Abstract
Introduction
Initiatives designed to improve the health of populations often rely on monitoring specific health conditions and characteristics. Goals and targets are set for assessing where progress is being realized or further efforts are needed. Evaluating the sum changes in the individual targets on the overall health of the population, however, requires cross-cutting measures of health. Summary measures of population health represent health in the broadest sense, and reflect both specific underlying health characteristics and impacts made by other nonhealth factors. The effect of the other sectors, such as education and employment, on health measures is an important factor that differentiates them from the more targeted measures related to health conditions, risk behaviors, and care. Understanding the relationship between the specific health measures and the summary measures used to describe health more broadly is important.
Initiatives to Monitor Attainment of Health Goals
Many countries have adopted initiatives to monitor achievement of health goals. Healthy People (HP) is an example of an initiative in the United States, designed to improve the health of all Americans, which incorporates a monitoring function. Initiated in 1979 by the U.S. Department of Health and Human Services, HP identifies 10-year health promotion and disease prevention objectives (U.S. Department of Health and Human Services, 1991). For each of the next three decades, HP has established targets for, and monitored progress toward, overarching goals envisioned at the start of the decade. For the current decade, Healthy People 2020 has four overarching goals:
Attain high-quality, longer lives free of preventable disease, disability, injury, and premature death.
Achieve health equity, eliminate disparities, and improve the health of all groups.
Create social and physical environments that promote good health for all.
Promote quality of life, healthy development, and healthy behaviors across all life stages.
The specific objectives set for each successive decade in Healthy People 2000, Healthy People 2010, and Healthy People 2020 have increased in scope and number over time, from 226 initially to roughly 1,200 today (National Center for Health Statistics, 2012). Summary measures of health, the most comprehensive measures for evaluating overall differences and trends in population health (Institute of Medicine, 1998), have been included in HP since its inception. This reflects the ongoing recognition of the importance of monitoring the overall health context of the population that the topic-specific objectives, which form the basis of HP, are designed to impact. However, throughout the course of the initiative, the summary measures have evolved. This evolution reflects the changes in our understanding of the complex nature of health, and the changing methodologies and measures for monitoring health.
Changing Summary Measures
The original 1979 HP publication included five broad measurable targets, reflecting the importance of enhancing life in each of the five major life stages (infants, children, adolescents and young adults, adults, and older adults) and monitored primarily by mortality (U.S. Department of Health, Education, and Welfare, 1979). Mortality by age (below 1 year, 1-14 years, 15-24 years, and 25-64 years) was used to monitor the first four goals. The fifth target, for the population 65 years of age and above, was a morbidity-based measure aimed at preserving independence and defined as difficulty in two or more activities of daily living. While these original five targets were retained in future HP decades, changes in how health is conceptualized and measured resulted in new measures being used in subsequent HP decades.
Drawing on experiences from these early HP efforts, Healthy People 2000 developed its own strategy to improve the health of all Americans and developed three guiding goals for the decade:
Increase the span of healthy life,
Reduce health disparities, and
Achieve access to preventive services.
Moreover, unlike Healthy People 1990 where the chosen targets were indicator specific and the goals of the decade mapped one to one to these specific indicators with well-defined targets, Healthy People 2000 began to incorporate both measures to describe more broadly changes in population health and indicator-specific measures to capture changes in particular health objectives.
Three summary measures of population health were used to address and monitor Goal 1 of the initiative—to increase the span of healthy life (Erickson, Wilson, & Shannon, 1995). The use of the three measures for Healthy People 2000 reflected the realization that length of life was no longer sufficient for describing overall health and the increased interest in quality of life. Goal 1 was monitored using life expectancy at birth (a measure of longevity), the percentage of people reporting that their general health status is fair or poor (a general, subjective health evaluation measure), and a newly created measure, expected years of healthy and unhealthy life (a summary measure developed which incorporates both mortality and morbidity, known as years of healthy life [YHL]; National Center for Health Statistics, 2012).
The new YHL measure estimates the average number of years expected to be lived in a healthy state. As a summary measure of population health, this measure incorporates mortality and morbidity into a single statistic. The mortality component utilizes total mortality, while the morbidity component is evaluated using two measures: respondent-assessed health and activity, and functional limitations due to chronic conditions (for further information on these measures, see Erickson, Wilson, & Shannon, 1995). In this way, the value for morbidity—typically called a health state—is calculated using two measures of health. Table 1 presents the range of health state values calculated, using respondent-assessed health and activity and functional limitations, for the beginning of the Healthy People 2000 decade. Values for the 30 health states form a continuum from 0.10 for the lowest health state (“poor” perceived health and unable to independently perform self-care activities of daily living) to 1.0 (“excellent” perceived health and not limited in functioning or activities). For example, a health state defined as being not limited in a major activity and in “good” health represents a person having 84% of full function for the year.
Source. Centers for Disease Control and Prevention, National Center for Health Statistics, National Health Interview Survey.
Note. Limited—Other = not limited in major activity, but limited in other activities; Limited—Major = limited in major activity; Unable—Major = unable to perform major activity; Limited in IADLs = limited in instrumental activities of daily living without the help of other persons; Limited in ADLs = limited in activities of daily living without the help of other persons (for further information, see Erickson, Wilson, & Shannon, 1995); IADLs = instrumental activities of daily living; ADLs = activities of daily living.
Activity limitation is based on data collected across a number of NHIS questions and represents ability to independently perform age-appropriate social roles such as working, keeping house, or going to school.
Respondent-assessed health is based on the question, “Would you say your health in general is excellent, very good, good, fair, or poor?”.
Based on life table methods, and a combination of information about the number of persons by age in the population, the health state values, and mortality rates, the resulting calculation of YHL is equivalent to the average years a person would live if he or she lived in a healthy state. During the course of the decade monitored by Healthy People 2000, the YHL increased (between 1990 and 1998) from 64.0 to 65.2 years. This was comparable with the rise in life expectancy (1.3 years during the same period; National Center for Health Statistics, 2001). Thus, average YHL as a percentage of total expected years of life remained largely unchanged.
The two overarching goals formulated for the next decade in Healthy People 2010 build on the first two from Healthy People 2000 but are more aspirational: increase the quality and YHL, and eliminate health disparities. The healthy life expectancy measures used in Healthy People 2010 were not, however, the same as those used in Healthy People 2000. This change was the result of a workshop, convened by the National Center for Health Statistics, aimed at selecting the best measures to include for the next decade of HP (Institute of Medicine, 1998). The recommendations included three new healthy life expectancy measures:
Expected years in good or better health,
Expected years free from activity limitations, and
Expected years free of selected chronic diseases.
While the 2000 YHL combined two measures of health (respondent-assessed health and activity limitations) into a single morbidity measure and then combined the result with mortality rates, the 2010 set of healthy life expectancy measures separate the morbidity constructs to better track change over time as a result of these individual components. Figure 1 shows trends from 2000 to 2013 in expected years in good or better health, expected years free from activity limitation, and life expectancy at birth. During this period, the average number of years individuals could expect to live free of activity limitation increased 1.2 years, from 65.6 to 66.8 years, but the increase was not statistically significant (p < .05). Expected years in good or better health also increased by 1.5 years (from 68.4 to 69.9 years; p < .05), and overall life expectancy improved by 2.0 years (from 76.8 to 78.8; p ≥ .05). The third Healthy People 2010 healthy life expectancy measure introduces summary information based on chronic disease and is defined as the average number of years a person can live without being diagnosed by a physician or health professional as having one or more selected chronic conditions. Specifically, the following chronic conditions are used in the calculations: arthritis, asthma, cancer, diabetes, heart disease, high blood pressure, kidney disease, or stroke.
Trends in life expectancy, expected years in good or better health at birth, and expected years without activity limitations at birth for both sexes: United States, 2000-2013.
During the development of Healthy People 2020, proposals were made to add additional summary measures to track overall changes in healthy life. In addition to the three measures used in 2010, measures of mental health and of a variety of health risk factors and determinants were proposed. While these indicators were of interest, adding a larger number of unrelated indicators would make it harder to summarize overall changes in health over time. The proposed indicators might be considered summary measures in their domain, but they were moving away from summary measures of health in general.
A Framework for Monitoring Overall Health
In response to the need to monitor overall health, a more structured approach to summary measures is proposed that presents a succinct number of broad measures of health in a coherent, visible, and conceptually consistent manner (Table 2).
Framework for Monitoring Overall Health.
Note. Measures included in the framework can be calculated at any age (e.g., at birth, at age 25, at age 65, etc.) and by sex.
A key concept of monitoring framework is the relationship between the measures included. The central premise maintains that the overall measures will change as a result of actions and interventions in the other, more specific health indicator goals or targets. The overall measures of health, by definition, are the most comprehensive and are important for documenting the overall performance of public health and clinical medicine, as well as the impact of other sectors that influence health. The summary indicators can be conceptualized as the apex of the monitoring system. Healthy life expectancies, because they combine measures of mortality and morbidity, are good candidates for the apex of the hierarchy. However, there are many choices for selecting the measures that comprise the health or morbidity dimensions used in constructing the healthy life expectancies. While having just one measure has advantages, a small set of measures would provide a more complete picture of health—but the measures should form an internal hierarchy with a logical structure.
A Hierarchical Approach
The proposed framework places healthy life expectancy measures at the top of the hierarchy—in Tier 1. Each of the four measures combines life expectancy and a summary measure of “health” (where health is defined broadly). The “health” components address limitation of activity, two functional or disability status, and respondent-assessed health status. There is a hierarchical relationship among the measures, with healthy life expectancy based on activity limitation at the top, followed by two healthy life expectancies based on disability (more severe and milder) and then healthy life expectancy based on “good or better” respondent-assessed health. Respondent-assessed health provides a summary of the more “biologic” aspects of health (disease, impairment, symptoms, etc.). It is also a determinant of functioning so is placed at the base of Tier 1. The expectancies in Tier 1 can be calculated at any age (e.g., at birth, at age 25, or at age 65).
The indicators in Tier 2 provide information on the component dimensions of the healthy life expectancy measures—life expectancy and the summary population health measures used to calculate the Tier 1 healthy life expectancies. Individually, the Tier 2 items are life expectancy, the percent free of activity limitation, the percent free of disability (assessed at the two levels of functioning), and the percent in “good or better” assessed health. The ordering of these measures mirrors the order in which they appear in Tier 1 of the hierarchy. As in Tier 1, the life expectancy component can be calculated at any age or by sex.
A final optional set, Tier 3, provides the percent with chronic conditions. Conditions are further down the causal chain in determining disability and activity limitation. These indicators are not summary indicators but can provide a bridge between the summary measures and the more specific indicators in monitoring initiatives. The usefulness of this tier will depend on the exact nature of the whole indicator set used to monitor the population.
Including Functioning in the Framework
Three of the four summary health measures capture aspects of functional status. Health is a complex multidimensional concept which in its broadest sense encompasses all aspects of physical and mental characteristics. An extremely large number of indicators would be necessary to adequately depict all aspects of an individual’s health, or the health of a population as is reflected in more than 1,200 objectives in the Healthy People 2020 initiative. Functional status provides a more comprehensive way to summarize the impact of the numerous and diverse aspects of health.
Functional status itself is complex and can be seen as the sum of risk factors, the medical and/or pathological aspects of health, symptoms, disease states, disease severity, and the use of health care. Furthermore, it involves multiple domains and levels of severity. Measures of functioning can address core domains—either without accommodation (“within the skin”) or with accommodation (e.g., with the use of assistive devices)—or on more complex functional domains. These complex domains incorporate capacity, along with environmental barriers and facilitators (such as health care and associated services, assistive devices, physical structures, social attitudes, and legal requirements), and their effect on social participation.
Measuring health through functional status makes it possible to summarize the impact of the medical and/or pathological aspects of health using a small number of indicators. Functioning-based measures are also of interest in the policy arena at local, national, and international levels. The United Nations Convention on the Rights of Persons With Disability (United Nations General Assembly, 2006) has at its core the requirement to eliminate participation restrictions due to limitations in functioning “within the skin” by dismantling barriers and creating facilitators that enhance full participation in all aspects of society. It is possible to create indicators from ongoing statistical systems that measure the level and the change in the level of functioning in the population without (and with) accommodation in key domains as well as the level of societal participation and the change in participation level and to relate the change to program and policy interventions.
The proposed framework includes two measures of functioning in core domains, labeled as disability in the framework. It also includes one measure of participation with accommodation, as it becomes increasingly difficult to measure functioning independent of environmental accommodations in complex functional domains like work and education.
Disability Indicators
The two measures of disability are included in the framework and are based on the work of the Washington Group on Disability Statistics (WG; Altman, 2016). The WG was organized in 2001 following the United Nations International Seminar on Measurement of Disability to address the need for statistical and methodological initiatives at an international level to facilitate the measurement of disability and the comparison of data on disability cross-nationally. All National Statistical Offices are eligible for membership in the WG. Currently, 113 National Statistical Offices are represented, as well as UN agencies, international organizations, and organizations that represent persons with disabilities. The main objective of the WG is the promotion and coordination of international cooperation in the area of health statistics by focusing on disability measures suitable for censuses and national surveys.
The WG first developed a Short Set (WG-SS) of questions for use in censuses and in national surveys to identify the subpopulation that is at a greater risk than the general population of experiencing restrictions in social participation. This is a population of great policy interest as it represents those who require some kind of accommodation. Risk is increased as a result of difficulties experienced in basic functional domains. The population identified by the WG-SS may or may not also experience restrictions in participation as participation will depend on the availability of assistive devices, a supportive environment, and other resources, and these will vary within the group identified as being at increased risk.
The WG-SS includes one question for each of six domains of functioning (vision, hearing, mobility, cognition, self-care, and communication) and asks whether the person has difficulty in that domain. Each question has the same four response categories: no difficulty, some difficulty, a lot difficulty, or unable to do. Milder disability is defined as having at least some difficulty in any one of the domains. More severe disability is defined as having at least a lot of difficulty in any one of the domains. The disability measures proposed for the framework, both the Tier 1 healthy life expectancy measures and the Tier 2 summary measures, are based on these milder and more severe categories. For most domains of functioning, these are evaluated without taking into account environmental factors that can change the effect of pathology on performance in these domains. 1 For example, in the case of walking, functioning is evaluated without the use of assistive devices even if these are usually used. While it is possible and useful to create indicators that evaluate walking when using assistive devices, these measures are not included in the framework to keep the number of indicators to a minimum.
Participation Indicator
Limitation of activity is used as indicator of less than full participation in society. The summary measure based on this indicator gives the expected years of life free of activity limitation. In the United States, a set of questions have been used in the National Health Interview Survey (NHIS) to measure the ability to independently perform age-appropriate social roles such as working, keeping house, or going to school, or in other activities due to a chronic condition. Other formulations that capture the same underlying concept are used in other countries, and there is much current interest in improving questions that capture health-related limitations in participation. The complex and encompassing nature of health and its potential effects on all aspects of life make crafting participation questions particularly challenging. However, functioning measured within complex domains and with accommodation defined broadly provides key information on overall quality of life, and when the causes of the limited participation in society are anchored in the health domain, the measures are of health-related quality of life. This outcome measure goes beyond the traditional definition of health and moves the focus to social participation—the ability to get an education, be employed, earn income, and participate in social, cultural, and civic activities which are key policy and program goals. Despite the challenges of obtaining this information and the need for better measures, current indicators of participation restriction using existing measures of activity limitation provide important policy-relevant information and continue to be collected.
Results
Figures 2 through 7 illustrate the healthy life expectancies, at ages 25 and 65 for both sexes, males and females, proposed for use in the framework. Using the four measures described and defined as above, the figures demonstrate the relationship between overall life expectancy and the healthy life expectancies, and show change in the United States from 2010 to 2013. Little change has occurred in the 3-year period, at both ages, in expected years of life overall, free of activity limitation/participation, and free of more severe disability. Expected years free of milder disability at age 25 have declined between 2010 and 2013, with statistically significant declines for both sexes (from 32.8 to 30.7; p < .05), males (from 32.3 to 30.3; p < .05), and females (from 33.3 to 31.2; p < .05). No such change occurred among older persons at age 65. Expected years of “good or better” health increased at both ages, and by sex, with the increases at age 65 statistically significant for both sexes (from 14.3 to 14.6; p < .05) and males (from 13.2 to 13.7; p < .05) only.
Life expectancies, both sexes, age 25: United States, 2010 and 2013.
Life expectancies, males, age 25: United States, 2010 and 2013.
Life expectancies, females, age 25: United States, 2010 and 2013.
Life expectancies, both sexes, age 65: United States, 2010 and 2013.
Life expectancies, males, age 65: United States, 2010 and 2013.
Life expectancies, females, age 65: United States, 2010 and 2013.
Discussion
Initiatives to improve health are often accompanied by monitoring systems that identify a set of indicators to be used to track progress in achieving the objectives in the initiative. The initiatives often contain a large number of specific indicators that are tracked to determine specific policy or program interventions that are needed to achieve program goals. These specific indicators are important for achieving the goals of improving health, but they do not provide an evaluation of whether health—defined broadly—is improving at the population level or for subgroups of interest. Summary measures are needed to attain this larger goal. The summary indicators should be related to the individual indicators in that they should reflect movement in the various parts of the system and provide an additional way to monitor the effectiveness of policy and programs to improve health. While having a single indicator has advantages for reporting purposes or for targeted program interventions, attempts to find the elusive single indicator to describe overall health have not been successful and multiple, unrelated indicators have been suggested. In some cases, the number of summary indicators has been so large that the original purpose of being able to summarize the overall status of the population is undermined.
We have proposed a framework based on a concise set of related measures that can be used to summarize changes in population health in the broadest sense and which can be incorporated into the monitoring component of initiatives to improve population health. The first tier is comprised of four measures of healthy life expectancy which combine life expectancy with four summary measures of health. The measures are purposefully ordered to form a hierarchy, with an indicator of participation in society at the top followed by two measures of disability and finally, a measure based on respondent-assessed health. The expectancies can be measured at any age, and examples are provided in Figures 2 through 7 using U.S. data at ages 25 and 65, by sex. The second tier of measures disaggregates the healthy life expectancies into the component parts to provide some more detail on changes over time. An optional third tier can be added which tracks the prevalence of key diseases to provide a clearer link to the specific indicators in the initiative.
There are many ways in which interventions can improve population health, including improved prevention of disease and injury and improved health and rehabilitation care aimed at controlling diseases and the effect they have on functioning. Improvements in prevention and health care are key to improving health. However, the goal of full participation in society, irrespective of disease, can be attained by instituting appropriate environmental features that ameliorate the impact of functional limitations in the core domains. The framework proposed is an attempt to provide a mechanism to evaluate whether programs and policies have been successful in achieving these broad policy goals.
Footnotes
Acknowledgements
The authors acknowledge David T. Huang, PhD, MPH, CPH, for the provision of materials related to the history of the Healthy People Initiative and Michael T. Molla, PhD, for analytic support.
Authors’ Note
The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the National Center for Health Statistics, Centers for Disease Control and Prevention.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.