OBJECTIVE. Despite the widespread use of the Maternal and Child Health Bureau definition of children with special health care needs, no published studies have considered the “at-risk” component of the definition. The purpose of this article is to present a conceptual model of risk for special health care needs.
METHODOLOGY. The conceptual model presented here was developed based on a comprehensive review of the literature on the determinants of population health and the etiologic literature for selected representative childhood chronic conditions.
RESULTS. Our conceptual model is built on 5 key pillars derived from the literature. First, determinants of health have been demonstrated to include genetic endowment, the physical and social environment, health-related behaviors, and the health care system. Second, the model recognizes that the relative importance of each of these domains in contributing to the presence of a special health care need is likely to vary across the major chronic conditions experienced by children. Third, these domains can be conceptualized as acting at the child, family, community, or societal level. Fourth, the model recognizes the presence of a complex interplay of causal factors influencing the development of chronic conditions and associated special health care needs. Fifth, the model incorporates a temporal aspect to the development of special health care needs.
CONCLUSIONS. The conceptual model presented here represents a starting point for thinking about the risk factors that influence the occurrence and severity of a special health care need. The model incorporates many of the important breakthroughs by social epidemiologists over the past 25 years by including a broad range of genetic, social, and environmental risk factors; multiple pathways by which they operate; a time dimension; the notion of differential susceptibility and resilience; and a multilevel approach to considering risk. Nevertheless, we recognize that the conceptual model represents an oversimplification of reality. The study of risk factors for special health care needs remains largely in its infancy and is ripe for additional development.
With large-scale improvements in sanitation, nutrition, health care, and general living conditions, child mortality rates have declined markedly over the past century.1–3 Although some infectious diseases have been eradicated and others have emerged, chronic conditions have come to dominate the illness burden of children.4–6 Children with special health care needs (CSHCN) are an important and highly vulnerable subset of the child population affected by chronic conditions.
Historically, the federal and state maternal and child health programs served this population. Although these programs initially focused on orthopedic impairments (consequences of polio being the most common), they gradually broadened their focus in response to the changing epidemiology of childhood chronic conditions.7–10 As the number and nature of the conditions treated by these programs expanded, the nomenclature used to describe the population shifted from “crippled children” in the 1930s to “handicapped children” in the 1960s and 1970s to CSHCN in the late 1980s.
Until 1998, no widely accepted definition of CSHCN existed. That year, the Maternal and Child Health Bureau (MCHB) published a new definition of CSHCN as those “children who have or are at increased risk for a chronic physical, developmental, behavioral, or emotional condition and who also require health and related services of a type or amount beyond that required by children generally.”11 This definition has 3 key attributes: (1) it is inclusive by incorporating all types of chronic conditions, whether physical, developmental, emotional, or behavioral; (2) it is consequence based or need based, in that the child's condition must result in an elevated service need to be included; and (3) it includes the population at increased risk for developing a special health care need. Including those without an existing special need but at increased risk of developing one reflects the public health orientation of MCHB. There was a consensus within the work group that developed the definition that it is better to prevent a child from developing a special health need than to treat the child after he/she has developed the special health need, hence the inclusion of at-risk children.11
The definition has been widely used in the field by state Title V Maternal and Child Health programs, Medicaid programs, health plans, and health services researchers. Tools have been developed to identify CSHCN based on this definition.12,13 A number of journal articles have also been published using the definition.14–22 Current estimates suggest that 13% to 18% of US children have an existing special health care need.14,15,23 In addition to the limitations related to their condition, CSHCN are at heightened risk for mental and behavioral health problems,11,24,25 bed days and school absence days,26 and having unmet health care needs.27 Economically, it has been estimated that the 13% to 18% of the child population with existing special health care needs account for ≥40% of all child-related health care costs.22,28 Although trend data are not available for the entire population of CSHCN, there is substantial evidence that the subgroup of CSHCN most severely affected by their conditions, those with disabilities, has grown considerably in the past 4 decades.2,4,5
Despite the widespread use of the MCHB definition of CSHCN, no published studies have considered the at-risk component of the definition. This is partly because resources are already stretched thin in serving children with existing special needs. However, another reason for ignoring this important component is that the population at increased risk has never been defined conceptually or described empirically. Until progress is made in this area, no inroads can be made in designing and implementing primary and secondary prevention programs to reduce the number of children developing special health care needs. By default, the present emphasis on treatment over prevention is perpetuated. The purpose of this article is to begin the process of conceptualizing the population at risk for special health care needs. Future work will encompass empirical estimation of the conceptual model described here.
Epidemiology of Risk
The science of risk assessment is well established in the epidemiological literature and provides an important foundation for conceptualizing the at-risk population. Generally speaking, risk assessment is the quantitative estimation of the likelihood of adverse effects that may result from exposure to specified health hazards or from the absence of beneficial influences.29,30 The process of risk assessment traditionally consists of 4 steps. Modified for the case of CSHCN, the 4 steps are described below.
Identification of Risk Factors and Modes of Resilience
This involves identifying the factors that contribute to increased risk of special health care needs, as well as protective factors that serve to ameliorate or reduce risk, the mechanisms by which they operate, the target population of children, and the conditions of exposure.
Characterizing Risk and Resilience
The characterization of risk and resilience involves describing the potential effects of the risk factors and the role of resilience on the presence of a special health care need, quantifying dose-effect and dose-response relationships where possible.
Assessment involves quantifying exposure to risk factors and the presence of resilience in the target population.
Risk estimation involves combining identified risk factors and mechanisms of resilience, characterizing risk and resilience, and assessing exposures to quantify the risk levels within the target population.
In this article, we are primarily concerned with identifying risk and resilience factors and their underlying causal mechanisms. Subsequent work will address the remaining components of a complete risk assessment.
Conceptualization of Risk and Resilience
The first step toward preventing children from developing special health care needs is to conceptualize and describe those factors that lead to an elevated risk of developing a special health care need and those factors that protect children from the same. Although there are no such conceptual models specific to CSHCN, much work has been devoted to assessing the determinants of health for the broader population. A second set of literature has developed around the etiology of specific diseases, such as asthma. We used both sets of literature in the development of the conceptual model presented here. The body of work on population health provides a particularly useful foundation for developing a conceptual model of the risk factors for developing a special health care need.
Determinants of health are broadly conceived as the medical and nonmedical factors that influence the health of individuals and communities.31–41 Within the research and public health community, the focus on nonmedical determinants is comparatively recent. Although the limited impact of medical care on population health and the importance of nonmedical factors, such as sanitation, has been known for many years,42,43 it was not until 1974 that policy-makers began to fully recognize the influence of nonmedical determinants of health in the development of national health policies. That year, the Canadian government issued a report entitled A New Perspective on the Health of Canadians,44 which stated the case for focusing on the nonmedical determinants of health as part of heath policy. The conclusions were grounded in research conducted in the United States and England showing that the major contributions to improved health in those countries came more from improvements in sanitation, the food supply chain, and income than from improvements in medical care.44,45
After the landmark Canadian effort, a number of reports, articles, and publications have been issued describing different conceptualizations of the determinants of population health.34,38,46–51 Most of these models focus on population health generally, but variants have been developed for children.35,36,51 Although the terminology varies from author to author, these conceptual models generally classify health determinants into 5 broad domains: genetic endowments, the social environment, the physical environment, health-influencing behaviors, and medical care.31,33,34,37,38,45,51,52 Critical to this multifaceted approach to thinking about population health is the understanding that the determinants do not act in isolation from each other. Rather, it is the complex interaction of these determinants and patterns of exposure that have an impact on the health of the population.53,54 It is also worth noting that the importance of these determinants may vary with the age and developmental status of the child.51
Another important observation concerns the mechanisms by which nonmedical factors are related to health outcomes. Frost55 noted in 1937 that it was not just the increased exposure to risk factors that resulted in higher rates of tuberculosis among the poor but hypothesized that there was also something about their susceptibility to disease after exposure that contributed to the higher rates. Cassel33 and Syme and Berkman56 have expanded on this argument with their observations that social conditions are linked to a broad spectrum of diseases and illnesses. They speculate that social conditions, broadly defined, influence the etiology of illness by creating a general susceptibility to disease. Hence, it is not simply the level of exposure to adverse risks in the physical and social environment that is important but also the extent to which an individual is susceptible or resilient that determines risk of illness.
Although the population health literature provided our starting point to examine the broad determinants of health, we also examined the clinical and epidemiological pediatric literature to assess whether the 5 broad domains that have been demonstrated to influence population health generally (genetic endowments, the social environment, the physical environment, health-influencing behaviors, and medical care) also hold for specific childhood chronic conditions. Indeed, given the varying etiology of individual chronic conditions, it seems likely that the relative importance of the 5 domains would vary from condition to condition. To investigate this hypothesis, we selected a broad range of specific childhood chronic conditions (asthma, depression, autism, attention-deficit/hyperactivity disorder [ADHD], learning disabilities, retinopathy of prematurity and obesity) for further review. These conditions were selected to ensure that we included examples of physical, developmental, behavioral, and emotional conditions that are encompassed in the MCHB definition.
We found that the current understanding of the initiation of the disease process is still poorly understood for most childhood chronic conditions and that the role of health determinants varies from condition to condition. Asthma is one of the most studied disease processes, with a number of authors pointing to the importance of a child's genetic endowment in combination with the physical environment. Specifically, a genetic susceptibility to atopic disease is thought to interact with different environmental factors (eg, dust mites, tobacco smoke, infections, diet, allergens, etc) to produce a child with asthma.57–60 Once symptoms are present, medical care is thought to influence the impact of the chronic condition on the child and the extent to which the child's activities may be limited.61
Behavioral and emotional conditions (represented in our literature review by ADHD, autism, learning disabilities, and depression) are also hypothesized to be caused by an interaction among genetic endowment, the social environment, and the physical environment.62–64 Through twin and relative studies, researchers have attempted to apportion the contributions of these various factors; estimates of the genetic component of ADHD, for example, generally start at 30% to 40% and go much higher.62,65,66
Specific environmental triggers for these conditions have not been as well identified as they have been in asthma and remain controversial.66 The impact of medical care on long-term outcomes is generally less well-established than in physical conditions, such as asthma, although it is important in some outcomes.67
Retinopathy of prematurity was selected as one example of the special health care needs that preterm and low birth weight infants are at an increased likelihood of developing.68–70 These infants continue to be a challenging public health issue in the United States; in 2003, 12.3% of infants were born at <37 weeks' gestation (an increase of 16% since 1990), and 1.4% of infants weighed <1500 g at birth.71 Risk factors for premature and low birth weigh infants may be found among virtually all 5 determinants of health; maternal smoking, low socioeconomic status, prepregnancy maternal health, and air pollution are just a few examples.72,73
Finally, obesity is influenced by health-influencing behaviors more than other medical conditions, although other determinants are clearly important as well. Diet, including level of consumption and composition of caloric intake, as well as the frequency and amount of exercise, are clearly linked to obesity.74–76 Other factors suspected as contributing to obesity include maternal smoking during pregnancy, low caloric intake during the first trimester of pregnancy, low levels of maternal education, and a history of parental obesity.77–79 Twin, adoption, and family studies indicate that inheritance accounts for 25% to 40% of interindividual difference in adiposity.79
Thus, the disease-specific literature that we reviewed confirms that each of the 5 key domains that determine population health (genetic endowments, the social environment, the physical environment, health-influencing behaviors, and medical care) also play a role in influencing the emergence and impact of childhood chronic conditions. We also observed that the relative importance of each domain in influencing emergence and impact varies from condition to condition, although the precise contribution of the domains is not well understood for any of the conditions reviewed here. These findings suggest that a single model incorporating the 5 major domains is appropriate for conceptualizing risk factors for special health care needs as long as it is recognized that the contributions of the domains will vary from condition to condition.
Developing a Conceptual Model for At-Risk Children
Taken together, these observations from the field of population health and the etiologic literature on specific childhood chronic conditions provide a foundation for conceptualizing risk of special health care needs. Using this knowledge base as a guide, we then developed the conceptual model shown in Fig 1. It builds directly on 5 important pillars. First, determinants of health have been demonstrated to include genetic endowment, the physical and social environment, health-related behaviors, and the health care system. Our model incorporates these domains; they are assumed to operate at all points along the disease pathway. Second, our model recognizes that the contribution or strength of association between the domains and the development of a special health care need will vary by type of chronic condition.
Third, as shown in Fig 1, these domains can be conceptualized as acting at the child, family, community, or societal level. Indeed, an individual child's risk of illness cannot be considered in isolation from the disease risk of the family and community to which she belongs. Variables in our model are expressed at the level of the individual child, the family, the community, and society as a whole. The community and societal level variables are either aggregated individual/family attributes (eg, the aggregate poverty rate in a community) or ecological in that they represent community and social attributes that cannot be derived directly from individuals/families (eg, characteristics of the health care system).
A fourth critical lesson concerns the complex interplay of the factors influencing health. Causality can be expressed directly and indirectly, and causal forces can also be mediated by other factors. Reciprocal causal relationships and feedback loops are also possible. The complexity of the causal relationships is reflected in Fig 1 by the overlapping ovals and by the intentional absence of arrows that would overly simplify reality. Our approach here builds directly on recent work on conceptualizing population health.50
Finally, population health researchers have demonstrated a temporal aspect to the development of illness. As we age, we do not remain static, but change in a number of significant areas: the prevalence of given diseases, the host response to a disease, susceptibility and resilience, and the impact of different determinants. This is a particularly important concept for children, because growth and development are so rapid.36,51 In our model, we incorporate this age-dependent aspect to the development of special care health needs.
While Fig 1 captures the multilevel nature of disease causation within the child population, Fig 2 presents a model for thinking about the development of a chronic condition or special health care need in a specific child. Figure 2 also contains an elaboration of 2 important theoretical constructs only touched on earlier.
One of these is the variable expression of illness. Disease outcomes are rarely binary, that is, simply present or absent.80 Rather, there are continuums of disease processes and severity. Indeed, the variable expression of chronic conditions in childhood provided an important rationale for the consequence-based approach taken in the MCHB definition of CSHCN.11 That is, whereas the presence of a chronic physical, developmental, emotional, or behavioral condition is a necessary condition for a special health care need, it is not sufficient. The chronic condition must also be expressed in the form of an elevated need for service to be considered as resulting in a special health care need. Empirical studies suggest that only approximately one third to one half of children with chronic conditions have elevated service needs.15 The final box in Fig 2 illustrates the concept of a spectrum of health, ranging from no condition to a chronic condition to a special health care need.
Next, the observation that some individuals are more susceptible to disease than others, despite exposure to similar risk factors, implies that a child's level of resilience is an important mediating factor in health outcomes; we have incorporated the notion of resilience into our child-specific model of risk. Resilience and susceptibility are factors that act to moderate a child's exposure and reaction to environmental influences and, thus, influence the development of a chronic condition or special health care need.
Despite the relative newness of the resilience concept, considering children as “active” contributors to their own health (or illness) and having “positive” outcomes is an important advance in modeling the development of illness. Most of the conceptual literature on the subject describes 2 component parts to resilience: that which is innate and that which is acquired.81–83 Aspects of innate resilience (also called assets) are represented by traits such as coping, interpersonal skills, and self-confidence. Recent literature has described a genetic component to the development of some of these assets.84 Parts of acquired resilience (also called resources) include family support, community role models, and public and private organizations accessible to the child. It is important to understand that resilience can only be looked for in a fraction of children: in order for resilience to be demonstrated, a child must first have been exposed to various risk factors. Susceptibility, the converse of resilience, predisposes a child to poorer outcomes than would be expected for exposure to a given set of risk factors. Although the literature on these topics concerns mostly behavioral and emotional conditions,81,85 the role of susceptibility and resilience for physical conditions, such as asthma and obesity, can be conceptualized in the same manner.
Opportunities for Prevention
Once risk factors have been identified, the conceptual models shown in Figs 1 and 2 suggest there are ≥2 intervention points in preventing children from developing special health care needs (see Fig 3). The first intervention point represents an opportunity for population-based primary prevention through reduction of exposures to adverse risk factors that would otherwise lead to the initiation of disease pathology. The second intervention point (secondary prevention) is coupled to ameliorating the factors that facilitate conversion of an existing chronic condition into a special health care need.
There are a number of levels where these interventions could take place. At the practitioner level, providers could choose to see children with identified risk factors more frequently and have a lower threshold for administering a screening test. Communities could provide referral programs to high-risk children to ameliorate a child's risk factors and strengthen his or her resilience. States could provide all children with sufficient risk factors public health insurance and other services.
Based on our review of the population health literature and the conceptual model presented in Figs 1 and 2, we would argue that a multifaceted approach to prevention policy is needed. The strategy should incorporate a focus on the child, family, and community and should consider a broad spectrum of medical and nonmedical influences on health. We know little at this point, however, about where the emphasis should be placed in developing an effective and efficient prevention approach. Given limited resources, would we be better off investing in primary or secondary prevention approaches or some combination? Similarly, would a prevention program that focuses on chronic conditions generally be more efficient than a set of programs organized around subsets of chronic conditions that share similarities in etiology? And are prevention programs more effective when initiated at the provider level or at another level? Identifying cost-effective strategies for preventive interventions requires a knowledge base of empirical findings. At present, the research needed to inform prevention policy is largely in a nascent stage of development.
Strengths and Weaknesses of the Conceptual Model
The conceptual model described and presented above represents a starting point for thinking about the risk factors that influence the appearance and severity of a special health care need. The model incorporates many of the important breakthroughs by social epidemiologists over the past 25 years, including a broad range of genetic, social, and environmental risk factors; multiple pathways by which they operate; a time dimension; the notion of differential susceptibility and resilience; and a multilevel approach to considering risk whereby risk factors may operate differentially at the child, family, and community level.
Our model, by incorporating current epidemiological thinking, provides a starting point for conceptualizing risk factors. Nevertheless, it remains a great oversimplification of reality. We know that reciprocal causal relationships exist among risk factors and between risk factors and the outcome variable. For example, a child exposed to poverty may be at greater risk of experiencing a special health care need because of the material deprivation associated with poverty. A special health care need may also result in impoverishing a family if parents must eliminate or cutback on employment to care for the child. A more realistic model would incorporate a number of circular processes and feedback loops, but doing so greatly increases the complexity of the model and the feasibility of testing it empirically. This tug of war between realism and feasibility creates a dilemma for researchers. On one extreme are models so complex that they ultimately cannot be modeled or even understood in the 2 or 3 dimensions normally used to conceptualize relationships among variables. On the other extreme are simple models that propose a direct cause-and-effect relationship between a single risk factor and an outcome. We have tried to take a middle ground approach by building a model that remains relatively simple but hopefully realistic enough to be useful.
Another limitation stems from the underlying science of risk factors for special health care needs. Any model is only as good as the science that serves as its foundation. To state the obvious, not all risk factors for special health care needs have been identified, and not all of the causal pathways are understood. Research continues to reveal new risk factors and unravel new causal pathways. The rapidity of scientific advancement in recent decades, particularly in genomics, underscores this point. In the not-too-distant future, our model may look as simplistic as the traditional epidemiological triad seems now.
Additional challenges become apparent when one considers empirically testing our conceptual model. Traditionally, chronic disease epidemiology has used a linear paradigm to examine causal relationships.86 In reality, few processes in nature seem to be linear, yet most empirical modeling techniques have, at their heart, the assumption of linearity. We also have presented a general model of the risk factors for special health care needs. While doing so, we recognized that the path to a special health care need will vary from condition to condition, such that the relative importance of genetic endowment, the physical and social environment, health behaviors, and the health care system will also vary across conditions. Ultimately, understanding patterns across individual conditions and identifying commonalities across those conditions may yield more valuable findings than an approach that focuses broadly on CSHCN as a group.
Although experimental designs are often viewed as the gold standard in health research, they are clearly impractical and unethical in the case of identifying risk factors for special health care needs. Alternative approaches include case-control and cohort designs. Case-control studies retrospectively assess differences in risk factor exposures for children with and without a particular disease outcome. The case-control approach is often used when only extant data sources are available. Cohort studies, where a group of children are followed prospectively over time, can provide a more powerful design. Exposures to risk factors are noted over time and then compared with outcomes. The main problem in conducting cohort studies is the cost of data collection. A large number of cases is needed to provide sufficiently large sample sizes for statistical analysis of risk factors, and the cohort must be followed over a sufficiently long period of time so that outcomes are apparent. The proposed National Children's Study, if funded, could provide an excellent basis for conducting a cohort study of risk factors for CSHCN. This longitudinal study is being designed to collect exposure information for a variety of genetic, social, and environmental risk factors for a large cohort of children from the prenatal period to adulthood. For now, though, retrospective case-control designs represent the only feasible option for testing the conceptual model. Although less powerful, they can play an important hypothesis-generating role by identifying risk factors that warrant additional investigation.
The study of risk factors for special health care needs is clearly in its infancy. The conceptual model presented here is meant to provide a basis for discussion among epidemiologists, practitioners, program administrators, and advocates for CSHCN. It is also intended to serve as foundation for empirical analysis of risk factors. Because of the limitations of the conceptual model and of extant data sets, empirical testing cannot be expected to verify with certainty the correctness of the conceptual model, but rather should serve the important purposes of hypothesis generation and conceptual model refinement.
This work was supported by the Federal Maternal and Child Health Bureau (grant 1R40MC03619).
- Accepted January 25, 2006.
- Address correspondence to Paul W. Newacheck, DrPH, Health Policy Institute for Health Policy Studies, 3333 California St, Suite 265, San Francisco, CA 94118. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
The opinions expressed in this article are those of the authors and do not necessarily reflect the views or policies of the institutions with which the authors are affiliated or the funding agency.
- ↵Berkman LF, Karachi I. Social Epidemiology New York, NY: Oxford University Press; 2000
- ↵Anderson RN. Deaths: leading causes for 2000. Natl Vital Stat Rep.2002;50 :1–85.
- ↵Wise PH. The transformation of child health in the United States. Health Aff.2004;23 :9– 25
- ↵Perrin EC, Newacheck P, Pless IB, et al. Issues involved in the definition and classification of chronic health conditions. Pediatrics.1993;91 :787– 793
- ↵Pless IB, Douglas JW. Chronic illness in childhood. I. Epidemiological and clinical characteristics. Pediatrics.1971;47 :405– 414
- ↵McPherson M, Arango P, Fox H, et al. A new definition of children with special health care needs. Pediatrics.102: 137–140, 1998
- ↵Newacheck PW, Strickland B, Shonkoff JP, et al. An epidemiologic profile of children with special health care needs. Pediatrics.1998;102 :117– 123
- Newacheck PW, McManus M, Fox HB, Hung YY, Halfon N. Access to health care for children with special health care needs. Pediatrics.2000;105 :760– 766
- Mayer ML, Skinner AC, Slifkin RT. Unmet need for routine and specialty care: data from the National Survey of Children With Special Health Care Needs. Pediatrics.2004;113(2) . Available at: www.pediatrics.org/cgi/content/full/113/2/e109
- ↵Bethell CD. Paper presented at: Academy Health Conference; June 26–28, 2005; Boston, MA
- ↵Dosa NP, Boeing NM, Ms N, Kanter RK. Excess risk of severe acute illness in children with chronic health conditions. Pediatrics.2001;107 ;499– 504
- ↵Wilkinson R, Marmot M, eds. Social Determinants of Health: The Solid Facts. 2nd ed. Copenhagen, Denmark: International Centre for Health and Society, Europe World Health Organization; 2003
- ↵Black SD, Townsend P, Davidson N, et al. Inequalities in Health: The Black Report and the Health Divide. New York, NY: Penguin Books; 1992
- Bunker JP, Frazier HS, Mosteller F. The role of medical care in determining health: creating an inventory on benefits. In: Amick BJ, Levine S, eds. Society and Health New York, NY: Oxford University Press; 1995:305–341
- ↵Cassel J. The contribution of the social environment to host resistance: the Fourth Wade Hampton Frost Lecture. Am J Epidemiol.1976;104 :107– 123
- ↵Evans RG, Barer ML, Marmor TR. Why Are Some People Healthy and Others Not? The Determinants of Health of Populations New York, NY: Aldine de Gruyter; 1994
- ↵Marmot MG, Wilkinson RG. Social Determinants of Health Oxford, United Kingdom, and New York, NY: Oxford University Press; 1999
- McKeown T. The Role of Medicine: Dream, Mirage, or Nemesis? 2nd ed. Princeton, NJ: Princeton University Press; 1979
- Institute of Medicine. The Interplay of Biological, Behavioral and Societal Influences Washington, DC: National Academy Press; 2001
- ↵LaLonde M. A New Perspective on the Health of Canadians Ottawa, Canada: Ministry of Health and Welfare; 1974
- ↵McGinnis MJ, Williams-Russo P, Knickman JR. The case for more active policy attention to health promotion. To succeed, we need leadership that informs and motivates, economic incentives that encourage change, and science that moves the frontiers. Health Aff (Millwood).2002;21 ;78– 93
- ↵Kaplan G. Searching for simplicity in a complex world. Paper presented at: Society of Epidemiologic Research Conference; June 16, 2004; Salt Lake City, UT
- Lynch JW, Smith GD, Kaplan GA, House JS. Income inequality and mortality: importance to health of individual income, psychosocial environment, or material conditions. BMJ.2000;320 ;1200– 1204
- Starfield B. Basic concepts in population health and health care. J Epidemiol Community Health.2001;55 ;452– 454
- Berkman LF. Social Integration, Social Networks, Social Support, and Health Oxford, United Kingdom, and New York, NY: Oxford University Press; 2000
- ↵National Committee on Vital and Health Statistics U. Shaping a health statistics vision for the 21st century Available at: www.ncvhs.hhs.gov/21st%20final%20report.pdf. Accessed July 25, 2005
- ↵National Academy Press. Committee on Evaluation of Children's Health: Children's Health, the Nation's Wealth: Assessing and Improving Child Health Washington, DC: National Academy Press; 2004
- ↵Link B, Phelan, J. Social conditions as fundamental causes of disease. J Health Soc Behav.1995(special issue) :80– 94
- ↵Stansfeld S. Social support and social cohesion. In: Marmot M, Wikinson RG, eds. Social Determinants of Health. Oxford, United Kingdom, and New York, NY: Oxford University Press; 1999:155–174
- ↵Susser M. Causal Thinking in the Health Sciences; Concepts and Strategies of Epidemiology New York, NY: Oxford University Press; 1973
- ↵Syme SL, Berkman LF. Social class, susceptibility and sickness. Am J Epidemiol.1976;104 :1– 8
- Halken S. Prevention of allergic disease in childhood: clinical and epidemiological aspects of primary and secondary allergy prevention. Pediatr Allergy Immunol.2004;15 (suppl 16) :4– 5, 9–32
- Lemanske RF, Jr. The childhood origins of asthma (COAST) study. Pediatr Allergy Immunol.2002;13(suppl 15) :38– 43
- ↵Liu AH, Szefler SJ. Advances in childhood asthma: hygiene hypothesis, natural history, and management. J Allergy Clin Immunol. Mar.2003;111(3 suppl) :S785– S792
- Berney TP. Autism–an evolving concept. Br J Psychiatry.2000;176 :20– 25
- ↵Hack M, Taylor HG, Klein N, Mercuri-Minich N. Functional limitations and special health care needs of 10- to 14-year-old children weighing less than 750 grams at birth. Pediatrics.2000;106 ;554– 560
- ↵Martin JA, Kochanek KD, Strobino DM, Guyer B, MacDorman MF. Annual summary of vital statistics–2003. Pediatrics.2005;115 ;619– 634
- Bowman SA, Gortmaker SL, Ebbeling CB, Pereira MA, Ludwig DS. Effects of fast-food consumption on energy intake and diet quality among children in a national household survey. Pediatrics.2004;113 :112– 118
- Reilly JJ, Armstrong J, Dorosty AR, et al. Early life risk factors for obesity in childhood: cohort study. BMJ.2005;330 :1357
- ↵Maffeis C. Aetiology of overweight and obesity in children and adolescents. Eur J Pediatr.2000;159(suppl 1) :S35– S44
- ↵Rose GA. The Strategy of Preventive Medicine. Oxford, England and New York, NY: Oxford University Press; 1992
- ↵Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science.2003;301 ;386– 389
- ↵Pless IB. The Epidemiology of Childhood Disorders New York, NY: Oxford University Press; 1994
- Copyright © 2006 by the American Academy of Pediatrics