In 2011, about 25 percent of the Medicare fee-for-service population had diabetes. Among Medicare beneficiaries with diabetes, approximately 14 percent had type 1, 85 percent had type 2 but did not use insulin, and less than 1 percent had type 2 diabetes and used insulin to manage their condition. Between 2007 and 2011, beneficiaries with type 2 diabetes who used insulin had the highest burden of comorbidity, hospitalization rates, and allowed payment, followed by those with type 1 diabetes. Most beneficiaries with diabetes had evaluation and management visits. Most also received needed preventive care, including HbA1c and LDL^* testing, and about half received an annual flu shot and eye exam. However, beneficiaries with type 2 diabetes using insulin had the lowest rates of receipt of preventive care. Most beneficiaries with diabetes visited both primary care and specialty providers. The number of providers with whom they had contact is high, indicating potential fragmentation in both primary and specialty care.

METHODS

Our analysis was performed on data from 100 percent of the Medicare beneficiaries represented in the Chronic Conditions Data Warehouse (CCW) data.¹² We identified Medicare beneficiaries who had diabetes between January 2007 and December 2011. (See Figures 1-4.) Beneficiaries were age 65 or older and had Part A and Part B Medicare coverage with no Medicare Advantage for their entire period of eligibility during the reference year.

We assessed trends in health care services provided to beneficiaries with diabetes type 1 or type 2 across demographic factors, including age and gender. These characteristics were identified in the beneficiary summary file during the reference year. We also assessed trends based on factors such as race, region, and urbanicity as defined below. (See Tables 1 and 2.)

Table 1

Number and percent distribution of Medicare FFS eligibles, by demography and type of diabetes, 2011.

Table 2

Number and type of comorbidities among Medicare FFS beneficiaries, by type of diabetes, 2011.

Diabetic type: Beneficiaries with diabetes were categorized as having diabetes type 1, diabetes type 2 with receipt of insulin, and diabetes type 2 without receipt of insulin. We used the CCW diabetes flags as the starting point for our classification. We limited our sample to persons who had at least one claim with a diagnosis of diabetes in the calendar year. Persons who had any ICD-9^* diagnosis of type 1 diabetes (250.01, 250.03, 250.11, 250.13, 250.21, 250.23, 250.31, 250.33, 250.41, 250.43, 250.51, 250.53, 250.61, 250.63, 250.71, 250.73, 250.81, 250.83, 250.91, and 250.93) were classified as type 1. All other persons with a CCW diabetes flag were considered to have type 2 diabetes. Persons with type 2 diabetes who had evidence of insulin use from HCPCS^† codes in either the Carrier or Durable Medical Equipment file (E0781, E0784, A4231, A4230, K0552, S5565, S5566, A4221, J1815, J1816, J1817, A4632, K0601, K0602, K0605, A4365, A5120, A4245, A4247, A6257, A6258, A4364, A4450, A9274) were considered insulin using. Thus, persons who were both type 2 diabetic and insulin using were defined as meeting this combination.

Race/ethnicity: Race and ethnicity were defined using the Research Triangle Institute Race Code, which applies a surname algorithm to assign Hispanic ethnicity.^13,14

Age: The age categories were defined 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. Beneficiaries who do not reside in a CBSA are considered to be rural.¹⁵

Dual status: CMS has an established algorithm for defining the annual dual eligibility status of each beneficiary, using the monthly State Reported Dual Eligibility Status Codes. We used these algorithms to categorize 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; Other/Partial Duals had Medicaid and participated in the Specified Low-Income Medicare Beneficiary Program during the most recent month of dual eligibility for the reference year, the Qualifying Individual Program, or the Qualified Disabled and Working Individuals Program; and Non-Duals had Medicare coverage only for the reference year.¹⁶

Comorbidities: We used the Chronic Condition Categories identified in the Medicare CCW to indicate presence of chronic conditions.¹⁷ Based on existing literature, we provide counts of comorbidities that were related to diabetes or unrelated to diabetes in their etiology and treatment.^18-21 Related conditions include acute myocardial infarction, chronic kidney disease, heart failure, ischemic heart disease, cataract, glaucoma, stroke, and atrial fibrillation. Unrelated conditions include Alzheimer's disease, related disorders, or dementia; chronic obstructive pulmonary disease; depression; hip/ pelvic fracture; osteoporosis; rheumatoid arthritis or osteoarthritis; and cancers (endometrial, breast, lung, and prostate).

Evaluation and management (E&M): E&M visits were defined using Berenson-Eggers type of service (BETOS) codes M1A or M1B for new or established E&M office visits.²² E&M visits were limited to one per beneficiary per day. (See Table 3.)

Table 3

Number and percentage of evaluation and management visits among Medicare FFS beneficiaries, by type of diabetes, 2011.

Preventive care: Preventive care services were identified in the Part B institutional and noninstitutional claims using CPT codes 90724, 90658, 90659, G0008, and 90656 to identify flu shots, and CPT codes 67028, 67030, 67031, 67036, 67038, 67039, 67040, 67101, 67105, 67107, 67108, 67110, 67112, 67121, 67141, 67145, 67208, 67210, 67218, 67220, 67221, 67227, 67228, 92002, 92004, 92012, 92014, 92018, 92019, 92225, 92226, 92230, 92235, 92240, 92250, and 92260 to identify eye exams. CPT codes 83036, 83037, 3044F, 3045F, 3046F, and 3047F were used to identify HbA1c tests, and CPT codes 80061, 83700, 83701, 83704, 83715, 83716, 83721, 3048F, 3049F, and 3050F were used to identify lipid tests. (See Table 4.)

Table 4

Percentage of annual preventive care services received among Medicare FFS beneficiaries, by type of diabetes, 2011.

Primary care visits: Primary care visits were defined as a visit to a primary care physician specialty by HCFA^* specialty codes as follows: general practitioner = 01, family practitioner = 08, internal medicine = 11, and geriatric medicine = 38.²³ Visits were limited to one per beneficiary/ provider pair per day. Individual providers were identified by National Provider Identifier (NPI).²⁴ (See Table 5.)

Table 5

Number and percentage of primary care visits and number of primary care providers seen annually among Medicare FFS beneficiaries, by type of diabetes, 2011.

Specialty care visits: Specialty care visits were defined as a visit to a non-primary care physician specialty by HCFA specialty codes as follows: podiatry = 48, optometry = 41, ophthalmology = 18, and endocrinology = 46. Visits were limited to one per beneficiary/provider pair per day. Individual providers were identified by NPI. (See Table 6.)

Table 6

Number and percentage of specialty care visits among Medicare FFS beneficiaries, by type of diabetes and specialty type, 2011.

Hospitalizations: Hospitalizations are defined as acute care stays. Hospitalizations with procedure are based on Medicare Severity Diagnosis Related Group (MS-DRG) type of SURG.

Emergency room visits: Inpatient emergency room (ER) visits are defined as an acute care stay with a 0450-0459 revenue center code.²⁵ Inpatient ER visits are limited to one per hospital stay per patient. Outpatient visits are defined as a claim type of 40 (outpatient) with a 0450-0459 revenue center code. Multiple outpatient ER visits by patient by date are included.

RESULTS

Among Medicare fee-for-service (FFS) beneficiaries, about 25 percent have diabetes. The percentage of all Medicare beneficiaries with diabetes varies by State. In 2011, New York had the highest percentage of beneficiaries with diabetes at 30 percent, followed by New Jersey and Michigan (at 29.9% and 28.9%, respectively; Figure 1). Colorado, Montana, and Wyoming had the lowest percentages (17%, 17.5%, and 18%). Puerto Rico, not pictured in Figure 1, had a higher percentage than any State at 42.3 percent.

Figure 1

Medicare FFS eligibles with diabetes, by State, 2011.

Among the population of Medicare beneficiaries who have diabetes, the distribution of diabetic types also varied by State. In the Medicare diabetic population in 2011, Alaska had the highest percentage classified as type 2 diabetes not using insulin, followed by Hawaii and Minnesota (92.9%, 92.2%, and 91.6%; Figure 2). The District of Columbia, New Jersey, and Rhode Island had the lowest percentages at 79.6, 80.6, and 81.2 percent, respectively.

Figure 2

Medicare FFS eligibles with type 2 diabetes not using insulin, by State, 2011.

From 2007 to 2011, the number of older Medicare beneficiaries who are classified as having diabetes rose from 6,599,119 to 7,031,644 (Figure 3).

Figure 3

Number of Medicare FFS eligibles with diabetes, by type, 2007-2011.

This rise was due to an increase in persons with type 2 diabetes not using insulin (5,399,954 to 5,992,778), which offset a decline in the number of persons classified as type 1 (1,163,241 to 999,066). The number and proportion of beneficiaries with type 2 diabetes using insulin remained stable over this period, while the proportion of beneficiaries classified as type 1 dropped from 17.6 percent in 2007 to 14.2 percent in 2011 (Figure 4).

Figure 4

Percentage of diabetic Medicare FFS eligibles, by type, 2007-2011.

Most beneficiaries with diabetes had type 2 diabetes without insulin use (85.2% in 2011), and the smallest proportion had type 2 diabetes with insulin use (0.57% in 2011; Table 1). There were no strong differences by age in the classification of beneficiaries with diabetes into the three subtypes. By race, however, there were no strong differences. The highest percentage of beneficiaries with type 1 diabetes was African American (18.9%) and the lowest percentage was American Indian/Alaska Native (9.9%). The highest percentage of beneficiaries with type 2 diabetes using insulin was non-Hispanic White (0.61%) and the lowest was Hispanic (0.39%; Table 1).

Women were slightly more likely to be classified as having type 1 diabetes (14.34% vs. 14.05%) and men were more likely to be classified as type 2 using insulin (0.66% vs. 0.49%).

The percentage of beneficiaries with type 1 diabetes was slightly higher in urban areas than rural areas (14.41% vs. 12.33%). Full dual-eligible beneficiaries were more likely to be classified as type 1 or type 2 using insulin than nonduals. Partial duals and others had intermediate levels of classification (Table 1).

The total annual mortality was 6.12 percent for beneficiaries with diabetes who died during 2011. Within this population, mortality was highest for beneficiaries with type 2 diabetes using insulin (12%), followed by beneficiaries with type 1 (7.8%). Beneficiaries with type 2 diabetes not using insulin had the lowest mortality rate (5.8%).

There were dramatically different comorbidity profiles across diabetic types. Overall, 84.8 percent of beneficiaries with type 1 diabetes and 83.9 percent of beneficiaries with type 2 using insulin had one or more comorbidities related to diabetes (CKD, AMI, AFIB, heart failure, ischemic heart disease, cataract, glaucoma, or stroke) compared with 72.7 percent of beneficiaries with type 2 diabetes not using insulin. While the frequency of diagnosis varied across diabetic types, for all three groups the most commonly noted comorbidity was ischemic heart disease, followed by chronic kidney disease and heart failure. The frequency of cataract diagnosis was comparable across all groups at 18.3 percent to 19.4 percent (Table 2).

The frequency of comorbidities unrelated to diabetes showed a similar pattern. About 68 percent of beneficiaries with type 2 diabetes who use insulin and with type 1 diabetes had one or more comorbidities. About 58 percent of beneficiaries with type 2 diabetes not using insulin had one or more comorbidities unrelated to diabetes (Table 2). Among these comorbidities, arthritis was the most common, followed by chronic obstructive pulmonary disease, depression, and Alzheimer's disease.

Almost all beneficiaries with diabetes regardless of type had contact with the health care system in the form of evaluation and management (E&M) visits across all providers (range 93%-95%).

The mean number of E&M visits per year was highest for beneficiaries with type 2 diabetes using insulin (16 visits) and lowest for beneficiaries with type 2 not using insulin (10 visits). Beneficiaries with type 1 had an intermediate number of visits (13; Table 3).

In addition to E&M visits, beneficiaries with diabetes consistently received at least some preventive care. Slightly more than half of all beneficiaries with diabetes received an annual flu shot (Table 4). Annual eye exams were most commonly received by beneficiaries with type 1 (58.3%) and less frequently received by beneficiaries with type 2 diabetes using and not using insulin (45.9% and 49.6%). Annual HbA1c testing was also highest among beneficiaries with type 1 diabetes (88.1%) and lowest among beneficiaries with type 2 using insulin (77.1%). Beneficiaries with type 2 not using insulin were intermediate at 82.7 percent. Annual LDL tests were least frequently received by beneficiaries with type 2 diabetes using insulin (69.0%) and were received at similar rates for beneficiaries with type 1 and type 2 not using insulin (77.3% and 78.4%). It is important to note that for every preventive care measure, beneficiaries with type 2 using insulin had the lowest rate of receipt of services (Table 4).

Use of primary care providers was high, with an average of 78.2 percent of all beneficiaries with diabetes having at least one visit (Table 5). The mean number of visits to a primary care provider was similar for beneficiaries with type 1 and type 2 diabetes using insulin (5.8 visits and 6.1 visits). Beneficiaries with type 2 not using insulin had a slightly lower average of 5.0 visits. While use of primary care providers may be high, beneficiaries are likely to see more than one primary care provider.

About 70 percent of beneficiaries with diabetes saw two or more primary care providers, while less than 10 percent of beneficiaries who saw a primary care provider saw only one. Across all groups, about 30 percent of beneficiaries with diabetes saw five or more different primary care providers (Table 5).

Visits with specialists were common but not as frequent as primary care visits. Overall, approximately 70 percent of beneficiaries with diabetes made one or more visits to any specialty provider (Table 6). Beneficiaries with type 1 diabetes had the highest percentage of visits to a specialty provider (82%), while those with type 2 diabetes using and not using insulin had similar rates of specialist use, with 68.6 percent and 67 percent visiting one or more specialists. Among those who had specialist visits, the mean number of visits was similar across groups at 5.8 visits for type 1, 6.1 visits for type 2 using insulin, and 5.0 visits for type 2 not using insulin (data not shown in table). Most Medicare beneficiaries with diabetes who saw a specialist saw more than one specialist—68.5 percent of beneficiaries with type 1 diabetes, 54.4 percent of type 2 using insulin, and 50.9 percent of type 2 using insulin saw more than one specialist in 2011 (Table 6).

The most frequently seen specialists were ophthalmologists or optometrists and cardiologists. Endocrinologists were infrequently seen. For example, only 21.3 percent of Medicare beneficiaries with type 1 diabetes had a visit with an endocrinologist and 12.6 percent of beneficiaries with type 2 diabetes who used insulin had a visit with an endocrinologist. The percentage of beneficiaries who saw only one specialist was considerably lower than the percentage of beneficiaries who saw someone representing only one specialty.

For example, 42.1 percent of beneficiaries saw providers representing only one specialty but 15.7 percent saw only one specialist, which indicates that beneficiaries were seeing multiple individual providers within a specialty group.

Consistent with their comorbidity burden, the highest hospitalization rates were for beneficiaries with type 2 diabetes using insulin (111 visits per 100 beneficiaries) and lowest for beneficiaries with type 2 diabetes not using insulin (43 visits per 100 beneficiaries; Table 7). Beneficiaries with type 1 were intermediate at 84 visits per 100 beneficiaries. Across all three groups, the rate of medical hospitalizations only (no procedure) was greater than the rate of hospitalizations involving a procedure, but hospitalizations for beneficiaries with type 2 diabetes using insulin were more likely to involve a procedure than for other groups (34 hospitalizations per 100 beneficiaries involve procedures for this group).

Table 7

Rate of hospitalizations and emergency room visits per 100 Medicare FFS beneficiaries, by type of diabetes, 2011.

Rates of emergency room use were similar for beneficiaries with type 1 diabetes and type 2 diabetes using insulin (133 per 100 beneficiaries and 143 per 100 beneficiaries; Table 7). Rates of use were substantially lower among beneficiaries with type 2 diabetes not using insulin at 75 visits per 100 beneficiaries. Beneficiaries with type 2 diabetes not using insulin had lower rates of ER visits leading to hospitalization (31 visits per 100 beneficiaries) while beneficiaries with type 1 and type 2 using insulin had higher rates of ER visits leading to hospitalization (64 visits per 100 beneficiaries and 71 visits per 100 beneficiaries).

Overall, allowed payment per beneficiary increased from 2007 to 2011 ($48,184 to $71,652) and is strongly related to diabetes type.

In 2011, the mean allowed payment for beneficiaries with diabetes was highest for those with type 2 diabetes using insulin ($203,446), followed by individuals with type 1 ($140,863; Table 8). The mean allowed payment for individuals with type 2 diabetes not using insulin was considerably lower at $59,238.This relative pattern was maintained across age, race, gender, urbanicity, and dual status category, with few exceptions.

Table 8

Mean allowed payment per beneficiary among Medicare FFS eligibles, by demography and type of diabetes, 2011 ($ thousands).

DISCUSSION

This study of Medicare beneficiaries with diabetes illustrates some important challenges to care and care coordination for this growing population. First, while most beneficiaries with diabetes have type 2 diabetes and do not use insulin, those with type 1 diabetes or with type 2 diabetes who use insulin have disproportionately high hospitalization, emergency room use, and mortality rates. Likewise, annual Medicare allowed payments are highest for beneficiaries with type 2 diabetes using insulin and lowest for those with type 2 diabetes not using insulin. Comorbidities experienced by beneficiaries with diabetes, both related to and unrelated to diabetes, show a similar pattern. Beneficiaries with type 1 diabetes or type 2 diabetes using insulin have more comorbidities than those with type 2 diabetes not using insulin. In the case of type 2 using insulin, these associations may be influenced by the fact that insulin is likely at least a third-level intervention for glucose control. Type 2 patients on insulin have likely had insufficient benefit from lifestyle changes and oral agents and may have especially poor diabetes control. This may be due to poor adherence or a longer course of treatment for type 2, where patients have simply arrived at needing insulin therapy over time.

In spite of the patterns found in comorbidities, mortality, and inpatient or emergency utilization, we do not see signs that beneficiarie with type 1 diabetes or type 2 diabetes using insulin receive any more preventive care than beneficiaries with type 2 diabetes not using insulin. Beneficiaries with type 2 diabetes using insulin have the lowest rates of eye exams, HbA1c testing, and LDL testing. There are two potential interpretations of these patterns. The first is that an opportunity for preventive care has been missed for those who are, in fact, most needy. Alternatively, it is possible that the higher level of comorbidity and associated health care use interferes with the beneficiaries' ability to manage their preventive care. Regardless, these findings point to the need and opportunity to more effectively provide preventive care for Medicare beneficiaries with diabetes, particularly those who use insulin.

Care coordination can be difficult to quantify on a population level. In this report we see signs that point to suboptimal care coordination. For example, we found that while most beneficiaries with diabetes have contact with primary care, most also interact with multiple primary care providers and with multiple specialists.

About 80 percent of beneficiaries who had contact with primary care saw more than one provider. About 35 percent of the insulin-using diabetics (type 1 or type 2) saw five or more primary care providers. Patterns in beneficiary use of specialists pointed to a similar lack of care coordination. While most beneficiaries with diabetes saw one or more specialists, between 20 and 40 percent of those beneficiaries saw five or more different specialists. Most beneficiaries saw providers representing only one or two specialties, indicating that multiple individual specialists are being visited within health care specialties.

In spite of the strengths of this analysis, some limitations must be acknowledged. First, our classification of diabetics is based on diagnosis codes and procedure codes related to insulin use. It is possible that there is some miscategorization between type 1 and type 2 diabetes classification, though we relied on widely used algorithms. Second, our classification of providers and provider specialties is based on provider IDs. While these also link to payment, errors are possible that would lead to an overcount of individual providers. Third, we present unadjusted numbers and do not conduct statistical testing of differences in proportions. Our sample includes the entire Medicare FFS population 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.²⁶ 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 and the reliance on well-established protocols supports the strengths of our conclusions.

CONCLUSION

Our examination of health care use and access to providers by Medicare beneficiaries with diabetes points to several important findings. First, use of preventive and acute care varies dramatically between beneficiaries with type 1 diabetes, type 2 using insulin, and type 2 not using insulin.

We found that most beneficiaries with diabetes have contact with both primary care and specialty providers. However, the number of distinct providers with whom they have contact is sufficiently high that it calls the reality of care coordination into question.

If programs related to encouraging care coordination are to be considered a success, reducing the number of different providers seen is a logical place to focus. In addition, it may be helpful for evaluations to take into account the number of providers seen when assessing whether attempts at care coordination lead to better preventive care and reductions in acute care use.

REFERENCES

1.

Centers for Disease Control and Prevention. Chronic Diseases and Health Promotion. 2012. [November 16, 2012]. Available at: http://www​.cdc.gov/chronicdisease​/overview/index.htm.

2.

Centers for Disease Control and Prevention. National diabetes fact sheet: national estimates and general information on diabetes and prediabetes in the United States, 2011. Atlanta, GA: Centers for Disease Control and Prevention; 2011.

3.

Geiss LS, Li Y, Kirtland K, et al. Increasing prevalence of diagnosed diabetes – United States and Puerto Rico, 1995-2010. MMWR. 2012;61(45):918–21. [PubMed: 23151951]

4.

American Diabetes Association. Living with diabetes: complications. [November 16, 2012]. Available at: http://www​.diabetes.org​/living-with-diabetes/complications/

5.

Russell LB, Suh DC, Safford MA. Time requirements for diabetes self-management: too much for many? J Fam Pract. 2005;54(1):52–6. [PubMed: 15623407]

6.

Winocour PH. Effective diabetes care: a need for realistic targets. BMJ. 2002;324(7353):1577–80. [PMC free article: PMC1123507] [PubMed: 12089099]

7.

Gendelman N, Snell-Bergeon J, McFann K, et al. Prevalence and correlates of depression in individuals with and without type 1 diabetes. Diabetes Care. 2009;32(4):575–9. [PMC free article: PMC2660458] [PubMed: 19171719]

8.

Maciejewski ML, Maynard C. Diabetes-related utilization and costs for inpatient and outpatient services in the Veterans Administration. Diabetes Care. 2004;27(suppl 2):b69–73. [PubMed: 15113786]

9.

Jiang JH, Stryer D, Friedman B, et al. Multiple hospitalizations for patients with diabetes. Diabetes Care. 2003;26(5):1421–6. [PubMed: 12716799]

10.

Liu CW, Einstadter D, Cebul RD. Care fragmentation and emergency department use among complex patients with diabetes. AJMC. 2010;16(6):413–20. [PubMed: 20560685]

11.

Vogeli C, Shields AE, Lee TA, et al. Multiple chronic conditions: prevalence, health consequences, and implications for quality, care management, and costs. JGIM. 2007;22(suppl 3):391–5. [PMC free article: PMC2150598] [PubMed: 18026807]

12.

Buccaneer, a General Dynamics Company. Chronic Conditions Data Warehouse. [January 3, 2013]. Available at: http://www​.ccwdata.org.

13.

Research Data Assistance Center. Race code. [November 2, 2012]. Available at: http://www​.resdac.org​/cms-data/variables​/Research-Triangle-Institute-RTI-Race-Code.

14.

Agency for Healthcare Research and Quality. Creation of new race-ethnicity codes and socioeconomic status (SES) indicators for Medicare beneficiaries. [November 2, 2012]. Available at: http://www​.ahrq.gov/qual​/medicareindicators​/medicareindicators2.htm.

15.

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/reference/gtc/gtc_cbsa.html.

16.

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.

17.

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

18.

Redelmeier DA, Tan SH, Booth GL. The treatment of unrelated disorders in patients with chronic medical diseases. NEJM. 1998;338(21):1516–20. [PubMed: 9593791]

19.

Struijs JN, Baan CA, Schellevis FG, et al. Comorbidity in patients with diabetes mellitus: impact on medical health care utilization. BMC Health Serv Res. 2006;6(1):1–9. [PMC free article: PMC1534031] [PubMed: 16820048]

20.

Piette JD, Kerr EA. The impact of comorbid chronic conditions on diabetes care. Diabetes Care. 2006;29(3):725–731. [PubMed: 16505540]

21.

Lagu T, Weiner MG, Hollenbeak CS, et al. The impact of concordant and discordant conditions on the quality of care for hyperlipidemia. JGIM. 2008;23(8):1208–13. [PMC free article: PMC2517982] [PubMed: 18465174]

22.

Centers for Medicare & Medicaid Services. Berenson-Eggers type of service (BE-TOS). [Feburary 2, 2013]. Available at: https://www​.cms.gov/Medicare​/Coding/HCPCSReleaseCode-Sets​/BETOS.html.

23.

Research Data Assistance Center. HCFA provider specialty table. [December 3, 2012]. Available at: http://www​.resdac.org/sites/resdac​.org/files​/HCFA%20Provider​%20Specialty%20Table.txt.

24.

Centers for Medicare & Medicaid Services. National Plan and Provider Enumeration System: National Provider Identifier. [November 29, 2012]. Available at: https://nppes​.cms.hhs​.gov/NPPES/StaticForward​.do?forward=static.npistart.

25.

Research Data Assistance Center. Revenue center code. [December 3, 2012]. Available at: http://www​.resdac.org​/cms-data/variables/revenue-center-code.

26.

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

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.

Suggested Citation: Virnig BA, Shippee ND, O'Donnell B, Zeglin J, Parashuram S. Use and access of health care by Medicare beneficiaries with diabetes: impact of diabetes type and insulin use, 2007-2011. Use of Diabetes Care. Data Points # 18 (prepared by the University of Minnesota DEcIDE Center, under Contract No. HHSA29020100013I ). Rockville, MD: Agency for Healthcare Research and Quality. January 2014. AHRQ Publication No. 14-EHC005-EF.