers to implementation.

The criminal justice infrastructure is a complex system housed in multiple settings that requires a variety of policies and approaches to successfully implement Medicaid enrollment programs. This is illustrated by the great variation in design of the sixty-four programs in the inventory. As these initiatives evolve, future research should examine how they influence health and criminal justice outcomes and should assess their role in larger efforts to reduce health disparities and improve population health outcomes.