Prevalence of Treated Bipolar Disorders and Associated Comorbidities in Managed Care and Medicaid Populations

Bipolar disorder (BPD) is characterized by cyclic mood swings of manic, hypomanic, and depressive episodes. BPD often begins in adolescence or early adulthood and continues throughout life, with recurring mood episodes. Estimated relapse rates during the first year after illness onset range between 37% and 44%.^1,2 Globally, BPD is a major cause of disability, suicide, and death.³

BPD affects more than 2.3 million adult Americans, or 1% of the population, by conservative estimates.^4,5 A number of epidemiologic studies have reported prevalence rates of BPD. Variability in reported prevalence rates may be attributed to differences in sampling between studies, as well as ascertainment. For example, the lifetime prevalence rate for BPD in the United States was 0.8% in the Epidemiological Catchment Area (ECA) study and 1.6% in the National Comorbidity Study (NCS).^6,7 The NCS surveyed only noninstitutionalized individuals aged 15 to 54 years who met the Diagnostic and Statistical Manual of Mental Disorders-Third Edition-Revised (DSM-III-R) criteria for BPD, whereas the ECA included persons aged 18 to 64 years with the full spectrum of DSM-III-R BPD.^6,7 Using medical claims data, the treated prevalence rate of BPD has been estimated at between 0.2% and 0.5%.^8-10

Higher prevalence rates have been reported when ascertainment includes the full spectrum of BPDs. (Ascertainment was used because of multiple subtypes of BPD that were not assessed consistently in published studies.) Studies based on survey questionnaires have reported a wide spectrum of prevalence rates of BPDs ranging from 3.0% to 6.5% in adult populations.^11-14 Using the Mood Disorder Questionnaire screening instrument, Hirschfeld and colleagues and Das and colleagues found that from 3.7% to 9.8% of American adults may suffer from a BPD,^14,15 with higher rates occurring in low-income populations.¹⁵

Psychiatric and medical comorbidities in patients with BPD have received increasing attention in recent years,^16-21 because of the possibility that specific bipolar comorbidity associations may have an impact on the treatment of BPDs.^7,17,22,23 High prevalence rates of certain psychiatric comorbidities among patients with BPD have been reported, including personality disorders (9%- 89%),^21,24,25 alcohol and substance abuse (17%- 60.7%),^14,26-29 and anxiety disorders (25.2%-30.6%).^15,19,29 These findings, however, were based on self-reported survey or interview data that might have been subject to limitations of incomplete recall or information bias.^30,31

Numerous health insurance plans in US managed care organizations (MCOs) cover mental health services indicated by multiple research reports.^{5,16,18,32,33} Although managed care databases provide data from medical records and claims about patient diagnoses and treatments, population-based clinical studies examining the prevalence of BPDs and related comorbidities are scarce. Furthermore, little information exists about the characteristics of patients with BPD in the managed care Medicaid population, despite an increasing number of Medicaid enrollees who are also enrolled in MCOs.³⁴ We therefore proposed to examine the monthly prevalence trend of patients treated for a BPD in a large US MCO population and to examine the occurrence of treated psychiatric disorders and medical comorbidities associated with BPDs.

Patients and Method
Study Design
A longitudinal, retrospective cohort study was conducted. The primary data source was a multistate managed care claims database (PharMetrics) from January 1998 to December 2002 (5 calendar years). This database includes more than 45 million lives enrolled in MCOs with 70 health plans, including managed care Medicaid programs, in 4 US regions: Midwest (34.1%), East (15.6%), South (23.9%), and West (26.4%). Population distributions in the database are similar to the US population distribution by age and sex distributions.³⁵ This geographically diversified claims database provides a large population perspective of health information, and the use of a managed care claims database to conduct epidemiologic studies has been well documented.^36-40

Patient names, insurance plan identification numbers, and other patient identifiers were deleted from the claims database to protect patient confidentiality. Randomized patient numbers and their birth years were used for identification and calculation of age, respectively. The research project was approved by the University of Cincinnati Medical Center Institutional Review Board.

Patient Cohort Selection
The PharMetrics managed care claims database included all pharmacy, medical, and institutional claims. Each medical claim was recorded with accompanying diagnostic codes (International Classification of Diseases, Ninth Revision [ICD-9]) that justified the medical service. Between 1998 and 2002, a total of 709,865 patients, including 6.2% Medicaid enrollees, had at least 1 diagnosis of an affective disorder (ICD-9 code 296.xx) or cyclothymia (ICD-9 code 301.13). Of these, 136,763 patients, including 9.8% Medicaid enrollees, had a bipolar diagnosis indicated by one of the following ICD-9 codes—296.0, 296.1, 296.4-296.8, or 301.¹³. The final group of patients with BPD consisted of 2 separate cohorts—123,292 managed care patients and 13,471 Medicaid patients. Patients with a diagnosis of depression only (ICD-9 codes 296.2x and 296.3x) or schizophrenia (ICD-9 code 295.xx) during the study period were excluded from this cohort. Medicaid recipients aged =65 years were excluded from this cohort, because they might also have had Medicare coverage, and therefore, their Medicaid claims might not have reflected a complete medical picture.

After identifying patients with a bipolar diagnosis, we further categorized those patients with possible diagnostic subcategories, including bipolar I disorder, manic or hypomanic (ICD-9 codes 296.0x, 296.1x, 296.4x), bipolar I disorder, mixed mania (ICD-9 code 296.6x), bipolar I disorder, depressed (ICD-9 code 296.5x), and bipolar II disorder (ICD-9 code 296.89). Furthermore, patients were categorized based on the presence of psychotic features, which was indicated by a "4" as the fourth digit of the ICD-9 diagnosis code. Because numbers of patients with only BPD, not otherwise specified, or cyclothymia, were less than 0.1%, patients with these diagnoses were not subcategorized.

Data Analytic Strategy
A bipolar diagnosis indicated by defined ICD-9 codes was used to estimate prevalence rates. The numerator included patients with a bipolar diagnosis in medical claims at any time during the study period up to that month. The denominator included all enrolled persons in either the MCO (non-Medicaid) or the Medicaid program for each month during the calendar year. The cumulative prevalence rate of BPD per 100 enrollees was calculated for each calendar month during the study period. Because of the dynamic enrollment and turnover rate in Medicaid and MCOs, we assessed monthly prevalence rates to capture the prevalence pattern and measure the prevalence trend more accurately. Confidence intervals ([CIs], 95%) were constructed to assess the average monthly prevalence of BPD. Graphs of monthly prevalence rates were constructed to depict the circular trend of the bipolar prevalence rate. The prevalence rate at the highest point (ie, plateau point) was identified and considered as the prevalence rate during the study period.

Prevalence rates of treated BPD in the MCO and Medicaid populations were stratified by sex and age categories. A patient's age was calculated as the number of years between the birth year and the index date of bipolar diagnosis. The index date of a bipolar diagnosis was the first date of a bipolar diagnosis indicated by defined ICD-9 codes during the study period. Age categories were <12 years, 12 to 17 years, 18 to 34 years, 35 to 49 years, 50 to 64 years, and ≥65 years.

Descriptive methods were used to examine background characteristics of both bipolar cohorts. Demographic characteristics and frequencies of comorbidities were compared between the managed care and Medicaid populations, using chi-square tests for categorical variables and t-tests for continuous variables. In addition, the following major psychiatric and medical comorbidities were evaluated:

1. Major psychiatric disorders: alcohol use disorder (ICD-9 code 303), substance use disorders (ICD-9 code 304), anxiety disorders (ICD-9 code 300), impulse control disorders (ICD-9 code 312), personality disorders (ICD-9 code 301), and eating disorders (ICD-9 code 307.5)
2. Key medical comorbidities: cerebrovascular diseases (ICD-9 codes 433-438), ischemic heart diseases (ICD- 9 codes 411-414), neoplasm/cancer (ICD-9 codes 140-208, except 173, 211.5, 230, 235, 239.0), arthritis (ICD-9 codes 711-716), obesity (ICD-9 codes 278, 278.0), diabetes mellitus (ICD-9 code 250), hypertension (ICD-9 code 401), and chronic obstructive pulmonary disease (COPD; ICD-9 code 496).

Results
Demographic Characteristics

Table

The total monthly enrollees in this managed care population increased from 6 million in 1998 to 14 million in 2002. The proportion of Medicaid enrollees in MCOs increased from 3.2% (N = 178,668) in January 1998 to 4.4% (N = 603,937) in January 2000, and to 10.8% (N = 1.35 million) in January 2002. Compared with the Medicaid cohort (N = 13,471), the managed care cohort (N = 123,292) had significantly longer periods of enrollment (19.4 months vs 27.3 months, respectively; P <.001), fewer female patients (64.2% vs 60.7%, respectively; P <.001), and a higher mean age (29.4 years vs 36.5 years, respectively; P <.001) (Table).

More children and young adult patients (aged 0-34 years) and fewer middle-aged adults (aged 35-64 years) were diagnosed with a BPD in the Medicaid cohort compared with the MCO cohort. There was a slightly higher frequency of psychosis in the MCO cohort than in the Medicaid cohort (8.2% vs 7.5%).

Treated Prevalence

Figure 1

Figure 2

The average monthly prevalence rates of patients treated for BPD were 0.31% (95% CI, 0.29-0.33) for the MCO population, with a plateau point of 0.5%, and 0.47% (95% CI, 0.45-0.49) for the Medicaid population, with a plateau point of 0.6%. Figure 1 and Figure 2 demonstrate the trend of the cumulative monthly prevalence rate, stratified by sex, in both populations.

The prevalence rate of BPD in females was higher than that in males. In the MCO population, the prevalence rate of BPD was 0.37% (95% CI, 0.34%-0.40%) in females, with a plateau point of 0.6%, and 0.24% (95% CI, 0.22%-0.26%) in males, with a plateau point of 0.4%. In the Medicaid population, the prevalence rate was 0.55% (95% CI, 0.53%-0.58%) for females, with a plateau point of 0.7%, and 0.36% (95% CI, 0.34%-0.38%) for males, with a plateau point of 0.5%.

Figure 3

Figure 4

Figure 3 and Figure 4 indicate that monthly prevalence rates increased with age, peaking at 35 to 49 years in both populations. The plateau points of monthly prevalence rates during the 5-year period by age categories in the MCO and Medicaid populations, respectively, were <12 years, 0.1% and 0.2%; 12 to 17 years, 0.6% and 0.8%; 18 to 34 years, 0.6% and 1.2%; 35 to 49 years, 0.7% and 2.1%; and, 50 to 64 years, 0.5% and 1.3%.

Current Key Psychiatric Disorders and Medical Comorbidities
In the MCO cohort, 44% of patients had at least 1 comorbid psychiatric disorder and 29% had at least 1 comorbid medical condition. In the Medicaid cohort, 46% of patients had at least 1 comorbid psychiatric disorder and 25% had at least 1 comorbid medical condition. The most frequent comorbid psychiatric disorder was an anxiety disorder (36% and 37% in the MCO and Medicaid populations, respectively), followed by an alcohol disorder (6.5% and 8.2%, respectively), a substance disorder (5.6% and 9.6%, respectively), and a personality disorder (3.8% and 4.5%, respectively).

The most frequent medical comorbidity was hypertension (13% and 18%, respectively), followed by diabetes mellitus (7.2% and 7.4%, respectively), obesity (6.1% and 7.9%, respectively), and COPD (3.3% and 4.0%, respectively) (Table).

Discussion
This is the first multistate, population-based study based on medical claims data that examines the prevalence of treated BPDs in MCO and Medicaid populations and assesses their related psychiatric and medical comorbidities during a 5-year period. The monthly prevalence of BPDs increased to a peak of 0.7% in the managed care population and 2.1% in the Medicaid population at ages 35 to 49 years, then declined to 0.5% in the managed care cohort and 1.3% in the Medicaid cohort at ages 50 to 64 years (Figures 3 and 4). This indicates that BPDs are most prevalent among young to middle-aged adults, most likely because individuals in these age-groups have passed through the period of greatest risk for illness onset.^6,7,13,41,42 Possible explanations for a lower BPD prevalence rate among persons aged 50 to 64 years may be attributed to deaths in patients with bipolar, changes in healthcare coverage, or a reduction in healthcare service utilization.

The trends seen in the data (Figures 1-4) demonstrate that the monthly prevalence rate of BPDs was cyclical, starting near zero in the first month, then reaching a peak every December. Each January, the trough of the monthly BPD prevalence rate was the result of changing health enrollment in managed care plans for some patients and starting enrollment for many new managed care recipients. The variations in monthly BPD prevalence rates most likely do not reflect an underlying circular annual trend in BPD diagnoses or bipolar episodes. Cyclic trends are primarily an artifact of the cumulative method used to identify BPDs and changing enrollment on a calendar year cycle.

The prevalence rates of patients treated for a BPD in the Medicaid population were higher than those in the managed care population. The higher BPD prevalence in the Medicaid population may be related to the low socioeconomic status characteristic of Medicaid programs. These findings are consistent with 2 recent surveys, which indicate that the lifetime prevalence of BPD was highest among participants with the lowest annual household income.^14,15,43 Other community-based studies have also shown that economically disadvantaged persons have higher rates of mental disorders compared with persons of higher socioeconomic status.^44,45

Unlike the managed care population, the monthly BPD prevalence rate in the Medicaid population declined in 2001 and 2002, because of a high attrition rate and a relatively short enrollment period. This decline was most likely a result of the increase in the total number of Medicaid managed care enrollees from 700,000 in July 2001 to 1.4 million in June 2002, as many state Medicaid programs transferred their beneficiaries into MCOs. This dramatic increase in the denominator of the BPD prevalence rate probably contributed to the lower prevalence rate seen during this period. The BPD prevalence rate likely returned to the plateau level when Medicaid enrollment became stable, as indicated in November and December 2002 in Figures 2 and 4.

There is an increasing prevalence of patients treated for BPD in Medicaid and MCO populations. Many possible factors contribute to this trend, such as patient education about mental illness, nationwide direct-to-consumer advertisement, and wide-spectrum mental health coverage in Medicaid and MCOs. Because of the nature of managed care enrollment policy, we were unable to observe the true plateau level of BPD prevalence. Instead, we obtained the plateau points of BPD prevalence rates. In this study, the BPD prevalence rate at plateau points for adults aged 18 to 64 years ranged from 0.6% to 2.1%.

The prevalence of treated BPDs (0.31%-0.47%) fell within the range of published rates, including those from the ECA study and NCS.^8-10 However, prevalence rates in age-specific categories reported by Hirschfeld and col-leagues^14,43 and Das and colleagues¹⁵ are higher than our estimated prevalence rates. These studies screened adults for lifetime BPDs.

Our estimation of BPD prevalence was solely based on retrospective medical claims from patients who were diagnosed and treated during a 5-year period. We were unable to capture patients who were diagnosed with a BPD before or after the study period, as well as those patients who remained untreated. The prevalence rate of treated BPDs was higher for females (0.6%-0.7%) than for males (0.4%-0.5%) in both populations. This result is consistent with treated prevalence rates reported by Un√ºtzer and colleagues (females, 0.5%; males, 0.3%) in an MCO.⁹

Patients diagnosed with BPDs showed high proportions of psychiatric comorbidities, including anxiety disorders, personality disorders, and alcohol and substance abuse. Treated psychiatric comorbidity rates in the present study are at the lower end of published lifetime psychiatric comorbidity rates, based on survey questionnaire and interview data. For example, the occurrence of personality disorders ranged from 9% to 89%,^21,24,25 alcohol/substance abuse ranged from 20% to 60%,^14,15,26-29 and anxiety disorders ranged from 30% to 93%.^7,28 This may be best explained by the underdiagnosis or underrecognition of psychiatric comorbidities in general medicine clinics.

Das and colleagues reported that only 68% of patients with BPD and only 49% of patients with depressive disorders had a documented mental disorder or psychiatric symptom in their medical records.¹⁵ Hirschfeld and colleagues reported prevalence rates of 7.3% for diabetes, 15.3% for hypertension, and 2.2% for COPD⁴³; these rates are consistent with our observed rates (diabetes, 7.2%-7.4%; hypertension, 13.0%-18.2%; COPD, 3.3%- 4.0%). This finding suggests that components of metabolic syndrome are likely to be present in patients with a BPD.^38-41 Clinicians should consider managing these general medical conditions at the same time they are treating bipolar symptoms.

In a separate economic study using the same patient population, the bipolar treatment cost was considerable, with an average charge of $22,110 per-patient per-year or an $11,641 reimbursed amount per-patient per-year.³⁹ The total treatment cost was attributed to the cost related to treatment of BPD (30%) and its comorbid conditions (70%). These total treatment costs were significantly associated with key comorbidities, such as drug use disorder, cerebrovascular disease, ischemic heart disease, diabetes mellitus, and hypertension.³⁹

Limited literature is available about the prevalence of treated patients with BPDs and associated comorbidities in managed care populations. Whereas most managed care health plans and state Medicaid programs cover alarge proportion of US patients with mental illnesses, the findings of this present study provide useful information for health and drug benefit policy, as well as for planning disease management strategies.

Limitations
Because the Medicaid population in this study was limited to those enrolled in MCOs, these findings may not be generalizable to the fee-for-service Medicaid population. Because of the limitations of medical claim databases, the diagnostic information extracted may not be always reliable or valid. Undercoding and/or overcoding can occur in diagnosis files. Undercoding can occur, because insurance claim forms limit the number of diagnoses that can be documented, or providers simply do not make the effort to record secondary diagnoses.

Furthermore, because of the retrospective nature of a claims database review, it was not possible to verify the accuracy of the ICD-9 codes that physicians used or to identify possible misclassification of bipolar I and II disorders, because of illness progression. It is possible that this study underestimated prevalence rates because of the following reasons:

1. The limited time window of managed-care enrollment, changes in patient Medicaid eligibility, and mental health services through community clinics not billed to Medicaid
2. The difficulty of recognizing hypomania in some patients¹³
3. The exclusion of some patients with depression only (ICD-9 codes 296.2 and 296.3), which may cause the present study to underestimate patients with ostensible unipolar depression.⁴⁶

It is difficult to assess the lifetime prevalence rate of BPDs or the long-term prognosis for patients with a BPD diagnosis using a medical claims database. If a patient with a BPD had no claim for treatment during the study period, that individual would not have been detected as a patient with lifetime BPD. Comorbid conditions were identified by diagnostic codes, without considering the combination of medications for diabetes, obesity, hypertension, and other diseases.

Conclusion
Prevalence rates of BPDs were higher in Medicaid compared with the MCO population. The prevalence of patients treated for a BPD is increasing in the US MCO and Medicaid populations. Our study shows that BPD patients have a host of comorbid conditions. Clinicians should consider managing these general medical comorbidities at the same time they are treating symptoms of BPD. It is important to manage these conditions when treating patients with a BPD, because comorbid psychiatric and medical conditions are associated with increased morbidity and mortality.

Acknowledgments
This project was conducted with an educational grant provided by the Bristol-Myers Squibb Pharmaceutical Research Institute. The opinions and conclusions expressed in this manuscript are solely those of the authors. The authors would like to thank Dongming Jiang, PhD, for his programming and statistical data analysis. We also appreciate Raymond Jang, PhD, Patricia K. Corey-Lisle, PhD, and William Carson, MD, for their support and expertise.

This study was presented at the Collegium Internationale Neuro-Psychopharmacologicum Biannual Meeting, Paris, France, June 20-24, 2004, and at the American Psychiatry Association Annual Meeting, Atlanta, GA, May 24-26, 2005.

Disclosure Statement
Dr Guo receives research grants from Baxter Healthcare, Bristol- Myers Squibb, Eli Lilly, Johnson & Johnson, and Novartis; Dr Patel has been a consultant for AstraZeneca, Johnson & Johnson, Merck, and Pfizer; Dr Li is an employee of Bristol-Myers Squibb; and Dr Keck receives research grants from Abbott, AstraZeneca, Bristol- Myers Squibb, Eli Lilly, GlaxoSmithKline, Janssen, Ortho-McNeil, Pfizer, and Shire.

References

1. Gitlin MJ, Swendsen J, Heller TL, Hammen C. Relapse and impairment in bipolar disorder. Am J Psychiatry. 1995;152:1635-1640.
2. O'Connell RA, Mayo JA, Flatow L, et al. Outcome of bipolar disorder on longterm treatment with lithium. Br J Psychiatry. 1991;159:123-129.
3. National Alliance on Mental Illness (NAMI). Understanding bipolar disorder and recovery: what you need to know about this medical illness. August 2008. www.nami.org/Content/ContentGroups/Helpline1/NAMI_Bipolar_Disorder_Aug08.pdf. Accessed June 30, 2007.
4. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, DSM-IV. Washington, DC: American Psychiatric Association; 1994.
5. Weissman MM, Bland RC, Canino GJ, et al. Cross-national epidemiology of major depression and bipolar disorder. JAMA. 1996;276:293-299.
6. Weissman MM, Bruce ML, Leaf PJ, et al. Affective disorders. In: Robins LN, Regier DA, eds. Psychiatric Disorders in America: The Epidemiologic Catchment Area Study. New York, NY: Free Press; 1990:53-80.
7. Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Arch Gen Psychiatry. 1994;51:8-19.
8. Johnson RE, McFarland BH. Treated prevalence rates of severe mental illness among HMO members. Hosp Comm Psychiatry. 1994;45:919-924.
9. Unützer J, Simon G, Pabiniak C, et al. The treated prevalence of bipolar disorder in a large staff-model HMO. Psychiatr Serv. 1998;49:1072-1078.
10. Bebbington P, Ramana R. The epidemiology of bipolar affective disorder. Soc Psychiatry Psychiatr Epidemiol. 1995;30:279-292.
11. Akiskal HS, Bourgeois ML, Angst J, et al. Re-evaluating the prevalence of and diagnostic composition within the broad clinical spectrum of bipolar disorders. J Affect Disord. 2000;59(suppl 1):S5-S30.
12. Hirschfeld RM. Bipolar spectrum disorder: improving its recognition and diagnosis. J Clin Psychiatry. 2001;62(suppl 14):5-9.
13. Angst J. The emerging epidemiology of hypomania and bipolar II disorder. J Affect Disord. 1998;50:143-151.
14. Hirschfeld RM, Calabrese JR, Weissman MM, et al. Screening for bipolar disorder in the community. J Clin Psychiatry. 2003;64:53-59.
15. Das AK, Olfson M, Gameroff MJ, et al. Screening for bipolar disorder in a primary care practice. JAMA. 2005;293:956-963.
16. Keck PE Jr, McElroy SL, Strakowski SM, et al. 12-month outcome of patients with bipolar disorder following hospitalization for a major or mixed episode. Am J Psychiatry. 1998;155:646-652.
17. Keck PE, McElroy SL. Bipolar disorder, obesity, and pharmacotherapy-associated weight gain. J Clin Psychiatry. 2003;64:1426-1435.
18. Hilty DM, Brady KT, Hales RE. A review of bipolar disorder among adults. Psychiatr Serv. 1999;50:201-213.
19. Wozniak J, Biederman J, Monuteaux MC, et al. Parsing the comorbidity between bipolar disorder and anxiety disorders: a familial risk analysis. J Child Adolesc Psychopharmacol. 2002;12:101-111.
20. Goodwin RD, Hoven CW. Bipolar-panic comorbidity in the general population: prevalence and associated morbidity. J Affect Disord. 2002;70:27-33.
21. George EL, Miklowitz DJ, Richards JA, et al. The comorbidity of bipolar disorder and axis II personality disorders: prevalence and clinical correlates. Bipolar Disord. 2003;5:115-122.
22. Frank E, Cyranowski JM, Rucci P, et al. Clinical significance of lifetime panic spectrum symptoms in the treatment of patients with bipolar I disorder. Arch Gen Psychiatry. 2002;59:905-911.
23. McElroy SL, Altshuler LL, Suppes T, et al. Axis I psychiatric comorbidity and its relationship to historical illness variables in 288 patients with bipolar disorder. Am J Psychiatry. 2001;158:420-426.
24. Koenigsberg HW, Kaplan RD, Gilmore MM, Cooper AM. The relationship between syndrome and personality disorder in DSM-III: experience with 2,462 patients. Am J Psychiatry. 1985;142:207-212.
25. Turley B, Bates GW, Edwards J, Jackson HJ. MCMI-II personality disorders in recent-onset bipolar disorders. J Clin Psychol. 1992;48:320-329.
26. Escamilla MA, Batki S, Reus VI, et al. Comorbidity of bipolar disorder and substance abuse in Costa Rica: pedigree- and population-based studies. J Affect Disord. 2002;71:71-83.
27. Feinman JA, Dunner DL. The effect of alcoholism and substance abuse on the course of bipolar affective disorder. J Affect Disord. 1996;37:43-49.
28. Winokur G, Coryell W, Akiskal HS, et al. Alcoholism in manic depressive (bipolar) illness: familial illness, course of illness, and the primary-secondary distinction. Am J Psychiatry. 1995;152:365-372.
29. Brady KT, Lydiard RB. Bipolar affective disorder and substance abuse. J Clin Psychopharmacol. 1992;12(suppl 1):17S-22S.
30. Dunn J, Ferri CP. Epidemiological methods for research with drug misusers: review of methods for studying prevalence and morbidity. Rev Saude Publica. 1999;33: 206-215.
31. Hantouche EG, Akiskal HS, Lancrenon S, et al. Systematic clinical methodology for validating bipolar-II disorder: data in mid-stream from a French national multisite study (EPIDEP). J Affect Disord. 1998;50:163-173.
32. Li J, McCombs JS, Stimmel GL. Cost of treating bipolar disorder in the California Medicaid (Medi-Cal) program. J Affect Disord. 2002;71:131-139.
33. Stender M, Bryant-Comstock L, Phillips S. Medical resource use among patients treated for bipolar disorder: a retrospective, cross-sectional, descriptive analysis. Clin Ther. 2002;24:1668-1676.
34. Heisler M, DeMonner SM, Billi JE, Hayward RA. Medicaid managed care: are academic medical centers penalized by attracting patients with high-cost conditions? Am J Manag Care. 2003;9:19-29.
35. PharMetrics. PharMetrics Patient-Centric Database: a large medical claims data source in US. Watertown, MA, 2004. www.pharmetrics.com. Accessed August 1, 2006.
36. Walkup J, Crystal S, Sambamoorthi U. Schizophrenia and major affective disorder among Medicaid recipients with HIV/AIDS in New Jersey. Am J Public Health. 1999;89:1101-1103.
37. Mullins CD, Cooke JL Jr, Wang J, et al. Disparities in prevalence rates for lung, colorectal, breast, and prostate cancers in Medicaid. J Natl Med Assoc. 2004;96:809-816.
38. Guo JJ, Keck PE Jr, Corey-Lisle PK, et al. Diabetes mellitus associated with atypical antipsychotic use among patients with bipolar disorder: a retrospective, population- based, case-control study. J Clin Psychiatry. 2006;67:1055-1061.
39. Guo JJ, Keck PE, Li H, Patel NC. Treatment costs related to bipolar disorder and comorbid conditions among Medicaid patients with bipolar disorder. Psychiatr Serv. 2007;58:1073-1078.
40. Guo JJ, Keck PE, Li H, Corey-Lisle PK, et al. Risk of diabetes mellitus associated with atypical antipsychotic use among Medicaid patients with bipolar disorder: a nested case-control study. Pharmacotherapy. 2007;27:27-35.
41. Loranger AW, Levine PM. Age at onset of bipolar affective illness. Arch Gen Psychiatry. 1978;35:1345-1348.
42. Burke KC, Burke JD Jr, Regier DA, Rae DS. Age at onset of selected mental disorders in five community populations. Arch Gen Psychiatry. 1990;47:511-518.
43. Hirschfeld RM, Holzer C, Calabrese JR, et al. Validity of the mood disorder questionnaire: a general population study. Am J Psychiatry. 2003;160:178-180.
44. Dohrenwend BP, Levav I, Shrout PE, et al. Socioeconomic status and psychiatric disorders: the causation-selection issue. Science. 1992;255:946-952.
45. Regier DA, Farmer ME, Rae DS, et al. One-month prevalence of mental disorders in the United States and sociodemographic characteristics: the Epidemiologic Catchment Area Study. Acta Psychiatr Scand. 1993;88:35-47.
46. Benazzi F. Underdiagnosis of bipolar II disorders in the community. J Clin Psychiatry. 2003;64:1130-1131.