The prevalence of diabetes is high in this mentally ill Medicaid population (11.8%), compared to the national prevalence of diagnosed diabetes (7.8%) estimated by the Centers for Disease Control and Prevention.31
Diabetes quality of care did not meet recommended standards. We found that 47% of psychiatric Medicaid patients with diabetes were given annual HbA1c testing, 56% received lipid testing, and 32% had eye examinations. In contrast, a recent VA study found that approximately 65% of all patients with diabetes had HbA1c measured in the last year.14
National self-report surveys from the early part of this decade suggest that among 18- to 75-year-olds with diabetes, 84.6% had an annual lipid profile and 67.7% had a dilated eye examination.29
Using the Behavioral Model of Health Services Utilization as our quality of care framework, we found that predisposing mental health (vulnerability) measures (low GAF and being prescribed SGA medication) were negatively associated with some diabetes quality of care indicators. Furthermore, some of the contextual measures, including receiving mental health services from a FFS psychiatrist, visiting a primary care physician, or having eye glass services, were positively associated with some quality of care indicators.
It may not be surprising that diabetes screening is less common among the handful of patients determined by mental health clinicians to have serious impairments in communication and judgment. It is disturbing that the use of SGA medication was significantly associated with a reduced likelihood of HbA1c testing, particularly given calls for increased metabolic screening among such patients.9
Indeed, psychiatrists have been encouraged to do their own primary care monitoring, including fasting glucose and lipid profiles, for patients with serious mental illness who are unable to secure such testing through traditional primary care mechanisms.30
Contextual factors, as measured by type of service or type of provider, did influence diabetes quality of care. It is not clear why patients who see a FFS psychiatrist are more likely to have diabetes preventive screenings. The FFS psychiatrists could have ordered the screenings or been more assertive in encouraging the preventive care. Alternatively, FFS psychiatric patients may be a self-selected group with better transportation or less severe psychiatric pathology.
Clearly, interventions to improve diabetes care for socially disadvantaged populations should be directed to health providers and the health system as well as to patients.32
Indeed, systems or contextual factors commonly associated with poor physical care of those with serious mental illness include different funding streams, different location of physical and mental treatment sites, and organizational difficulties in communication.33
Poor medical management of psychiatric patients in California was documented years ago.34
In fiscal year 2000–2001 alone, California county mental health programs served more than 197,000 adults with a serious mental illness35
in a system that is carved-out from physical health care. Consistent with experiences in other states,36
as California’s Medicaid managed care (for both physical and mental health care) was implemented in the 1990s, Memorandums of Understanding were signed between the physical and mental health plans, which clarified who pays for what (such as medications) but did not organize meaningful service integration.
Regardless of carve-out financing, physician organizations providing primary care services are more likely to use organized care management processes (case management, physician feedback, disease registries, clinical practice guidelines, and programs for teaching self-management skills) for chronic diseases if there are substantial external financial incentives, such as contracts, and adequate clinical information technology capacity.37
A recent survey found that many physician and provider organizations heavily involved in California’s Medicaid program are extensively engaged in preventive and chronic care management programs.38
In fact, a southern California study involving Medicaid patients found that a diabetes case management program was successful in improving glycemic control.39
Medicaid psychiatric patients receiving SGA prescriptions in California were more likely to have lipid testing than patients in several other states.9
Other states have tried aggressive approaches to overcoming the legal, financial, and organizational barriers of behavioral health carve-outs. In Michigan’s Washtenaw County, the state, the county, a university, and a Medicaid managed care plan collaborated to create a new entity to provide coordinated care.40
In Oregon, a Medicaid-only plan experimented with two different models of implementing the Robert Wood Johnson Foundation’s “Depression in Primary care” initiative.36
Unfortunately, there are not published data showing improvement in diabetes care.
Further study within this county using more detailed data could examine geographical access issues, such as physical co-location of mental and primary care services, because there are more than 20 county-operated and contracted outpatient clinics spread across 30 cities and towns, with some in urban centers and others in more remote desert or mountain communities. It is possible that psychiatric FFS providers may have been located closer to other health care providers than were mental health clinics.
This study has the basic limits of administrative data, which include lack of clinical specificity, possible miscoding, and possible incomplete coding.41
The prevalence of diabetes was likely undercounted given its reliance on International Classification of Diseases, Ninth Revision
diagnosis codes in claims data. Other studies have considered the presence of antidiabetic drug prescriptions9
and other clinical laboratory values.12
The observed diabetes prevalence of 11.8% in this sample is lower than the 15.2% reported in Medicaid patients from Ohio;42
however, it is comparable to a similar study that identified diabetes in 9% of psychiatric Medicaid patients in California9
and a study that found diabetes claims for 11.1% of beneficiaries in a southern state.4
Although chart reviews often are used to assess quality of care,11
such an approach was not feasible given the organizational barriers between medical and mental care providers. Furthermore, unlike researchers in the VA12
or private managed care organizations,13
we were unable to examine laboratory values to determine whether screening examinations had occurred.29
FFS Medi-Cal claims are fairly complete and accurate, however, because the data are linked to payment.41
Although managed care organizations do not submit all medical/outpatient encounter data, data quality is roughly comparable between managed care and FFS.41
Some researchers have suggested that GAF scores are of limited value, particularly for predicting mental health related outcomes.43
It is not possible to assess the GAF score’s validity in this dataset; however, the fact that low scores were associated with lower quality of care does suggest that the score may have relevance for non-mental health outcomes. Finally, there may be limited generalizability of findings because results are from one county in California. Nonetheless, there are many similarities, such as patient characteristics and benefits structure, in Medicaid-funded mental health systems within California and across all 50 states.44
In summary, retrospective analysis of California Medicaid claims for one county suggests that compliance with standard diabetes screening measures was lower among psychiatric patients who received care in public mental health clinics. Predictors of worse diabetes care were poor psychological and social functioning and anti-psychotic medication usually reserved for refractory schizophrenia and bipolar illness. Diabetes care was improved when patients were seen by primary care physicians or eye specialists. Further study is needed to better understand why patients who received specialty mental health services through private-practice FFS psychiatrists received better diabetes care than those patients receiving specialty mental health care only in county mental health clinics, and to identify realistic methods for reducing those disparities.