In Massachusetts, 97% of the population has health insurance as a result of Massachusetts General Law Chapter 58 (1), “An Act Providing Access to Affordable, Quality, Accountable Healthcare.” The insurance coverage provisions of the Patient Protection and Affordable Care Act, including expanded Medicaid, subsidies for low-income participants, insurance exchanges, and the individual mandate, are modeled on Massachusetts General Law Chapter 58.

Previous studies conducted in Massachusetts after implementation of Chapter 58 have indicated that patients who are engaged in behavioral health services (for mental disorders, substance use disorders, or both) are uninsured at higher rates than persons in the general population (2; personal communication, Botticelli M, Bureau of Substance Abuse Services, Massachusetts Department of Public Health, 2010). Why are so many patients who seek behavioral health services in a state with near-universal coverage not enrolled in insurance plans at the time they seek services? And what can be done to reduce the number of patients who are eligible for coverage but who remain uninsured and to promote greater continuity of coverage? This qualitative study conducted in June and July 2011 sought to better understand these questions, which are important to state policy makers designing both expanded Medicaid and state insurance exchange programs.

At face value, it is not surprising that factors frequently associated with behavioral health disorders, including loss of employment, frequent change of address or lack of stable housing, and unconventional behaviors, correlate with a loss of or lack of health insurance. An analysis of population survey data showed that 22.6% of adults with frequent mental distress were uninsured, which was higher than the rate of 17.7% among adults with frequent physical distress (3). A study in Massachusetts that was conducted after implementation of Chapter 58 estimated that between 25% and 30% of behavioral health patients were uninsured at the point of service (2). Another point-of-service survey undertaken by the Massachusetts Department of Public Health found that 25% of patients seeking detoxification services were primarily uninsured 18- to 25-year-old males (personal communication, Botticelli M, Bureau of Substance Abuse Services, Massachusetts Department of Public Health, 2010). Anecdotal evidence from chief executive officers (CEOs) of community mental health centers and addiction treatment agencies have confirmed these estimates; these individuals stated that more than 20% of their patients are uninsured at the point of needing emergency acute services.

A study conducted in Massachusetts four years after implementation of Chapter 58 found that most (66%) of the general population of adults who remained uninsured were employed but were not insured because either their employer did not offer insurance or they could not afford subsidized insurance (4). Essentially, uninsured persons in Massachusetts have either not enrolled in available insurance options or have become disenrolled. When a person becomes disenrolled from his or her health insurance and is otherwise eligible to reenroll in the same or a different health insurance plan, the disenrollment-reenrollment process is called “churn” (5). Previous studies have described the general size and scope of the issue of churn; however, with the exception of children, these studies have not focused on specific subpopulations, such as persons with diagnoses of behavioral health disorders (6,7). Churn and other forms of uninsurance have an impact on the health care system at all levels: patient health is compromised, providers cannot be reimbursed, and the additional administrative burden of disenrolling and reenrolling patients is costly and inefficient for subsidized insurance plans and government agencies.

A report from the Robert Wood Johnson Foundation’s Maximizing Enrollment Initiative noted that the lack of current standardized metrics for churn makes describing the magnitude of the occurrence of eligible-but-not-enrolled persons challenging (8). The report proposed several measures to track the magnitude and impact of churn. Other health insurance studies have discussed “continuity ratios,” which divide the average monthly number of enrollees during a fiscal year by the total number of unduplicated persons enrolled at any time over the year. For all categories of Medicaid eligibility, the continuity ratio was found to be 78% nationally and 82% for Massachusetts (9). For the low-income, childless, adult population, the same report noted continuity ratios of 68% nationally and 78% for Massachusetts. A 2003 report stated that 38% of the population under 65 experienced a break in health insurance coverage over a three-year period (7). Given these estimates of insurance discontinuity in the overall low-income population, it is not surprising that the scope of churn is greater for vulnerable low-income individuals with mental and substance use disorders and related circumstances such as homelessness.

MassHealth is the state’s Medicaid program, and Commonwealth Care is the subsidized insurance offered through the state’s insurance exchange. Massachusetts invested in a substantial outreach initiative to enroll residents in health insurance. Outreach and enrollment support from MassHealth included $11.5 million over four years (2008–2012), automatic enrollment of certain low-income and uninsured groups, contracts to enrollment assistance centers, and grants to other nonprofit agencies for navigators to reach targeted groups defined by language, geography, race, and immigration status (10). Health Care for All, a statewide consumer advocacy organization, helps individuals locate and access affordable health insurance coverage and is funded by the state to field 40,000 calls annually to its help line. This assistance also provided mini-grants to 50 organizations (emergency departments, community health centers, and patient advocacy organizations) that all have access to the “Virtual Gateway,” the state’s online consumer portal for accessing services. In addition, the Blue Cross Blue Shield (BC/BS) Foundation of Massachusetts and other area health foundations fund approximately 23 additional Connecting Consumers With Care grants ($20,000–$25,000 per grantee) that together assist from 1,800 to 8,000 individuals per month (10).

Collectively, these “navigators” assist beneficiaries with initial submission of enrollment documents and follow-up tracking, interpretation and response to verification notices, reapplication and reenrollment processes, plan and provider selection, appeal and adjudication of eligibility determination decisions, and transitions between plans. Stand-alone behavioral health organizations are not eligible for or represented in either the MassHealth or the BC/BS grants.

An application for MassHealth coverage can be initiated by direct patient application, navigator-assisted application, or enrollment center–assisted application. A documented “user call” on June 16, 2011, to the fully automated MassHealth help line required from six to ten transactions to generate the mailing of a 33-page membership booklet containing 13 pages of application forms. The booklet explains nine MassHealth programs, eligibility criteria, application and appeals processes, and health plan and provider choices, along with other information.

Two major drivers—the information technology (IT) infrastructure and the operating policies and procedures—shape the systems that enroll, disenroll, and reenroll beneficiaries in MassHealth and Commonwealth Care. A uniform application form and the same infrastructure and system are employed for both programs. The IT software infrastructure for applications is based on DOS programming language, has been modified over the years, and is no longer easily adapted. At the time of the study in 2011 a new system was being developed that would operate like those in the 32 states that have Web-based real-time application systems (11). Currently, applicants cannot directly complete and submit an electronic application for insurance. Beneficiaries can access the Virtual Gateway to view their current coverage and to update personal information after eligibility is established. Only state-designated providers can submit electronic applications on behalf of an individual through the Virtual Gateway. The electronic application requires supplemental paper forms to complete the submission. Individuals can submit paper-based applications with or without the aid of a navigator, independent of the Virtual Gateway system.

These fixed drivers that shape the system—the IT infrastructure and the operating policies and procedures—are complemented by a multitude of programs, people, contracts, practices, and other components that constitute the day-to-day functions involved in enrolling, disenrolling, and reenrolling beneficiaries. [Two figures illustrating the steps in the enrollment and reenrollment processes for these two programs are available as an online data supplement to this article.]

This study used a qualitative approach to understand current practices and barriers that result in otherwise eligible and often previously enrolled patients becoming uninsured. The study focused on two publicly subsidized Massachusetts health care programs, MassHealth and Commonwealth Care, the subsidized insurance offered through the Health Connector, the state’s insurance exchange, which also offers other approved plans. The study enjoyed support and cooperation from the state directors of mental health, addiction, and Medicaid behavioral health units, as well as the behavioral health provider association and related consumer advocacy organizations.

Structured discussions were conducted in June and July 2011 with a purposive sample of 15 individuals representing the state mental health authority, the single state addictions agency, the Medicaid behavioral health division, the state’s insurance exchange (Health Connector), and two different enrollment assistance and consumer advocacy organizations. During the same period we held a total of four focus groups: one group with 14 CEOs of community-based addiction and mental health service providers and three groups with 33 patients of two different agencies in three different settings. Focus group participants represented convenience samples. The CEOs were members of the Association for Behavioral Health, who volunteered to add one hour to a standing meeting. The patients responded to recruitment postings seeking volunteers at two of the state’s largest community behavioral health agencies, located in central and southeastern Massachusetts. Patients were all low-income adults between age 20 and 55, primarily white and equally divided between men and women. Patients were provided with a $35 grocery gift card for participating.

Discussion guides were developed and used for all interviews and focus group meetings. The same individuals (study authors and team members) conducted all interviews and focus group meetings. Data were also gathered by direct use of the telephone application system and of the Virtual Gateway. Previous reports on and studies of enrollment and disenrollment in health insurance provided context for the study.

Eleven of the 33 behavioral health patients in the focus groups were currently uninsured. Half of the patients (N=17) reported being uninsured during the past year. Only one of 33 patients had never enrolled in one of the subsidized public plans.

With few exceptions, the patients in the focus groups described the system of applying, responding to questions, and reapplying as complex and confusing. Patients reported that it took from 45 minutes to two hours to complete the enrollment forms, with most reporting more than one hour. They reported additional time required for gathering, copying, and mailing or delivering verification materials, such as pay stubs, birth certificates, and proof of state residence. Most who had been disenrolled reported being unaware that they had become uninsured until they sought services. The reasons offered for not knowing about their disenrollment included not opening the mailed notification for fear of the implications of receiving a “government notice,” inability to understand the notice content, and never seeing or receiving the notice.

The most common reasons reported for disenrollment included change of address and “not receiving notices,” change in income from Supplemental Security Income, change in Medicare status, and miscalculation of income (for example, counting certain U.S. Department of Veterans Affairs benefits as income). Many of the patients in the focus groups reported relying on the effort of safety-net providers, help line staff, and personnel of the MassHealth enrollment center to navigate application, disenrollment, and reenrollment. The case of one individual, with reenrollment assistance, took from April 2010 until May 2011.

The patient focus groups and the group that included policy makers and advocates identified the following disenrollment opportunities: missing and incomplete documentation; acute and chronic conditions, especially severe mental and addiction disorders; long-term disabilities that interfere with a patient’s ability to respond to complex notices; and lack of awareness among beneficiaries of the consequences of changes in income, numbers and ages of dependents, and other conditions that trigger termination from the enrolled program and create eligibility for and need to transfer to either another Medicaid program, Commonwealth Care, or Medicare. Possible approaches to addressing these issues are discussed below.

Scope of the challenge

There are 1.3 million beneficiaries in MassHealth (12), 160,000 beneficiaries in Commonwealth Care, and an estimated 35,000 persons who are eligible for Commonwealth Care but not enrolled. MassHealth staffs an eligibility central processing unit and four enrollment centers. Each day, an enrollment center receives approximately 160 to 200 families in person, 800 to 1,200 calls, and 800 pieces of mail. Performance metrics for this work focus on the number of and length of time that applications and supporting documents are in queue.

The functions required to establish eligibility, enroll new members, annually recertify 1.3 million existing members, randomly reverify income of existing members, and reenroll members whose coverage was terminated for other reasons involve a huge number of transactions for both the state system and the beneficiaries. Each transaction represents an opportunity for missing, incomplete, or incorrect data responses, most of which can trigger disenrollment, which increases churn. The numbers are significant: MassHealth closes about 34,000 cases per month for administrative purposes; almost 11,000 of these individuals are reenrolled within 90 days; the estimated cost to MassHealth of closing and reopening a case is $200 (5).

Reasons for and consequences of disenrollment

The policies and procedures that drive the MassHealth and Commonwealth Care systems are built on a default assumption of ineligibility or disenrollment until the applicant or beneficiary takes responsibility to establish eligibility and complete enrollment requirements. For example, if mail is returned because of an address change, eligibility is terminated until the beneficiary establishes a new address. (The exception to this practice is for individuals who indicate homelessness on the application.) Eligibility may also be administratively terminated when the beneficiary does not provide information in response to notices that are automatically generated when inconsistent beneficiary data are found by matching the MassHealth database and the database of the Massachusetts Department of Revenue (weekly match) or the database of the federal Social Security Administration (daily match).

Safeguards are built into the system to maintain health services. To promote continuous access to health care, Massachusetts retained an uncompensated care program (Health Safety Net) to defray the cost of care to uninsured individuals that is delivered in health centers and hospitals (behavioral health centers are ineligible). Another safeguard takes the form of a check-off box on the application that indicates homelessness or nursing home resident status. This provision also allows the community provider or navigator to be designated as “eligibility representative designee” and receive information on behalf of the beneficiary and monitor his or her coverage status. The homeless designation ensures enrollment continuity for 12 months. MassHealth also provides monthly member recertification notices to managed care organizations and primary care program physicians. Theoretically, this allows organizations to follow up with their enrollees to complete the recertification forms. However, most behavioral health clients receive these services in managed behavioral health carve-outs that do not receive the notices.

The consequences of disenrollment and churn are especially significant for beneficiaries. Patients with chronic illnesses, including asthma, diabetes, hypertension, and psychiatric conditions, who experience interruptions in health insurance coverage are hospitalized with greater frequency than those with continuous coverage (13,14). When the patient is aware of the coverage gap, he or she often delays or defers use of needed services. For the behavioral health patient, this can mean that continuing rehabilitation services are not available after detoxification, for example, or that medication compliance is not monitored or that counseling is not available to complement medication therapies.

When health insurance coverage is interrupted, negative consequences also accrue to the provider that must subsidize the care, to the navigator-support system that bears the cost of reenrolling the beneficiary, and to the insurer-payer that bears the administrative costs of enrollment and disenrollment processes as well as the downstream medical costs of deferred or delayed care.

Addressing the challenge

Changes in both the information system infrastructure and the operating policies and procedures involved in outreach, enrollment, disenrollment, and reenrollment can improve continuity and retention in health insurance, especially for behavioral health patients. For example, reducing the number of human transactions and increasing the use of prepopulated real-time electronic systems can speed up enrollment and reduce opportunity for disenrollment, error, data gaps, and inefficient redetermination processes. Real-time, online, electronic application and reenrollment systems are currently employed in 32 states (10). Rules proposed by the Centers for Medicare and Medicaid Services encourage electronic enrollment systems with “express-lane eligibility” capabilities (prepopulated with existing state data) and waiver applications that allow express-lane capacity for adults as well as children (15).

Although outreach and navigator assistance are most needed at initial enrollment, continuous assistance is necessary to maintain enrollment. It can be provided at key locations that target the most disabled and vulnerable populations, such as offices of the Temporary Assistance for Needy Families program, community mental health and specialty addiction treatment centers, legal aid offices, community health centers, and hospital emergency departments. System improvement is also required to address the fact that a portion of postenrollment navigation work represents a less-than-efficient mechanism to fill systemic information gaps, update data, or correct human errors for patients whose life circumstances are fluid at the same time that their essential eligibility characteristics (for example, poverty and chronic illness) remain relatively stable. Enrollment, recertification, and reenrollment outreach that includes clear, concise, and easy-to-understand multilingual communication with applicants and beneficiaries can be conducted by designated third-party representatives that identify and reach the most disabled individuals.

Federal Medicaid policy requires state recertification at six- or 12-month intervals. To verify continuing eligibility, Massachusetts matches Medicaid databases with other Federal and state employment and revenue collection databases (for example, Social Security and state revenue). Presumed eligibility and exemption from “random-match results” can be extended for defined categories of permanent or long-term beneficiaries, such as those who have severe and persistent mental illness, those who are disabled and dually eligible for Medicaid and Medicare, children under 18, and those who were previously covered and currently seeking emergency services. Presumed eligibility for these beneficiary categories that are most likely to be continuously eligible supports continuity of care and avoids wasted administrative processing time and resources, the inefficient use of navigator time, and uncompensated care for providers. In addition the default assumption of ineligibility when a patient fails to initially respond to a notice could be changed to “continued eligible” pending full verification of circumstances or until ineligibility status is demonstrated. Program integrity can be maintained by ongoing sampling analysis of cases and by assigning representative benefactors to verify beneficiary status.

The ability of policy and administrative personnel to shape and target improvement practices with the highest potential return in terms of quality and efficiency is hindered by the absence of standardized data that captures enrollment and disenrollment performance. The need for meaningful data to better understand the consequences or costs of churn is highlighted in a Robert Wood Johnson Foundation–sponsored report that describes three categories of specific metrics: enrollment-disenrollment, retention-transition, and indicators of coverage cessation (8). Enrollment metrics should focus on numbers of applications and reapplications in the queue, length of processing time, rates of disenrolled beneficiaries who become reenrolled within specified time frames (for example, one to three months), and the disenrollment period for eligible but not enrolled beneficiaries. Further efforts are required to define metrics that capture the costs of utilization and of unwarranted navigator and administrative work and uncompensated care associated with disenrolling and reenrolling otherwise eligible beneficiaries.

Finally, the misalignment between Medicaid, subsidized exchange, and Medicare plans of enrollment effective dates (the beginning and end of the month, annually, and so forth) and of benefit continuity and eligibility categories often results in periods of discontinuity for beneficiaries whose circumstance requires migrating between these plans. Enrollment effective-date periods, benefit continuity, and eligibility categories within and between these programs should be aligned.

Study limitations

This study had several limitations. Obvious limits were imposed by the study focus and design. First, the study represented one state, Massachusetts, which has a near-universal, mandate-based health insurance plan. However, because of the similarity in coverage provisions between the Patient Protection and Affordable Care Act and Massachusetts General Law Chapter 58, lessons from the study may be generalized to other states. Second, the study is limited by the absence of standardized metrics and accompanying data to provide a context and basis for comparison over time within and between states. Unfortunately, this gap appears to be applicable to all states. Third, the study intentionally focused on vulnerable patients who were receiving care and thus may not accurately reflect the experience of all populations who were not receiving treatment. This population, however, is significant both in size and in its utilization of health services. Despite these limitations, the study attempted to accurately and completely document the systems and processes in one state that reach out to, determine eligibility for, enroll, and disenroll individuals and that advocate for members of subsidized insurance plans.

The promise of broader if not universal coverage of individuals who previously lacked health insurance is especially welcomed by consumers, providers, and advocates involved with preventing and treating mental and substance use disorders. Coverage for this population holds the promise of better access to care for untreated persons, greater continuity of care for those in treatment, and more opportunity to address related health conditions. For providers, barriers to the provision of prevention and treatment services that result from insufficient funding could be mitigated by insurance payments.

The experience in Massachusetts raises concerns. The ability to provide treatment for these vulnerable and chronically ill patients and for other similar groups may be compromised by insurance churn. Direct understanding of both the architecture (information system technology) and processes (policies and procedures) used to reach, enroll, and maintain enrollment for patients is key to achieving the promise of high-quality continuous care. A “walk through” (16) (a direct personal experience of the enrollment process) by key Medicaid leaders of a different component of the enrollment system each year—for example, the initial application, including the initial request for application telephone call; response to a verification letter; or response to the annual recertification request—is one way policy makers can understand the obstacles that individuals experience in enrollment. Another approach is to develop and make available to all parties and the public an annual “process map” that illustrates the flow of an application from initial request to recertification, including verification processes. Such practices together with the improvements discussed in this article can reduce the number of patients with mental and substance use disorders who are eligible but unenrolled and ensure that needed continuous care is provided.

The study on which this article is based was supported by prime contract HHSS283200700017I/HHSS28300001T TO HHSS28300002T from the Substance Abuse and Mental Health Services Administration to Deloitte Consulting LLP.

The authors report no competing interests.