Within the Medicare population, the majority of beneficiaries who have a psychiatric diagnosis also have one or more medical diagnoses. Use of multiple medications is common in this population. In particular, the number of unique medications and number of unique medication groups is greatest for those with both medical and psychiatric diagnoses. This pattern is seen at all Part D benefit phases.

METHODS

We used Medicare enrollment data, 100 percent FFS Parts A and B institutional/ noninstitutional claims data, and 100 percent Part D claims data for reference years 2009 through 2011 from the CMS Chronic Conditions Warehouse (CCW). We identified Medicare beneficiaries who had Part A and Part B Medicare coverage with no Medicare Advantage (i.e., managed care enrollment) during their period of eligibility. We determined whether they had a psychiatric or a selected medical diagnosis during the reference year.

Psychiatric diagnoses, identified by International Classification of Diseases-9th edition (ICD-9) codes, included: Schizophrenia (295.x), Bipolar disorders (296.0-296.1, 296.4-296.8, 300.13), Major depression (296.2x,296.3x), Dysthymia (300.4), Other depression (298.0, 309.0-309.1, 311), Generalized anxiety disorder (300.2), Panic disorder with or without agoraphobia (300.01, 300.21), Post-traumatic stress disorder (PTSD, 309.81), Obsessive-compulsive disorder (OCD, 300.3), Social phobia (300.23), Other anxiety (300.00, 300.09), Delusional disorders (297.x), Other non-organic psychoses (298.1-298.9), and Other psychiatric conditions (e.g., personality disorders and any other disorders excluding developmental, organic, or substance use disorders; ICD-9 codes: 296.90, 296.99, 300.1, 300.20, 300.22, 300.29, 300.5-300.9, 301, 307, 309 [except 309.81-PTSD, 312.0-312.4, 312.89-312.9]). Medical diagnoses analyzed in this report include: Alzheimer's disease, Dementia, Cancers (Colorectal, Endometrial, Female breast, Lung, and Prostate), Chronic kidney disease (CKD), Chronic obstructive pulmonary disease (COPD), Diabetes, Heart Failure, Ischemic heart disease (IHD), Arthritis, and Stroke or Transient ischemic attack (TIA).

The medical conditions were identified using the CCW condition definitions.²⁰

We assessed trends in diagnosis and medication use across years and demographic factors, including age and gender (identified in the beneficiary summary file during the year the procedure took place), as well as race/ethnicity, region, and urbanicity as defined below.

Race/ethnicity: We defined race and ethnicity using the Research Triangle Institute Race Code, which applies a surname algorithm to assign Hispanic ethnicity.²¹

Age: We defined age categories using the age of the beneficiary at the end of each reference year.

Urban/Rural: We used the Core Based Statistical Area (CBSA) of the beneficiary to identify the urban region in which the beneficiary resides. We categorized beneficiaries who do not reside in a CBSA as rural.²²

Dual Status: We used CMS' established algorithm for defining the annual dual eligibility status of each beneficiary using the monthly State Reported Dual Eligibility Status Codes. The algorithm categorizes beneficiaries into four groups: Full Duals had full Medicare and Medicaid coverage (including prescription drugs) during the most recent month of dual eligibility for the reference year; QMB beneficiaries had Medicaid and participated in the Qualified Medicare Beneficiary Program during the most recent month of dual eligibility for the reference year; and Other/Partial Duals had Medicaid and participated in the Specified Low-Income Medicare Beneficiary Program, the Qualifying Individual Program, or the Qualified Disabled and Working Individuals Program during the most recent month of dual eligibility for the reference year.

Nonduals had Medicare coverage only for the reference year.²³

Part D Status: We categorized Part D status into three groups. All Part D indicates continuous Part D coverage during the reference year for all months of Medicare eligibility (i.e., months where the beneficiary was alive and covered by Medicare). Any Part D indicates at least one month of continuous or noncontinuous Part D coverage during the reference year. No Part D indicates no Part D coverage during the reference year. We identified the number of coverage months using the monthly Part D plan contract identifier.

Part D Coverage Phase: We identified the number of days each beneficiary spent in the deductible, pre-initial coverage limit (pre-ICL), initial coverage limit (ICL), and catastrophic coverage phase using the “Benefit Phase” indicator variable included in the CCW Part D Event files.²⁴ We calculated the number of days in each stage using the dates each stage went into effect.

We used the Medi-Span classification system to classify a variety of aspects of prescribed medications.²⁵ The system consoli-dates National Drug Codes (NDCs) into a single drug title, with a total of 2,523 drugs in the classification system. Drugs are consolidated into 95 drug groups, and drug groups are consolidated into 18 drug categories (See Appendix A).

Antipsychotic use: We considered beneficiaries to be using antipsychotic or antimanic drugs if they used any drug in the Medi-Span drug group “Anti-psychotic/Antimanic Agents.” We used the indicator to find individuals who received these medications but who did not have a psychiatric diagnosis captured in our analysis, as undiagnosed mental conditions are common among Medicare beneficiaries.

Antipsychotics are one of the groups of critical medications covered by Part D.¹⁶ We did not conduct a parallel analysis with antidepressants (another critical group) because they have nonpsychiatric indications such as pain or insomnia.^26-28

Generics and brands: We used the “Brand/Generic” indicator variable developed by CMS to classify each drug as name brand or generic. (Note: This indicator is a CMS internal variable not generally released in files sent to external researchers.)

Unique prescribers: We counted unique prescribers using the “CCW_PRSCRBR_ID” variable included in the CCW Part D Event files.²⁴

RESULTS

The distribution of psychiatric and medical diagnoses among the approximately 33 million Medicare beneficiaries from 2009 to 2011 is shown in Table 1. Approximately 30 percent of the population had no psychiatric or medical diagnosis captured by our measures, and approximately 50 percent had chronic medical diagnoses only (though this percentage decreased somewhat from 49.3% in 2009 to 47% in 2011). About 17 to 18 percent of beneficiaries had both psychiatric and medical diagnoses, and 5 percent had a psychiatric diagnosis only.

Table 1

Percentage of total Medicare FFS eligibles by diagnosis type, 2009-2011.

The population of 33.7 million Medicare beneficiaries with full Part A and B fee-for-service (A and B FFS) coverage in 2011 overall and by psychiatric and medical morbidity categories is described in Tables 2 and 3 (overall totals and proportions shown in Tables 2 and 3, population numbers shown in Appendix B). Results from 2009 and 2010 show no significant variation from 2011, so we present only results for 2011 here.

Table 2

Number and percent distribution of Medicare A and B FFS beneficiaries by demography and diagnosis type, 2011.

Table 3

Number and percent distribution of Medicare A and B FFS beneficiaries by diagnosis type and comorbidity, 2011.

Among age groups, disability-eligible individuals (age <65) were much more likely to have psychiatric diagnoses and were more likely to only have psychiatric conditions (Table 2, Figure 1). Among racial and ethnic groups, Asian/Pacific Islanders had notably lower percentages of psychiatric diagnoses only or psychiatric diagnoses with medical comorbidity than other groups. Relatively high psychiatric and/or psychiatric-medical morbidity was evident among females; beneficiaries with dual Medicaid and Medicare status; and beneficiaries with Part D coverage (Table 2).

The most common of the specific psychiatric diagnoses we investigate were other depression (3.7 million beneficiaries), other anxiety (2.7 million), and major depression (1.4 million; Table 3). The most common chronic medical diagnoses within this population were ischemic heart disease (IHD, 9.7 million), arthritis (9.4 million), and diabetes (9 million). One-quarter to one-half of those with chronic medical diagnoses were also diagnosed with a psychiatric condition (Table 3). A majority of individuals diagnosed with each psychiatric condition had psychiatric-medical comorbidity, ranging from 50% of beneficiaries diagnosed with social phobia who also have a medical comorbidity to 90% of beneficiaries with other non-organic psychoses who have a medical comorbidity (Table 3).

The population of 17.6 million Medicare beneficiaries who had Part D coverage for all months of Medicare enrollment in 2011 is described in Tables 4 and 5 (overall totals and proportions shown in Tables 4 and 5, population numbers shown in Appendix C). Among those with Part D coverage, the largest morbidity category overall was medical only (48%).

Table 4

Number and percent distribution of Medicare Parts A, B, and D FFS beneficiaries by demography and diagnosis type, 2011.

Table 5

Number and percent distribution of Medicare Parts A, B, and D FFS beneficiaries by diagnosis type and comorbidity, 2011.

However, the percentage with no psychiatric or medical diagnosis was nearly equal to that with psychiatric-medical comorbidity (approximately 23% for both). A greater percentage of Part D beneficiaries (6.4%) also had psychiatric-only diagnoses compared to the overall A and B FFS group (5%). Other than often higher percentages for morbidity groups, most other patterns mirrored those in the overall A and B FFS population (Table 5).

Slightly over 50 percent of Medicare A and B FFS beneficiaries had Part D coverage for all months of Medicare eligibility in 2011, and a greater percentage of beneficiaries with a psychiatric diagnosis (67%) had Part D coverage for all eligible months than other diagnosis groups (Figure 2).

Table 6 displays co-occurrence of psychiatric condition categories with chronic medical conditions and categories among the whole A and B FFS population. The most common combinations of psychiatric and medical conditions were among depressive and/or anxiety disorders linked with cardiovascular and arthritis conditions. The most common combinations experienced in this population were depression-cardiovascular (2.2 million beneficiaries); depression-arthritis (2.1 million); depression-diabetes (1.6 million); anxiety-cardiovascular (1.5 million); and anxiety-arthritis (1.4 million). Depressive disorders, anxiety disorders, and the combination of both a depressive and anxiety condition were the most prevalent categories of psychiatric conditions, and generally had higher numbers of beneficiaries with concurrent medical conditions than other psychiatric conditions. The same analysis among beneficiaries with Parts A, B, and D coverage for all eligible months is shown in Table 7. The overall patterns were consistent between the Part D and whole Parts A and B populations.

Table 6

Number and percentage of Medicare A and B FFS beneficiaries by concurrent psychiatric and medical diagnosis category, 2011.

Table 7

Number and percentage of Medicare Parts A, B, and D FFS beneficiaries by concurrent psychiatric and medical diagnosis category, 2011.

The most common combinations for the population with Parts A, B, and D coverage were depression-cardiovascular (1.5 million beneficiaries); depression-arthritis (1.4 million); depression-diabetes (1.1 million); anxiety-cardiovascular (1.0 million); and anxiety-arthritis (0.9 million; Table 7).

An analysis of the time in, time to reach, and percentage of beneficiaries reaching each Part D benefit phase for A and B FFS beneficiaries with Part D coverage for all eligible months is shown in Table 8. Within this population, all diagnosis groups spent the longest mean time in the pre-ICL (pre-coverage gap) phase. Beneficiaries with psychiatric, chronic medical, or psychiatric-medical conditions spent fewer days in the deductible and pre-ICL phases and reached these phases more quickly than beneficiaries with no diagnosed conditions. Those with psychiatric-medical comorbidity had the shortest time in deductible and pre-ICL phases with 63 and 217 mean days in phase compared to 163 and 263 days for the no-diagnosis group. Conversely, those with psychiatric-medical comorbidity had the longest mean time in the ICL (coverage gap) phase (127 days), just slightly longer than the medical diagnosis group (124 days). While those with none of the measured diagnoses had the most time in the catastrophic phase (163 days), followed by those with psychiatric diagnoses (152 days), and those with psychiatric-medical diagnoses (140 days), they were also the least likely to reach this phase.

Table 8

Number of days to reach and number of days in Part D benefit phase among Medicare parts A, B, and D FFS beneficiaries by diagnosis category.

Only 2.4 percent of beneficiaries with none of our selected diagnoses reached the catastrophic stage, while 17.2 percent of beneficiaries with psychiatric-medical diagnoses reached this stage (Table 8).

The analysis of mean days to reaching each benefit phase shows that the psychiatric-only, medical-only, and psychiatric-medical groups reached the deductible, pre-ICL, and ICL phases more quickly than the no-diagnosis group. However, possibly owing to the shorter time spent by the no-diagnosis group in the ICL coverage gap, this group moved from the ICL to the catastrophic stage the most quickly, with 196 days to reach the catastrophic stage compared to between 210 and 237 days for the other diagnosis groups (Table 8).

Table 9 shows the number of medications, groups, prescribers, and percentage of prescription claims for generic drugs in each benefit phase by diagnosis category for the population with Parts A, B, and D coverage for all eligible months in 2011. Claims were submitted for the highest average number of unique medications during the pre-ICL phase for all diagnostic categories, reflecting the longest time in phase for pre-ICL (7.3 medications prescribed in 4.1 drug groups by 2.7 prescribers for the overall population). In almost all phases, the psychiatric-medical comorbidity group had the highest mean unique medications (10.3 in the pre-ICL phase and 5.8 in the ICL phase); the highest mean number of drug groups (5.6 in the pre-ICL and 3.1 in the ICL); and the highest number of prescribers (3.5 in pre-ICL and 1.8 in the ICL).

Table 9

Prescription-related claims among Medicare parts A, B, and D FFS beneficiaries by diagnosis category, 2011.

The only exception was a mean of 1.2 prescribers in the deductible phase for the psychiatric-only group versus 1.0 in the psychiatric-medical group. In the same phase, the psychiatric-only group's mean medication count (2.3) was also close to that of the psychiatric-medical category (2.5; Table 9).

All diagnostic categories' percent generic fills decreased across coverage phases (Table 9). All diagnostic categories had the highest percentage of generic prescription fills in the deductible phase, with the total population having 91 percent of prescriptions filled as generics. The pre-ICL phase showed similar trends to the deductible phase, with all categories having approximately 80 percent of fills as generics. In the ICL (coverage gap) phase, the percentage of generic fills was higher among the psychiatric-medical (72%) and psychiatric-only (69%) groups and lower among medical-only (67%) and no-diagnosis (62%) groups. This pattern continued in the catastrophic phase, with a higher percentage of generic fills for psychiatric diagnoses alone (63%) or with chronic medical comorbidities (66%) than for medical-only (61%) or no-diagnosis groups (56%; Table 9).

DISCUSSION

Our work based on an examination of the full population of beneficiaries in the Medicare fee-for-service system reveals important patterns about the prevalence of psychiatric and medical diagnoses in this group. Psychiatric diagnoses are prevalent and variable with age, gender, and race. Perhaps more importantly, a majority of persons with psychiatric diagnoses also have co-occurring chronic diseases: 77 percent of beneficiaries with one psychiatric diagnosis and 80 percent of beneficiaries with two or more psychiatric diagnoses also have a diagnosis of a chronic medical condition (Table 3). The most prevalent condition combinations generally occurred among common psychiatric and medical condition categories, particularly depression and/or anxiety disorders, cardiovascular-related conditions, and arthritis. Our estimate of the co-occurrence of medical and psychiatric conditions includes nine common chronic medical conditions. This may underestimate the full burden of psychiatric and medical comorbidities in this population, since many types of cancer and less common comorbidities such as liver disease, anemia, HIV/AIDS, dyslipidemia, and others are not counted.

While coverage phase patterns varied, individuals with psychiatric-medical comorbidity generally spent the fewest days in the early Part D coverage phases, but more days in the ICL/coverage gap phase than any other group.

They moved through benefit phases more quickly than most other diagnosis groups, and 17% of beneficiaries with psychiatric-medical comorbidity reached the final catastrophic benefit phase, a much higher proportion than among the other diagnosis categories (Table 8). Beneficiaries with psychiatric-medical conditions also were prescribed the greatest number of unique medications across more groups by more prescribers than other diagnosis groups in nearly every benefit phase (Table 9). Along with research cited previously, these results suggest that the joint occurrence of medical and psychiatric diagnoses increases the number of involved health care providers, and may increase the risk of poorer health care outcomes and problems related to polypharmacy.

This analysis also compared the population with Parts A and B FFS Medicare coverage with the population with Parts A and B FFS coverage that also received Part D coverage during their months in Medicare. The Medicare Part D benefit provides both an opportunity and a challenge to studying pharmacy use among beneficiaries.

Roughly 50 percent of Medicare beneficiaries with A and B FFS coverage received pharmacy benefits through the Part D program for all months of Medicare eligibility in 2011 (Figure 2). However, beneficiary profiles differ between the full population and the population receiving Part D benefits. For example, about 16 percent of the entire Parts A and B FFS Medicare population are designated as “Full Duals” receiving both Medicare and Medicaid coverage (Table 2).

When we examine the population receiving Part D benefits, the proportion of Full Duals jumps to 30 percent, indicating that those covered by Part D may also be a more vulnerable population (Table 4). There is also a greater burden of psychiatric and psychiatric-medical comorbid conditions among the Part D population. For example, looking at the burden of psychiatric diagnoses within each population, we find that 9.1% of the whole A and B FFS population compared to 12.3% of the population with Parts A, B, and D coverage have two or more psychiatric diagnoses (calculations based on Tables 3 and 5). Thus, for the purposes of understanding psychiatric diagnoses in the Medicare program, Part D data may be very useful. However, it is important to note that Part D comparison populations who do not have psychiatric diagnoses are likely sicker than the average Medicare population.

When interpreting findings related to time until, time in, and percentage of benefi-ciaries reaching Part D benefit phases, it is important to note that many plans under Part D do not include a deductible phase, meaning some people have no days in this phase before reaching the pre-ICL phase.²⁹ Many plans also may have varying cost sharing, covered drugs, premiums, and related policies beyond minimum “standard” or actuarially equivalent coverage, which is also affected by low income status or dual eligibility with Medicaid.¹⁶

Additionally, any effect of the ICL/coverage gap phase on patterns shown here is subject to change as the Federal Government phases out this gap through 2020.³⁰ This suggests a need for further study on the impact of variations in coverage and costs to beneficiaries and approaching changes in Part D under the Affordable Care Act on patients with different health condition profiles.

In 2011, more than 5 million beneficiaries with psychiatric diagnoses received benefits through Medicare's FFS and Part D pharmacy programs (Table 4). While this accounts for approximately one-quarter of beneficiaries with this type of coverage, the sheer size of this population deserves attention. For example, there are 950,000 beneficiaries with a diagnosis of major depression, 2.4 million with other depression, 1.5 million with cardiovascular-depression comorbidity, and 1.1 million with diabetes-depression comorbidity in this population (Tables 5 and 7). Although most beneficiaries in Medicare are not represented in this population, this group still represents a large number of beneficiaries who are experiencing complicated conditions. Thus, understanding the joint role of psychiatric and medical diag-nosis within this population is important to maintaining the health of the Medicare population.

In spite of the strengths of this analysis, some limitations must be acknowledged. First, as we have previously noted, we include analysis of a limited number of chronic medical conditions only, which may underestimate the full extent of the psychiatric-medical burden of comorbidity in the Medicare population. Second, we present unadjusted numbers and do not conduct statistical testing of differences in proportions. However, our analysis is based on a 100 percent sample of Medicare FFS beneficiaries for all study years, the full population of interest.

With the large size of the full Medicare FFS population, it is likely that even small, unimportant trends would be measured as being statistically different.³¹ Therefore, we chose to omit statistical testing and leave conclusions about the importance of patterns to the reader. Despite these limitations, we believe that the use of the full population supports the strength of our conclusions, and that the results of this inquiry provide important insights about medication use among beneficiaries with medical and psychiatric conditions over time and across demographic groups.

CONCLUSION

Within the Medicare population, the majority of beneficiaries who have a psychiatric diagnosis also have one or more medical diagnoses. While the beneficiaries who suffer from psychiatric or combined psychiatric-medical conditions are a minority of all Medicare beneficiaries, they still represent a large population with high rates of unique drug and drug group use and catastrophic coverage use. The joint occurrence of medical and psychiatric diagnoses places these beneficiaries at risk for poorer health outcomes and increases the risk of problems related to polypharmacy. This group accounts for 6 million beneficiaries in the Medicare Parts A and B FFS population, and is important to consider in future research and decisionmaking regarding best practices in medication and health care management.

REFERENCES

1.

Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Ann Intern Med. 2002;162(20):2269–76. [PubMed: 12418941]

2.

Centers for Medicare & Medicaid Services. Chronic conditions among Medicare beneficiaries, chartbook: 2012 edition. Baltimore, MD: Centers for Medicare & Medicaid Services; 2011.

3.

Niefeld MR, Braunstein JB, Wu AW, et al. Preventable hospitalization among elderly Medicare beneficiaries with type 2 diabetes. Diabetes Care. 2003;26(5):1344–9. [PubMed: 12716786]

4.

Laursen TM, Munk-Olsen T, Gasse C. Chronic somatic comorbidity and excess mortality due to natural causes in persons with schizophrenia or bipolar affective disorder. PLoS ONE. 2011;6:e24597. [PMC free article: PMC3173467] [PubMed: 21935426]

5.

Yoon J, Yano EM, Altman L, et al. Reducing costs of acute care for ambulatory care-sensitive medical conditions: the central roles of comorbid mental illness. Medical Care. 2012;50(8):705–13. [PubMed: 22437618]

6.

Unützer J, Patrick DL, Diehr P, et al. Quality adjusted life years in older adults with depressive symptoms and chronic medical disorders. Int Psychogeriatr. 2000;12(01):15–33. [PubMed: 10798451]

7.

Weiner M, Warren L, Fiedorowicz JG. Cardiovascular morbidity and mortality in bipolar disorder. Ann Clin Psych. 2011;23(1):40–7. [PMC free article: PMC3190964] [PubMed: 21318195]

8.

Hughes LD, McMurdo MET, Guthrie B. Guidelines for people not for diseases: the challenges of applying UK clinical guidelines to people with multimorbidity. Age Ageing. 2012;42(1):62–9. [PubMed: 22910303]

9.

Mojtabai R, Olfson M. National trends in psychotropic medication polypharmacy in office-based psychiatry. JAMA Psych. 2010;67(1):26–36. [PubMed: 20048220]

10.

Gandhi TK, Weingart SN, Borus J, et al. Adverse drug events in ambulatory care. New Engl J Med. 2003;348:1556–64. [PubMed: 12700376]

11.

Devlin JW, Mallow-Corbett S, Riker RR. Adverse drug events associated with the use of analgesics, sedatives, and antipsychotics in the intensive care unit. Crit Care Med. 2010;38:S231–43. [PubMed: 20502176]

12.

Stagnitti MN. Statistical brief #240. Rockville, MD: Agency for Healthcare Research and Quality; 2009. Prescription drug estimates for Medicare beneficiaries, 2005 and 2006.

13.

Congressional Budget Office. CBO's March 2008 baseline: Medicare. Washington, DC: Congressional Budget Office; 2008.

14.

Cheng LI, Rascati KL. Impact of Medicare Part D for Medicare-age adults with arthritis: prescription use, prescription expenditures, and medical spending from 2005 to 2008. Arthritis Care Res. 2012;64(9):1423–29. [PubMed: 22508663]

15.

Donohue JM, Zhang Y, Lave JR, et al. The Medicare drug benefit (Part D) and treatment of heart failure in older adults. Am Heart J. 2010;160(1):159–65. [PMC free article: PMC2898511] [PubMed: 20598987]

16.

O'Sullivan J. Medicare Part D primer. Washington, DC: Congressional Research Office; 2008.

17.

Huskamp HA, West JC, Rae DS, et al. Part D and dually eligible patients with mental illness: medication access problems and use of intensive services. Psych Serv. 2009;60(9):1169–74. [PMC free article: PMC2768558] [PubMed: 19723730]

18.

Schneeweiss S, Patrick AR, Pedan A, et al. The effect of Medicare Part D coverage on drug use and cost sharing among seniors without prior drug benefits. Health Aff. 2009;28(2):w305–16. [PMC free article: PMC2683877] [PubMed: 19189990]

19.

Zhang Y, Baik SH, Zhou L, et al. Effects of Medicare Part D coverage gap on medication and medical treatment among elderly beneficiaries with depression. Arch Gen Psychiatry. 2012;69(7):672–9. [PMC free article: PMC3390754] [PubMed: 22752233]

20.

Buccaneer, a General Dynamics Company. Chronic condition data warehouse: chronic condition categories. [February 13, 2013]. Available at: http://www​.ccwdata.org​/chronic-conditions/index.htm.

21.

Research Data Assistance Center. Research Triangle Institute (RTI) race code. [January 7, 2013]. Available at: http://www​.resdac.org​/cms-data/variables​/Research-Triangle-Institute-RTI-Race-Code.

22.

United States Census Bureau. Geographic terms and concepts – core based statistical areas and related statistical areas. [January 7, 2013]. Available at: http://www​.census.gov​/geo/www/2010census/gtc/gtc_cbsa.html.

23.

Centers for Medicare & Medicaid Services. Medicare savings programs. [November 2, 2012]. Available at: http://www​.medicare.gov​/your-medicare-costs​/help-paying-costs​/medicare-savings-program​/medicare-savings-programs.html.

24.

Buccaneer, a General Dynamics Company. Chronic conditions data warehouse: data dictionaries. [March 28, 2013]. Available at: http://www​.ccwdata.org​/web/guest/data-dictionaries.

25.

Wolters Kluwer Health. Medi-Span. [March 28, 2013]. Available at: http://www​.medispan.com/index.aspx.

26.

Petty DR, House A, Knapp P, et al. Prevalence, duration, and indications for prescribing of antidepressants in primary care. Age Ageing. 2006;35(5):523–26. [PubMed: 16690637]

27.

Patten SB, Esposito E, Carter B. Reasons for antidepressant prescriptions in Canada. Pharmacoepidem Dr S. 2007;16(7):746–52. [PubMed: 17286318]

28.

Mojtabai R, Olfson M. Proportion of antidepressants prescribed without a psychiatric diagnosis is growing. Health Aff. 2011;30(8):1434–42. [PubMed: 21821561]

29.

Centers for Medicare & Medicaid Services. Costs for Medicare drug coverage: yearly deductible for drug plans. [June 20, 2013]. Available at: http://www​.medicare.gov​/part-d/costs/deductible​/drug-plan-deductibles.html.

30.

Centers for Medicare & Medicaid Services. Closing the coverage gap. Baltimore, MD: Centers for Medicare & Medicaid Services; 2013. [June 20, 2013]. Available at: http://www​.cms.gov/Outreach-and-Education​/Outreach/Partnerships​/downloads/11522-P.pdf.

31.

Snedecor GW, Cochran WB. Statistical methods. 8th ed. Ames, IA: Iowa State University Press; 1989.

Acknowledgments

The authors wish to thank the DEcIDE mental health consortium for their leadership in establishing the scope and objectives of this report.

This project was funded under Contract No. HHSA29020100013I from the Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services, as part of the Developing Evidence to Inform Decisions about Effectiveness (DEcIDE) program. The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services. This project has been approved by the University of Minnesota Institutional Review Board.

APPENDICES

Suggested Citation: Shippee ND, Hall K, Zeglin J, O'Donnell B, Virnig BA. Medication use among Medicare beneficiaries with medical and psychiatric conditions. Medication Use in Medicare. Data Points # 19 (prepared by the University of Minnesota DEcIDE Center, under Contract No. HHSA29020100013I ). Rockville, MD: Agency for Healthcare Research and Quality. April 2014. AHRQ Publication No. 14-EHC028-EF.