National Health Interview Survey

The National Center for Health Statistics has sponsored the NHIS annually since 1957. It describes the health status of the civilian, noninstitutionalized population of the United States (1). Data are collected through a series of personal household interviews and include a core questionnaire and supplemental questions about current health topics. In the core questionnaire, each person is asked about the presence or absence of specific chronic conditions, including “Have you ever been told by a doctor or other health professional that you have diabetes or sugar diabetes?” and the frequency of medical contacts in the past year, including hospital- and non-hospital-based physician office visits, emergency room visits, and telephone contacts. From this, the frequency of contacts with the medical care system for people with diabetes is calculated. The questionnaire also asks about age at diagnosis of diabetes and treatment for diabetes. From this, people with probable type 1 diabetes are distinguished from people with probable type 2 diabetes. Type 1 diabetes is defined as age at onset <30 years and current use of insulin. All others are considered to have type 2 diabetes. In 2003, a supplemental questionnaire was administered to all individuals who self-reported physician- diagnosed diabetes. It included questions about whether the respondent had one physician he/she usually sees for diabetes, the frequency of contact with that person, and more details regarding the care received.

Self-reported information, such as that provided by the NHIS, has limitations. Self-report of physician-diagnosed diabetes may be incorrect, although the accuracy of self-report has been shown to be very good (2,3). In addition, recall bias may impact the reporting of remote events. For this reason, the NHIS only asks about physician visits and hospitalizations in the past 12 months. Finally, a proxy respondent (generally a spouse or other responsible household adult) may answer questions in the core questionnaire for sample persons who are not available at the time of the interview or are unable to answer for themselves (i.e., children, older adults, and those who are mentally incompetent). In 2010, 95.4% of people with diabetes responded for themselves, and 4.6% had proxy respondents. In general, proxy responses are considered to be of equal quality as non-proxy responses (4).

When looking at data trends over time, it is also important to consider how data collection instruments have changed. In 1997, the National Center for Health Statistics redesigned the sampling frame of the NHIS. This impacted the percentage of people estimated to have diabetes and the comparability of some of the data before and after that date (5).

National Ambulatory Medical Care Survey

The National Center for Health Statistics has conducted the NAMCS annually since 1989. It provides a national probability sample of visits to non-federally employed, office-based physicians who are primarily engaged in direct patient care. Visits to specialists in anesthesiology, pathology, and radiology are excluded (6). The survey assesses patients’ demographic characteristics, vital signs, and up to three complaints, symptoms, or other reasons for the visit and the corresponding physicians’ diagnoses, diagnostic testing, health education, and the medication and non-medication treatments ordered. The NAMCS codes each physician visit according to the diagnosis most associated with the patient’s primary complaint (primary diagnosis). Two additional diagnoses (secondary diagnoses) can be coded corresponding to other reasons for the patient’s visit for a maximum of three diagnoses per visit. Medical data and drug coding are performed centrally with rigorous quality control procedures. Physicians’ diagnoses are coded according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). Data are presented for people with diabetes (ICD-9-CM codes 250 [all], 357.2, 362.0, 366.41, 648.0, and 775.1) listed as the primary (first-listed) diagnosis or, separately, as any listed diagnosis.

Rates of outpatient visits for people with diabetes were calculated by dividing the number of outpatient visits with diabetes as the primary or any listed diagnosis (in 2010) by the estimated number of people with physician-diagnosed diabetes in the United States in 2010 as determined by the NHIS.

Starting in 2005, check boxes were added to the NAMCS to record the presence of specific chronic medical conditions, including diabetes. People with diabetes can be identified in one of two ways. The first is to identify people with diabetes using the three physicians’ diagnoses. This method likely underestimates the true prevalence of diabetes in patients receiving ambulatory care because the diagnoses listed reflect the reason for the visit rather than the patient’s full range of medical conditions. The second method is to identify people with diabetes as those with diabetes checked as a chronic medical condition. Unfortunately, there are no published studies validating information provided by the check box, and it is unclear whether a checked box reflects a physician diagnosis, patient self-report, or other less methodologically sound evidence. For analyses presented in this chapter, people with diabetes were defined using the first method. Even though this may miss some people with diabetes seeking care for what are likely to be nondiabetes-related medical issues, it is more likely to capture people with physician-diagnosed diabetes who are seeking medical care for diabetes. It also makes these estimates comparable to those reported in earlier editions of Diabetes in America (7). Because NAMCS data are visit-based and not patient-based, they may overrepresent people who take insulin for their diabetes, since these individuals have more frequent ambulatory care visits than those who are not treated with insulin (7).

Medical Expenditure Panel Survey

The Agency for Healthcare Research and Quality, through the U.S. Department of Health and Human Services, has conducted the MEPS annually since 1996. There are two major components to the MEPS: the Household Component, which collects data from individual households and their medical providers, and the Insurance Component, which collects employer-based health insurance data from employers (8). The Household Component uses a sample of families and individuals drawn from a nationally representative subsample of households that participated in the prior year’s NHIS. Data are collected during a series of household interviews covering two full years and include demographic characteristics, health conditions, health status, use of medical services, charges and source of payment, access to care, satisfaction with care, health insurance coverage, income, and employment. Diabetes status in the MEPS is based on self-reported diabetes diagnosis or a medical visit related to diabetes in the past year using ICD-9 codes 249–250, 790.2, 791.5–791.6, V45.85, V53.91, and V65.46.

When looking at data trends over time, it is important to remember that the MEPS started in 1996. Data prior to that time are not available. In addition, while the MEPS represents a subset of the previous year’s NHIS respondents, the MEPS takes the additional steps of asking respondents to prospectively record their utilizations and expenditures, thus allowing comparision of the self-reported information with actual medical record and insurance claim data.

National Health Interview Survey

In addition to asking about ambulatory care, as described in the preceding section on ambulatory medical care for people with diabetes, the NHIS asks questions about hospitalization, including whether the respondent was hospitalized overnight in the past year, frequency of hospitalization, and number of nights spent in the hospital during each hospitalization.

National Hospital Discharge Survey

The NHDS, conducted between 1965 and 2010, was designed to gather information on a national sample of inpatients discharged from non-federal, short-stay hospitals (9). Data were collected by abstracting hospital records or by reviewing electronically available data files and then extrapolated using statistical techniques to represent all discharges in the United States. Demographic characteristics of the patients, admission and discharge dates, diagnoses (one primary and up to seven [for 1965–2009] or 15 [for 2010] secondary), and procedures were obtained, and medical data were coded according to the ICD-9-CM.

Data are presented for hospitalizations in which diabetes (ICD-9-CM codes 250 [all], 357.2, 362.0, 366.41, 648.0, and 775.1) was listed as the primary (first-listed) diagnosis or, separately, for hospitalizations in which diabetes was listed as any diagnosis. Other conditions were ascertained using ICD-9-CM codes: diabetic ketoacidosis (DKA) or hyperosmolar coma (250.1 or 250.2); stroke (430–438); myocardial infarction (410); congestive heart failure (428); and cardiovascular disease (410–414, 426–438, or 440–448). Procedures were ascertained using ICD-9-CM procedure codes: carotid endarterectomy (38.12); cardiac catheterization (37.21, 37.22, or 37.23); percutaneous coronary intervention (36.01, 36.02, 36.05, 36.06, or 36.07); coronary artery bypass grafting (36.1x); and lower extremity vascular bypass or stenting (39.25 or 39.29).

Rates of hospital discharges for people with diabetes were calculated by dividing the number of hospitalizations with diabetes as the primary or any listed diagnosis (in 2010) by the estimated number of people with physician-diagnosed diabetes in the United States in 2010 as determined by the NHIS.

National hospital discharge data are visit-based, not patient-based. A consequence of visit-based analysis is that if the same person is hospitalized more than once, he or she is counted more than once. Based on the NHIS 2011, multiple hospital admissions in a year were reported by approximately 29% of people who reported being hospitalized. This may be important when examining discharges for events that cluster in individuals, such as DKA. Nevertheless, NHDS data are useful to ascertain the characteristics of people admitted to the hospital and their reasons for admission.

If a patient’s admission is not related to his or her diabetes status, then it is unlikely that diabetes will be coded as a diagnosis. Approximately 40% of people with diabetes who are hospitalized do not have diabetes listed on their hospitalization discharge record (10). As a result, it is impossible to ascertain the prevalence of diabetes among people admitted to the hospital. Even given these limitations, NHDS remains the best source of data to estimate diabetes-related hospitalizations, and while the prevalence of diabetes is likely underestimated, the underestimation likely remained constant from 1965 to 2009, allowing one to estimate changes over time.

In 2010, the National Center for Health Statistics redesigned the NHDS to collect data on up to 15 diagnoses, whereas between 1965 and 2009, it collected only seven diagnoses. In this section, when only 2010 data are presented, all 15 diagnoses are used. However, when trends over multiple years are presented, only the first seven diagnoses from 2010 are included to be consistent with earlier surveys.

Only a small percentage of NHDS discharges that list diabetes classify it as the primary diagnosis. The percentage of NHDS discharges with diabetes as any listed diagnosis that list diabetes as the primary diagnosis has declined over time for all age groups (11). Fifteen percent of such discharges listed diabetes as the primary diagnosis in 1990 (7), and 12% listed diabetes as the primary diagnosis in 2010.

Ascertainment of race from the NHDS is problematic because a substantial percentage of discharges are missing racial classification and because the sampling technique does not allow for accurate estimation for some minority groups. In 2010, race was unknown for 14% of all hospital discharges in the NHDS.

Medical Expenditure Panel Survey

In addition to asking questions about ambulatory care, as described in the preceding section on ambulatory medical care for people with diabetes, the MEPS asks questions about hospitalization, including whether the respondent was hospitalized overnight in the past year, the frequency of hospitalization, and the number of nights spent in the hospital during each hospitalization.

A limitation of using the MEPS data for hospital discharges for people with diabetes is the small sample size. When MEPS data are presented in this section, they should be interpreted as the yearly average for the time period 2008–2012.

2010 Hospital Discharge Data

In 2010, approximately 622,000 hospitalizations in the United States recorded diabetes as the primary diagnosis on the hospital discharge record (2% of all hospitalizations in the United States), and 6.76 million hospitalizations recorded diabetes as any listed diagnosis (21% of all hospitalizations in the United States). Hospitalizations with a primary diagnosis of diabetes accounted for 2.84 million hospital days, and those with diabetes as any listed diagnosis accounted for 34.67 million hospital days. Approximately 73% of hospitalizations with diabetes as the primary diagnosis occurred in people age ≥45 years (Figure 40.24). For hospitalizations with diabetes as any listed diagnosis, the number of hospitalizations increased with age; approximately 90% were for people age ≥45 years.

Diabetes was listed as the primary diagnosis for 2% of all hospitalizations in the United States. The hospital discharge rate with diabetes as the primary diagnosis (calculated as the number of discharges with diabetes listed as the primary diagnosis per 100 people with diabetes as determined from the NHIS) was more than double for people age 18–44 years compared to people age 45–64 or ≥65 years (Figure 40.25). Younger women tended to have a higher rate of hospitalizations with the primary diagnosis of diabetes compared to younger men. Rates of hospitalization were similar for the two older age groups by sex (Figure 40.26).

Diabetes was included as any listed diagnosis for 21% of all hospitalizations in the United States. The hospitalization rate for diabetes as any listed diagnosis increased with age and was more than double for persons age ≥65 years compared to the youngest age group (Figure 40.25). After stratifying by sex, the hospitalization rate for diabetes as any listed diagnosis increased with age for men. Women in the two younger age groups had similar rates, and both were approximately half the rate in the oldest age group (Figure 40.26).

FIGURE 40.24

Number of Hospital Discharges With Diabetes as Primary or Any Listed Diagnosis Among Adults Age ≥18 Years, by Age, U.S., 2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (more...)

FIGURE 40.25

Rate of Hospital Discharges With Diabetes as Primary or Any Listed Diagnosis Per 100 Diabetic Population, by Age, U.S., 2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 250 (all), (more...)

Figure 40.27 shows rates of hospitalization with diabetes as the primary or any listed diagnosis by age and race/ethnicity. Rates of hospitalization with diabetes as the primary diagnosis were highest in non- Hispanic blacks. For diabetes as the primary diagnosis, non-Hispanic whites and non- Hispanic blacks in the youngest age group had the highest rates of hospitalization. These differences may in part be due to differences in the prevalence of type 1 diabetes among these racial/ethnic groups. For diabetes as any listed diagnosis, non- Hispanic whites showed rates of hospitalization that increased with age. Non-Hispanic blacks and Hispanics both showed the highest rates of hospitalization in the ≥65 years age group.

FIGURE 40.26

Rate of Hospital Discharges With Diabetes as Primary or Any Listed Diagnosis Per 100 Diabetic Population, by Age and Sex, U.S., 2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (more...)

FIGURE 40.27

Rate of Hospital Discharges With Diabetes as Primary or Any Listed Diagnosis Per 100 Diabetic Population, by Age and Race/Ethnicity, U.S., 2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (more...)

Time Trends in Hospital Discharges Listing Diabetes

Between 1990 and 2010, the proportion of hospitalizations that listed diabetes as the primary diagnosis out of the number of hospitalizations that listed diabetes as any diagnosis decreased from 14.4% to 11.6%. Between 1990 and 2010, the number of hospital discharge records that listed diabetes as the primary diagnosis increased 53% from 406,000 to 622,000. The increase was constant over time, at approximately 2% per year (Appendix 40.1). When the number of hospital discharge records with diabetes listed as the primary diagnosis was stratified by age, the middle age group (45–64 years) had the greatest increase over time, almost doubling between 1990 and 2010. The increases in the other two age groups were less pronounced (Figure 40.28). When stratified by sex, men had a greater increase over time (70%) than women (41%) (Figure 40.29, Appendix 40.1). When stratified by race, whites had the smallest increase (27%), and blacks had a 90% increase (Figure 40.30). Race was missing in 25% of the NHDS 1990–2010 data, so caution must be taken when looking at hospitalization rates by race (Appendix 40.2).

The rate of hospital discharges with diabetes listed as the primary diagnosis per 100 diabetic individuals has decreased over time for all age groups (Figure 40.31). Although the rates for men and women have both decreased over time, women had a slightly greater decrease in hospital discharge rate with diabetes listed as the primary diagnosis compared to men (Figure 40.32). Although rates were higher in blacks than whites, whites and blacks had similar decreases over time (Figure 40.33). The fact that the number of hospital discharges is increasing while the rates of hospital discharges are decreasing indicates a greater increase in the size of the denominator (i.e., the total population of people with diabetes).

The number of hospitalizations with diabetes as any listed diagnosis increased 88% during 1990–2010 from 2.8 million to 5.3 million (using the first seven listed diagnoses). The increase from 1990 to 2003 averaged 4.5% per year. After 2004, the percent increase remained stable at approximately 1% per year (Appendix 40.3). The oldest age group (≥65 years) had the largest number of hospital discharges with diabetes as any listed diagnosis. The middle age group (45–64 years) had the greatest increase over time, more than doubling between 1990 and 2010 (Figure 40.34). The increases in the other age groups were less pronounced. Men had a greater increase than women over time (200% vs. 175%), but both sexes showed a steady increase from 1990 to 2003 and then a smaller increase through 2010 (Figure 40.35, Appendix 40.3). Whites had a 70% increase and blacks had a 225% increase in hospitalization with diabetes as any listed diagnosis (Figure 40.36, Appendix 40.4).

FIGURE 40.28

Time Trend in the Number of Hospital Discharges With Diabetes Listed as the Primary Diagnosis Among Adults Age ≥18 Years, by Age, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical (more...)

FIGURE 40.29

Time Trend in the Number of Hospital Discharges With Diabetes Listed as the Primary Diagnosis Among Adults Age ≥18 Years, by Sex, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical (more...)

FIGURE 40.30

Time Trend in the Number of Hospital Discharges With Diabetes Listed as the Primary Diagnosis Among Adults Age ≥18 Years, by Race, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical (more...)

The rate of hospital discharges with diabetes as any listed diagnosis per 100 diabetic individuals has decreased over time for all age groups, and this decrease is more pronounced as age increases (Figure 40.37). Although the rates for men and women have both decreased over time, women had a slightly greater decrease in the rate of hospital discharges with diabetes as any listed diagnosis compared to men (Figure 40.38). Whites and blacks had similar decreases over time (Figure 40.39). While the Other race category also showed a decrease in the rate over time, this category again had substantial random variation due to small sample size. Similar to when diabetes is listed as the primary diagnosis, when diabetes is listed as any diagnosis and the number of hospital discharges are increasing while the rates of hospital discharges are decreasing, this implies a rapidly growing denominator (i.e., the total population of people with diabetes).

FIGURE 40.31

Time Trend in the Rate of Hospital Discharges With Diabetes Listed as the Primary Diagnosis Among Adults Age ≥18 Years, by Age, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical (more...)

FIGURE 40.32

Time Trend in the Rate of Hospital Discharges With Diabetes Listed as the Primary Diagnosis Among Adults Age ≥18 Years, by Sex, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical (more...)

FIGURE 40.33

Time Trend in the Rate of Hospital Discharges With Diabetes Listed as the Primary Diagnosis Among Adults Age ≥18 Years, by Race, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical (more...)

FIGURE 40.34

Time Trend in the Number of Hospital Discharges With Diabetes Listed as Any Diagnosis Among Adults Age ≥18 Years, by Age, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

FIGURE 40.35

Time Trend in the Number of Hospital Discharges With Diabetes Listed as Any Diagnosis Among Adults Age ≥18 Years, by Sex, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

FIGURE 40.36

Time Trend in the Number of Hospital Discharges With Diabetes Listed as Any Diagnosis Among Adults Age ≥18 Years, by Race, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

FIGURE 40.37

Time Trend in the Rate of Hospital Discharges With Diabetes as Any Listed Diagnosis Among Adults Age ≥18 Years, by Age, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

FIGURE 40.38

Time Trend in the Rate of Hospital Discharges With Diabetes as Any Listed Diagnosis Among Adults Age ≥18 Years, by Sex, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

FIGURE 40.39

Time Trend in the Rate of Hospital Discharges With Diabetes as Any Listed Diagnosis Among Adults Age ≥18 Years, by Race, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

Hospital Discharges Listing Complications of Diabetes

Figure 40.40 and Table 40.22 show time trends in the number of hospital discharges listing complications of diabetes. DKA and hyperosmolar coma are common causes of hospitalization for people with diabetes. In 1990, 95,000 hospital discharges listed DKA or hyperosmolar coma as any diagnosis, and in 2010, 191,000 hospital discharges listed DKA or hyperosmolar coma as any diagnosis. Macrovascular disease remains the most common complication of diabetes. There was a steady increase in hospital discharges listing both diabetes and cardiovascular disease, defined as any listed diagnosis of transient ischemic attack (TIA) or stroke, myocardial infarction, or congestive heart failure, from 809,000 discharges in 1990 to 1,373,000 discharges in 2002. Thereafter, the number of hospital discharges listing both diabetes and cardiovascular disease decreased each year and was the same in 2010 as in 1994. In contrast, the number of discharges with TIA or stroke and diabetes has continued to increase slowly over time.

The rates of hospital discharges listing complications of diabetes have also changed over time. The rate of hospital discharges with any diagnosis of DKA or hyperosmolar coma was approximately 1.5 discharges per 100 persons with diabetes per year until 1997, and since then, it has declined to approximately 0.9 discharges per 100 persons with diabetes per year (Figure 40.41, Table 40.23). Similar trends were seen for TIA or stroke, myocardial infarction, and congestive heart failure, where rates of hospital discharges in 2010 were less than half of what they were in 1990. Rates of discharges for cardiovascular disease (TIA or stroke, myocardial infarction, and congestive heart failure combined) and diabetes reached their peak in 1996 and have declined since, with the most dramatic decline occurring since 2004.

Similar results have been reported by Gregg et al. (12) who found that the rates of five major diabetes complications (stroke, acute myocardial infarction, amputation, end-stage renal disease, and death from hyperglycemic coma) have declined significantly between 1990 and 2010.

FIGURE 40.40

Time Trend in the Number of Hospital Discharges Listing DKA or Hyperosmolar Coma, TIA/Stroke, Myocardial Infarction, Congestive Heart Failure, and Total Cardiovascular Disease as Any Diagnosis Among Adults Age ≥18 Years, U.S., 1990–2010. (more...)

The number of discharges among people with diabetes that listed carotid endarterectomy more than doubled between 1990 and 1995 but has been relatively stable since 1995 (Table 40.24). The number of discharges among people with diabetes that listed cardiac catheterization has increased over time with a slight decrease since 2002 (Figure 40.42, Table 40.24). The number of discharges among people with diabetes that listed coronary artery bypass graft increased from 1990 to 1999 but decreased from 1999 to 2010. For percutaneous coronary intervention, the number of discharges listing this procedure in people with diabetes more than tripled since 1990 but reached its zenith in 2003 and has declined since. Some of these trends likely reflect changes in clinical practice and whether procedures are performed on an inpatient or outpatient basis. The number of discharges listing lower extremity vascular bypass or stenting procedures in people with diabetes increased between 1990 and 1998 but has decreased since then. Between 1990 and 2010, the rate of hospital discharges per 100 people with diabetes listing both a diagnosis of diabetes and at least one procedure code for carotid endarterectomy has been relatively stable between 0.4 to 0.1 (Table 40.25). The rate of coronary artery bypass graft has decreased over time and was at a low of 0.2 procedures per 100 persons with diabetes in 2010 (Figure 40.43, Table 40.25). The rate of percutaneous coronary intervention was the same in 2010 as in 1990, and although there has been some fluctuation over time, it has never been higher than 1.3 procedures per 100 persons with diabetes. Procedures for cardiac catheterization and lower extremity vascular bypass or stenting, which remained stable until about 2005, have decreased since then by about half.

TABLE 40.22

Time Trend in the Number of Hospital Discharges Listing DKA or Hyperosmolar Coma, TIA/Stroke, Myocardial Infarction, Congestive Heart Failure, and Total Cardiovascular Disease as Any Diagnosis Among Adults Age ≥18 Years, U.S., 1990–2010. (more...)

FIGURE 40.41

Time Trend in the Rate of Hospital Discharges Listing DKA or Hyperosmolar Coma, TIA/Stroke, Myocardial Infarction, Congestive Heart Failure, and Total Cardiovascular Disease as Any Diagnosis Among Adults Age ≥18 Years, U.S., 1990–2010. (more...)

TABLE 40.23

Time Trend in the Rate of Hospital Discharges Listing DKA or Hyperosmolar Coma, TIA/Stroke, Myocardial Infarction, Congestive Heart Failure, and Total Cardiovascular Disease as Any Diagnosis Among Adults ≥18 Years, U.S., 1990–2010.

TABLE 40.24

Time Trend in the Number of Hospital Discharges Listing Both a Diagnosis of Diabetes and at Least One Cardiovascular Procedure Among Adults Age ≥18 Years, U.S., 1990–2010.

FIGURE 40.42

Time Trend in the Number of Hospital Discharges Listing Both a Diagnosis of Diabetes and at Least One Cardiovascular Procedure Among Adults Age ≥18 Years, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, (more...)

TABLE 40.25

Time Trend in the Rate of Hospital Discharges Listing Both a Diagnosis of Diabetes and at Least One Cardiovascular Procedure Among Adults Age ≥18 Years, U.S., 1990–2010.

FIGURE 40.43

Time Trend in the Rate of Hospital Discharges Listing Both a Diagnosis of Diabetes and at Least One Cardiovascular Procedure Among Adults Age ≥18 Years, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, (more...)

Proportion of Adults Hospitalized in the Past Year

Using self-reported data from the NHIS, Figure 40.44 and Appendix 40.5 show the percentages of adults with and without diabetes who reported being hospitalized at least once during the past year. In 2011, 20.8% of people with diabetes and 8.3% of those without diabetes reported being hospitalized. The percentage of people with or without diabetes who reported being hospitalized increased with age. People with diabetes were 2.5 times more likely to report being hospitalized in the previous year than people without diabetes. The difference was greatest for people age 18–44 years and decreased with age such that in the oldest age group, the rate in people with diabetes was only 1.7 times that in people without diabetes. With increasing age, the proportion of nondiabetic adults who reported being hospitalized in the past year increased 2.2-fold, from 6.8% at age 18–44 years to 15.2% at age ≥65 years. The proportion of diabetic adults who reported being hospitalized in the past year increased 1.6-fold, from 16.4% at age 18–44 years to 25.7% at age ≥65 years. MEPS data showed similar trends for 2008–2012 (Figure 40.45).

Hospitalization of Adults With Type 1 and Type 2 Diabetes

Figure 40.46 and Appendix 40.5 show the percentage of adults with type 1 and type 2 diabetes who reported being hospitalized at least once during the past year. In each age group, the percentage of adults who reported being hospitalized in the past year was highest for people with type 1 diabetes, followed by people with type 2 diabetes treated with insulin, and lowest for people with type 2 diabetes not treated with insulin. Among people with type 1 diabetes, the proportion of adults hospitalized tended to increase with age.

FIGURE 40.44

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Diabetes Status and Age, U.S., 2011. Data are based on self-report. Error bars represent 95% confidence intervals.

FIGURE 40.45

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Diabetes Status and Age, U.S., 2008–2012. Diabetes status is based on self-reported diabetes diagnosis or a medical visit related to diabetes in the past (more...)

FIGURE 40.46

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Type of Diabetes and Age, U.S., 2011. Type 1 diabetes is defined as diabetes diagnosed at age <30 years and current insulin use; all other persons with (more...)

Percentage of Adults Hospitalized by Diabetes Status, Age, Sex, and Race/Ethnicity

Figure 40.47 and Appendix 40.5 show the percentage of adults with and without diabetes who reported being hospitalized at least once in the past year by age and sex. Those with diabetes were more likely to report being hospitalized than those without diabetes, and older people were more likely to report being hospitalized than younger people. Women age 45–64 years both with and without diabetes were an exception. Women age 18–44 years were more likely to report being hospitalized than women age 45–64 years. This finding may reflect more frequent hospitalization for pregnancy and childbirth in women in the youngest age group. Trends in rates of hospitalization were similar in MEPS 2008–2012 data; however, the percentages were lower in each age/sex group in the MEPS compared to the NHIS 2011 data (Figure 40.48).

People with diabetes were more likely to report being hospitalized than people without diabetes across all racial/ethnic groups (Figure 40.49, Appendix 40.5). Trends in hospitalization by race/ethnicity were similar in the NHIS 2011 and the MEPS 2008–2012 data; however, the percentages were higher in the NHIS data than in the MEPS data in each age, race, and ethnic group (Figure 40.50). Women of all races/ethnicities in all age groups were more likely to report being hospitalized than men (Appendices 40.6 and 40.7).

FIGURE 40.47

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Diabetes Status, Age, and Sex, U.S., 2011. Data are based on self-report. Error bars represent 95% confidence intervals.

FIGURE 40.48

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Diabetes Status, Age, and Sex, U.S., 2008–2012. Diabetes status is based on self-reported diabetes diagnosis or a medical visit related to diabetes in the (more...)

FIGURE 40.49

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Diabetes Status, Age, and Race/Ethnicity, U.S., 2011. Data are based on self-report. Error bars represent 95% confidence intervals. * Mexican American is a subset (more...)

FIGURE 40.50

Percent of Adults Age ≥18 Years Who Report Being Hospitalized in the Past Year, by Diabetes Status, Age, and Race/Ethnicity, U.S., 2008–2012. Diabetes status is based on self-reported diabetes diagnosis or a medical visit related to diabetes (more...)

Hospitalization by Duration of Diabetes

For people with type 1 diabetes, hospitalization tended to be most frequent for those with duration <5 years (Table 40.26). For people with type 2 diabetes, hospitalization was most frequent for those with duration ≥15 years (Table 40.26). Among people with type 2 diabetes who were hospitalized in the past year and treated with insulin, 15% had a duration of diabetes <5 years, 37% had a duration of 5–14 years, and 48% had a known duration of ≥15 years (Figure 40.51). People with type 2 diabetes who were hospitalized in the past year but were not treated with insulin did not differ significantly with respect to diabetes duration. Thirty-six percent of those with type 2 diabetes who were were hospitalized and not treated with insulin had known duration of diabetes <5 years, 38% had a duration of 5–14 years, and 26% had a duration of ≥15 years.

TABLE 40.26

Percent of Adults With Diabetes Age ≥18 Years Who Report Being Hospitalized Overnight at Least Once in the Past Year, by Type and Duration of Diabetes, U.S., 2011.

FIGURE 40.51

Insulin Use and Duration of Diabetes Among Adults With Type 2 Diabetes Age ≥18 Years Who Have Been Hospitalized in the Past Year, U.S., 2011. Duration of diabetes is calculated as years since diagnosis of diabetes. Type 1 diabetes (excluded from (more...)

Multiple Hospitalizations Among Adults With Diabetes

Using data from the NHIS 2011, of all adults with diabetes who were hospitalized, 71% reported only one hospitalization, 17% reported two hospitalizations, and 12% reported three or more hospitalizations. The proportion of adults who reported only one hospitalization, or multiple hospitalizations, was similar across age groups (Figure 40.52). MEPS 2008–2012 data also showed that the proportion of adults who reported only one, or multiple hospitalizations, was similar across age groups; however, the percentage of people with only one hospitalization was higher in the MEPS than in the NHIS for each age group (Figure 40.53).

Length of Hospital Stay

In the NHIS 2011, people with diabetes age <65 years reported slightly longer lengths of stay (LOS) than those without diabetes (Table 40.27). In the oldest age group (≥65 years) people with diabetes had an average LOS that was 1 day shorter than those without diabetes, which could be due to in-hospital mortality. In the MEPS 2008–2012, people with diabetes reported an average LOS that was 1 day longer than those without diabetes. This was consistent across all age groups (Table 40.28). In both people with diabetes and those without diabetes, average LOS increased strongly with age. In the NHDS 2010, average LOS was strongly associated with age, whether diabetes was listed as the primary discharge diagnosis or as any listed diagnosis (Table 40.29). For many years, the average LOS as assessed by the NHDS has been decreasing. Based on NHDS discharges with diabetes as the primary diagnosis, the average LOS decreased by 42% (from 8.0 days to 4.6 days) between 1990 and 2010; for diabetes as any listed diagnosis, the average length of stay declined by 40% (from 8.6 days to 5.1 days) between 1990 and 2010 (Figure 40.54).

FIGURE 40.52

Distribution of the Number of Hospitalizations in the Past Year Among Adults With Diabetes Age ≥18 Years Who Report Being Hospitalized, by Age, U.S., 2011. Data are based on self-report. Error bars represent 95% confidence intervals. ¹ Relative (more...)

FIGURE 40.53

Distribution of the Number of Hospitalizations in the Past Year Among Adults With Diabetes Age ≥18 Years Who Report Being Hospitalized, by Age, U.S., 2008–2012. Data are based on self-report. Error bars represent 95% confidence intervals. (more...)

TABLE 40.27

Mean Length of Stay Per Hospitalization Among Adults Age ≥18 Years Who Report at Least One Hospitalization in the Past Year, by Diabetes Status and Age, U.S., 2011.

TABLE 40.28

Mean Length of Stay Per Hospitalization Among Adults Age ≥18 Years Who Report at Least One Hospitalization in the Past Year, by Diabetes Status and Age, U.S., 2008–2012.

TABLE 40.29

Number of Hospitalizations and Average Length of Hospital Stay Among Adults Age ≥18 Years With Diabetes as the Primary or Any Listed Diagnosis, by Age, U.S., 2010.

FIGURE 40.54

Time Trends in the Average Length of Stay for Hospital Discharges Listing Diabetes, U.S., 1990–2010. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 250 (all), 357.2, 362.0, (more...)

National Nursing Home Survey

The NNHS was conducted by the National Center for Health Statistics in 1973–1974 and again in 1977, 1985, 1995, 1997, 1999, and 2004. It surveyed a nationally representative sample of U.S. nursing homes and their services, staff, and residents (14). Data were collected from nursing homes with at least three beds that were either certified by Medicare or Medicaid or had a state license to operate as a nursing home. The NNHS collected data on the providers of services and the recipients of care. For recipients of care, data were obtained from medical records and included demographic information, health status, diagnoses at time of admission and at the time of interview, medication use, services received, and sources of payment.

In 2004, the NNHS was redesigned to collect data on one admission diagnosis (with one or two corresponding ICD-9-CM codes) and up to 16 possible diagnoses at the time of interview (with one or two corresponding ICD-9-CM codes each). Previously, the NNHS collected data for six ICD-9 diagnosis codes at both the time of admission and time of interview. For the data presented in this section, all listed diagnosis codes at the time of admission and the time of interview were used.

Data are reported for people with diabetes (ICD-9-CM codes 250, 357.2, 362.0, 366.41, 648.0, and 775.1) listed as any diagnosis at the time of admission or at the time of the interview. For some analyses, people with diabetes are compared to those without diabetes. Other conditions were ascertained by their ICD-9-CM codes: neoplasm (140–239); mental disorders (290–319); diseases of the nervous system and sense organs (320–389); essential hypertension (401); diseases of the circulatory system (390–459); heart disease (391–392.0, 393–398, 402, 404, 410–416, or 420–429); cerebrovascular disease (430–438), diseases of the respiratory system (460–519); diseases of the digestive system (520–579); diseases of the genitourinary system (580–629); and diseases of the musculoskeletal system and connective tissue (710–739). Analyses were restricted to recipients of care age ≥55 years.

A major limitation of the NNHS is that it used a current-resident sample, which describes people residing in the nursing home on the night before data collection. Residents with short nursing home stays are less likely to be on the nursing home rolls on a given night and are thus underrepresented in the sample. The current-resident sample is also more likely than a discharge sample to include long-term nursing home residents. In addition, for these analyses, a diagnosis of diabetes must be listed at the time of admission or at the time of interview since, by definition, residents had not been discharged.

Although the NNHS does not exclude people based on age, the majority of nursing home residents are age ≥65 years (88% in 2004). To be consistent with the previous editions of Diabetes in America, these new analyses of NNHS data have been limited to residents age ≥55 years.

National Home and Hospice Care Survey

The NHHCS was conducted by the National Center for Health Statistics in 1992, 1993, 1994, 1996, 1998, 2000, and 2007. It surveyed a nationally representative sample of U.S. home health and hospice agencies and their services, staff, and patients (15). Agencies that were either certified by Medicare and/or Medicaid or licensed by a state to provide home health and/or hospice services were surveyed. Agencies that provided only homemaker or housekeeping services, assistance with instrumental activities of daily living, or durable medical equipment and supplies were excluded. The NHHCS collected information about both the providers of services and the recipients of care. For recipients of care, data were obtained from medical records of the home health care patients and hospice discharges and included demographic information, health status, diagnoses at time of admission (one primary) and at the time of interview (one primary and up to 15 secondary), medication use, and services received. Analyses included only home health care patients. People receiving hospice care were excluded.

Data are reported for people with diabetes (ICD-9-CM codes 250, 357.2, 362.0, 366.41, 648.0, and 775.1) listed as any diagnosis either at the time of admission or the time of interview. For some analyses, people with diabetes are compared to those without diabetes. Other conditions were ascertained using ICD-9-CM codes: diseases of the blood and blood-forming organs (280–289); mental disorders (290–319); diseases of the nervous system and sense organs (320–389); diseases of the circulatory system (390–459); diseases of the respiratory system (460–519); diseases of the digestive system (520–579); diseases of the genitourinary system (580–629); diseases of the musculoskeletal system and connective tissue (710–739); and post-hospital aftercare (V42–V46, V52, V53.3–V53.7, or V54–V58).

A limitation of the NHHCS is that it used a current-patient sample, which describes people receiving home health care on the night before data collection. The current-patient sample is more likely than a discharge sample to include long-term home health care patients. Home health care patients receiving short-term care are less likely to be on the patient list on a given night and are thus underrepresented. In addition, for these analyses, a diagnosis of diabetes must be listed at the time of admission or at the time of interview since, by definition, patients had not been discharged.

Medical Expenditure Panel Survey

In addition to asking about ambulatory and inpatient care, as described in the preceding sections, the MEPS asks questions about home health care, including whether the respondent received home health care from an agency, a paid non-agency provider, or an unpaid informal caregiver, such as a spouse or child.

A limitation of using MEPS data for home health care for people with diabetes is the small sample size. When MEPS data are presented in this section, they should be interpreted as the yearly average for the time period 2008–2012. It is also important to note that the MEPS has a much broader scope than the NHHCS data. The MEPS includes three types of home health care providers, as specified above, whereas the NHHCS is much more selective, and eligible agencies had to be either certified by Medicare or Medicaid or licensed by a state to provide services.

Nursing Home Environment

In 2004, there were 16,100 licensed nursing home facilities in the United States with almost 1.5 million residents (16). There were, on average, 108 beds per nursing home with an occupancy rate of 86%. More than two-thirds (68%) of nursing homes were in metropolitan counties, and most (67%) were located in the Midwest and South. Sixty-two percent of nursing homes were owned by for-profit organizations, 31% were operated as voluntary nonprofit facilities, and 8% were owned by government and other entities.

Certified nursing assistants represented the majority of staff employed in nursing homes. Other staff included registered nurses, licensed practical nurses, nurses’ aides, and orderlies. Nursing care was usually provided according to the orders of physicians who visited the facilities. Supervising physicians included private physicians from the community (86%), physicians from contracted physician group practices (30%), or directly employed staff physicians (20%) (16). Pharmacy services and medical director services were most often provided under formal contract with outside providers (84% and 84%, respectively) (16). Other services commonly provided by outside sources included hospice, physical and occupational therapy, podiatry, dental, and diagnostic services.

Demographic Characteristics of Nursing Home Residents

According to the NNHS, 358,500 people with diabetes age ≥55 years resided in nursing homes in the United States in 2004. Table 40.30 shows the number of nursing home residents with diabetes by age and sex. Approximately 25% of nursing home residents had diabetes listed as any diagnosis at either the time of admission or the time of interview. The percentage of nursing home residents with diabetes decreased with age. Nursing home residents age ≥85 years were approximately half as likely to have diabetes as those age 55–64 years. The prevalence of diabetes was slightly higher for men (29%) than women (24%). Women tended to have higher prevalence of diabetes in the younger age groups (55–74 years), and men tended to have a higher prevalence of diabetes in the older age groups (≥75 years) (Figure 40.55).

Figures 40.56 and 40.57 show time trends in the numbers and percentage of nursing home residents age ≥55 years with diabetes by age. For the youngest age group (55–64 years), the number of people with diabetes in nursing homes increased from 10,600 in 1995 to 34,200 in 2004 (3.2 times increase). Between 1995 and 2004, the number increased 1.4 times for all other age groups. In 2004, the number of people with diabetes in nursing homes increased with age from 34,200 people age 55–64 years to 123,200 people age ≥85 years. The percentage of people with diabetes in the youngest age group (55–64 years) increased over time from 17% in 1995 to 37% in 2004 (2.2 times increase). The percentage of people with diabetes increased 1.5 times for all other age groups. In 2004, the percentage of people with diabetes in nursing homes decreased with age from 37% in people age 55–64 years to 18% in people age ≥85 years.

TABLE 40.30

Number of Adults With Diabetes Age ≥55 Years Living in Nursing Homes, by Age and Sex, U.S., 2004.

FIGURE 40.55

Prevalence of Diabetes Among Adults Age ≥55 Years Living in Nursing Homes, by Age and Sex, U.S., 2004. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 250, 357.2, 362.0, 366.41, (more...)

FIGURE 40.56

Trends in the Numbers of Nursing Home Residents With Diabetes Age ≥55 Years, by Age, U.S., 1995–2004. Diabetes is defined as ICD-9-CM codes 250, 357.2, 362.0, 366.41, 648.0, and 775.1. For 1995, 1997, and 1999, there were six ICD-9-CM (more...)

FIGURE 40.57

Trends in the Percent of Nursing Home Residents With Diabetes Age ≥55 Years, by Age, U.S., 1995–2004. Diabetes is defined as ICD-9-CM codes 250, 357.2, 362.0, 366.41, 648.0, and 775.1. For 1995, 1997, and 1999, there were six ICD-9-CM (more...)

Table 40.31 shows the characteristics of nursing home residents with and without diabetes compared to the general population with and without diabetes. Nursing home residents with diabetes were more likely to be younger (age 55–84 years), men, and nonwhite than nursing home residents without diabetes. Women with diabetes in nursing homes were less likely to be widowed, but more likely to be divorced or separated, than women without diabetes in nursing homes. Men with diabetes in nursing homes were more likely to be married or never married than men without diabetes in nursing homes. Compared to people with diabetes in the general population, nursing home residents with diabetes were more likely to be women, of non-Hispanic origin, and widowed or never married. Many of these differences are likely due to the considerably older age of people with diabetes in nursing homes compared with people with diabetes in the general population.

Living Arrangements Before Nursing Home Admission

Table 40.32 shows the living arrangements of nursing home residents before admission by diabetes status, after adjusting for age and sex. Nursing home residents with diabetes were more likely to be admitted from acute care hospitals and less likely to be admitted from private residences or assisted living/group homes than nursing home residents without diabetes. Forty-two percent of nursing home residents with diabetes were admitted from acute care hospitals, 28% from private residences, and 12% from another nursing home.

TABLE 40.31

Demographic Characteristics of Nursing Home Residents Age ≥55 Years and the General Population Among Adults Age ≥55 Years, by Diabetes Status, U.S., 2004.

TABLE 40.32

Age- and Sex-Standardized Living Arrangements Before Nursing Home Admission Among Nursing Home Residents Age ≥55 Years, by Diabetes Status, U.S., 2004.

Chronic Medical Conditions in Nursing Home Residents

Table 40.33 shows the prevalence of selected chronic medical conditions in nursing home residents age ≥55 years by diabetes status. Diseases of the circulatory system were common in residents with and without diabetes but were more common in nursing home residents with diabetes compared to those without diabetes. Eighty-nine percent of nursing home residents with diabetes also had diseases of the circulatory system (64% hypertension, 52% heart disease, and 27% cerebrovascular disease) compared to 78% of nursing home residents without diabetes (53% hypertension, 43% heart disease, and 23% cerebrovascular disease). Residents with diabetes were more likely to also have genitourinary disorders than residents without diabetes (26% vs. 18%). Residents with diabetes were less likely to have comorbid mental disorders (63% vs. 70%), diseases of the nervous system and sense organs (39% vs. 46%), and diseases of the musculoskeletal system and connective tissue (39% vs. 49%) than residents without diabetes. These differences remained after adjusting for age and sex (Figure 40.58). The prevalences of neoplasms, diseases of the respiratory system, and diseases of the digestive system were similar among residents with and without diabetes.

Figure 40.59 and Table 40.33 also show the number of chronic medical conditions (excluding diabetes) among nursing home residents. Nursing home residents with and without diabetes had similar numbers of chronic medical conditions, even after adjusting for age and sex (Figure 40.59). Forty-one percent of nursing home residents with diabetes had four or more chronic medical conditions, excluding diabetes.

TABLE 40.33

Prevalence of Selected Chronic Medical Conditions at Admission or Time of Interview Among Nursing Home Residents Age ≥55 Years, by Diabetes Status, U.S., 2004.

FIGURE 40.58

Age- and Sex-Standardized Prevalence of Selected Chronic Medical Conditions Among Nursing Home Residents Age ≥55 Years, by Diabetes Status, U.S., 2004. Diabetes is defined as ICD-9-CM codes 250, 357.2, 362.0, 366.41, 648.0, and 775.1. Other conditions (more...)

Limitations in Activities of Daily Living in Nursing Home Residents

Approximately 6% of nursing home residents with diabetes required only limited assistance with bathing, dressing, toileting, or transferring (moving from bed to a chair or back again), while 23% required limited assistance with eating and 49% require limited assistance with walking (Table 40.34). Fifty-eight percent of nursing home residents with diabetes were unable or required extensive assistance to dress themselves, 55% were unable or required extensive assistance to bathe themselves, and 19% were unable or required extensive assistance to eat by themselves; in addition, about 47% of nursing home residents with diabetes were unable or required extensive assistance to perform one or more of transferring, toileting, or walking (Figure 40.60). Forty-seven percent of nursing home residents with diabetes were unable to bathe themselves, 38% were unable to dress themselves, approximately 30% were unable to toilet or transfer by themselves, 9% were unable to eat by themselves, and 18% were unable to walk by themselves (Table 40.34). The frequencies of limitations in activities of daily living in nursing home residents were similar in every category for persons with and without diabetes.

Nursing home residents with diabetes were less likely to have difficulty controlling urine compared to nursing home residents without diabetes, but urinary incontinence was prevalent in both groups (approximately 50%). Fecal incontinence was only slightly less common than urinary incontinence (about 42%) and was similar for residents with and without diabetes.

FIGURE 40.59

Age- and Sex-Standardized Number of Chronic Medical Conditions (Excluding Diabetes) Among Nursing Home Residents Age ≥55 Years, by Diabetes Status, U.S., 2004. Data are standardized to the overall National Nursing Home Survey population using (more...)

TABLE 40.34

Frequency of Limitations in Activities of Daily Living Among Nursing Home Residents Age ≥55 Years, by Diabetes Status, U.S., 2004.

Time Since Admission

The mean and median number of days since admission for people with diabetes were 754 days and 424 days compared to 853 days and 486 days for people without diabetes, respectively. The shorter mean and median length of stays for people with diabetes are likely due to the higher competing risk of death for people with diabetes. About half of residents were institutionalized for at least the entire year before the interview (Table 40.35). Forty-eight percent of people age 55–64 years with diabetes were residents ≥1 year but were less likely to be institutionalized ≥1 year than residents without diabetes. The proportion increased for each age group, so that 62.5% of people age ≥85 years with diabetes were residents ≥1 year. Because all individuals were, by definition, current nursing home residents, time from admission does not reflect average length of stay.

FIGURE 40.60

Prevalence of Limitations in Activities of Daily Living Among Nursing Home Residents With Diabetes Age ≥55 Years, U.S., 2004. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (more...)

TABLE 40.35

Time Since Admission Among Nursing Home Residents Age ≥55 Years, by Diabetes Status and Age, U.S., 2004.

Sources of Payment and Expenditures

Table 40.36 shows the sources of payment for the first month of nursing home care and the month of care immediately before the interview for residents with and without diabetes. For both the first and most recent months of care, residents with diabetes were less likely than those without diabetes to have private sources of payment. For the first month of care for residents with and without diabetes, the most frequent non-private source of payment was Medicare. For the most recent month of care, the most frequent non-private source of payment was Medicaid.

Table 40.37 shows the distribution of annual expenditures for nursing home care paid by an individual or family by diabetes status. Self or family payment was shifted toward no payment or smaller amounts of payment for people with diabetes compared with those without diabetes. Figure 40.61 shows the mean annual expenditures for nursing home care by self/family or by other sources by diabetes status. Total expenditures were similar for residents with diabetes ($43,175) and without diabetes ($42,927). However, the proportion of expenditures paid by self/family was less for residents with diabetes (20%) compared to residents without diabetes (28%), and the absolute annual cost burden to self/family was less for diabetic residents ($8,647) than for nondiabetic residents ($12,091).

TABLE 40.36

Sources of Payment Among Nursing Home Residents Age ≥55 Years, by Diabetes Status, U.S., 2004.

TABLE 40.37

Annual Expenditures for Nursing Home Care Paid by Individuals or Families Among Residents Age ≥55 Years, by Diabetes Status, U.S., 2004.

FIGURE 40.61

Distribution of Annual Expenditures for Nursing Home Care Among Residents Age ≥55 Years, by Diabetes Status and Source of Payment, U.S., 2004. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (more...)

Home Health Care Environment

Home health care agencies provide health care in a home setting. Home health care can vary from rehabilitation care after surgery for an otherwise healthy person to end-of-life care for a terminally ill person. Home health care agencies provide long-term care for an increasing number of people. For many, home health care provides an attractive alternative to nursing home admission.

In 2007, there were 14,500 licensed home health and hospice care agencies in the United States, with most (75%) providing only home health care (17); approximately 15% provided only hospice care; and 10% provided both home health and hospice care. There were an average of 109 current home health care patients for agencies providing only home health care and an average of 178 current home health care patients for agencies providing both home health and hospice services. More than three-quarters (76%) of agencies providing home health care only and 58% of agencies providing both home health care and hospice services were in metropolitan counties. Seventy-six percent of agencies that provided home health care only were owned by for-profit organizations. Only 26% of agencies providing both home health care and hospice services were proprietary. The majority of home health care agencies provided skilled nursing services (95%), physical therapy (79%), and assistance with activities of daily living (83%). Other services commonly provided included speech therapy, occupational therapy, medical social services, and homemaker services.

Characteristics of Home Health Care Patients

According to the NHHCS, 456,395 people with diabetes age ≥18 years received home health care in 2007. Nearly one-third of all home health care recipients had diabetes. Table 40.38 shows the number and prevalence of diabetes in home health care recipients by age and sex. The prevalence of diabetes increased with age, reaching its highest prevalence (45%) in home health care recipients age 65–74 years. Thereafter, the prevalence of diabetes decreased in home health care recipients. The prevalence of diabetes was similar among men (35%) and women (32%) receiving home health care in all age groups (Figure 40.62).

According to the MEPS, the yearly average number of people with diabetes age ≥18 years receiving home health care between 2008 and 2012 was over 2 million, and 30% of all home health care recipients had diabetes. Table 40.39 shows the number of people with diabetes and prevalence of diabetes in home health care recipients by age and sex. The data from the MEPS are similar to the NHHCS data. Most people with diabetes receiving home health care receive care provided by someone who works for an agency (78%), by an informal/unpaid home health care provider (15%), or a non-agency home health care worker paid directly by the client (7%) (Table 40.40). The NHHCS and the MEPS show a similar prevalence of diabetes among men and women receiving home health care in each age group (Figures 40.62 and 40.63).

TABLE 40.38

Prevalence of Diabetes Among Current Home Health Care Recipients Age ≥18 Years, by Age and Sex, U.S., 2007.

FIGURE 40.62

Prevalence of Diabetes Among Adults Age ≥18 Years Receiving Home Health Care, by Age and Sex, U.S., 2007. Diabetes includes the following International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes: 250 (all), (more...)

TABLE 40.39

Prevalence of Diabetes Among Current Home Health Care Recipients Age ≥18 Years, by Age and Sex, U.S., 2008–2012.

TABLE 40.40

Prevalence of Diabetes Among Current Home Health Care Recipients Age ≥18 Years, by Type of Home Health Care Provider, Age, and Sex, U.S., 2008–2012.

FIGURE 40.63

Prevalence of Diabetes Among Adults Age ≥18 Years Receiving Home Health Care, by Age and Sex, U.S., 2008–2012. Diabetes status is based on self-reported diabetes diagnosis or a medical visit related to diabetes in the past year using International (more...)

Figure 40.64 shows the time trend in the percentage of home health care recipients who had diabetes by age. The prevalence of diabetes among home health care recipients has varied year-to-year but has generally increased over time for all age groups.

Table 40.41 shows the demographic characteristics of home health care recipients with and without diabetes and the general population with and without diabetes. Home health care recipients with diabetes were more likely to be age 55–84 years and nonwhite than home health care recipients without diabetes. The proportion of home health care recipients with and without diabetes who were married or widowed did not differ by diabetes status, but home health care recipients with diabetes were more likely to be divorced or separated, and a smaller proportion tended never to have married compared to people without diabetes. Home health recipients with diabetes were more likely to be women and to be unmarried compared to people with diabetes in the general population. Many of these differences are likely due to the considerably older age of people with diabetes receiving home health care compared to people with diabetes in the general population.

FIGURE 40.64

Trends in the Prevalence of Diabetes Among Patients Age ≥18 Years Receiving Home Health Care, by Age, U.S., 1996–2007. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (more...)

TABLE 40.41

Demographic Characteristics Among Home Health Care Recipients Age ≥18 Years and the General Population Age ≥18 Years, by Diabetes Status, U.S., 2007.

Living Arrangements of Home Health Care Recipients

Table 40.42 shows the living arrangements of home health care recipients before they received home health care and while they were receiving home health care, by diabetes status. People with and without diabetes did not differ according to their types of living arrangements before they received home health care. The majority of home health care recipients with diabetes (75%) and without diabetes (75%) were admitted to an emergency room or hospital before they received home health care. Other recipients lived in nursing homes or rehabilitation facilities before receiving home health care. While receiving home health care, most recipients lived in a private residence (93% of recipients with diabetes and 90% without diabetes).

Medical Conditions of Home Health Care Recipients

Figure 40.65 and Table 40.43 show the prevalence of selected chronic medical conditions in home health care recipients at the time of admission or interview. Diseases of the circulatory system were the most common chronic condition in recipients with and without diabetes but were more common in home health care recipients with diabetes (77% vs. 61%). Home health care recipients with diabetes were more likely to also have diseases of the respiratory system than home health care recipients without diabetes (21% vs. 17%). Home health care recipients with diabetes were less likely to be receiving post-hospital aftercare. This difference is likely due to otherwise healthy home health care recipients without diabetes needing care for rehabilitation or recovery after surgery, whereas home health care recipients with diabetes are more likely to use home health care as an alternative to nursing home or hospital admission. Home health care recipients with diabetes had similar numbers of chronic conditions (excluding diabetes) as home health care recipients without diabetes (Figure 40.66, Table 40.43).

TABLE 40.42

Living Arrangements Among Home Health Care Recipients Age ≥18 Years, by Diabetes Status, U.S., 2007.

FIGURE 40.65

Prevalence of Selected Chronic Medical Conditions Among Home Health Care Recipients With Diabetes Age ≥18 Years, U.S., 2007. Diabetes is defined as ICD-9-CM codes 250, 357.2, 362.0, 366.41, 648.0, and 775.1. ICD-9-CM codes for other conditions (more...)

TABLE 40.43

Number of Selected Chronic Medical Conditions Among Home Health Care Recipients Age ≥18 Years at Admission or Interview, by Diabetes Status, U.S., 2007.

FIGURE 40.66

Number of Chronic Medical Conditions (Excluding Diabetes) Among Home Health Care Recipients Age ≥18 Years, by Diabetes Status, U.S., 2007. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (more...)

Limitations in Activities of Daily Living

With the exception of toileting, home health care recipients with and without diabetes had similar needs for assistance with activities of daily living (Figure 40.67, Table 40.44). Home health care recipients with diabetes were less likely to need assistance toileting (36%) than those without diabetes (44%). Almost three-quarters of home health care recipients with diabetes (74%) needed assistance bathing, and 70% required help walking or climbing stairs. In addition, 66% of home health care recipients with diabetes who needed help walking also needed help with at least one other activity of daily living. Sixty-two percent required assistance dressing themselves, and 57% required assistance transferring out of a bed or chair. Few home health care recipients with diabetes required assistance with eating (15%). Among home health care recipients who needed any assistance in activities of daily living, home health agencies provided assistance on average 45% of the time (average percentage of all specific activities listed). This was true for persons with and without diabetes.

Approximately 43% of home health care recipients with and without diabetes had difficulties with urinary incontinence. Eleven percent of home health care recipients with diabetes and 19% of home health care recipients without diabetes had problems with fecal incontinence.

FIGURE 40.67

Prevalence of Limitations in Activities of Daily Living Among Home Health Care Recipients With Diabetes Age ≥18 Years, U.S., 2007. Diabetes is defined as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (more...)

TABLE 40.44

Frequency of Limitations in Activities of Daily Living Among Home Health Care Recipients Age ≥18 Years, by Diabetes Status, U.S., 2007.

Length of Time With Home Health Care

People with diabetes receiving home health care had been receiving it for an average of 227 days with a median of 66 days. People without diabetes receiving home health care had been receiving it for an average of 225 days with a median of 68 days. For those receiving home health care, people without diabetes tended to be older than people with diabetes (Table 40.45).

TABLE 40.45

Length of Time in Home Health Care Among Adults Age ≥18 Years Receiving Home Health Care, by Diabetes Status and Age, U.S., 2007.

Direct Medical Costs

Figure 40.68 and Table 40.46 show total health care expenditures (direct medical costs) in the United States in 2012 (20). Health care expenditures totaled over $1,334 billion, with $306 billion (23%) incurred by people with diabetes. The excess costs attributable to diabetes were $176 billion. The proportion of expenditures attributable to diabetes increased with age.

Approximately 8% of the costs attributable to diabetes were incurred by people age <45 years, 33% by people age 45–64 years, and 59% by people age ≥65 years (Figure 40.69) (20). Despite a slightly larger population of men, women had higher health care expenditures attributable to diabetes ($91 billion for women vs. $84 billion for men). Across racial/ethnic groups, non-Hispanic whites had the highest expenditures attributable to diabetes, driven in large part by the size of the population relative to the other groups.

Institutional Care. Inpatient, nursing home, and hospice care accounted for the largest proportion of the total direct medical costs of health care in the United States in 2012 (>$576 billion). Inpatient care accounted for 36%, nursing home care for 7%, and hospice care for 1% of the direct medical costs (Figure 40.70) (20). Approximately 26% of hospital inpatient costs, 32% of nursing home costs, and 13% of hospice costs were incurred by people with diabetes, and 16%, 17% and <1% of these costs were attributable to diabetes, respectively (Table 40.47) (20). The proportion of the excess costs attributable to diabetes for inpatient, nursing home, and hospice care increased with age. Approximately 6% of direct medical costs attributable to diabetes were incurred by people with diabetes age <45 years, 28% by people age 45–64 years, and 66% by people age ≥65 years. The overall pattern of excess costs attributable to diabetes by sex and race/ethnicity groups was consistent across types of institutional care.

Outpatient Care. Outpatient care accounted for the second largest proportion of total direct medical costs in the United States (nearly $446 billion) (Figure 40.70) (20). Outpatient care includes visits to physician offices, hospital outpatient clinics, and emergency departments, as well as ambulatory services, home health care, and podiatry. Approximately 16% of the total dollars spent for outpatient care were incurred by people with diabetes ($70 billion), and 46% ($32 billion) of those costs were attributable to diabetes.

People with diabetes spent about $43 billion for outpatient physician visits ($31.4 billion for physician office visits and $11.4 billion for hospital outpatient visits), $453 million for ambulance services, >$14 billion for emergency department visits, >$11 billion for home health care, and $458 million for podiatry (Table 40.48) (20). The proportion of outpatient costs attributable to diabetes increased with age, but the overall pattern was not as marked as it was for institutional care, as shown in Table 40.47 (20). People age <45 years were much more likely to incur costs related to emergency department use and podiatric care compared to other outpatient services use for that age group. For ambulance services, most of the cost attributable to diabetes was incurred by the 45–64 years age group (77%). Health care expenditures for outpatient care attributable to diabetes were similar for women and men, although expenditures for hospital outpatient and emergency department care were slightly higher for women.

TABLE 40.46

Direct Medical Costs, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

FIGURE 40.69

Total Annual Health Care Expenditures Attributable to Diabetes, by Age, Sex, and Race/Ethnicity, U.S., 2012. Hisp, Hispanic; NHB, Non-Hispanic black; NHO, Non-Hispanic other; NHW, Non-Hispanic white.

FIGURE 40.70

Direct Medical Costs of Health Care, by Type, U.S., 2012.

TABLE 40.47

Health Care Expenditures for Institutional Care, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

TABLE 40.48

Health Care Expenditures for Outpatient Care, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

Outpatient Medications and Supplies. Outpatient medications and supplies accounted for 23% of the total direct medical costs in the United States in 2012 (almost $310 billion) (Figure 40.70) (20). Approximately $83 billion (27% of the total) was incurred by people with diabetes, and $53 billion (17% of the total) was attributable to diabetes. Most of the cost associated with medications and supplies was for prescription medications (Table 40.49) (20). People with diabetes incurred $59 billion or 22% of the total cost of prescription medications and $3.6 billion or 15% of the total cost for other equipment and supplies, including eyewear, orthopedic items, hearing devices, prostheses, etc. This accounts for over one-quarter (28%) of all health care expenditures attributed to diabetes (data not shown) (20).

When examining health care expenditures for outpatient medications and supplies attributable to diabetes by age, the proportion of total dollars attributable to diabetes increased with age for each category, especially for prescription medications, where 8% was incurred by the <45 years age group, 33% by the 45–64 years age group, and 60% by the ≥65 years age group. As seen for inpatient and outpatient care, health care expenditures attributable to diabetes tend to be split fairly evenly between women and men across categories of outpatient medications. With respect to race/ethnicity, non-Hispanic whites had the highest proportion of health care expenditures for medications, largely due to the larger population size.

TABLE 40.49

Health Care Expenditures for Outpatient Medications and Supplies, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

TABLE 40.50

Health Care Expenditures for Antihyperglycemic Medications Among Adults With Diabetes Age ≥18 Years, by Age, Sex, and Race/Ethnicity, U.S., 2012.

MEPS 2012 data were analyzed to describe expenditures for antihyperglycemic, antihypertensive, and cholesterol-lowering medications by age, sex, and race/ethnicity. Table 40.50 shows that in 2012, $15.3 billon was spent on insulin and $6.7 billion was spent on non-insulin antihyperglycemic medications by adults with diabetes age ≥18 years. For insulin, 18% of the $15.3 billion was spent by those age 18–44 years, 47% by those age 45–64 years, and 33% by those age ≥65 years. For non-insulin antihyperglycemic medications, the proportion of total expenditures increased with age, with 52% of expenditures by those age ≥65 years. Men spent slightly more than women on antihyperglycemic medications, and non-Hispanic whites had the highest proportion of health care expenditures for antihyperglycemic medications.

In 2012, people with diabetes spent $8.0 billion on antihypertensive medications. Approximately $15.9 billion was spent by people without diabetes (Table 40.51). Approximately half of all expenditures for antihypertensive medications for people with and without diabetes were among those age ≥65 years. Men and women with diabetes had about equal expenditures for antihypertensive medications. Among those without diabetes, expenditures for antihypertensive medications were higher for women. Non-Hispanic whites had the highest proportion of expenditures for antihypertensive medications.

In 2012, people with diabetes spent $5.2 billion on statins and $2.7 on other cholesterol-lowering medications (Table 40.52). Most of the health care expenditures for statins and other cholesterol-lowering medications were for people age ≥45 years, with somewhat higher expenditures among men than women. Again, non-Hispanic whites had the highest proportion of health care expenditures for statins and other cholesterol-lowering medications. These trends were similar to those seen for people without diabetes.

Additional details on the types of medications used by people with diabetes can be found in Chapter 39 Medication Use and Self-Care Practices in Persons With Diabetes.

TABLE 40.51

Health Care Expenditures for Antihypertensive Medications Among Adults Age ≥18 Years, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

TABLE 40.52

Health Care Expenditures for Statins and Other Cholesterol-Lowering Medications Among Adults Age ≥18 Years, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

Per Capita Health Care Expenditures Attributable to Diabetes. Figure 40.71 shows the annual excess per capita health care expenditures attributable to diabetes in 2012 by age, sex, and race/ethnicity (20). Overall, $7,888 was spent each year per person with diabetes for care attributable to diabetes, its complications, and comorbidities. This figure was driven by the high per capita expenditures for people age ≥65 years ($11,825). Women had slightly higher per capita expenditures than men. Women spent on average $873 more per person per year than men. Although non-Hispanic whites had the highest health care expenditures attributable to diabetes overall (Tables 40.46–40.51), non-Hispanic blacks had the highest annual per capita expenditures ($9,540). Non-Hispanic whites had the next highest per capita annual expenditures ($8,101), and non-Hispanic others ($6,459) and Hispanics ($5,930) had the lowest annual per capita expenditures.

People with diabetes are older than people without diabetes and, thus, have higher average annual health care expenditures. Table 40.53 shows that after adjusting for age and sex differences between people with and without diabetes, people with diabetes have annual per capita health care expenditures that are 2.3 times higher than those without diabetes ($13,741 vs. $5,853) (20). This age- and sex-adjusted ratio is highest for hospital inpatient care (2.6) and prescription medications (2.2), lowest for other equipment and supplies (1.4), and slightly less than 2.0 for all other types of services and medications.

Table 40.54 shows the annual per capita excess health care expenditures attributable to diabetes by type of service and by age (20). For institutional care, the average annual per capita health care expenditure attributable to diabetes was $3,404 for hospital inpatient care, $662 for nursing home/residential facility care, and $1 for hospice care. For each of these categories, per capita expenditures increased with age, with the ≥65 years age group having the highest average per capita cost.

FIGURE 40.71

Annual Per Capita Health Care Expenditures (in Dollars) Attributable to Diabetes, by Age, Sex, and Race/Ethnicity, U.S., 2012. Hisp, Hispanic; NHB, Non-Hispanic black; NHO, Non-Hispanic other; NHW, Non-Hispanic white.

TABLE 40.53

Annual Per Capita Health Care Expenditures, by Diabetes Status and Type of Service, U.S., 2012.

The highest annual per capita expenditure for outpatient care was for office visits ($683 for physician office visits and $226 for hospital outpatient visits). This spending is two to three times the annual per capita expenditure for the next two categories: emergency department ($299) and home health care ($200). This likely reflects the high rates of office visits necessary for the treatment of diabetes and its complications. Ambulance services and podiatry had very low annual per capita expenditures (approximately $10 each). Per capita expenditures increased with age for physician office visits and home health. For emergency department visits, the highest per capita expenditures were for the youngest (<45 years) age group. For all other categories of outpatient care, the per capita expenditures were similar across age groups.

After hospital inpatient care, the next largest contributor to the excess annual per capita expenditure attributable to diabetes was for prescription medications. Average annual per capita expenditures for insulin and other antihyperglycemic medications were $276 and $544, respectively, and expenditures for diabetes supplies were $103. Average annual per capita expenditures attributable to diabetes for other prescription medications were $1,423. These represent expenditures for prescription medications (other than insulin and other antihyperglycemic medications) to treat the complications and comorbidities of diabetes. Average annual per capita expenditures for medications and supplies increased with age with the ≥65 years age group having almost three times higher annual per capita costs than the <45 years age group. The youngest age group (<45 years) had the highest average annual per capita expenditures for insulin, which is not unexpected as this age group has the highest proportion of people with type 1 diabetes.

Table 40.55 shows the annual per capita expenditures for antihyperglycemic medications among adults with diabetes (among those taking insulin or non-insulin antihyperglycemic agents) age ≥18 years by age, sex, and race/ethnicity based on analyses of MEPS 2012 data. The annual per capita cost of insulin among those taking insulin was $2,358, and the annual per capita cost of non-insulin antihyperglycemic agents among those taking these drugs was $416. The annual per capita expenditure for insulin decreased with increasing age, while the annual per capita expenditure for non-insulin antihyperglycemic medications increased with age. Per capita expenditures for antihyperglycemic medications were similar for men and women and across all race/ethnicity groups, with the exception of the non-Hispanic other group which had a lower than average per capita expenditure for antihyperglycemic medications.

Based on analyses of MEPS 2012 data, people with diabetes who were taking an antihypertensive medication had an excess per capita expenditure of $108 for antihypertensive medications after adjusting for age, sex, and race/ethnicity (Table 40.56). The excess per capita expenditure for antihypertensive medications was greatest in the 45–64 years age group ($142) and least in the age 18–44 years age group (-$42). Women had a slightly lower excess per capita expenditure than men, and the non- Hispanic other race/ethnicity group had the lowest excess per capita expenditure of all race/ethnicity groups (-$24).

TABLE 40.54

Annual Per Capita Health Care Expenditures Attributable to Diabetes, by Age and Type of Service, U.S., 2012.

TABLE 40.55

Annual Per Capita Health Care Expenditures for Antihyperglycemic Medications Among Adults With Diabetes Age ≥18 Years, by Age, Sex, and Race/Ethnicity, U.S., 2012.

Analyses of MEPS 2012 data showed that people with diabetes who were taking a statin or other cholesterol-lowering medication also had an excess per capita expenditure for statins and other cholesterol-lowering medications of $22 after adjusting for age, sex, and race/ethnicity (Table 40.57). The excess per capita expenditure was highest for the 18–44 years age group, followed by the 45–64 years age group and the ≥65 years age group. Women with diabetes had a higher excess per capita expenditure than men with diabetes. Excess per capita expenditure for cholesterol-lowering medications was highest among non-Hispanic blacks and lowest among non-Hispanic other, but this was likely driven by the low sample size of the non-Hispanic other group.

TABLE 40.56

Annual Per Capita Health Care Expenditures for Antihypertensive Medications Among Adults Age ≥18 Years, by Diabetes Status, Age, Sex, and Race/ Ethnicity, U.S., 2012.

TABLE 40.57

Annual Per Capita Health Care Expenditures for Statins and Other Cholesterol-Lowering Medications Among Adults Age ≥18 Years, by Diabetes Status, Age, Sex, and Race/Ethnicity, U.S., 2012.

Health Care Expenditures Associated With Complications and Comorbidities Attributable to Diabetes. Table 40.58 shows the annual excess health care expenditures in 2012 that were attributable to diabetes in billions of dollars for individual complications and comorbidities; expenditures are further stratified by hospital inpatient visits, physician office and hospital outpatient visits, and emergency department visits (20). Overall, the care of other general medical conditions attributable to diabetes accounted for approximately 50% of expenditures. In addition to diabetes itself, cardiovascular complications, including peripheral vascular disease, were the largest contributors to health care expenditures. Together, diabetes, cardiovascular and peripheral vascular comorbidities, and general medical conditions accounted for 85% of hospital inpatient expenditures attributable to diabetes, 77% of expenditures for physician visits and 85% of hospital outpatient care attributable to diabetes, and 88% of expenditures for emergency department care.

Figure 40.72 shows the percentage of total U.S. expenditures in 2012 for complications and comorbidities attributable to diabetes (20). Expenditures for hospital inpatient, hospital outpatient, physician office, and emergency department visits, and prescription medications are combined within categories of complications/comorbidities. For five conditions (neurological, cardiovascular, peripheral vascular, renal, and other), >25% of total expenditures were attributable to diabetes. Twenty-one percent of ophthalmic costs were attributable to diabetes, followed by 11% of metabolic and 7% of general medical costs.

TABLE 40.58

Health Care Expenditures Attributable to Diabetes, by Medical Condition and Type of Service, U.S., 2012.

FIGURE 40.72

Percent of Total U.S. Expenditures for Complications and Comorbidities Attributable to Diabetes, U.S., 2012. * Includes all other health care that is not a known comorbidity of diabetes.

Indirect Costs

Total indirect costs attributable to diabetes in the United States in 2012 were >$68 billion (Table 40.59) (20). Presenteeism (reduced performance at work), reduced labor force participation due to disability, and premature mortality each accounted for approximately 30% of the total. Both absenteeism (work days absent) and reduced productivity days for those not in the labor force accounted for only small percentages of the total indirect costs of diabetes (7% and 4%, respectively).

In 2012, an estimated 25 million days of work were lost due to diabetes. The cost of missing work due to diabetes was estimated to be $5 billion. The reduced performance at work due to diabetes was estimated to represent 113 million lost work days per year, accounting for a cost burden of $20.8 billion. If people with diabetes not in the workforce had similar numbers of days during which they were unable to work because of their diabetes, this would have accounted for 20 million days of reduced productivity and would have cost $2.7 billion. People with diabetes were more likely to leave the workforce because of disability related to their diabetes, and in 2012, 130 million days of productivity were lost at a cost of $21.6 billion.

Using estimates of premature mortality from the ADA publication, of all deaths in the United States in 2012, 246,000 were attributable to diabetes. This total includes 73,000 deaths with diabetes as the primary cause of death, 25,000 deaths with renal disease as the primary cause of death, 38,000 deaths with cerebrovascular disease as the primary cause of death, and 110,000 deaths with cardiovascular disease as the primary cause of death (Table 40.60) (20). The average cost of premature death declines with age, because there are fewer remaining expected working years. Nevertheless, mortality attributable to diabetes accounts for $18 billion in indirect costs, or $75,100 per death. Additional information on mortality associated with diabetes is provided in Chapter 35 Mortality in Type 1 Diabetes and Chapter 36 Mortality Trends in Type 2 Diabetes.

TABLE 40.59

Productivity Losses and Indirect Costs of Diabetes, U.S., 2012.

TABLE 40.60

Deaths and Indirect Costs of Mortality Attributable to Diabetes, U.S., 2012.

Per Capita Productivity Loss. Table 40.61 shows the average annual excess indirect cost in 2012 per person with diabetes by age, sex, and race/ethnicity (20). Annual per capita indirect cost was $2,322 for women and $3,813 for men with diabetes. For men and women alike, the total annual burden of productivity loss per person with diabetes was highest in the middle age groups (35–64 years). This is likely because they are most likely to be employed and more likely to have higher earnings. Fewer young and old people are in the workforce. There are also differences across racial/ethnic groups, with non-Hispanic blacks with diabetes having the highest annual per capita productivity loss ($3,719), followed by non-Hispanic whites ($3,155), non-Hispanic others ($2,727), and Hispanics ($2,271).