Lichstein, Jesse C. MSPH*; Domino, Marisa E. PhD*,†; Beadles, Christopher A. MD, PhD*,†,‡; Ellis, Alan R. PhD, MSW†; Farley, Joel F. PhD§; Morrissey, Joseph P. PhD*,†; Gauchat, Gordon W. PhD†; DuBard, C. Annette MD, MPH∥; Jackson, Carlos T. PhD∥
One of the major challenges facing the US healthcare system today is appropriately caring for patients with multiple chronic conditions (MCC), defined as ≥2 concurrent chronic conditions. The prevalence of MCC nationally is >25% and rising.1 Patients with MCC experience poor quality of life,2 poor functional status,3–5 decreased disease control,6,7 and increased mortality.1,8 In addition, patients with MCC generally have greater healthcare needs, requiring both more care and more complex care,1 typically from multiple providers. On the basis of this evidence, researchers and policy makers have suggested that patients with MCC need coordinated, high-quality healthcare to preserve functional status and improve disease control.1,9
Comorbid chronic conditions are particularly common among patients with a severe mental illness (SMI) such as schizophrenia. Patients with SMI have a high prevalence of hypertension, cardiovascular disease, hyperlipidemia, and diabetes,10 which when combined with SMI are more detrimental to overall health than in the general population.11 Furthermore, preventable medical conditions are the leading cause of premature death among patients with SMI.12,13 These difficulties are compounded by the fact that many patients with SMI do not receive regular care, resulting in undiagnosed, untreated, or poorly treated physical health conditions.14,15 Even those receiving care from psychiatric specialists often do not obtain primary care, and care coordination among multiple providers is lacking.16
The medical home approach, currently at the forefront of national and state healthcare reform, has the potential to create more efficient utilization of primary care. Based in primary care settings, medical homes promote accessible, comprehensive, patient centered, and coordinated care with systems-based quality improvement.17 Recent studies suggest that medical homes improve primary care utilization and quality of care for patients with chronic physical illnesses.18,19 However, it is still unclear to what degree patients with MCC, particularly those with comorbid SMI, participate in and utilize medical homes. The objective of this study was to examine participation in and utilization of primary care medical homes in a large sample of Medicaid patients with MCC with and without SMI. Describing utilization patterns and highlighting potential issues, such as differential utilization for patients with SMI, will benefit medical home planners, administrators, and evaluators as they seek to implement and improve medical homes.
We examined medical home participation and utilization in Community Care of North Carolina (CCNC), North Carolina’s medical home program for the state’s Medicaid population. CCNC began as a pilot program in 1998 and has since expanded to a statewide program with 14 community networks of physicians, nurses, hospitals, and other community organizations.20 The program links enrollees to primary care medical homes, implements disease-specific quality improvement initiatives, and provides case management for high-risk patients.21 Patient enrollment in CCNC is generally mandatory, but some Medicaid categories (eg, foster children) have optional enrollment and some categories (eg, residents of skilled nursing facilities) are excluded; people in the optional and excluded categories tend to have different diagnoses than mandatory enrollees. The 14 nonprofit networks receive a modest per member per month (PMPM) payment (ranging from $3 to $13.50) from the state to provide care management services as well as training, coordination, medical support staff, pharmacists, and psychiatrists.22 Participating primary care providers also receive a PMPM payment for managing and coordinating care, but otherwise are reimbursed for services on a fee-for-service basis. The program includes over 1600 primary care practices in North Carolina, which together manage the care of approximately 1.2 million of the state’s Medicaid recipients.23
We used the North Carolina Integrated Data for Researchers, a unique data source containing North Carolina Medicaid claims data linked to data from state psychiatric hospitalizations, state-funded mental health services, and a 7-county regional behavioral health carve-out.24 The data included demographic information and monthly data on Medicaid enrollment, CCNC medical home enrollment, diagnoses, and medical care utilization for fiscal years 2008 through 2010. The linked data source allowed for improved detection of mental illness, as mental health diagnoses were available from all 4 administrative data systems. For example, if a patient received a mental health diagnosis through state-funded mental health services but not through services covered by Medicaid, we were able to correctly assign a mental health diagnosis to that individual when analyzing Medicaid data.
Study Design and Sample
To examine participation in and utilization of medical homes by patients with and without comorbid SMI, we constructed 2 retrospective cohorts of patients with MCC: (1) children aged 6–17 years and (2) adults aged 18–64 years. Children and adults were examined separately due to differences in Medicaid eligibility criteria and different patterns of diagnosis and treatment. Children under the age of 6 were excluded because of a low prevalence of MCC. Age was determined at the beginning of each fiscal year.
The study samples included patients with ≥2 of the following 8 chronic health conditions: major depressive disorder, psychosis (schizophrenia for adults), hypertension, diabetes, hyperlipidemia, seizure disorder, asthma, and chronic obstructive pulmonary disease (COPD). We defined individuals as having the above chronic conditions if they had at least 2 outpatient or emergency department visits or at least 1 inpatient stay with the relevant diagnosis in any of our administrative data sources. Because these 8 conditions are chronic, we defined disease indicators on a time-invariant basis.
For each year of data, we included only those individuals who contributed at least 1 month of medical home enrollment, and thus enrolled in Medicaid for at least a month during the fiscal year. Annual medical home enrollment across the 3 years was approximately 78% for children and 65% for adults. We excluded months in which an individual was hospitalized for at least half of the month or was dually eligible for Medicaid and Medicare. Data were then aggregated to the person-year level, yielding up to 3 observations per person. We used annualized measures to aid in interpretability. Sensitivity analyses were conducted at the person-month level; results were consistent with those of the main analyses and are not presented here.
The main outcome measures for this study were participation in and utilization of a medical home. We defined participation as a binary indicator of having at least 1 visit to the patient’s assigned CCNC provider, among those enrolled; and we defined utilization as the number of visits to the CCNC provider among those who participated in CCNC. In combination, these 2 measures allow a detailed description of the association between SMI and healthcare use. For example, 2 patient groups may have the same probability of participating in a medical home, but 1 group may have much lower utilization of a medical home. If participation and utilization are influenced by different factors—for example, if participation is influenced more by patient-level factors than provider-level factors—understanding differences at both levels will be important and informative for policy makers.
Patients were considered enrolled in a medical home if monthly management fees to both the primary care provider and the CCNC network were identified in the claims. We calculated visits attributed to a medical home provider by matching the billing provider number on the claim for the outpatient visit to the billing provider number of the CCNC payments to the provider. Because of potential error in matching provider identification numbers, we conducted sensitivity analyses defining participation as having at least 1 outpatient (primary care or medical specialist) visit reimbursed by Medicaid during a month when the patient was enrolled in the medical home program. Results from these analyses were consistent with those found in the primary analyses and therefore are not reported separately.
The main explanatory variable of interest was diagnosis of SMI, either major depressive disorder (ICD-9-CM codes 296.2X, 296.3X, 300.4X, 311.XX) or psychosis (ICD-9-CM code 295.XX), which includes schizophrenia for adults. SMI was defined as a categorical variable with 3 categories: psychosis (with or without concurrent major depressive disorder), major depressive disorder without concurrent psychosis, and no record of psychosis or depression. Major depressive disorder was defined without concurrent psychosis to examine differences in medical home use for the 2 mental health diagnoses.
Control variables included age at start of fiscal year, sex, race (categorized as white, black, or other), ethnicity (Hispanic, not Hispanic), number of months enrolled in the medical home during the year, total number of chronic conditions (between 2 and 8 inclusive), and 6 diagnosis indicators for diabetes, asthma, hypertension, hyperlipidemia, seizure disorder, and COPD. Year-fixed effects were also included to control for time trends. To determine whether differences across CCNC networks biased our results on included characteristics, such as race or ethnicity, we ran additional models with indicators for each of the 14 networks (results on model covariates were very similar to the main analyses and are available from the lead author upon request).
We first conducted bivariate analyses, comparing values of medical home participation and use, as well as control variables for individuals with or without SMI. We then examined 2 multivariate models, each corresponding to one of our outcome measures, to assess the association between SMI and medical home use. The participation model utilized generalized estimating equation (GEE) models with binomial distribution, logit link, and exchangeable within-group correlation structure. The utilization model was a GEE model with reintroduced zeros, negative binomial distribution, log link, and exchangeable within-group correlation structure; models with Poisson distributional assumptions yielded similar results. GEE models were used to account for clustering of repeated individual observations over time. We controlled for the number of months enrolled in a medical home (requiring Medicaid enrollment) using a quadratic form, as a greater proportion of time in a medical home provides a greater opportunity for participation and visits, and the association is expected to diminish over time. To improve model fit, final analyses also included age in quadratic form, and interactions between total number of chronic conditions (between 2 and 8) and the study diagnoses. All models were estimated in Stata 12. Results are presented as average marginal effects (AME) for ease of interpretation, although the model is not intended to be causal as omitted characteristics correlated with disease indicators could be contributing to differences in participation.
The study protocol was approved by the University of North Carolina at Chapel Hill Institutional Review Board.
The final sample included 8759 children and 105,542 nonelderly, nondual (ie, eligible for Medicaid but not Medicare) adults with ≥2 study conditions who were enrolled in medical homes. Annual medical home participation was 75.8% for children and 73.3% for adults (Table 1). On average, children and adults both had 3.5 medical home visits per year, including enrolled individuals without visits. Nearly 60% of children were diagnosed with 1 of the 2 mental illnesses examined here, compared with approximately 50% of adults. Children had an average of 2.2 of the 8 target chronic conditions examined, with asthma being the most prevalent (72.8%); adults had an average of 2.9 of the 8 target chronic conditions, with hypertension being the most prevalent (76.3%). Among both cohorts, most patients were female (51.7%–67.1%), white (46.7%–51.4%), and non-Hispanic (95.8%–97.8%). Relative to their proportions in the overall population, white individuals were overrepresented among patients with depression, black individuals were overrepresented among patients with psychosis, and Hispanic individuals were underrepresented in both categories.
In unadjusted analyses, children with major depressive disorder and those with psychosis were less likely to participate in the medical home, had fewer months enrolled in the medical home, and had lower utilization (P<0.01), compared with those without depression or psychosis (Table 1). Children with depression or psychosis were older than children without depression or psychosis and were less likely to be male, Hispanic, or of other race. Although the prevalence of each chronic physical condition was lower among children with depression or psychosis, on average these children had more chronic illnesses in total (including the depression or psychosis) than other children.
Adults with psychosis exhibited patterns similar to those of children with psychosis, except that adults with psychosis were younger and had more months of medical home enrollment (P<0.01) than adults without depression or psychosis. However, in contrast to the patterns among children, adults with major depressive disorder had greater utilization (P<0.01) than adults without depression or psychosis.
In the multivariate analysis, children with major depressive disorder were significantly less likely to participate in a medical home (AME=−5.0 percentage points, P<0.05) (Table 2). However, among those who visited a medical home, utilization was not significantly different (AME=−0.22, P=0.18) from that of children without depression or psychosis.
Diagnosis of psychosis was negatively associated with both participation and utilization. Children with psychosis had a 12.2 percentage point lower probability of participating in a medical home (P<0.01), and had 0.92 fewer medical home visits (P<0.01) among those participating.
Of the chronic physical conditions, only asthma was significantly and positively associated with both outcomes. Increasing age had a significant negative association with both outcomes, as did male sex and other race. Black race also had a strong negative association with participation and medical home utilization. Compared with white children, black children had a 12.6 percentage point lower probability of participation in a medical home (P<0.01) among those enrolled, and had 1.5 fewer medical home visits among those participating (P<0.01). In addition, number of months in the medical home and the year-fixed effects had significant positive associations with all outcomes.
Adults with major depressive disorder did not have a significantly different rate of participation or number of medical home visits, compared with adults without depression or psychosis. Compared with adults without depression or psychosis, adults with psychosis had an 8.2 percentage point lower probability of participation (P<0.01) and 1.0 fewer medical home visits per year (P<0.01).
All physical illness diagnoses except seizure disorder were positively associated with participation in the medical home. Seizure disorder was negatively associated with utilization (P<0.01), whereas COPD was positively associated with utilization. Black race again had a strong negative association with participation in a medical home and utilization. Hispanic ethnicity and male sex were also negatively associated with participation and utilization. Number of months in the medical home was again positively associated with both outcomes; however, the year-fixed effects were negatively associated with medical home utilization, indicating a decrease in the average number of visits per year over the study period.
Given the high prevalence and burden of physical comorbidity in patients with SMI, primary care participation and utilization as well as coordination of care are of great importance to these patients. The medical home has the potential to increase the use of primary care, and overall, medical home participation and utilization were relatively high across diagnosis groups. We found that most Medicaid recipients with MCC who were enrolled in a medical home made ≥1 visits to that medical home, even those patients with SMI. Consistent with concerns that patients with SMI may not participate in primary care medical homes as readily as other patients with MCC,25,26 we found that rates of medical home participation were lower among children with depression and among both children and adults with psychosis than among their counterparts without either condition. These differences were more pronounced for patients with psychosis than for those with depression. In addition, among patients who participated in medical homes, those with psychosis had fewer total visits.
Given our findings, it may be beneficial to develop strategies to further increase participation in and utilization of the medical home model among patients with SMI. Primary care providers may be unfamiliar with the population with SMI and/or unprepared to provide primary care in coordination with mental health specialists. Strategies could include provider incentives such as enhanced PMPM payments for clinicians treating patients with SMI. Patients may benefit from outreach, incentives, and education surrounding the benefits of the medical home and of primary care. Furthermore, it may be necessary to implement a combination of these provider-level and patient-level strategies to increase both participation and utilization. For example, patient-targeted strategies may be useful for encouraging participation in care in the first place, and provider-targeted strategies may be important for improving the process of care and encouraging patients to return for continued care.
Importantly, our findings suggest that patients with SMI and medical comorbidities are not a homogenous group with regard to their participation in and utilization of medical homes. Several factors may be contributing to this observed heterogeneity. For example, primary care providers are generally more comfortable treating depressive disorders than psychosis-related disorders,27,28 and depression is more commonly treated in the primary care setting.29 This may translate into greater medical home participation and/or utilization among patients with depression than among those with psychosis.27,28 Therefore, fewer enhancements to the primary care medical home may be required to improve participation and utilization among patients with depression and comorbid medical conditions. Some observers have argued that for those with schizophrenia and other forms of psychosis, participation and utilization may be increased more easily by locating health homes in the specialty mental health sector rather than in primary care settings.25,26
In enhancing medical home participation and utilization among patients with SMI, another distinction that merits consideration is the degree of medical comorbidity. To illustrate, patients with high behavioral health needs and high physical health needs (sometimes referred to as Quadrant IV patients) may have different health priorities than patients with high behavioral health needs and low physical health needs (Quadrant II patients).30 Patients with few medical comorbidities (Quadrant II) may be best served in the specialty mental health sector, with a reverse co-location approach,31 where embedded medical providers address physical health needs. Similarly, patients with greater physical health needs (Quadrant IV) may receive optimal care with on-site behavioral healthcare in the primary care setting. Therefore, strategies to improve participation and utilization for patients with SMI should consider not only provider-level and patient-level strategies, but also optimal location of services.
In addition, disease-specific initiatives may also provide avenues to improve care for patients with SMI. During its existence, CCNC has implemented several disease-specific initiatives to target prevalent conditions that require a coordinated, system-wide approach to care. These programs identify and recruit patients and improve quality of care through standardization of care, best practices, patient and provider education, tracking the process and outcomes of care, and community partnerships.32 One of the first disease-specific initiatives at CCNC was targeted toward Medicaid recipients with asthma. In our study, patients with asthma exhibited high rates of medical home participation and utilization compared with the other populations examined (Table 2). Given the success of this program, a similar targeted approach might improve the participation and utilization among patients with specific mental illnesses. CCNC recently implemented such an approach for behavioral health in general. Beginning in 2010, a Behavioral Health Integration Initiative was launched to support the integration of primary care, mental health, and substance abuse services in primary care practices across North Carolina.33 Results of these management strategies should be evaluated for patients with SMI and medical comorbidities.
There are several limitations to our study. First, some of the observed differences in participation and medical home visits may be due to differences in disease severity or other unobserved factors rather than diagnosis. Second, chronic diagnoses examined here were treated as time invariant; differential rates of diagnosis between medical homes and other providers could affect the interpretation of these results. Third, although we examined diagnoses of psychosis or major depressive disorder, we did not measure the presence of other mental health conditions, such as bipolar disorder or anxiety disorder. Despite these limitations, this study provides an important insight into the degree to which patients with MCC and comorbid SMI participate in and utilize medical homes as well as the heterogeneity among SMI patients.
As medical homes are implemented, it will be important for policy makers to understand how diverse populations participate in and utilize medical homes to create strategies to better serve these populations. This study provides a preliminary examination of medical home participation and utilization by patients with MCC, with and without SMI. Across diagnosis groups, we found relatively high rates of medical home participation. However, participation and utilization were lower among patients with SMI, particularly for those with psychosis. Targeted strategies may be necessary to increase medical home participation and utilization among patients with SMI.
1. Anderson G .Chronic Conditions: Makingthe Case for Ongoing Care. 2010 .Princeton, NJ:Robert Wood Johnson Foundation and Johns Hopkins University.
2. Fortin M, Bravo G, Hudon C, et al .Relationship Between Multimorbidity and Health-related Quality of Life of Patients in Primary Care.Qual Life Res. 2006; 15:83–91.
3. Bayliss EA, Bayliss MS, Ware JE, et al .Predicting declines in physical function in persons with multiple chronic medical conditions: what we can learn from the medical problem list.Health Qual Life Outcomes. 2004; 2:47
4. Kadam U, Croft P. N.S.G.C. Group .Clinical multimorbidity and physical function in older adults: a record and health status linkage study in general practice.Fam Pract. 2007; 24:412–419.
5. Marengoni A, von Strauss E, Rizzuto D, et al .The impact of chronic multimorbidity and disability on functional decline and survival in elderly persons. A community-based, longitudinal study.J Intern Med. 2009; 265:288–295.
6. Wong ND, Lopez V, Tang S, et al .Prevalence, treatment, and control of combined hypertension and hypercholesterolemia in the United States.Am J Cardiol. 2006; 98:204–208.
7. Kerr EA, Heisler M, Krein SL, et al .Beyond comorbidity counts: how do comorbidity type and severity influence diabetes patients’ treatment priorities and self-management? J Gen Intern Med. 2007; 22:1635–1640.
8. Menotti A, Mulder I, Nissinen A, et al .Prevalence of morbidity and multimorbidity in elderly male populations and their impact on 10-year all-cause mortality: the FINE study (Finland, Italy, Netherlands, Elderly).J Clin Epidemiol. 2001; 54:680–686.
9. Boyd C, Fortin M .Future of multimorbidity research: how should understanding of multimorbidity inform health system design? Public Health Rev. 2010; 451–474.
10. Nasrallah HA, Meyer JM, Goff DC, et al .Low rates of treatment for hypertension, dyslipidemia and diabetes in schizophrenia: data from the CATIE schizophrenia trial sample at baseline.Schizophr Res. 2006; 86:15–22.
11. Hennekens CH, Hennekens AR, Hollar D, et al .Schizophrenia and increased risks of cardiovascular disease.Am Heart J. 2005; 150:1115–1121.
12. Osborn DP .The poor physical health of people with mental illness.West J Med. 2001; 175:329–332.
13. Parks J, Svendsen D, Singer P, et al .Morbidity and Mortality in People With Serious Mental Illness. 2006 .Alexandria, VA:National Association of State Mental Health Program Directors.
14. Salsberry PJ, Chipps E, Kennedy C .Use of general medical services among Medicaid patients with severe and persistent mental illness.Psychiatr Serv. 2005; 56:458–462.
15. Lambert TJ, Newcomer JW .Are the cardiometabolic complications of schizophrenia still neglected? Barriers to care.Med J Aust. 2009; 190:suppl S39–S42.
16. Dombrovski A, Rosenstock J .Bridging general medicine and psychiatry: providing general medical preventative care for the severe mentally ill.Curr Opin Psychiatr. 2004; 17:523–529.
17. Iglehart JK .No place like home—testing a new model of care delivery.N Engl J Med. 2008; 359:1200–1202.
18. Peikes D, Zutshi A, Genevro JL, et al .Early evaluations of the medical home: building on a promising start.Am J Manag Care. 2012; 18:a105–a122.
19. Jackson GL, Powers BJ, Chatterjee R, et al .The patient-centered medical home.Ann Intern Med. 2013; 158:169–178.
20. Willson CF .Community care of North Carolina: saving state money and improving patient care.N C Med J. 2005; 66:229–233.
21. Dobson LA, Hewson DLevis .Community Care of North Carolina: an enhanced medical home model.N C Med J. 2009; 70:219–224.
22. Steiner BD, Denham AC, Ashkin E, et al .Community Care of North Carolina: improving care through community health networks.Ann Fam Med. 2008; 6:361–367.
23. Dobson LA Jr, Hewson DL .Community Care of North Carolina in 2013.N C Med J. 2013; 74:S S12–S15.
25. Alakeson V, Frank RG, Katz RE .Specialty care medical homes for people with severe, persistent mental disorders.Health Aff. 2010; 29:867–873.
26. Smith TE, Sederer LI .A new kind of homelessness for individuals with serious mental illness? The need for a “mental health home”.Psychiatr Serv. 2009; 60:528–533.
27. Williams JW Jr, Rost K, Dietrich AJ, et al .Primary care physicians’ approach to depressive disorders. Effects of physician specialty and practice structure.Arch Fam Med. 1999; 8:58–67.
28. Goldman LS .Medical illness in patients with schizophrenia.J Clin Psychiatry. 1999; 60:suppl 21 10–15.
29. Spitzer RL, Kroenke K, Linzer M, et al .Health-related quality of life in primary care patients with mental disorders. Results from the PRIME-MD 1000 Study.JAMA. 1995; 274:1511–1517.
30. Mauer B .Behavioral Health/Primary CareIntegration: The Four Quadrant Model and Evidence-Based Practices. 2006 .Rockville, MD:National Council for Community Behavioral Healthcare.
31. Collins C, Hewson DL, Munger R, et al .Evolving Models of Behavioral Health Integration in Primary Care. 2010 .New York:Millbank Memorial Fund.