Population by Sex and Age

Between 1975 and 2015, the U.S. population grew from 216.0 million to 321.4 million and the percentage aged 65 and over increased for both males and females.

The aging of the population has important consequences for the health of the nation (18). The increase in the fraction of the population aged 65 and over, particularly those aged 85 and older (the oldest old), suggests a growing number of older Americans living longer with chronic health conditions, placing pressure on both the acute and long-term care delivery systems, and public and private payers. The rectangularization of the population pyramid occurring between 1975 and 2015 reflects longer life spans and lower birth rates, although the U.S. decline in fertility has not been as extensive as other developed countries (19).

From 1975 to 2015, the number of Americans aged 65 and over more than doubled from 22.6 million to 47.8 million (See data table for Figure 1). By 2030, it is projected that one in five Americans will be 65 or older (20). Within the population aged 65 and over, those 85 and over have experienced the most rapid growth, particularly among women. While the proportion of both men and women 85 and older increased between 1975 and 2015, in both time periods there are more women than men aged 85 and older. By 2015, 2.5% of the female population was 85 and older (4.1 million) compared to 1.4% of the male population aged 85 and over (2.2 million).

Figure 1

Population, by sex and five-year age groups: United States, 1975 and 2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig01 NOTES: Resident Population. See data table for Figure 1.

Population by Race and Ethnic Group

Between 1980 and 2015, the U.S. population became more diverse as the percentage of the population in racial and ethnic minority groups (including non-Hispanic black, Hispanic, non-Hispanic Asian, and all other non-Hispanic persons) grew among all age groups.

The racial and ethnic composition of the population has implications for the health care system because many risk factors, including health behaviors, disease prevalence, mortality rates, health insurance coverage, and access to and utilization of health services, differ substantially by race and ethnicity (Health, United States, 2016 Trend Tables). The U.S. population is becoming more ethnically and racially diverse (3). In 1980, 20.1% of the population identified as racial or ethnic minorities; in 2015, 38.4% of the population identified as racial or ethnic minorities (Figure 2). This growing diversity also has implications for the health care workforce given the importance of providing culturally competent care to all race and ethnic groups (21).

Figure 2

Population, by race and Hispanic origin and age: United States, 1980, 1990, 2000, and 2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig02 NOTES: Resident population. Persons of Hispanic origin may be of any race. Race data for (more...)

Non-Hispanic whites remain the largest racial and ethnic group, although the population of people who identified as non-Hispanic white declined among all age groups from 1980 to 2015. Diversity is highest among children and adolescents aged 0–17: between 1980 and 2015, the non-Hispanic white share of the population aged 0–17 decreased by 31% and the Hispanic share nearly tripled. By 2015, just over one-half of the child and adolescent population was non-Hispanic white and one-quarter was Hispanic. Over the same time frame, among adults aged 18–64, the non-Hispanic white share of the population decreased by 24% and the Hispanic share of the population nearly tripled. By 2015, 61.5% of the 18–64 population was non-Hispanic white and 17.3% was Hispanic. Among adults aged 65 and over, the non-Hispanic white share of the population decreased by 12% between 1980 and 2015 and the Hispanic share of the population nearly tripled. By 2015, just over three-quarters of the older adult population was non-Hispanic white, 8.8% was non-Hispanic black, and 7.9% was Hispanic.

Foreign-Born Population

Since 1970, the foreign-born share of the population residing in the United States has nearly tripled, increasing from 4.7% of the population to 13.5% of the population in 2015.

The increasing racial and ethnic diversity in the U.S. population is due in part to the immigrant population who are more likely to be from racial and ethnic minority groups. Foreign-born immigrants are often younger and healthier than native-born Americans, explained in part by the healthy immigrant effect—that those who immigrate are healthier than those who do not (22,23). However, addressing the health and health care needs of an increasingly diverse immigrant population can be challenging given different socioeconomic circumstances, immigration statuses, and federal and state policies related to access to health care (24).

Since 1970, the foreign-born share of the population residing in the United States has nearly tripled, from 4.7% to 13.5% in 2015. The distribution of country of origin of immigrants has changed between 1970 and 2015. In 1970, 61.7% of immigrants living in the U.S. were from Europe, 19.4% were from Latin America, and 8.9% were from Asia. By 2015, European immigrants made up a smaller share of the foreign-born population (11.1%), while Latin Americans accounted for more than one-half (51.1%) and Asians accounted for nearly one-third (30.6%) of all immigrants.

Figure 3

Foreign-born population: United States, selected years 1970–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig03 NOTES: Resident population for 1970–2000 and civilian noninstitutionalized population for 2010 and (more...)

Population by Poverty

During 1975–2015, children under age 18 were more likely to live in poverty than adults aged 18–64, and 65 and over.

The relationship between socioeconomic status and health is well established (25,26). Although in some cases illness can lead to poverty, more often poverty is associated with poor health (25–28). Poverty and low income are related to many aspects of health, including access to care, preventive health care, insurance coverage, and health status (26). Children and adults with income below or near 100% of poverty experience worse health outcomes compared with those with higher income levels (Tables 35, 38–46, 102–105).

During 1975–2015, the percent of children under age 18 living in poverty reached a high of 22.7% in 1993, declined to 16.2% in 2000, rose to 22.0% in 2010, and then declined to 19.7% in 2015. Between 1975 and 2015, the percent of adults aged 18–64 living in poverty increased by 35%, from 9.2% in 1975 to 12.4% in 2015. In contrast, adults aged 65 and over experienced a 42% decline in poverty, from 15.3% in 1975 to 8.8% in 2015—due in part to Social Security’s cost-of-living adjustments (COLAs), which went into effect in 1975 (29).

In 2015, the percent of persons living below 200% of the poverty level differed significantly by age and gender. Approximately 42% of children, 28% of adults aged 18–64, and 31% of adults aged 65 and over were living below 200% of poverty. Among adults aged 18–64 and 65 and over, a higher percent of women were living below 200% of poverty compared to men in 2015. In contrast, the percent of those living below 200% of poverty were similar for both boys and girls under age 18.

Figure 4

Population, by percent of poverty level and age: United States, 1975–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig04 NOTES: Percent of poverty level is based on family income and family size and composition using (more...)

Population by Urban-Rural Residence

Between 1970 and 2015, the rural (nonmetropolitan) share of the population declined almost by one-half, from 26.8% to 14.4%, while the suburban share of the population doubled from 12.3% to 24.8%.

Health differences between urban and rural communities have persisted over time (4). The rural population is generally older and poorer than the urban population (28,30). Rural residents have higher injury and smoking rates, and are more likely to lack health insurance, among other health differences (4; Table 105). In addition, the supply of physicians and other health care services differs in urban and rural areas (30). Physicians in rural areas are less likely to be specialists, and mental health providers are less available in rural areas (30). Within metropolitan areas, most of the population growth has occurred in the suburbs of large (population of 1 million or more) metropolitan areas. The population in suburbs are generally wealthier and healthier than inner city areas in large metropolitan areas (4).

Between 1970 and 2015, the percentage of the population living in metropolitan (urban) counties increased by 17% to 85.6% in 2015 and the percentage living in nonmetropolitan (rural) counties declined by 46% to 14.4% in 2015. During this same time period, the percentage of Americans living in suburbs (large fringe metropolitan counties) doubled from 12.3% to 24.8%. In 2015, more than one-half (55.6%) of the U.S. population resided in the inner cities or suburbs of large metropolitan statistical areas of 1 million or more population. Thirty percent of the population lived in medium or small metropolitan areas. An additional 8.5% of the population lived in nonmetropolitan areas with a small city or town and 5.9% lived in the most rural areas.

Figure 5

Population, by urbanization level: United States, 1970, 1980, 1990, 2000, and 2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig05 NOTES: The categories micropolitan and noncore were not used in 1970, 1980, and 1990; therefore, (more...)

Life Expectancy at Birth

The gap in life expectancy at birth between white persons and black persons persists, but has narrowed since 1975; in 2015, life expectancy was longer for Hispanic persons than for non-Hispanic white and non-Hispanic black persons.

Life expectancy is a measure often used to gauge the overall health of a population. Life expectancy at birth represents the average number of years that a group of infants would live if the group were to experience the age-specific death rates present in the year of birth (31). Differences in life expectancy among various demographic subpopulations, including racial and ethnic groups, may reflect differences in a range of factors such as socioeconomic status, access to medical care, and the prevalence of specific risk factors in a particular subpopulation (32).

During 1975–2015, life expectancy at birth in the United States increased from 68.8 to 76.3 years for males and from 76.6 to 81.2 years for females (Table 15 and data table for Figure 6). During this period, life expectancy at birth for males and females was longer for white persons than for black persons. Racial disparities in life expectancy at birth persisted for both males and females in 2015, but continued to narrow.

Life expectancy at birth was 7.1 years longer for white males than for black males in 1975, and 4.4 years longer for white males than for black males in 2015. In 1975, life expectancy at birth was 6.0 years longer for white females than for black females, and 2.8 years longer for white females than for black females in 2015.

In 2015, Hispanic males and females had the longest life expectancy at birth (79.3 and 84.3, respectively) and non-Hispanic black males (71.8) and females (78.1) had the shortest. In 2015, life expectancy at birth was 7.5 years longer for Hispanic males than for non-Hispanic black males and 6.2 years longer for Hispanic females than non-Hispanic black females.

Figure 6

Life expectancy at birth, by sex, race and Hispanic origin: United States, 1975–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig06 NOTES: Life expectancy data by Hispanic origin were available starting in 2006 and were (more...)

Infant Mortality

Between 1975 and 2015, the infant mortality rate declined 63% to reach an historic low, but the rate of decline has slowed since the mid-1990s.

Infant mortality, the death of a baby before his or her first birthday, is an important indicator of the health and well-being of a country. It is used as an indicator of maternal health, community health status, and the availability of quality health services and medical technology (33,34).

Between 1975 and 2015, the infant mortality rate decreased 63% to 5.90 infant deaths per 1,000 live births in 2015. The infant mortality rate has declined more slowly since the mid-1990s, with average annual declines of 0.8% per year during 1996–2007 and 1.8% per year during 2007–2015, compared with 4.7% per year during 1975–1982.

During 1975–2015, the neonatal mortality rate (death rate among infants under 28 days, a subset of infant mortality) declined 66% to 3.93 infant deaths per 1,000 live births in 2015, and the postneonatal mortality rate (death rate among infants 28 days through 11 months, a subset of infant mortality) declined 56% to 1.96 infant deaths per 1,000 live births in 2015. The neonatal mortality rate has declined more slowly since the mid-1990s, similar to the overall infant mortality rate, while the postneonatal mortality rate declined during 1975–1998, followed by periods of stability and decline during 1998–2015.

Although the infant mortality rate has declined for all racial and ethnic groups (Table 10), rates for infants of non-Hispanic black (10.93 in 2014) and non-Hispanic American Indian or Alaska Native (7.66 in 2014) mothers remain higher than the rates for infants of non-Hispanic white (4.89 in 2014), Hispanic (5.01 in 2014), and non-Hispanic Asian or Pacific Islander (3.68 in 2014) mothers.

Figure 7

Infant mortality rates, by infant age at death and race and Hispanic origin of mother: United States, 1975–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig07 NOTES: Infant (under 1 year of age), neonatal (under 28 days), (more...)

Leading Causes of Death

Heart disease and cancer have remained the top two leading causes of death for the past 40 years.

Cause of death rankings present the most frequently occurring causes of death and illustrate the relative burden of a specific cause of death compared to other causes (35). Because rankings measure the mortality burden relative to other causes, a rank for a specific cause of death may decline over time, even if its death rate has not changed, or its rank may remain the same over time even if its death rate is declining. Rankings may vary by age, sex, and racial and ethnic group (Tables 19 and 20).

In 1975, the five leading causes of death were heart disease, cancer, stroke, unintentional injuries, and influenza and pneumonia. In 2015, the five leading causes of death were heart disease, cancer, chronic lower respiratory diseases, unintentional injuries, and stroke. Throughout 1975–2015, heart disease and cancer remained the top two leading causes of death. The age-adjusted death rate for heart disease declined during 1975–2011, and then stabilized during 2011–2015 (23.4% of deaths in 2015). The age-adjusted death rate for cancer increased during 1975–1990, followed by periods of stability and decline during 1990–2000, and finally a steady decline during 2000–2015 (22.0% of deaths in 2015). Greater declines in heart disease than cancer mortality during 1975–2015 have narrowed the gap between heart disease and cancer deaths.

Between 1975 and 2015, stroke shifted from being the third to the fifth leading cause of death (5.2% of deaths in 2015). The age-adjusted death rate for stroke declined by 70% during 1975–2015. Unintentional injuries was the fourth leading cause of death in both 1975 and 2015 (5.4% of deaths in 2015). The age-adjusted death rate for unintentional injuries declined between 1975 and the mid-1990s, followed by increasing death rates through 2015. Between 1975 and 2015, influenza and pneumonia shifted from being the fifth to the eighth leading cause of death (2.1% of deaths in 2015). Chronic lower respiratory diseases was the third leading cause of death in 2015 (5.7% of deaths). The age-adjusted death rate for chronic lower respiratory disease decreased steadily during 2000–2015. The remaining top ten leading causes of death in 2015 include Alzheimer’s disease (4.1%); diabetes mellitus (2.9%); nephritis, nephrotic syndrome, and nephrosis (1.8%); and suicide (1.6%).

Figure 8

Leading causes of death in 1975 and 2015: United States, 1975–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig08 NOTES: Underlying causes of death are based on the International Classification of Diseases, 8th Revision (more...)

Birth Rates

Between 1975 and 2015, the birth rates among women aged 15–19, 20–24, and 25–29 declined, with the smallest percent decline among women aged 25–29, while birth rates among women aged 30–34, 35–39, and 40–44 increased.

Changing patterns in social and cultural norms, such as increases in educational attainment and contraceptive use, have contributed to the decline in birth rates among younger women and the increase in birth rates among older women (36).

The birth rate among teens aged 15–19 declined by 64% between 1991 and 2015, after a period of increasing birth rates during 1987–1991. By 2015, the teen birth rate was 22.3 live births per 1,000 females, a record low for the country. While the birth rate among women aged 20–24 has also declined during much of the period 1975–2015 to 76.8 births in 2015, the birth rate among women aged 25–29 had stable, declining, and increasing periods during 1975–2006, followed by a decline during 2006–2015 (104.3 births in 2015).

Between 1975 and 2015, the birth rate declined by 60% for teens aged 15–19, 32% for women aged 20–24, and 4% for women aged 25–29. Between 1975–2015, the birth rate among women aged 30–34 doubled (101.5 births in 2015) and the birth rate among women aged 35–39 and 40–44 more than doubled (51.8 births and 11.0 births, respectively, in 2015).

In 2015, nearly 30% of the first live-births to non-Hispanic American Indian or Alaska Native mothers (27.3%) and one-fifth of the first live-births to non-Hispanic black (19.5%) and Hispanic (21.2%) mothers occurred under age 20. More than one-half of the first live-births to Asian or Pacific Islander mothers (52.5%) and more than one-third of first live-births to non-Hispanic white mothers (34.8%) occurred at the age of 30 years and over. Non-Hispanic Asian or Pacific Islander mothers were the least likely to have their first live-birth under age 20 (2.4%) and the most likely to have their first live-birth at the age of 40 and over (3.2%).

Figure 9

Birth rates, by age of mother and age at first live-birth: United States, 1975–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig09 NOTE: See data table for Figure 9.

Current Cigarette Smoking

The prevalence of smoking declined among men and women aged 25 and over at each education level between 1974 and 2015; however, men and women with no high school diploma were more than four times as likely to smoke as those with a bachelor’s degree or higher in 2015.

Smoking is the leading cause of preventable disease, disability, and death in the United States (37). It is associated with an increased risk of heart disease, stroke, lung and other types of cancers, and chronic lung diseases (38). Between 1974 and 2015, the age-adjusted prevalence of smoking among persons aged 25 and over decreased from 36.9% to 15.6% (Table 48 and data table for Figure 10). Smoking declined among men and women at each education level during 1974–2015; however, the magnitude of annual decline in the prevalence of smoking has been smaller since the early 1990s among men with a high school education or higher and among women with some college or higher education.

In 2015, the age-adjusted percent of men aged 25 and over who were current smokers ranged from 6.6% of those with a bachelor’s degree or higher (21.7 percentage points lower than 1974) to 28.6% of those with no high school diploma (23.7 percentage points lower than 1974). Among women aged 25 and over, the age-adjusted percent of current smokers in 2015 ranged from 5.3% of those with a bachelor’s degree or higher (20.6 percentage points lower than 1974) to 22.6% of those with no high school diploma (14.0 percentage points lower than 1974).

In 2015, both men and women aged 25 and over with no high school diploma were more than four times as likely to smoke as those with a bachelor’s degree or higher. Although the difference in smoking prevalence between men and women has narrowed since 1974, men were more likely than women to smoke at each education level, except for those with some college where the prevalence of smoking was similar.

Figure 10

Cigarette smoking among adults aged 25 years and over, by sex and education level: United States, 1974–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig10 NOTES: Current cigarette smokers are defined as ever smoking (more...)

Children, Adolescents, and Adults with Obesity

The percentage of children and adolescents aged 2–19 with obesity increased from 1988–1994 to 2003–2004 and then remained stable through 2013–2014; for adults aged 20 and over, the age-adjusted percentage with obesity increased from 22.9% in 1988–1994 to 37.8% in 2013–2014.

Excess body weight in children is associated with excess morbidity during childhood and excess body weight in adulthood (39–42). Among adults, obesity is a significant risk factor for numerous chronic diseases and conditions including cardiovascular disease, diabetes, and cancer (43–46). Obesity is a major public health challenge for the United States and many other countries (47–49). During 1988–1994 through 2003–2004, the percentage of children and adolescents aged 2–19 with obesity increased from 10.0% to 17.1%, and then was stable from 2003–2004 to 2013–2014. The percentage of children and adolescents with obesity was 17.2% in 2013–2014.

For adults aged 20 and over, the age-adjusted percentage with obesity increased steadily from 22.9% in 1988–1994 to 37.8% in 2013–2014. For adults aged 20 and over, the age-adjusted percentage with grade 1 obesity increased from 14.8% in 1988–1994 to 20.0% in 2003–2004, and then stabilized. In 2013–2014, the age-adjusted percentage of adults with grade 1 obesity was 20.7%. The age-adjusted percentage of adults with grade 2 obesity increased steadily from 5.2% in 1988–1994 to 9.5% in 2013–2014. The age-adjusted percentage of adults with grade 3 obesity more than doubled from 2.9% in 1988–1994 to 7.6% in 2013–2014.

Figure 11

Obesity among children and adolescents aged 2–19 and adults aged 20 years and over: United States, 1988–1994 through 2013–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig11 NOTES: For children and adolescents (more...)

Untreated Dental Caries

Untreated dental caries increases with decreasing family income for children and adolescents aged 5–19 and for adults aged 20 and over.

Oral health is integral to an individual’s overall health and well-being (50). The presence of dental caries, or tooth decay, is one of the most common chronic conditions in the United States (50). Possible causes of dental caries include high levels of bacteria in the mouth and frequent consumption of high-sugar foods and drink (51). Preventive measures include community water fluoridation, limiting consumption of high-sugar foods and drink, use of dental sealants, and regular oral care (50,52,53). Utilization of dental visits varies with family income, those with lower family income being less likely to have had a recent dental visit (Table 78) and more likely to delay needed dental care due to cost (Table 63) than those with higher family income.

The percentage of children and adolescents aged 5–19 with untreated dental caries declined steadily from 1988–1994 to 2011–2014 for those living below 100% of the poverty threshold, while remaining stable from 1988–1994 to 1999–2004, and then declining for those at 100%–199% of poverty. The percentage of children and adolescents aged 5–19 with untreated dental caries remained stable over the entire time period in the two highest income groups. In 2011–2014, among children and adolescents aged 5–19, the percentage with untreated dental caries ranged from 9.1% for those at 400% or more of poverty to 24.7% for those living below poverty. Between 1988–1994 and 2011–2014, the difference between the highest and lowest percentage of untreated dental caries narrowed from 28.6 percentage points to 15.6 percentage points (16).

The age-adjusted percentage of adults with untreated dental caries was similar in 1988–1994 and 2011–2014 for all income levels. Similar to the pattern for children and adolescents, untreated dental caries among adults is inversely related to family income. During 1988–1994 through 2011–2014, among adults aged 20 and over, the age-adjusted percentage of untreated dental caries was highest among those living below 100% of the poverty threshold (49.7% in 2011–2014) and lowest among those at 400% or more of poverty (13.3% in 2011–2014). The disparity in untreated dental caries between the highest and lowest poverty groups remained stable between 1988–1994 and 2011–2014 (36.4 percentage points in 2011–2014).

Figure 12

Untreated dental caries among children and adolescents aged 5–19 years and adults aged 20 and over, by percent of poverty level: United States, 1988–1994, 1999–2004, and 2011–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig12 (more...)

Diabetes

In both 1988–1994 and 2011–2014, the prevalence of diabetes was higher among non-Hispanic black and Mexican origin adults compared to non-Hispanic white adults.

Diabetes is a group of conditions in which insulin is not adequately secreted or utilized by the body (54). Long-term complications of high glucose levels and diabetes include cardiovascular disease, renal failure, nerve damage, and retinal damage (54,55). Diabetes was the 7th leading cause of death in 2015 (Figure 8). Treatment guidelines for diabetes include recommendations for dietary modifications, physical activity, weight loss (if overweight), and the use of medication (54,55). In order to manage the disease and avoid or delay long-term complications, ongoing medical care is recommended (56).

Between 1988–1994 and 2011–2014, the overall prevalence of total diabetes increased 35% to 11.9%, but among racial and ethnic groups the increase was significant only for non-Hispanic white adults. The prevalence of physician-diagnosed diabetes increased among non-Hispanic white and black adults from 1988–1994 to 2011–2014 and did not increase significantly among Mexican origin adults. The prevalence of undiagnosed diabetes decreased among non-Hispanic white and black adults over this time frame while remaining stable among Mexican origin adults.

In 2011–2014, approximately 1 in 9 adults in the United States had diabetes. The prevalence of total diabetes was higher among non-Hispanic black, non-Hispanic Asian, and Mexican origin adults than non-Hispanic white adults. The prevalence of both diagnosed and undiagnosed diabetes also was higher among non-Hispanic black, non-Hispanic Asian, and Mexican origin adults than non-Hispanic white adults.

Figure 13

Diabetes prevalence among adults aged 20 years and over, by diagnosis status and race and Hispanic origin: United States, 1988–1994 and 2011–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig13 ^‡ Estimates (more...)

Uncontrolled Hypertension

During 1988–1994 through 2011–2014, the age-adjusted percentage of uncontrolled high blood pressure among adults aged 20 and over with hypertension decreased for all adults and for men and women separately; in 2011–2014, the prevalence of uncontrolled high blood pressure among men and women with hypertension varied by age.

Hypertension is an important risk factor for cardiovascular disease, stroke, kidney failure, and other health conditions (57,58) and can lead to premature death (59,60). In 2011–2014, 84.1% of adults with hypertension were aware of their status, and 76.1% were taking medication to lower their blood pressure (61).

During 1988–1994 through 2011–2014, the age-adjusted percentage of adults aged 20 and over with hypertension who had uncontrolled high blood pressure decreased steadily from 77.2% to 52.8%. Among men with hypertension, the age-adjusted percentage with uncontrolled high blood pressure decreased steadily from 83.2% in 1988–1994 to 58.1% in 2011–2014. Among women with hypertension, the age-adjusted percentage with uncontrolled high blood pressure declined from 68.5% in 1988–1994 to 45.5% in 2011–2014. However, for women, the magnitude of the decline changed over this period. During 1988–1994 through 2011–2014, among those with hypertension, the age-adjusted percentage of uncontrolled hypertension was higher for men than for women.

In 2011–2014, the pattern by age for uncontrolled high blood pressure among those with hypertension differed among men and women. Among men with hypertension, the prevalence of uncontrolled high blood pressure was higher for those in the two youngest age groups than for those of other ages. Among women with hypertension, the prevalence of uncontrolled hypertension was higher for women aged 75 and over than for women aged 35–44, 45–64, and 65–74 years; the prevalence of uncontrolled hypertension was not significantly different for women in the youngest and oldest age groups.

Figure 14

Uncontrolled high blood pressure among adults aged 20 and over with hypertension, by sex and age: United States, 1988–1994 through 2011–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig14 NOTES: Uncontrolled (more...)

Prescription Drugs

In 2013–2014, 36.5% of adults aged 18–44, 69.6% of adults aged 45–64, and 90.8% of those aged 65 and over took a prescription drug in the past month—up from levels in 1988–1994.

Prescription drug use over the past 40 years has been affected by many factors, including medical need, prescription drug development, increased direct-to-consumer advertising, and expansions in health insurance and prescription drug coverage (62–64). Even though Americans are now living longer lives, a greater fraction of older Americans are living with several chronic conditions that may require multiple medications. As prescription drug use increases, however, so do concerns about polypharmacy. Polypharmacy—which is commonly defined as taking five or more drugs—increases the risk of drug interactions, adverse drug events, nonadherence, and reduced functional capacity (65).

Between 1988–1994 and 2013–2014, the use of at least one prescription drug in the past 30 days increased 5.2 percentage points for adults aged 18–44, 14.8 percentage points for adults aged 45–64, and 17.2 percentage points for adults aged 65 and over. For adults aged 45–64, use of at least one prescription drug during the past 30 days increased throughout the period, while for adults aged 18–44 and 65 and over, use initially increased before remaining stable in recent years. For adults aged 18–44, use of at least one prescription drug remained stable from 2007–2008 to 2013–2014, while for adults aged 65 and over, use of at least one prescription drug remained stable from 2003–2004 to 2013–2014.

Between 1988–1994 and 2013–2014, the percent of adults reporting the use of five or more prescription drugs in the past 30 days rose—by 2.7 percentage points for adults aged 18–44, 12.8 percentage points for adults aged 45–64, and 28.4 percentage points for adults aged 65 and over. In contrast, the percentage of adults reporting the use of one to four prescription drugs between these two periods remained stable for adults aged 18–44 and 45–64, while decreasing for adults aged 65 and over.

Figure 15

Prescription drug use in the past 30 days among adults aged 18 and over, by age and number of drugs taken: United States, 1988–1994 through 2013–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig15 NOTES: Respondent-reported (more...)

Health Provider Visits

Between 1997 and 2015, the percent of persons with a health care visit in the past 12 months increased for each of five provider types among children and adults aged 65 and over. Among adults aged 18–64, the percent with a visit to a dentist remained stable and the percent with a visit to the other four provider types increased.

Visits to health providers are influenced by a variety of factors including patient characteristics, supply and distribution of providers, and health care affordability. Since the 1990s, shifts in disease prevalence have increased the need for chronic care management, while expansions to health insurance coverage have rendered health services more affordable for those previously uninsured (66,67). Increases in the supply of selected provider types have also increased the potential for utilization, although geographic differences in the distribution of providers may create disparities in utilization by urban/rural status (68,69).

In 1997, 2006, and 2015, the percent of persons with one or more visits to generalist physicians in the past year was higher than the percent with visits to specialist physicians, eye doctors, or mental health providers among children aged 2–17, adults aged 18–64, and adults 65 and over. The percent of persons with a mental health provider visit was lower than that for any other type of provider; in 2015, 8.7% of children, 8.8% of adults aged 18–64, and 4.8% of those 65 and over reported a visit to a mental health provider in the past year.

Among all three age groups, the percent with a health care visit in the past year increased overall between 1997 and 2015 for each of the provider types shown with the exception of dental visits among adults aged 18–64, which remained stable. After a decrease in the percent of adults aged 18–64 with a dental visit between 1997 and 2006 (from 64.1% to 62.4%), the percent increased to 64.0% between 2006 and 2015.

Figure 16

Health care visits in the past 12 months among children aged 2–17 and adults aged 18 and over, by age and provider type: United States, 1997, 2006, and 2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig16 NOTE: See data (more...)

Emergency Departments

The percent of children and adults aged 18–64 with an ED visit in the past year decreased during all or recent parts of the period 1997–2015, regardless of insurance coverage, with the exception of ED use among adults without health insurance, which remained stable.

Emergency departments (EDs) provide services to individuals with emergency health needs or who are seeking after-hours care. Additionally, EDs provide a safety net for those who have difficulty accessing alternative sources of care (70,71). Over the past 40 years, many health policies have impacted ED use. For instance, the 1986 Emergency Medical Treatment and Labor Act required providers to administer emergency care regardless of the patient’s ability to pay, while the 1997 Balanced Budget Act instructed some managed care plans to pay for emergency services as long as a “prudent layperson” standard was met (72,73).

Among children, the percent with an ED visit in the past year has decreased in recent years, although the year in which the decrease began varied by type of insurance. For children with Medicaid, ED use decreased an average of 1.0 percentage point per year from 2009–2015; for children with private insurance, ED use decreased an average of 0.5 percentage points per year from 2002–2015; and for children without health insurance, ED use decreased an average of 0.2 percentage points per year from 1997–2015.

Among adults aged 18–64, the percent with an ED visit decreased for those with Medicaid coverage as well as those with private insurance. Among adults with Medicaid, ED use remained stable from 1997–2010 before decreasing an average of 1.0 percentage point per year from 2010–2015, and among adults with private insurance, ED use decreased an average of 0.1 percentage points per year from 1997–2015. In contrast, ED use among adults without health insurance remained stable throughout 1997–2015.

From 1997–2015, children and adults with Medicaid coverage were more likely than those with private or no health insurance to have an ED visit in the past year. In 2015, 22.8% of children with Medicaid had a recent ED visit compared with 14.3% of children without health insurance and 12.5% of children with private insurance. For adults aged 18–64 in 2015, 34.8% of those with Medicaid had a recent ED visit, compared with 14.0% of those with private insurance and 18.2% of those without health insurance.

Figure 17

Emergency department visits in the past 12 months for persons under age 65, by age and type of coverage: United States, 1997–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig17 NOTE: See data table for Figure 17.

Hospital Stays

The percentage of persons with an overnight hospital stay was lower in 2015 than in 1975 for males and females under age 75, and was not significantly different in 2015 than in 1975 for males and females aged 75 and over.

The U.S. population is aging, and there is a concurrent increase in the prevalence of chronic conditions, both of which suggest an increasing demand for hospital care (74) (Figure 1; Table 39). However, payment changes by private insurers and government programs have tended to reduce the number and length of hospitals stays. These efforts include Medicare’s prospective payment system which reduced payments, and managed care policies, such as physician financial incentives and utilization review, designed to reduce the use of hospital care (75) (Figure 25). In addition, technological innovations and changes in practice patterns and patient preferences have shifted some formerly hospital inpatient procedures to outpatient settings, further reducing inpatient hospital stays (66,76).

Among males in all age groups and females aged 18–44 and 65–74 the percentage of persons with a hospital stay had periods of stability and decline during 1975–2015. Among females aged 1–17 the percentage with a hospital stay declined throughout 1975–2015, while among females aged 45–64 and 75 and over the percentage with a hospital stay increased for specific time periods during 1975–2010; however, focusing on more recent trends, the percentage with a hospital stay has declined since 2010 for women aged 45–64 and since 2003 for women aged 75 and over. For men and women under age 75 the percentage with a hospital stay was lower in 2015 than in 1975; for men and women aged 75 and over, the difference between the percentage with one or more hospital stays in 1975 and 2015 was not statistically significant.

In 2015, the percentage of the population with at least one hospital stay in the past 12 months was similar for boys and girls under 18 and for men and women aged 45–64. Hospital use was higher among men than women, aged 65–74 and aged 75 and over. Women of childbearing age (18–44) were almost three times as likely to have had a hospital stay in the past year as men in this age group.

Figure 18

Overnight hospital stays in the past 12 months, by sex and age: United States, 1975–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig18 NOTES: Hospitalizations include those relating to deliveries. Because persons who (more...)

Use of Mammography and Colorectal Tests and Procedures

While use of mammograms and use of colorectal cancer tests and procedures have increased for all racial and ethnic groups, disparities in utilization persist in 2015.

Although breast cancer and colorectal cancer remain among the leading causes of cancer deaths in the U.S., advancements in and increased utilization of cancer screening tests have, in part, contributed to decreasing cancer death rates since the 1980s (77,78). During the 1990s, initiatives such as the National Breast and Cervical Cancer Early Detection Program increased educational outreach and financial subsidies for breast and cervical cancer screening among the population’s most vulnerable women (79). Periodic screening increases the likelihood of detecting cancers at earlier stages and enables less invasive treatment, though the recommendations—which include the age of initiation and interval between screenings—have changed over time (80–82).

During 1987–2015, the percent of women aged 40–74 with a mammogram in the past two years increased through the mid-1990s for all four racial and ethnic groups before either continuing to increase (non-Hispanic Asian women), decreasing (non-Hispanic white women), or stabilizing (Hispanic and non-Hispanic black women) in recent years. In 2015, the percent of women aged 40–74 with a mammogram in the past two years ranged from approximately 63% for non-Hispanic Asian women and Hispanic women to 72.3% for non-Hispanic black women. (See data table for Figure 19.)

The percent of adults aged 50–75 with a colorectal cancer test or procedure—either a fecal occult blood test in the past year, a sigmoidoscopy in the past 5 years with fecal occult blood test (FOBT) in the past 3 years, or a colonoscopy in the past 10 years—approximately doubled between 2000 and 2015 for all four racial and ethnic groups. However, racial and ethnic disparities in utilization remain; in 2015, 52.1% of non-Hispanic Asian and 47.4% of Hispanic adults aged 50–75 had a colorectal cancer test, compared with 60.3% of non-Hispanic black and 65.6% of non-Hispanic white adults aged 50–75.

Figure 19

Mammography use and colorectal cancer testing use, by race and Hispanic origin: United States, selected years 1987–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig19 NOTES: Prior to 1999, data were tabulated according (more...)

Hospitals

The number of community hospital beds per 1,000 resident population fell by almost one-half and average length of stay fell by almost one-third between 1975 and 2014 to 2.5 beds per population and 5.5 days per hospital stay in 2014.

Hospital care has undergone substantial changes in the past 40 years (76). This transformation has been driven by a variety of forces, including payment system reforms, technological advancements in medical care, modifications to practice patterns, and changes in consumer preferences (66). In 1983, Medicare implemented a prospective payment system for hospital care which paid a flat rate per case based on diagnosis instead of reimbursing costs (75). Increased managed care penetration, especially in private plans, put further pressure on hospitals to manage costs. Hospitals responded to managed care penetration by consolidating (83). In the past four decades, technological innovations have reduced recovery times and allowed some formerly hospital inpatient procedures, such as laparoscopic surgery and cataract removal, to take place in outpatient settings (66).

Changes in practice patterns and consumer preferences have also shifted some care from hospitals to alternative sites, such as post-acute care settings, nursing homes, and home health care (66). These forces have combined to encourage efficiency in hospital care and reduced incentives for longer hospital stays.

Between 1975 and 2014, community hospital beds per resident population, average length of stay, and occupancy rate declined. As a result of closures and consolidation, the number of community hospitals has also declined by 16%, from 5,875 in 1975 to 4,926 in 2014 (Table 89). During the same time frame, the number of community hospital beds per 1,000 resident population was almost cut in half from 4.6 to 2.5 (Figure 20). The average length of stay declined from 7.7 days in 1975 to 5.5 days in 2014. Occupancy rates for community hospitals averaged 62.8% in 2014—a 16% decline from 1975, although much of the decline took place during the mid-1980s.

Figure 20

Community hospital beds, average length of stay, and occupancy rate: United States, selected years 1975–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig20 NOTE: See data table for Figure 20.

Active Primary Care Generalist and Specialist Physicians

Between 1975 and 2013, the number of primary care generalist physicians per 10,000 population increased by 53% and the number of specialists per 10,000 population increased by 90%.

An adequate supply of physicians is necessary to provide quality health care. In addition to the number of physicians, the geographic distribution and specialty mix affect the adequacy of the supply. The need for physician services is growing, due to population growth, higher utilization due to higher rates of insurance coverage, and the aging population (84). Research suggests that the supply of primary care physicians is not keeping up with this demand (85–87).

Since 1975, the overall supply of active physicians in the United States nearly doubled, from 15.5 to 27.0 physicians per 10,000 population in 2013. Between 1975 and 2013, the number of primary care generalist physicians per 10,000 population increased by 53% and the number of specialists per 10,000 population increased by 90%. Because of the faster growth in specialist physicians compared with primary care generalist physicians during this period, the percentage of physicians who were specialists increased from 57.4% in 1975 to 62.6% in 2013.

Figure 21

Active primary care generalist and specialist physicians, by self-designated specialty: United States, selected years 1975–2013. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig21 NOTES: Primary care generalist physicians (more...)

Nursing Homes

From 1977 to 2014, the number of nursing home residents aged 65 and over per 1,000 population aged 65 and over fell by about one-half.

Long-term care (LTC) services fill a crucial role by delivering needed health care, personal care, housing, and supportive services to those with chronic conditions, disabilities, and, especially, frail older persons with age-related conditions (66). Long-term care services are available in several different settings, including the home and other residential care settings, the community, and institutions. Long-term care services are provided by adult day care centers and home health agencies; residential care settings, such as assisted living or continuing care communities; in-home or in-facility hospice care organizations; and by nursing homes (88). Until recently, nursing home care has been a key component of LTC, especially for older adults. However, although the U.S. population aged 65 and over increased from 10.6% to 14.9% in 2015 (18,89–91; Figure 1), use of nursing home care began to decline as early as 2000 (Figure 22)(66,92). A variety of factors likely contributed to this ongoing decline, including changes in consumer care preferences and the availability of additional long-term care options with the growth of residential care communities, such as assisted living (88,92,93).

Figure 22

Nursing home beds, residents, and occupancy rate: United States, selected years 1977–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig22 NOTE: See data table for Figure 22.

As a result of closures and consolidation, the number of nursing homes declined by 15% between 1977 and 2014 (94,95). From 1977 to 2014, the number of nursing home beds per 1,000 population aged 65 and over, and the number of nursing home residents aged 65 and over per 1,000 population aged 65 and over declined, as did the nursing home occupancy rate (that is, number of nursing home residents per 100 nursing home beds). The number of nursing home beds per 1,000 population aged 65 and over decreased by 40%, from 59.7 in 1977 to 36.0 beds per 1,000 population aged 65 and over in 2014. The number of nursing home residents aged 65 and over per 1,000 population aged 65 and over declined by 47% from 47.1 in 1977 to 25.2 residents per 1,000 population aged 65 and over in 2014. The nursing home occupancy rate fell by 11% to 82.3%.

Source of Funds and Type of Expenditure

Between 1975 and 2015, U.S. health care spending changed in terms of who paid for care and the type of care that was paid for.

The United States spent $2.7 trillion on personal health care for an average of $8,468 per person in 2015; in contrast in 1975, the United States spent $113 billion, on average $514 per person (96). In 2015, 17.8% of the U.S. Gross Domestic Product (GDP) was spent on national health care—more than twice the percentage in 1975 (7.9% of GDP) (96). More is spent on health care in the United States, in terms of a percentage of GDP, than any other developed country for which data are collected by the Organisation of Economic Cooperation and Development (OECD) (97).

Between 1975 and 2015, the share of personal health care expenditures paid by private health insurance increased from 24.5% to 34.8% and the share paid by Medicare increased from 13.8% to 22.3%. In addition, the share the federal government paid for Medicaid increased from 6.2% to 11.6% and the share that states paid for Medicaid increased from 5.1% to 6.7%. During the same period, the share paid out of pocket by consumers decreased from 32.9% to 12.4%, and the remaining share paid by other sources decreased from 17.5% to 12.2%.

Between 1975 and 2015, the share of personal health care expenditures paid for hospital care decreased from 45.3% to 38.1%, the share for nursing care facilities and continuing care retirement communities decreased from 7.1% to 5.8%, and the share for dental services decreased from 7.1% to 4.3%. During the same period the share of personal health care expenditures paid for prescription drugs increased from 7.1% to 11.9%, for home health care increased from 0.5% to 3.3%, and for other types of care increased from 10.6% to 13.2%. During this period, about one-quarter (22.4%–23.4%) of personal health care expenditures were paid for physician and clinical services.

Figure 23

Personal health care expenditures, by source of funds and type of expenditure: United States, 1975, 1995, and 2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig23 NOTES: Personal health care expenditures are outlays for goods (more...)

Mental Health and Substance Use Disorder Expenditures

Between 1986 and 2014, the share of mental health and substance use disorder spending for inpatient services decreased, while the share for outpatient services and retail prescription drugs increased.

Mental health and substance use disorders are serious, potentially disabling or fatal, and costly on a personal and societal level (98–101; Tables 27 and 30). Mental illness and substance use disorders affect a significant segment of the U.S. population (102,103). Mental health and substance use disorders are amenable to treatment (104–107), however, treatment can be costly and not always readily available (108–113). The Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008 (114) and the ACA (115) included provisions to expand access to mental health and substance abuse coverage and treatment. In 2013, 14.6% of noninstitutionalized adults received any type of mental health treatment and 1.3% received any type of substance use treatment (102).

In 2014, $186 billion was spent on mental health treatment, representing 6.4% of all health spending^* (116). Between 1986 and 2014, the share of mental health expenditures paid for inpatient care decreased from 41% to 16%; residential treatment decreased from 22% to 12%; outpatient treatment increased from 24% to 35%; and retail prescription drug spending increased from 8% to 27%.

In 2014, $34 billion was spent on substance use disorder treatment, representing 1.2% of all health spending^*. Between 1986 and 2014, the share of substance use disorder expenditures paid for inpatient care decreased from 50% to 19%; outpatient treatment increased from 27% to 40%; residential treatment increased from 17% to 27%; and retail prescription drug spending increased from less than 1% to 5%.

Figure 24

Mental health and substance use disorder expenditures, by type of expenditure: United States, selected years, 1986–2014. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig24 NOTE: See data table for Figure 24.

Footnotes

*

All health spending is defined here as health consumption expenditures (national health expenditures minus investments for structures and research). See Table 94.

Medicare Managed Care Enrollment

In 2015, more than three in ten (31.3%) Medicare beneficiaries were enrolled in Medicare managed care plans, with participation varying substantially across states.

Medicare provides health insurance coverage to most persons aged 65 and over, in addition to those under age 65 with long-standing disabilities and certain serious medical conditions. In 2015, the Medicare program (117) covered 55.3 million persons (118), with the majority enrolled in the traditional Medicare program (Part A and Part B) (119–121). Annually, nearly all Medicare beneficiaries have the choice to enroll (or disenroll) in Medicare managed care programs offered in their area (originally referred to as Medicare Part C or Medicare + Choice, and now called Medicare Advantage [MA]). MA plans provide the same coverage as the traditional Medicare program, generally through a restricted care network, and may offer extra coverage including vision, dental, and prescription drug coverage.

Enrollment in MA plans is affected by the availability of a plan in the enrollee’s county (which was limited in the early years of the program, especially in rural areas) and by having premiums and cost sharing low enough, or benefits generous enough, to attract beneficiaries from traditional Medicare (122–124). Increased participation in MA programs in the mid-to-late 1990s mimicked the general growth in managed care participation and then in the late 1990s MA participation fell with wide-spread managed care “backlash” against managed care cost containment practices, such as prior authorization (123,125). In 2003, MA payments were increased and enrollment grew (120). In recent years, MA plan choices have become increasingly varied and include health maintenance organizations (HMOs), preferred provider organizations (PPOs), private fee-for-service plans (PFFS), special needs, HMO point-of-service plans, and Medical Savings Account plans (126).

Nationwide between 1994 and 1999, the percentage of Medicare enrollees participating in managed care plans more than doubled from 7.9% to 18.2%, and then decreased steadily to 13.0% by 2004. Between 2004 and 2015, the percentage of enrollees in managed care plans more than doubled, increasing from 13.0% to 31.3%.

In 2015, the percentage of Medicare enrollees participating in managed care plans varied across states, ranging from 1.0% in Alaska to 53.8% in Minnesota. The 13 states in the top quartile with the highest percentage of Medicare managed care enrollees in 2015 were: Tennessee, Rhode Island, Colorado, New York, Wisconsin, Arizona, Florida, Pennsylvania, California, Ohio, Oregon, Hawaii, and Minnesota.

Figure 25

Medicare enrollees in managed care: United States, 1994–2015. Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig25 NOTE: See data table for Figure 25.

Coverage Among Children Under Age 18

Between 1978 and September 2016 (preliminary data), the percentage of children under age 18 who were uninsured decreased; the percentage with Medicaid coverage increased, while the percentage with private coverage decreased.

Children and adolescents need regular and ongoing healthcare to provide routine preventive care, such as age-appropriate vaccinations; to offer health and developmental guidance; to screen for health conditions; to control and to treat acute and chronic conditions; and to provide injury care (127). Historically, children have been more likely than adults (Figure 27) to have coverage primarily because they have been more likely to qualify for Medicaid, enacted in 1966 (128). Starting in 1997, the Children’s Health Insurance Program has provided coverage to eligible low-income, uninsured children who do not qualify for Medicaid (129).

Between 1978 and September 2016 (130,131) the percentage of uninsured children decreased 7.0 percentage points from 12.0% to 5.0%; Medicaid coverage increased 27.9 percentage points from 11.3% to 39.2% while private coverage decreased 21.6 percentage points from 75.1% to 53.5%; private workplace coverage, which is the largest component of private coverage, decreased 18.4 percentage points from 67.6% to 49.2%. For children, the percentage uninsured increased 0.4 percentage points per year on average during 1978–1990, and then decreased 0.4 percentage points per year on average during 1990–September 2016. The decrease in the percentage uninsured starting in 1990 was primarily due to increases in Medicaid coverage during most of the period 1990–2012. Increases in Medicaid coverage were larger than decreases in private coverage during that period.

Figure 26

Health insurance coverage among children under age 18, by type of coverage: United States, selected years 1978–September 2016 (preliminary data). Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig26 NOTES: Preliminary estimates (more...)

Coverage Among Adults Aged 18–64

Between 2010 and September 2016 (preliminary data), the percentage of adults aged 18–64 who were uninsured decreased, while both private and Medicaid coverage increased.

Adults aged 18–64 were historically more likely to be uninsured than children and adolescents because they were less likely to qualify for public coverage, primarily Medicaid (Figure 27) (128,132,133). Passage of the Affordable Care Act (ACA) in 2010 (115) authorized states to expand Medicaid eligibility to low income adults to 138% of the federal poverty level (134) and established the health insurance marketplace in 2014. The health insurance marketplace offered deductible and copayment subsidies for people at 139%–250% of the federal poverty level and tax credits for people at 139%–400% of the federal poverty level without offers of affordable coverage through an employer. However, in states that did not expand Medicaid, subsidies and premiums are offered to those at 100%–250% and 100%–400% of the federal poverty level, respectively.

For adults aged 18–64, the percentage uninsured increased from 11.9% in 1978 to 18.8% in 1993 and 22.3% in 2010, followed by a decline of 10 percentage points to 12.3% in September 2016. Both private and Medicaid coverage increased during 2010–September 2016. Overall, between 1978 and September 2016, private coverage decreased 12.4 percentage points from 81.4% to 69.0%; private workplace coverage, the largest component of private coverage, decreased 10.5 percentage points from 71.4% to 60.9%; and Medicaid coverage increased 9.7 percentage points from 4.4% to 14.1%. Despite changes in the percent uninsured during 1978–September 2016, the percent uninsured was similar in 1978 and September 2016 (11.9% and 12.3%, respectively).

Figure 27

Health insurance coverage among adults aged 18–64, by type of coverage: United States, selected years 1978–September 2016 (preliminary data). Excel and PowerPoint: http://www.cdc.gov/nchs/hus/contents2016.htm#fig27 NOTES: Preliminary estimates (more...)

Data Sources

Data for the Health, United States, 2016, Chartbook come from many surveys and data systems and cover a broad range of years. Detailed descriptions of the data sources included in the Chartbook are provided in Appendix I. Data Sources. Additional information clarifying and qualifying the data is included in the table notes and in Appendix II. Definitions and Methods.

Data Presentation

Many measures in the Chartbook are shown for people in specific age groups because of the strong effect of age on most health outcomes. In some cases, age-adjusted rates and age-adjusted percentages are computed to eliminate differences in observed rates that result from age differences in population composition (see Appendix II, Age adjustment). Age-adjusted rates and age-adjusted percentages are noted as such in the text; rates and percentages without this notation are crude rates and crude percentages. For some charts, data years are combined to increase sample size and the reliability of the estimates. Some charts present time trends, and others focus on differences in estimates among population subgroups for the most recent time period available. Trends are generally shown on a linear scale to emphasize absolute differences over time.

Point estimates and standard errors for Chartbook figures are available either in the Trend Table and Excel spreadsheet specified in the note below the chart, or in the Chartbook data tables section. Chartbook data tables may include additional data that were not graphed because of space considerations.

Reliability of Estimates

Overall estimates generally have relatively small sampling errors, but estimates for certain population subgroups may be based on small numbers and have relatively large sampling errors. Numbers of deaths obtained from the National Vital Statistics System represent complete counts and therefore are not subject to sampling error. They are, however, subject to random variation, which means that the number of events that actually occur in a given year may be considered as one of a large series of possible results that could have arisen under the same circumstances. When the number of events is small and the probability of such an event is small, considerable caution must be observed in interpreting the conditions described by the charts. Estimates that are unreliable because of large sampling errors or small numbers of events have been noted with an asterisk. The criteria used to designate or suppress unreliable estimates are indicated in the notes to the applicable tables or charts.

For NCHS surveys, point estimates and their corresponding variances were calculated using the SUDAAN software package, which takes into consideration the complex survey design (135). Standard errors for other surveys or data sets were computed using the methodology recommended by the programs providing the data, or were provided directly by those programs.

Statistical Testing

Data trends can be analyzed in many ways. The approaches used in the Chartbook to analyze trends in health measures depend on the data source and the number of data points. Trend analyses of data from the National Vital Statistics System and the National Health Interview Survey are based on aggregated point estimates and their standard errors. Trend analyses of data from the National Health and Nutrition Examination Survey are based on record-level data. If data from at least seven time points were available, then one objective of the trend analysis was to identify time points when changes in trend occurred.

For trend analyses of data on birth, infant mortality, and death rates from the National Vital Statistics System (Figures 7–9) and data from the National Health Interview Survey (Figures 10, 17–19, 26, 27), increases or decreases in the estimates during the entire time period shown are assessed by the weighted least squares regression method in the National Cancer Institute’s Joinpoint software (with Grid search and either permutation model selection for 10 or more time points or BIC criterion model selection for fewer than 10 time points). Joinpoint software identifies the number and location of joinpoints when changes in trend have occurred. The maximum number of joinpoints searched for was limited to four or fewer, based on the number of available time points. For more information on Joinpoint, see: http://surveillance.cancer.gov/joinpoint. Trend analyses using weighted least squares regression were carried out on the log scale for birth, infant mortality, and death rates so that results provide estimates of annual percent change. For the charts based on the National Health Interview Survey, trend analyses were carried out on the linear scale and results provide estimates of annual percentage point change. A limitation of using aggregated data and Joinpoint software alone for trend analysis of the National Health Interview Survey is that this approach does not account for year-to-year correlation or use the proper degrees of freedom for statistical testing.

For trend analyses from the National Health and Nutrition Examination Survey (NHANES) (Figures 11 and 15), with nine time points, increases or decreases in the estimates during the entire period shown were assessed using polynomial regression (SUDAAN PROC REGRESS). Linear trends were tested separately and quadratic trends were tested with both linear and quadratic terms in the models. If a quadratic trend was significant, Joinpoint software was used to find an inflection point, and the difference in slopes between the two segments on either side of the inflection point was assessed using piecewise linear regression (SUDAAN PROC REGRESS). For trend analyses based on three to five time points from the NHANES, linear and quadratic trends were assessed using polynomial regression. If a quadratic trend was significant, pairwise differences between percents were tested using z-tests to obtain additional information regarding changes in the trend.

For other charts either the difference between two points was assessed for statistical significance using z-tests or the statistical testing methods recommended by the data systems were used. For analyses that show two time points, differences between the two points were assessed for statistical significance at the 0.05 level using two-sided significance tests (z-tests) without correction for multiple comparisons. For data sources with no standard errors, generally relative differences greater than 10% are discussed in the text. Chartbook data tables include point estimates and standard errors, when available, for users who would like to perform additional statistical tests.

Terms such as “similar,” “stable,” and “no difference” used in the text indicate that the statistics being compared were not significantly different. Lack of comment regarding the difference between statistics does not necessarily suggest that the difference was tested and found to be not significant. Because statistically significant differences or trends are partly a function of sample size (the larger the sample, the smaller the change that can be detected), they do not necessarily have public health significance (136).