A Measured Approach to Child Health

Ruth E. K. Stein, MD
From the Department of Pediatrics, Preventive Intervention Research Center for Child Health, Albert Einstein College of Medicine and Childrens Hospital at Monteore Medical Center, Bronx, NY Presented in part at the 50th anniversary celebration, Pediatric Academic Societies Annual Meeting, Vancouver, British Columbia, May 2010. Address correspondence to Ruth E. K. Stein, MD, Albert Einstein College of Medicine, Childrens Hospital at Monteore Medical Center, Department of Pediatrics, Preventive Intervention Research Center for Child Health, 6B27 Van Etten, 1300 Morris Park Avenue, Bronx, New York 10461 (e-mail: ruth.stein@einstein.yu.edu). Received for publication October 9, 2010; accepted November 27, 2010.

ABSTRACT
Childrens health and its measurement have gained increasing attention in the face of advances in treating disease, and the growing recognition of long-term implications of child health for adult health and the nations economy. Advances in measurement are aided by new conceptualizations, including a dynamic denition of child health and model of how it evolves. This paper discusses challenges in measurement of child health, the role of large-scale data sets, how to select a measure, 2 promising measurement frontiers, and the role of the Academic Pediatric Association in promoting a measured approach to child health.

KEYWORDS: child health; cultural issues; health; health

services research; health status; measure; outcomes

ACADEMIC PEDIATRICS 2011;11:240246

IT IS APPROPRIATE to take stock of the progress in the

measurement of child health on the Academic Pediatric Associations (APA) 50th anniversary, and to be proud of the role of our organization, its members, and all of pediatrics in making the last half century a time of signicant achievement. In the early 20th century, children and their health were a relatively low priority,1 and for much of the century child health was measured primarily by rates of mortality and morbidity. The inception of the APA marked a period of renewed efforts to improve child health on both the individual and population levels, and this required an increasing focus on how to measure it. There is renewed interest in the development of more accurate measures of child health, and this paper examines the current status of work in this area. The discussion below includes a review of the most comprehensive and forward-looking denition and model of child health and considers the models use in understanding the evolution of child health. This is followed by a brief discussion of large-scale data sets, the importance of measuring child health, key elements in selecting a child health measure, and why measurement of child health is so challenging. Finally, the discussion features 2 promising new frontiers that are breaking ground in measurement of child health and a discussion of ways in which the APA can help to promote better measurement of child health.

DEFINITION
A committee of the Institute of Medicine and the National Research Council of the National Academies, which I was privileged to co-chair, developed a denition that built on the work of many thoughtful scholars and the new science summarized in the landmark book, Neurons
ACADEMIC PEDIATRICS Copyright 2011 by Academic Pediatric Association

to Neighborhoods.2 In the committee report, Childrens Health, The Nations Wealth,3 childrens health was dened as the extent to which individual children or groups of children are able or enabled to (a) develop and realize their potential, (b) satisfy their needs, and (c) develop the capacities that allow them to interact successfully with their biological, physical, and social environments.3 This denition builds on and goes beyond the World Health Organization4 and Ottawa Conference5 denitions in 4 ways: a) by including a view of health as a positive construct, b) by emphasizing development during childhood and its implications for long-term outcomes, c) by implying an interaction between the child and his or her environment, and d) by acknowledging that multiple inuences interact with biology over time.3 In addition, the report suggests a need to go beyond traditional measures of diseases and disorders to assess the functioning of children in their daily lives and environments and their positive capacity to be healthy. Functioning is the best way to assess the nal common pathway of the effects of childrens conditions and treatments on their daily lives.6 The report calls the positive capacity to be healthy health potential, but others have referred to it as resilience or well-being. Although there are many measures that assess decits and some that assess functioning, health potential represents the largest unexplored measurement frontier in child health and is one that needs to be tackled to fully understand and help foster child health in the future. This has become of even more importance in the face of new emphasis on preventive measures featured in the Patient Protection and Affordable Care Act.7 In the face of this new priority, it will become increasingly important to be able to measure levels of health beyond the absence of disorders or functional limitations.
Volume 11, Number 3 MayJune 2011

240

ACADEMIC PEDIATRICS

A MEASURED APPROACH TO HEALTH CARE

241

THE MODEL
Modern science has unraveled many mysteries of biology and its role in child health, but we are only now on the cusp of understanding the complex ways in which biology is altered by environment over time. Many risk and protective factors have been identied, which the report refers to as inuences, because it is known that in many instances the same thing, such as family, peers, or a substance (eg, iron or iodine), can be either a source of risk or of protection, depending on the circumstances. Inuences were divided into 6 somewhat arbitrary groupings (Figure 1). They include aspects of biology (which encompass biologically incorporated prenatal factors and exposures), behavior, social environment, and physical environment, which all exist in the context of services and policies. Unlike some models, such as Healthy People 2010,8 in this model the terms, services, and policies are not restricted to health services or health policies. Rather, here they refer to all services and policies that may directly or indirectly affect health outcomes of children. This is because many, if not most, services and policies affect child health, even when that is not their direct intent. For example, the policies that govern the built environment, a component of physical environment, may increase or limit the risks of injury and those that inuence mining, trafc, subsidies, or taxes, have both direct and indirect effects on the developing childs health. A similar case can be made for services in the community that directly or indirectly affect a childs life and environment. The diagram in Figure 1 is a representation at one point in time, but each circle consists of many subdomains that together form even more complex patterns of interactions and may have differing levels of saliency at different times of life. For example it is generally thought that family is

a primary social environmental component of infants in a way that is different from its role for adolescents who, although still inuenced heavily by their family, may also be more involved with peers and community. All these inuences affect child health in a transactional or interactive manner and are mediated or moderated by other inuences in complex ways that are only now being understood. The inuences overlap, so that the result is not a sum, but a result of their interactionsof how they affect one another and overall health. However, although child health status is often measured in a cross-sectional way as represented in Figure 1, this representation is only a snapshot at one point in time. In reality, childrens health and the effects of multiple inuences evolve over time and throughout development as shown on the 2 axes in Figure 2. Health results from the dynamic intersection of these interactions in a model that is best represented by a kaleidoscope. With each turn the patterns change, altering health, incorporating the previous elements, including the childs former health, and affecting health into adulthood. Inuences change in relative importance during different stages of development and across time. Thus, the relative size of each circle and the subcomponents within itthat is, their relative salienceare dynamic. The result is not random: as in a kaleidoscope, results are determined by the previous patterns or settings and how much change is produced by the next alteration. So too the specic inuences and combination of health inuences change and interact over time and throughout development in ways that can be predicted if the prior state is understood. At some stages the turns are rapid, representing substantial developmental change, and at others less so. Each turn incorporates the previous elements and the childs prior health and casts them in a new light, affecting

Figure 1. Cross-sectional model of child health and the inuences of child health. Source: Childrens Health, The Nations Wealth: Assessing and Improving Child Health.3 Reprinted with permission from the National Academies Press, Copyright 2004, National Academy of Sciences.

Figure 2. Longitudinal model of child health and inuences of child health over time and through development. Source: Childrens Health, The Nations Wealth: Assessing and Improving Child Health.3 Reprinted with permission from the National Academies Press, Copyright 2004, National Academy of Sciences.

242

STEIN

ACADEMIC PEDIATRICS

the childs current and future health. There may also be some critical and sensitive periods during which children and their health may be permanently affected by particular inuences. Understanding this model helps us explain how diverse early experiences set the stage for children to respond in distinctive ways to subsequent challenges and how their health diverges over time. An example of this is a landmark paper that demonstrated that the developmental trajectory of low birth weight babies from differing socioeconomic backgrounds diverged, with 1 group maintaining its developmental trajectory and another diverging form it.9 This type of interactive model has been replicated many times with differing inuences and is a key to explaining diverse health outcomes and perhaps many, if not all, health disparities. The model is useful in many ways. Although it may not be possible to study all the parts of it simultaneously, it provides an important conceptual framework for studying components, their interactions, and the patterns of outcomes. It may also help unravel what sometimes appear to be contradictory research ndings, which may reect differences in the pre-existing patterns of samples, and it helps to explain how social determinants become incorporated in biology and in health outcomes. One key issue is the importance of measuring child health longitudinally, because just like interpreting weight or height, it matters where the measure is on a trajectory. It is not enough to know that a child weighs 10 kg. The measure has different meaning depending on whether the child is a huge 4-month-old, an underweight 3-year-old, or a more typical 12-month-old. Child health advocates and providers need longitudinal assessments, and to consider where the childs health is relative to where we expect it to be or where the child is on a health trajectory. Although there are a number of short-term longitudinal studies that focus on specic issues and age groups, a key decit in the study of child health in the United States has been the absence of a large-scale, nationally representative, long-term, longitudinal sample of children. This is a decit that hopefully will be remedied by the National Childrens Study, now being piloted.10,11

health. Failure to do so has enduring repercussions for productivity and contributions to the economy and public welfare. It is now indisputable that early events and conditions of childhood serve as precursors of a wide array of adult diseases, such as cardiovascular diseases,12 cancer, and mental health conditions,13 and that intervention early in life has enormous lasting payoffs.14 Maximizing childrens health is the only way to prevent excessive morbidity and continuing costs to the individual and society. Most of the countries who are members of the Organisation for Economic Co-operation and Development have realized these implications and devoted considerable resources to monitoring and improving child health. However, the United States has lagged far behind other nations in making child health a priority as reected by the fact that we now rank an appalling 24th among the 30 nations in the Organisation for Economic Co-operation and Development in child health and safety.15

SELECTING A MEASURE
Child health measures are needed for many purposes, including population assessment, monitoring of subclinical populations, clinical intervention trials, assessment of clinical outcomes, research to understand child health and the effects of various inuences, and individual patient monitoring. It would be nice if there were one all-purpose measure of child health. The closest thing that exists is a question on the overall health of children, which has been used in many national and local surveys. It asks a respondent, usually the parent, a version of the following: Compared to other children, how would you rate your childs overall health? Excellent, very good, good, fair, poor? Although this question has been shown to have good predictive reliability on a population level, it is quite inadequate on an individual level, as some parents report a child to be in excellent health despite the presence of signicant and sometimes serious health conditions or considerable impairments. In addition, parents of differing racial and ethnic backgrounds also systematically differ in rating their childs health. It is not entirely clear why this is so. Our research group has previously reported that these patterns are not explained by rates of chronic or acute health conditions, use of health services, insurance status, and parental health, and only partly explained by social factors.16 In the absence of a one-size-ts-all solution to the measurement of child health, different types of measures are needed for different purposes. Monitoring an individual patient requires far greater precision and is usually more time consuming than is practical for assessing population health. So deciding on the appropriate measure requires choices. However, regardless of the purpose and application, all measures must meet certain psychometric criteria of reliability and validity; that is, they must measure what you think you are measuring in an accurate and reproducible way. The rst things to consider when setting out to measure child health is the target or aspect of child health you want

IMPORTANCE OF MEASURING CHILD HEALTH

Some might question why it is necessary to measure child health. From a public policy perspective, it is evident that unless something is monitored, it gets little attention.3 Without an agreed upon metric, it is difcult to assess whether progress is being made. Therefore, society needs valid, reliable information to monitor progress effectively toward maximizing child health. Measures are also important for clinical and research purposes. Although those devoted to child health and development may believe that children have intrinsic value and that it is morally right to try to maximize their health, others may question this. These doubters need to be reminded that children represent the future of any society, and that it is in every societys long-term self-interest to protect and nurture its childrens

ACADEMIC PEDIATRICS
Table. Steps in Selecting a Measure 1. Determine the target domain/aspect of health to be measured. 2. Determine the level of precision required. 3. Determine the age range. 4. Assess the availability of instruments and whether they are psychometrically sound. 5. Has the instrument been used with similar samples? 6. Determine if the instrument can be adapted for your purpose. 7. Determine its psychometric properties with your sample.

A MEASURED APPROACH TO HEALTH CARE

243

to assess (Table). Are you aiming to assess diseases and disorders, functioning, or health potential? If you are interested in the latter domains, are you focusing on overall, physical, or mental health, or some other aspect? Once this has been decided, there are further considerations. For example if assessing the severity of a disorder, do you want to measure the worst, best, or average impact on the individual? If you want to measure functioning, are you interested in what the person usually does on a daily basis or in what he has the capacity to do and is with or without extreme effort or assistance? These distinctions may be particularly critical in assessing the health of children with chronic disorders or impairments. Once you determine the aspect or domain of health you want to measure, or your target, it is important to consider how much precision is needed. As mentioned above, more precision is needed if you are evaluating the health of individuals than if you are assessing the health of populations. Are you more interested in constructs of health that are likely to be stable or those most likely to uctuate? The needs may be different when you are assessing acute intercurrent illness or long-term health conditions. Here again, your objectives will determine the nature of the optimal measure for your purpose. You also need to determine the age group of the children you wish to assess. Few measures span the full age range, and different measures have been validated for different age groups. This can pose a problem when following children across an age span that does not coincide with the validated age range of a single instrument, because it can be hard to determine whether measured differences reect true differences in health or merely measure variation. Next it is important to determine whether there are any appropriate psychometrically sound instruments available. If there are, then are they validated for the type of sample you propose to assess? This step maximizes the likelihood that it will work for your purpose. If an instrument or measure has not been used in similar samples, it may be necessary to do some recalibration or modication of an existing measure and to re-examine its psychometric properties (reliability and validity) to be sure that it will work for your purposes. There are established standards for checking whether an instrument has stable psychometric properties in the new environment and with the new sample. This process is analogous to moving an old-fashioned baby scale and recalibrating it in its new location. If your sample includes more than one language or culture, there are additional steps that should be considered to determine stability of the measure.17 These additional steps can be costly and complex and complicate much of

the international work done to compare health across nations. It often forces use of cruder measures than might be optimal, because these basic measures are less affected by language and culture. Having determined the answers to these questions, you are ready to proceed. However, in some instances there will not be a suitable measure and you may have to start the daunting and time-consuming task of developing and validating one. The development of child health measures is an area that needs much more work, but development and validation of measures is labor intensive and costly. There are relatively few validated, psychometrically sound child health measures, especially ones that go across the full age span, relative to the diversity of needs. The measurement gaps are greatest in younger age groups and those without high-cost morbidities. The result is that most of the health measurement work has been done in adults, where it is believed to have the largest potential for rapid payoff. It is often assumed that adult measures can be scaled down to apply to children, but adaptations modeled on adults often ignore the issues and roles that are important in assessing child health. Until recently, measurement of child health has been viewed as less important, because it has little immediate effect on current health economics where per capita spending on child health care is far less than adult health care. The new and compelling evidence that the precursors of adult disease begin in childhood, and the unraveling of the human genome, have combined to begin to change the tide. Today there is growing appreciation of the fact that getting a handle on early determinants of chronic conditions is the only way to contain long-term health care costs, and as a result, there is beginning to be more attention to the measurement of child health beyond the inventory of diseases and impairments.18

CHALLENGES IN MEASURING CHILD HEALTH

Another reason for the focus on adult health measurement is that the measurement of child health is more challenging. Measurement of any construct or concept is a complicated undertaking, and even more so when the construct is as abstract, multifaceted, and dynamic as child health. The very characteristics that make children special also complicate the measurement of their health. Much has been written about the key difculties briey summarized here. DEVELOPMENT The hallmark of pediatrics is developmentan ever changing process with a wide range of normal. Pediatric measures have different meaning under different circumstances. One result of the need for different ways of measuring constructs or domains at different stages of development is that the same questions or tests cannot be used across an age span. For example, the same test cannot be used to assess the ne motor coordination of a 9-monthold and a 9-year-old. When the measure itself is not stable, it is hard to prove if the same aspect of health is being

244

STEIN

ACADEMIC PEDIATRICS

assessed and whether the instrument has the same degree of precision. Use of different items at various ages often leaves open the question of whether detected variation in scores reects true changes in the underlying phenomenon (ie, the health of the child) or are simply discrepancies in the characteristics of the measure. DEPENDENCY A basic metric in assessing adult functioning across cultures and settings is self-sufciency or independence in activities of daily living (ADLs) and instrumental activities of daily living (IADLs). However, children, especially young ones, are by their nature normally dependent and cannot be judged by these functional and self-sufciency measures. They normally acquire these skills over time, and this makes one of the classical ways of measuring adult healththe use of ADLs and IADLsuseless in young children and of limited use in older children. Moreover, children are expected to acquire ADLs and IADLs at varying ages across cultures. Changing levels of dependency makes stable measurement very challenging. For example, lack of toilet training or poor self-regulatory behavior manifested by temper tantrums has different meaning when the childs age is 2, 7, or 17 years, as well among 2-year-olds in different cultures. When a concept is expressed differently at different ages and across settings and cultureswhen it is dynamicit is tricky to measure it and to show consistency of the measure across stages and samples. A second aspect of dependency is the need to rely on proxy responses, especially for children who lack cognitive capacity to understand and answer complex questions. There are always skeptics who question the validity of proxy respondents, although there are many instances in which these have been shown to be reliable.1921 DIVERSITY With 48% of new US births to minority group members, it is key to consider the roles of language and culture in denitions of illness and disorder and how families assess and rate their childs health. Culture plays an enormous part in expectations about the role of children, attainment of developmental milestones, denitions and perception of the illness and disorder, and in notions of when a childs health is in jeopardy. This is an area where there is a paucity of research to guide us. Instead we have dramatic examples of health-related constructs that do not translate, such as those depicted in The Spirit Catches You and You Fall Down.22 This book dramatically demonstrates the cultural schism in the ways that a Hmong family in California and medical team assess the progressive seizure disorder of a girl with Lennox-Gastaut syndrome and reminds us that not all families accept our Western medical model when assessing their childs health. Diversity also characterizes the epidemiology of disorders of child health, which is distinctly different from those of adult health. Childrens health is much more subject to acute intercurrent conditions, whereas many more adults have long-term health conditions, and these tend to cluster

such that a few diagnoses cover a large proportion of those with chronic disorders. In contrast, a far higher proportion of child chronic health issues involve a range of relatively rare diseases (or combination of disorders), something that has pushed many of us to advocate for using noncategorical or generic measures of child health.2325

USE OF LARGE-SCALE DATA SOURCES National surveys have been among the single greatest sources of measures of child health used in policy decisions. The portfolio of national surveys contains important indicators of health at the population level, but for the most part they rely on single items, rather than scales or instruments, for the measurement of child health and thus lack range and sensitivity to the intricacies of different aspects of child health. The Committee on the Evaluation of Child Health that developed the denition and model described above3 looked at length at the extensive data that the United States collects through a variety of national surveys and concluded that the data are very useful for many purposes at the aggregate level, but they have several important shortcomings. Among these are the lack of development, standardization, coordination, and validation of data across agencies; the inability to apply data to states and smaller geographic units; and the paucity of longitudinal data. Equally, if not more, important is the fact that surveys tend to focus on limited sets of inuences on health and the inability to link health data with other sources of data, including educational, health care delivery, and environmental data. In addition, few surveys include complex measures of health, and most depend on indirect, often single-item elements as indicators of health. These limitations taken together severely curtail the ability to enhance knowledge about the most effective ways to improve health at the population and subpopulation levels. Insurance les, electronic medical records, and other information systems that are designed to monitor health care delivery are becoming increasingly central to the measurement of child health outcomes, but all contain data that is subject to considerable methodological challenges, and much work is needed to standardize data collection and to validate the data they contain. They often lack information on geographic identiers, race/ethnicity, and socioeconomic classication that may be important in understanding child health and disparities. In addition, they (as most national data sets) are likely to continue to rely on specic indirect indicators of child health rather than comprehensive measures, and perhaps most critically, are liable to be driven by issues of cost containment among users of health care, rather than by overall concerns about population health, preventive issues, and understanding inuences that affect overall child health. NEW FRONTIERS IN THE MEASUREMENT OF CHILD HEALTH
Despite the challenges, progress is being made on many fronts, and the descriptions below of 2 measurement endeavors offer hope of dramatic progress in the coming

ACADEMIC PEDIATRICS

A MEASURED APPROACH TO HEALTH CARE

245

years. One is an important investment in developing measures of individual child health, and the other is a population-based effort to rally communities to improve childrens health, functioning, and health potential. PROMIS II The National Institutes of Health Roadmap Project included a large effort to develop measures of patient reported outcomes for clinical trials. This Patient Reported Outcomes Measurement Information System (PROMIS) project uses item response theory (IRT), instead of the traditional psychometric tests, to develop standardized measures. In IRT, all items from existing measures are rewritten to a consistent format and then tested on large samples to determine how they relate to one another, so that they can be rank ordered from easiest to hardest. John Ware, a measurement expert, compares current health care measurement to the development of thermometers, which initially went in different directions and all used different metrics, until cold was standardized as low and hot as high and the units of measurement were agreed upon.26 At present, we have measures that do not map onto one another. Converting their content and format and testing them through the IRT process permits assessment of the relationship of items from different measures to one another and determination of which items are most sensitive. The ultimate purpose of IRT is to allow a few items from an item bank to ascertain the health of individuals with precision. This implementation process is somewhat like using a developmental screen, in which one starts with an item from the item pool, and depending on the answer, moves up or down the scale, to other items, to determine where the child ts. Up to now the focus of PROMIS has been on development of adult measures, except for a small project at North Carolina led by Darren Dewalt.27 Recently, in round 2 of PROMIS (PROMIS II), several new pediatric sites were funded. This is the rst major federal investment in the development of new child health measures in many years and one we should watch closely. EARLY DEVELOPMENT INSTRUMENT AND AUSTRALIAN EARLY DEVELOPMENT INDEX On a population health front, several communities in Canada began monitoring the health of young children by using an Early Development Instrument (EDI) that assesses their functioning in terms of school readiness. Information on capacity in each of 5 domains (physical health and well being, social competence, emotional maturity, language and cognitive development, and communication skills and general knowledge) is assessed individually for each child by a teacher after several months of classroom observation.28 These ratings are used to indicate the proportion of children having difculty in each domain and to prole the strengths and decits of the communitys school entry population. The observations are then geocoded to identify neighborhoods in which EDI scores indicate children are functioning suboptimally in one or more domains and to map neighborhood resources. The informa-

tion is shared with a broad constituency of individuals and groups within a community. The ultimate goal is to prole the functioning of children in each community and to bring together multiple players to enhance the communitys effort to support positive development at the population level. To date, >660 000 Canadian children have been evaluated, and data presented at the 2010 annual meeting of the Pediatric Academic Societies reported on ratings on an impressive 97.5% of all rst-grade children in Australia by using an Australian adaptation of this tool, the Australian Early Development Index.29 Information will be used to determine how to invest resources to help communities raise healthier children. Similar efforts are underway in communities in the United States and around the world. The focus on improving population-level child health is a truly remarkable development being undertaken by communities and nations committed to the importance of childrens health for their own futures. This type of effort needs to be spread geographically as well as be developed for the assessment of child health across the entire age spectrum.

THE ROLE OF THE APA

There are many ways in which the APA can contribute to and promote the development of improved child health measures. In terms of its educational mission, it can play a vital role in informing members and trainees about the issues in child health measurement. Knowledgeable users and critical thinking about the nature of measures and measurement challenges can help to promote informed interest in the appropriate use of child health measures. The research component of the APA mission clearly could provide collaborative research opportunities to help test and validate child health measures. In terms of policy, we can speak both to the need for specic types of measures and the appropriate application of measures in the development and monitoring of policies. We can also use child health measures to identify areas of child health that need to be improved and to advocate for steps to improve outcomes. Advocating for improved funding of initiatives to increase the development of needed measures is also a continuing need. In terms of health care delivery, we can and must use child health measures to assess systems of care and the ways in which they improve child health. This is important both in the treatment of acute conditions and in the long-term management of children with complex health needs. As we develop integrated systems of care and accountable care organizations, it is important that the needs of children be addressed in each of the components of these systems. Finally, an infusion of funds through the new Agency for Healthcare Research and Qualitys Pediatric Healthcare Quality Measures Program Centers of Excellence U19 cooperative agreements will focus on the development of new measures of child health under the Childrens Health Insurance Program Reauthorization Act initiatives. These centers promise to make some important contributions in these areas, and it is likely that many members of the

246

STEIN

ACADEMIC PEDIATRICS
DC: US Government Department of Health and Human Services; 2000. Escalona SK. Babies at double hazard: early development of infants at biologic and social risk. Pediatrics. 1982;70:670676. National Childrens Study. Available at: http://www.nationalchildrens study.gov/Pages/default.aspx. Accessed October 5, 2010. Landrigan PJ, Trasande L, Thorpe LE, et al. The National Childrens Study: a 21-year prospective study of 100,000 American children. Pediatrics. 2006;188:21732185. DeBoo HA, Harding JE. The developmental origins of adult disease (Barker) hypothesis. Aust N Z J Obstet Gynaecol. 2006;46:414. Dube SR, Felitti VJ, Dong M, et al. The impact of adverse childhood experiences on health problems: evidence from four birth cohorts dating back to 1900. Prev Med. 2003;37:268277. Berrueta-Clement JR. Changed Lives: The Effects of the Perry Preschool Program on Youths Through Age 19. Ypsilanti, Mich: High Scope Press; 1984. OECD. Comparative child well-being across the OECD. In: Doing Better for Children OECD. 2009. Available at: http://www.oecd. org/dataoecd/19/4/43570328.pdf. Accessed August 25, 2010. Bauman LJ, Siegel MJ, Stein REK. Are children rated in poorer health less healthy? Differences by race/ethnicity. Poster presented at the Annual Meeting of the Pediatric Academic Societies, May 17, 2005; Washington, DC. Assessing health status and quality-of-life instruments: attributes and review criteria. Qual Life Res. 2002;11:193205. Heckman JJ. Policies to foster human capital. Res Econ. 2000;54: 356. Bauman LJ, Epstein SG, Gardner JD, Walker D. How well does the QuICCC identify individual children who have chronic conditions? Arch Pediatr Adolesc Med. 2000;154:447452. Riley AW. Evidence that school-age children can self-report on their health. Ambul Pediatr. 2004;4:371376. Eiser C, Morse R. Can parents rate their childs health-related quality of life? Results of a systematic review. Qual Life Res. 2001;10: 347357. Faidman A. The Spirit Catches You and You Fall Down: A Hmong child, Her American Doctors and the Collision of Two Cultures. New York, NY: Farrar, Straus and Giroux; 1997. Stein REK, Bauman LJ, Westbrook LE, et al. Framework for identifying children who have chronic conditions: the case for a new denition. J Pediatr. 1993;122:342347. Perrin EC, Newacheck P, Pless IB, et al. Issues involved in the denition and classication of chronic health conditions. Pediatrics. 1993; 91:787793. Stein REK. Measurement of childrens health. Ambul Pediatr. 2004;4: 365370. University of Massachusetts Ofce of Communications. Available at: http://www.umassmed.edu/Content.aspx?id85520. Accessed October 5, 2010. Irwin DE, Stucky BD, Thissen D, et al. Sampling plan and patient characteristics of the PROMIS pediatrics large-scale survey. Qual Life Res. 10:585594. Offord Centre for Child Studies. Available at: http://www.offordcentre. com/readiness/. Accessed October 5, 2010. Australian Early Development Index. Available at: http://www.rch. org.au/aedi/index.cfm?doc_id13051. Accessed December 27, 2010. Guyer B, Ma S, Grason H, et al. Early childhood health promotion and its life course consequences. Acad Pediatr. 2009;9:142149.

APA will participate in and benet from this initiative. The Patient Protection and Affordable Care Act of 20107 also will call for more rigorous assessment of child health and may afford additional opportunities. However, it is likely that both types of initiatives will be largely driven by the need for shorter-term cost containment rather than theoretical issues or priorities in the improvement of long-term health and in fostering the science of child health measurement.

9. 10. 11.

12. 13.

CONCLUSION
With all the progress to date and the new opportunities, it is an exciting time to be involved in the measurement of child health. This is because we now have a well-conceived child-based denition and theoretical model that incorporates the many inuences that affect child health and allow for the inclusion of both genetics and epigenetics. Moreover, the long-term economics of health care are nally forcing nations to focus on childrens health to prevent the epidemics of adult diseases that are programmed in early childhood.30 It is in this context that we all should enthusiastically embrace the challenge of measuring child health as we move into the next half century of APA.

14.

15.

16.

17. 18. 19.

ACKNOWLEDGMENT
I thank Ellen J. Silver, PhD, for her thoughtful comments on this manuscript.

20. 21.

REFERENCES
1. Hutchins VL. A history of child health and pediatrics in the United States. In: Stein REK, ed. Health Care for Children: Whats Right, Whats Wrong, Whats Next. New York, NY: United Hospital Fund; 1997. 2. Shonkoff JP, Phillips D, eds. From Neurons to Neighborhoods: The Science of Early Childhood Development. Washington, DC: National Academies Press; 2000. 3. National Research Council and Institute of Medicine. Childrens Health, The Nations Wealth: Assessing and Improving Child Health. Washington, DC: National Academies Press; 2004. 4. Preamble to the Constitution of the World Health Organization as adopted by the International Health Conference, New York, 1922 June 1946; signed on 22 July 1946 by the representatives of 61 States (Ofcial Records of the World Health Organization, no. 2, p. 100) and entered into force on 7 April 1948. 5. Ottawa Charter for Health Promotion, First International Conference on Health Promotion, Ottawa, 21 November 1986WHO/HPR/HEP/ 95.1 Available at: http://who.int/hpr/NPH/docs/ottawa_charter_hp. pdf. Accessed February 7, 2005. 6. Stein REK, Gortmaker S, Perrin E, et al. Severity of illness: concepts and measurement. Lancet. 1987;2:15061509. 7. H.R. 3590: The Patient Protection and Affordable Care Act. Available at: http://www.govtrack.us/congress/bill.xpd?billh111-3590. Accessed November 15, 2010. 8. U.S. Department of Health and Human Services. Healthy People 2010: Understanding and Improving Health. 2nd ed. Washington,

22.

23.

24.

25. 26.

27.

28. 29.

30.