Community Responses to Diabetes and Mental Health Care for the Uninsured Population in Michigan

Abstract

Objective

To examine how the safety net in Michigan is responding to the health care needs of their uninsured population with diabetes and/or mental illness.

Methods

We used a multiple-site case study design. Data were collected through interviews of key informants in five Michigan communities. Analytic patterns and themes were identified, and compared across communities and by organizational type.

Results

Informants reported they are managing to meet the needs of uninsured diabetics but are having great difficulty caring for the uninsured with mental illness. Specialty care for diabetes is obtainable, but is resource-intensive. Mental health services available for uninsured patients are severely limited. The presence of a county health plan (CHP) appears insufficient to improve access, especially to mental health services.

Conclusions

The safety net for Michigan’s uninsured population with diabetes and mental illness is weak. Processes including referrals and care coordination are of poor quality in some communities. The value of integrating mental health services into primary care should be examined.

Mental illness and diabetes confront many in Michigan’s uninsured population.^1,2 Treating these conditions effectively requires a mix of primary care and specialty care providers. For instance, based on the availability of effective interventions, the American Diabetes Association³ recommends annual screening for retinopathy in diabetic patients. Similarly, the American Psychiatric Association’s⁴ treatment recommendations for depression include psychotherapy and pharmacotherapy, which aim to prevent such complications as lost income and suicide. With adequate primary care and accessible specialty care, many complications of these disorders can be prevented.^5–7

According to the most recent Current Population Survey,⁸ 10.5% of all people in Michigan lack health insurance, and this rate is expected to rise. Participants in a recent study of Michigan’s health care safety net noted that patients increasingly present with multiple chronic diseases, in particular diabetes and mental illness.⁹ During the same time period, funding for mental health services and primary care for the uninsured in Michigan has been declining. For example, since 1991, 11 of Michigan’s 15 state-operated psychiatric hospitals have closed.¹⁰

In this article, we report the results of a project examining how Michigan’s safety net is responding to the health care needs of their uninsured populations with diabetes and/or mental illness. Our objective was to describe the specialty referral and care coordination processes used in the communities studied and to identify barriers to these services.

Methods

We used a multi-site case study approach focusing on health care safety net dynamics in selected communities.¹¹ We defined the community as the organizations that make up the safety net for the uninsured in a specific county. We sought to examine how organizations within a community organize to meet the mental health and chronic care needs of the uninsured population (with diabetes and depression as the index cases). These disorders were selected because (1) safety net organizations in Michigan have identified diabetes and mental illness as common;⁹ (2) both of these conditions are major causes of morbidity and mortality; and (3) there are accepted national standards guiding clinical care.

Site selection

We selected five counties for in-depth interviews. Criteria for site selection included the following: geographic diversity within Michigan (i.e., sites from the Upper and Lower Peninsulas, and sites from urban and rural areas); diversity of organizational types; and diversity of population groups served.

Specifically, we attempted to include communities with and without a county health plan (CHP); CHPs are designed to provide access to health care for uninsured and underinsured individuals and are available in most Michigan counties. Some plans resemble insurance models by reimbursing for covered services, while not providing referral or coordination activities. The second type of plan acts as a care coordinator on behalf of the patient by arranging for primary and specialty care within a network of volunteer providers or free clinics.

Data collection

Within the selected counties, we conducted in-depth interviews with key informants who have decision-making authority at safety net organizations. At each site, the executive director or chief executive officer and the chief medical officer or clinical director were selected for interviews. We developed standard semi-structured interview protocols to guide the interviews, but allowed for open discussion. Question topics included availability of specialty services for diabetes and mental illness, referral processes and networks, the role of CHPs, strengths and weakness of current system, and policy recommendations. The protocols were pre-tested at a free clinic and revised, where appropriate. The interviews were conducted in 2005–2006; each lasted between 1 and 1.5 hours. All interviews were led by one of two lead investigators (VKD, PDJ). Extensive notes from the interviews were later used to prepare case summaries for each community.

Data analysis

The two lead investigators (VKD, PDJ) reviewed case summaries for each community. Using these case reports, the investigators used content analysis¹² to identify patterns and themes, across communities and by organizational type.

Results

Sample characteristics

Five of six contacted communities agreed to participate in the study. Participating counties reflected the regional diversity of Michigan. The general demographic and health indicator information for the participating communities are presented in Tables 1 and 2. The number and type of organizations participating in each community are shown in Table 3. In our sample, two communities have the traditional insurance type of CHP (1 and 3), two counties have care coordinator type health plans (2 and 4) and one county (5) lacks a free health plan.

Table 1.

BASIC DEMOGRAPHIC INFORMATION FOR SELECTED COUNTIES

County	Populationa	Poverty Rateb	Mortalityc	Unemploymentd
1	596,666	10.0%	8.0 ± 1.8	6.5
2	83,971	7.8%	6.0 ± 4.8	5.3
3	278,592	12.7%	7.9 ± 2.9	7.6
4	64,760	10.5%	n/ae	5.7
5	1,998,217	16.50	10.6 ± 1.2	9.7

^a2005 Estimate, U.S. Census Bureau Fact Sheets, 2005. Available at quickfacts.census.gov/qfd/states/26000.html.

^b2003 Data, U.S. Census Bureau Fact Sheets, 2005. Available at quickfacts.census.gov/qfd/states/26000.html.

^c2004 Data, Michigan Department of Community Health, Number of Infant Deaths, Live Births and Infant Death Rates by County of Residence, 2004. Available at www.mdch.state.mi.us/pha/osr/InDxMain/Tab3.asp.

^dJuly 2006 Unemployment Rate. Available at www.michlmi.org/LMI/lmadata/laus/2005/laus2006.htm.

^eToo few events to calculate a confidence interval of statistical value.

Table 2.

DIABETES AND MENTAL HEALTH INDICATORS FOR SELECTED COUNTIES

County	Diabetes Ratea (per 100)	Diabetes Related Mortalityb (per 100,000)	Suicide Ratec (per 100)
1	7.4	16.8	9.3 ± 2.5
2	7.2	23.0	10.4 ± 3.9
3	7.4	22.5	7.8 ± 3.3
4	7.2	43.2	7.8 ± 3.3
5	7.4	29.0	9.1 ± 1.3

^aMichigan Department of Community Health, Diabetes Fact Sheets by County, 2002 Data, August 2004. Available at www.michigan.gov/mdch/0,1607,7-132-2940_2955_2980-13768--,00.html.

^bDiabetes as Cause of Death. Available at www.mdch.state.mi.us/pha/osr/CHI/Deaths/frame.html.

^cData for 2004. Available at www.mdch.state.mi.us/pha/osr/chi/cri/frame.html.

Table 3.

ORGANIZATIONS INTERVIEWED

County	Types of Organizations* (# individuals interviewed)
1	CHP(1), CMHA(2), FC(1), HOS(5)
2	CHP(2), CMHA(2), FC(2), HOS(3)
3	CHP(1), CMHA(1), FC(1), HOS(1), HD(3)
4	CHP(2), CMHA(1), HD(1)
5	CMHA(1), FC(2), FQHC(1), HOS(1)

CHP = county health plan

CMHA = community mental health agency

FC = free clinic

FQHC = federally-qualified community health center

HOS = hospital

HD = health department

In total we interviewed 35 individuals at 20 organizations including health departments, community mental health agencies (CMHA), county health plans (CHP), free clinics, hospitals, and community health centers (CHC). Our analysis revealed several themes within four patterns regarding our study communities’ ability to care for uninsured populations with diabetes and/or mental illness. The following patterns were identified: CHPs; diabetes care; mental health care; and integrating mental health services in primary care. Each of these patterns incorporated several themes, which are presented below.

County health plans

Our informants’ perception of CHPs’ value varied greatly. In communities with a traditional insurance type CHP (1 and 3), respondents did not perceive that the health plan improves access to care for uninsured populations. In one community, respondents complained that the plan does not provide reliable coverage for uninsured individuals, enrollment constantly opens and closes, and prescription availability is inconsistent. Conversely, in the communities with the care coordination type of CHP (2 and 4), all informants viewed the CHPs as key in their effort to improve access to primary and specialty care for the uninsured.

Diabetes care

In all communities studied, most diabetes care is reliably provided through the primary care system, rather than through sub-specialists. Specialty care for diabetics (including ophthalmology, podiatry, endocrinology, and nephrology services) is obtained through referral networks. Overall, our informants believed this is the optimal service model for diabetes care.

Study participants can obtain specialty referrals for uninsured diabetics, but such referrals often rely on loose networks of volunteer physicians and may depend on the clinic directors’ ability to encourage area physicians to help. For example, at one free clinic in community 3, the director has made connections with specialists in the region whom he could call for a volunteer referral. In other communities (2 and 4) the CHP reliably assists in arranging specialty referrals to willing providers.

Clinics working within a health system appear to have easier access to specialty services than those outside the system. For instance, in a particularly troubled community (5), the only clinic that reported access to specialty services has a formal affiliation with the community’s main health system. We found similar examples in communities 1 and 3. The interviews suggested that clinics without either formal referral systems (such as through a CHP or affiliations with local health systems) depend solely on volunteerism.

Mental health care

The CMHAs in Michigan provide care only to those patients presenting with severe and persistent mental illness (i.e., those with psycho-affective disorders who are not amenable to primary care intervention). For those who are not severely and persistently mentally ill, the CMHAs act as a triage mechanism and refer those patients back to the community.

Our interviews consistently revealed a much bleaker situation for mental health services than for diabetes care. Respondents in our sample were in virtually unanimous agreement that there are few treatment options for uninsured patients who are not severely and persistently mentally ill. Very few mental health providers in any of our study communities are willing to offer uncompensated care, either through the CHPs or as a consultation. Therefore, people who are not severely and persistently mentally ill, which we call the gap population, depend on other safety net providers for mental health services. But few safety net providers are prepared to provide these services.

None of the CHPs in our sample presently include mental health providers on their volunteer lists or reimburse for mental health services. One executive director of a health plan described the situation as “approaching a crisis” and reported that it is almost impossible to find mental health services for the gap population. Another respondent described the situation as a “horrendous catch 22”: the CMHA is unable to help people until they become disabled, at which point they qualify for Medicaid. That respondent continued, “We have created a class of disabled people to sustain ourselves.” When asked where the gap population receives services, no one knew for certain, but most respondents assumed that primary care providers in the community take on these patients.

Consequently, conflict between CMHAs and primary care providers is present in all five communities. Primary care respondents reported that CMHAs should do more for the gap population. Some expressed skepticism about how accurate CMHAs are in determining eligibility. One respondent bluntly accused a CMHA of “a systematic dumping of uninsured patients to the community.” Although we found no evidence of such practices, some respondents pointed out that CMHAs have a conflict of interest when it comes to determining mental illness severity.

We did not specifically ask our CMHA respondents about these concerns, but most acknowledged the challenges they face and the limited resources at their disposal. They were also aware that referrals back to the primary care safety net for the gap population often result in no treatment. Even so, one CMHA respondent argued that the gap population is the safety net’s responsibility, not the CMHA’s, given the specific state mandate for CMHAs to treat the severely and persistently mentally ill Medicaid population.

In sum, our study participants are pessimistic about the current system of providing care for the uninsured with mild to moderate mental illness. Sources of care are extremely limited and there is little agreement within communities about who should be serving these patients.

Integrating mental health services in primary care sites

When asked how the gap population would be best served, we found division between CMHAs and primary care providers. Many mental health providers believed this population should be cared for in the medical care network, largely though primary care providers. However, primary care respondents reported that they lack the resources and skill to manage mental illness, and doubted they would receive the required support from mental health providers. Still, some primary care providers agreed the gap population might benefit from the integration of mental health services into primary care. One respondent noted, “The separation of mental illness into a separate system makes it hard to treat illness at the right stage.”

Despite continuing debate, two communities appeared to be moving toward such an integrated model. In these communities, the CHPs are adding referral sources for mental health into their medical care network. A third community has already developed formal agreements with CMHA and private practitioners to staff primary care clinics several times a month. These arrangements begin to address the struggles between CMHA and primary care over their roles in serving uninsured patients, but have not yet been evaluated.

Discussion

Primary care safety net providers in Michigan believe they are meeting the basic medical needs of their uninsured diabetic patients, but find specialty referrals, including hospital care, require a great deal of time. Uninsured patients needing mental health services face much bleaker prospects. Access through either the CMHAs or the primary care safety net is extremely limited. It appears that CHPs have not bridged these gaps.

Nationally, uninsured patients have great difficulty with access to and quality of mental health care.^13,14 Several factors contribute to limited access. First, there is a severe shortage of mental health professionals, particularly those willing to accept uninsured patients. Second, the current mental health system for low-income people has left a huge number of uninsured patients without access to services. Cunningham et al.¹⁵ reported similar conditions among their study communities, and others have shown community health centers nationally report difficulty obtaining referrals for mental health services.¹⁶

In all communities studied, referral challenges make providers reluctant to take on uninsured patients. Other studies indicate that specialty referrals are hard to obtain. For instance, 40% of faculty members at publicly funded academic centers reported either rarely or never being able to obtain outpatient mental health services for their uninsured patients.¹⁷ Further, Gusmano et al.¹⁸ concluded that recent strategies to increase the number of community health centers to address growing need will not increase access to offsite services such as inpatient care, testing, or specialty referrals.

Using formal referral structures such as partnerships with local health systems or a CHP with care coordination services may result in fewer barriers than attempting to find volunteer providers or hospital-based care on a case-by-case basis. The potential advantages of health system membership for community health centers were recently described by Denver Health.¹⁹ After integrating the Denver area community health centers with Denver Health, access to specialty and hospital services became easier.

The number of CHPs we examined was too small to draw any conclusions about the effectiveness of CHPs in improving access to care. However, we found consistently positive views of the CHPs engaged in care coordination. The value of care coordination services has been examined in a number of other communities around the country.^20,21 Many report an improvement in access to services, although it is not at all clear that these strategies have the capacity to meet existing need. Still, much of Michigan’s current system depends largely on an informal, dynamic network of volunteers that is vulnerable to changes in funding and patient volume,^20,22,23 heightening concerns about the ability of Michigan’s safety net to meet rising needs.

Integrating mental health services into the existing primary care safety net may be an effective strategy to address the needs of the gap population by increasing access, reducing fragmentation, and facilitating coordination between various care providers. Presently, however, patients presenting to primary care providers with mental illness often do not receive recommended care.²⁴ To ensure that high quality care can be provided, many current care models require community-wide change. A recent study of one such project concluded that after implementing multidisciplinary, collaborative models of care, mental health services were improved among a group of primary care clinics, including publicly funded health centers.^25,26

Our case study methodology has several limitations. First, the findings cannot be applied to other states or populations. It is also possible our results would have been different had we selected a different set of study communities. On the other hand, because Michigan’s population is similar to that of the nation in terms of distribution by age, race/ethnicity, poverty rates, and health indicators, similar findings are likely in other areas.

The experiences of these Michigan communities indicate that the safety net struggles to provide adequate care to the uninsured with chronic illness, especially mental illness. The care coordination efforts of some CHPs, as well as providing care within a health system, appeared to relieve some of these challenges but it is clear these strategies are insufficient to deal with the needs of the uninsured. Serious consideration should be given to integration of mental health services with primary care. Although such a strategy would require considerable system change, it might well improve chronic illness care for the uninsured more broadly.