Measuring Healthy Life Expectancy for the U.S. Population in Man-Huei Chang, MPH Division of Epidemiologic and Analytic Methods for Population Health (Proposed) Epidemiology and Analysis Program Office Centers for Disease Control and Prevention Presentation at the 2012 National Conference on Health Statistics Washington, DC August 8, 2012 Office of Surveillance, Epidemiology, and Laboratory Services Epidemiology and Analysis Program Office

Outline  Objectives  Data and Methods  Results  Conclusions

CDC Reporting Population Summary Measures  In 1982, Hahn et al. published YPLL-65 and suggested alternative measures (e.g., DALY,QALY,YHL,YAAL, and HLE)  CDC published Years of Potential Life Lost before age 65 (YPLL-65) and leading causes of death in MMWR  In 2000, a CDC Burden of Disease Workgroup recommended routinely reporting summary measures in MMWR  In December 2010, the Epidemiology and Analysis Program Office (EAPO) began working on HLE to monitor trends and disparities  EAPO and NCHS collaborative effort  Regular reporting of summary measures  Goals  To promote public awareness of ‘premature’, ‘preventable’, and ‘unnecessary’ mortality; and  To develop and implement effective interventions for good health and longevity

Objectives  To estimate HLE for the U.S population and for smaller population subgroups  To assess disparities in HLE between different segments of the U.S. population

Data Sources  Data:  Mortality rates (NCHS Vital Statistics System)  Self-reported health status rates  NHIS (national, regional, divisional)  BRFSS (state-level)  Population segments  Total population (national)  4 regions  9 divisions  50 states and the District of Columbia  Demographic subgroups  Age groups (5-year age intervals), sex, race (whites, blacks), and ethnicity (Hispanics, non-Hispanics)

Health Status Data  NHIS question on self-assessed health status: “Would you say your health in general is excellent, very good, good, fair, or poor?”  1 Excellent  2 Very good  3 Good  4 Fair  5 Poor  7 Refused  9 Don't know  BRFSS question is the same as NHIS Healthy Unhealthy

HLE by Age and Sex – U.S

HLE by Age and Race – U.S

HLE by Age and Ethnicity - U.S

LE by Selected Demographics - U.S a t Birth

LE by Selected Demographics - U.S a t Age 65

HLE at Age 65 by Sex and Region - U.S

HLE at Age 65 by Race and Region - U.S

HLE at Age 65 by Ethnicity and Region - U.S

HLE at Age 65 by Race and Division - U.S

State-Specific HLE at Age 20 U.S. Adults, MaleFemale US average: 49.0 US average: 51.8

State-Specific HLE at Age 20 U.S. Adults, WhiteBlack US average: 51.1US average: 43.7

State-Specific HLE at Age 20 U.S. Adults, HispanicNon-Hispanic US average: 48.3US average: 50.6

Significant Disparities in HLE Age All Races Race Ethnicity MaleFemale Disparities WhiteBlack Disparities Hispanic Non- Hispanic Disparities Δ in HLEp-value Δ in HLEp-value Δ in HLEp-value < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < <0.001 SOURCES: CDC/NCHS, National Vital Statistics System, National Health Interview Survey, and the U.S. Census Bureau.

Summary  Disparities in HLE and LE observed among population segments and geographical locations  Differences in HLE among subpopulations result from the combined effects of mortality and morbidity (impacted by demographic, socioeconomic, and environmental factors)  Use HLE to compare the health of populations, monitor trends in the health of a population, and identify health inequalities within populations

Conclusions  HLE can be readily used by public health officials, healthcare providers, and policy makers to understand the health status of populations  These results can be used as baseline to routinely monitor the health of the U.S. population and to identify health disparities in populations

Acknowledgements  Co-authors  Michael Molla (NCHS)  Benedict Truman (NCHHSTP)  Heba Athar (EAPO)  Ramal Moonesinghe (OMMHE)  Paula Yoon (EAPO)  Special Thanks  Stephanie Zaza (EAPO)  Sukhjeet Ahuja (NAPSIS)  Julia Homes (NCHS)  Joyce Arbertha (NCHS)  Stephanie Robinson (NCHS)  Ajay Yesupriya (NCHS)  Carl Kinkade (EAPO)  Brenda Le (NCEH)  Sigrid Economou (EAPO) Disclaimer: The findings and conclusions in this presentation are those of the author, and do not necessarily represent the official position of the U.S. Centers for Disease Control and Prevention.

Thank you! Contract Information

For more information please contact Centers for Disease Control and Prevention 1600 Clifton Road NE, Atlanta, GA Telephone, CDC-INFO ( )/TTY: Web: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. Thank you Contact information Man-Huei Chang

Disclaimer The findings and conclusions in this presentation are those of the author, and do not necessarily represent the official position of the U.S. Centers for Disease Control and Prevention.

Background  Life expectancy (LE) has been increasing dramatically due to a large decline in death rates  Average LE rose from 47.3 years in 1900 to 77.0 years in 2000, an increase of 62.8 percent in the U.S 1  Dramatic change in the age structure of the population  U.S. population aged 65 and older accounted for 4.1% of the total population in 1900 and 12.4% of the population by the year 2000, a three- fold increase 2  U.S. average LE increased from 47.3 years in 1900 to 77.0 years in 2000, an increase of 62.8%; changes in age structure 2 SOURCE: 1. Molla 2001; 2. Molla 2010.

Gaps in LE  Not all populations have aged uniformly  Changes in mortality vary among subgroups  Differences in LE  By demographic characteristics (e.g., age, sex, and race/ethnicity); and  By social and environmental factors (e.g., unhealthy housing, limited access to health care, and quality of healthcare)

Mortality Changes in U.S. SOURCE: Centers for Disease Control and Prevention, Naitonal Center for Health Statistics. National Vital Statistics System (mortality); National Health Interview Survey (noninstitutionalized population) Reduction in mortality: white males >> black males,

LE Changes in U.S. SOURCE: Centers for Disease Control and Prevention, Naitonal Center for Health Statistics. National Vital Statistics System (mortality); National Health Interview Survey (noninstitutionalized population) Increase in LE: white males >> black males,

What is Healthy Life Expectancy (HLE)?  WHO defines HLE as the “average number of years that a person can expect to live in ‘full health’ by taking into account years lived in less than full health due to disease and/or injury”  Combines information on mortality and self-reported health status to estimate years of life expected in good health  Estimates the “aggregated’ effect of diseases or conditions, and evaluates the “overall” health conditions of a population  Expected years in good or better health

Why HLE?  Assesses the combined effects of mortality and morbidity on the overall length of healthy life of a population  Provides an overview of the health status of a population  Tool to monitor trends and disparities in difference populations  Indicator of the total impact of socioeconomic factors and other determinants of health  Meaningful and intuitive measure that can be readily used by public health officials, healthcare providers, and policy makers

Life Expectancy Summary Measure of Population Health Healthy Life Expectancy Population Mortality Age-, sex, race-, -ethnic- specific death rates Population Morbidity Age-, sex, race-, -ethnic- specific self-assessed health Nation, Region, Division State Basic schematic life table model SOURCE: Modified from Molla 2001

A schematic framework for life table model SOURCE: Modified from Molla 2001 NCHS National Regional Divisional estimates State estimates NCHS OSELS U.S. Bureau of the Census Mortality data Midyear population estimates NHIS health data BRFSS health data Age-, sex, race-, ethnic- specific death rates (Dx) Self-assessed health (poor or fair) (π x ) Self-assessed health (good or better) (1-π x ) Life Expectancy Healthy Life Expectancy Data sources Data LE and HLE Calculation

Sullivan Life Table exex SOURCE: Sullivan DF. A single index of mortality and morbidity. HSMHA Health Rep 1971;86(4):

HLE Calculation Source: Molla MT, Wagener DK, Madans JH. Summary measures of population health: methods for calculating healthy life expectancy. Healthy People 2010 Stat Notes 2001(21):1-11.

Age-Specific HLE variance  Estimated under the assumption of the binomial distribution as stated below: S 2 ( n π x )=[ n π x (1– n π x )] / n N x Where S 2 ( n π x )is the variance of HLE in the age interval (x, x+n) n N x is the number of persons in the survey who are in the age interval (x, x+n) 1– n π x is the proportion of population in the age interval (x, x+n) reporting good or better health Source: Molla MT, Wagener DK, Madans JH. Summary measures of population health: methods for calculating healthy life expectancy. Healthy People 2010 Stat Notes 2001(21):1-11.

Test of Disparity in HLE Source: Molla MT, Wagener DK, Madans JH. Summary measures of population health: methods for calculating healthy life expectancy. Healthy People 2010 Stat Notes 2001(21):1-11.

HLE at Age 65 by Sex and Division - U.S

HLE at Age 65 by Ethnicity and Division - U.S

Limitations  Misclassification in reporting and bridging multiple-race population estimates  Different population data (i.e., intercensal vs.vintage postcensal)  Did not correct misclassification for race and Hispanic origin on death certificates  Did not adjust for unknown age on death certificates  Health data different in NHIS and BRFSS (difference in data source, sampling scheme, and data collection methods)  Small death counts in Hispanics and blacks in selected age group and geographic areas