Introduction

Not long ago, scientists conceived of infants as unadulterated by historic behaviors or environmental insults but now understand that babies are record keepers of maternal and paternal exposures to environmental conditions [1–3]. We propose that this record keeping at least partially occurs through a biological process known as developmental programming whereby racial disparities in health are not only propagated [4–7] but, when combined with persistent environmental exposures, are magnified from one generation to the next.

Developmental programming, or developmental origins of health and disease, refers to the changes in fetal structure and physiology that result from adverse early life exposures that increase susceptibility to poor health outcomes later in life [8]. Programmed characteristics are believed to be largely stable after the postnatal period and have been associated with increased risk of obesity, cancer, and heart disease, among others [9]. The maternal stressors that serve as exposures associated with developmental programming include environmental pollutants [10, 11], social stress [12, 13], and dietary restriction or high-calorie malnutrition [14–17].

Conceptual Framework

Decades of social epidemiologic research demonstrate that race and ethnic minorities face greater exposure to adverse social and environmental health determinants, resulting in health disparities. These disparities are persistent across generations and have been studied in numerous domains (e.g., environmental injustice, neighborhood deprivation, lower quality educational opportunities), yet the mechanisms underlying their proliferation remain unknown. We propose that developmental programming is a critical, previously overlooked factor in the propagation of health disparities from one generation to the next.

Figure 1 illustrates a paradigm of developmental programming on health, according to ongoing environmental exposures, across three generations. Health status in the first (F0) generation is determined by a woman’s lifetime accumulation of adverse exposures; D0 denotes the health status for women in generation F0 who experience adverse exposures, and A0 denotes the health status of a woman whose exposure profile is healthier. It is well known that adverse exposures are more common at greater levels of social and economic disadvantage. Therefore, we characterize trajectories stemming from exposure likelihood in the baseline generation as representing disadvantaged (D) and advantaged (A) trajectories.

Fig. 1
figure 1

Theoretical role of developmental programming in propagation and magnification of health disparities over three generations. A gx and D gx denote health status Y gx for generation g and exposure status x. β g and β g* denote the difference in health on Ygx between advantaged and disadvantaged trajectories. α g denotes the magnitude of disparity in any in generation g

Full size image

If no further harmful exposures were incurred in subsequent generations, health trajectories could be expected to remain somewhat constant, traveling along the solid lines in the figure, resulting in positions A10, A20, D10, and D20. Because disadvantage accumulates over time, one would expect the health of disadvantaged women to decline over their lifetimes. Furthermore, the weathering hypothesis, originally conceived to explain the premature worsening of health by African-American women as a consequence of cumulative social disadvantage, can be applied here as the acceleration of disease progression within a woman’s life course [18]. αg denotes the difference in health between the advantaged and disadvantaged trajectories in generation g.

β represents the decrement in health caused by an ongoing adverse exposure; both advantaged and disadvantaged women move to a position of worse health when so exposed. The F1 generation communicates this health status to any developing offspring in generation F2, who are then born from the combined baseline maternal risk (A10 or D10) supplemented by effects of any adverse exposures accumulating prior to or during pregnancy (β1 or β1*). The scenario in which adverse exposures persist from one generation to the next produces an increasing health gap over time (α1, represented by the difference between A11 and D11; and α2, the difference between A21, and D21). Such persistent exposures observed in the social epidemiology literature include the intergenerational transmission of poverty, low educational expectations, and unhealthy coping behaviors.

One important element of Fig. 1 is that the effect β is not constant but instead varies according to exposure status in the previous generation. This effect difference, by advantaged or disadvantaged trajectories (Ag1 minus Ag0 or Dg1 minus Dg0), is illustrated by comparing the exposure effects β1 versus β1* and β2 versus β2*. The magnitude of effect, β, may be influenced by the biological processes underlying developmental programming. That is, maternal prenatal exposures may alter the physiology of the gestating offspring such that the offspring is more sensitive to the effects of postnatal exposures such as environmental pollutants, stress, or high-fat diet.

Postnatal exposures, referred to above as ongoing adverse exposures, comprise second hits to biological predispositions determined during gestation. We propose that this heightened sensitivity is a mechanism through which developmental programming exacerbates health disparities across generations. The increasing magnitude of disparity that results from differential exposure effects can be seen by comparing α1 with α2. Not only are women in disadvantaged trajectories at increased risk of poor health in the absence of ongoing exposures, they, and by extension their offspring, may experience a greater health “penalty” to ongoing postnatal insults. By contrast, advantaged offspring will be less likely to experience detrimental exposures later in life but, more importantly, will be more resilient than their less advantaged counterparts. In this review, we focus on evidence that effects of key postnatal exposures differ according to maternal exposure status (β versus β*). One enduring question in the social epidemiologic literature is why health disparities persist in light of targeted health promotion endeavors. One part of the health disparities’ intransigence may be that disadvantaged populations produce a larger than expected response to ongoing social-environmental insults in part, because they have been biologically primed to do so.

Objectives

We first review evidence that maternal exposure to air pollution, psychosocial stress, and overnutrition induces exaggerated offspring responses to subsequent social environmental exposures. We then discuss the social patterning of social environmental stressors and how each type of gestational exposure may contribute to the construction and transmission of health disparities. While existing reviews describe how gestational exposures impact offspring health (e.g., [16, 17, 19]), this paper focuses on how developmental programming can exacerbate the effects of socially patterned adverse environmental conditions, like those studied by social epidemiologists.

Methods

We performed an electronic search of publications on Google Scholar, PubMED, and EBSCO Host databases. The following search terms were used: “fetal priming,” “fetal exposures,” “fetal programming,” “early life exposures,” “developmental origins,” and “intergenerational,” in combination with a second group of key words such as, “adult health outcomes,” “adult health,” “offspring,” “influences on health,” “disease in later life,” “second hit,” and “child development.” To search for relevant papers in domains of interest, we included additional terms including “air pollution” and “stress”; For maternal overnutrition exposures, we searched for (obesity or “high-fat” or “fat”) and (maternal or gestation or pregnancy) and (offspring or fetal). In addition to using key word combinations, we employed the “related articles” electronic prompts provided with the search results and examined references from relevant papers. We retained original human and animal studies that included a gestational exposure and a childhood or adulthood second hit exposure. While the papers presented in the following tables do not represent the universe of possible manuscripts, they are representative of the second hit literature as of August 2014. This review does not discuss fetal outcomes resulting from gestational exposures, even though this work is relevant to future adult health.

Results

Environmental Pollutants

Poor environmental quality, like other social epidemiologic phenomena, has its greatest impact on those whose health status is already at risk [20], such as individuals living on the disadvantaged trajectory in our conceptual model. Exposure to environmental pollutants, inadequate buffering of environmental exposures, and increased susceptibility to environment-induced poor health owing to existing comorbidities are differentially distributed [21–24]. We propose developmental programming as an additional explanation for the disparate effects of environmental exposures by considering the evidence that gestational exposures to air pollution predispose offspring to worse physiological responses to subsequent air pollution experiences.

Evidence from animal studies suggests that exposure to air pollutants by neonates (e.g., mouse pups) during gestation is associated with an exaggerated reaction to air pollution exposure during child- or adulthood (Table 1). In studies by Auten and colleagues [25••, 26], pups born to moms who were exposed to diesel exhaust while gestating, who were then exposed to ozone during their childhood, developed persistent airway hyper-reactivity. Effects of childhood ozone exposure were stronger in pups born to dams exposed to diesel exhaust while gestating, than in pups born to dams that were instead exposed to filtered air [25••]. This paper follows earlier work, which reported that maternal diesel exposure induced inflammatory cytokine responses in the offspring that presumably contribute enhanced airway hyper-responsiveness to postnatal ozone exposure [26].

Table 1 Studies illustrating gestational air pollution priming followed by air pollution second hit
Full size table

A handful of observational epidemiologic studies examining gestational and later-life exposures investigated if fetal priming of later response is generalizable to humans. Perera [27] and colleagues collected polycyclic aromatic hydrocarbon (PAH) data using a personal air monitor during the third trimester of pregnancy. Postnatal interviews were administered at 6 months postdelivery and annually thereafter to determine changes in PAH exposure. Youth exposed to high amounts of PAH in utero exhibited lower intelligence quotient scores, but postnatal exposure did not exacerbate this effect. A second epidemiologic effort to tease apart effects of gestational and adult exposures comes from Montgomery and Ekbom [28]. They considered maternal smoking, a form of gestational air pollution, and the offspring’s own smoking at age 16 as the second air pollution hit related to diabetes development. After adjustment for maternal smoking, cigarette smoking during adolescence was independently associated with increased risk of diabetes. No evidence of hyper-sensitization was observed, but the authors went on to suggest mechanisms by which this observed effect is biologically plausible.

While the number of studies considering gestational air pollution exposures followed by subsequent air pollution is relatively small, the results suggest that gestational exposure primes the child or adult body for a more severe reaction to subsequent exposures. This finding is likely to be highly relevant for human health given the ubiquity of air pollution. Further, because social stratification results in non-randomly distributed adverse air quality, those most vulnerable to the effects of poor air quality are most likely ensured of ongoing adverse exposures.

Psychosocial Stress

Lack of autonomy within a social hierarchy, inadequate income, discrimination, and inadequacy of time and resources are sources of chronic psychosocial stress in disadvantaged populations and are linked with adverse health outcomes. Maternal stress induces elevations in the glucocorticoid hormone, cortisol [29]. Excessive maternal cortisol can overwhelm placental capacity to block its passage; fetal exposure to cortisol induces widespread developmental alterations that increase susceptibility to disease [12]. Notably, glucocorticoids are a common pathway through which a wide range of adverse maternal exposures induces developmental programming [30]. We reviewed evidence that maternal stress during pregnancy induces exaggerated response to stress experienced by the offspring.

In animal studies, offspring born to dams exposed to restraint stress, who then experienced stress during adolescence, have altered immune, neuroendocrine, or glucocorticoid systems, compared to those animals not exposed to prenatal maternal restraint (Table 2). Pascuan and colleagues [31] found that prenatal stress altered the intrinsic regulation response to acute stress in adult mice, while Xu and colleagues [32] found that prenatal restraint of pregnant rats led to a hyper-responsiveness of the HPA axis, and increased anxiety-like behavior in rats. Van den Hove and colleagues found that, compared with control rats, chronic variable mild stress in adolescent rats that had been exposed to maternal restraint stress during gestation partially offset the effects of prenatal stress, suggesting that responses to developmental stress may depend on later life exposures [33••]. In contrast, Schopper and colleagues [34] showed that intermittent light stress induced elevated glucocorticoid levels in maternal guinea pigs but that their offspring had an attenuated cortisol response over adult development and subsequent stress exposures.

Table 2 Studies illustrating gestational stress priming followed by stress second hit
Full size table

Human evidence for the effect of second hit stress exposures following prenatal stress is scant. One study collected the peripheral blood mononuclear cells from 34 young women whose mothers reported negative life events during their pregnancy (and from a comparison group whose mothers did not experience negative life events). The cytokine response to a potent T cell stimulator, or a phytohemagglutinin (PHA) second hit, was significantly altered in women born to mothers who were stressed while gestating [35], suggesting that prenatal stress exposure may induce an exaggerated immune response in adult women. Regardless of the exact source of social stress, maternal stress while gestating likely induces physiologic changes in the developing fetus, such that subsequent postnatal (lifetime) stress elicits perturbations in the offspring’s stress response.

Nutrition

“Western-style” diets, characterized by high fat and high refined sugar content, are ubiquitous in Western populations and even more pronounced in racial minority and low-income populations [36, 37]. The racial and neighborhood-level patterning of poor-quality food environments is a significant area of social epidemiologic research. We reviewed evidence for the hypothesis that prenatal Western-style diet (high fat, high sugar, or high-total energy) induces exaggerated offspring responses to Western-style diet in postnatal life (Table 3). A growing body of animal research supports this hypothesis. For example, Bruce and colleagues examined the combined effects of maternal high-fat diet during preconception, gestation, and lactation in mice and high-fat diet consumed by the offspring upon weaning [38]. Postweaning high-fat diet induced accelerated weight gain, higher adiposity, cholesterol and insulin resistance, and development of non-alcoholic fatty liver disease characteristics. Notably, these effects were exacerbated by maternal high-fat diet. Fan and colleagues conducted a similar experiment in non-human primates and evaluated prenatal and postnatal diet impacts on metabolic and cardiovascular outcomes [39••]. Compared to controls, insulin resistance was observed only among monkey offspring who consumed high-fat diets during both prenatal and postnatal life. Several markers of atherosclerotic risk such as intima-media thickness were more responsive to postnatal high-fat diet among offspring of high-fat diet-fed mothers, compared to offspring of control mothers.

Table 3 Selected studies illustrating gestational diet priming (overnutrition) followed by dietary second hit
Full size table

In other studies using mice, rats, sheep, and non-human primates, adverse measures of adiposity, liver triglyceride content, fatty acid content, leptin, glucose, and insulin were more pronounced in offspring exposed to both prenatal and postnatal high-fat or high-sugar diet compared to offspring having only prenatal dietary exposures [40–47]. Such augmented sensitivity could be due to changes in insulin production [48], perturbation to central appetite control [43, 47, 49, 50], altered circadian regulation of liver metabolism [51], or changes in the microbiome [52]. The effects of postnatal diet among offspring exposed to prenatal Western-style diet, compared to offspring of control-fed dams, varied within study by gender [42, 45, 46]. In other studies, the effects of postnatal overnutrition were similar across prenatal diet groups [53, 54] or, for body weight outcomes, even weaker [55, 56]. In many studies, it was difficult to distinguish between cumulative versus synergistic effects of prenatal and postnatal diets, as investigation of “priming” or “hyper-responsivity” to postnatal exposures was not the primary aim of the research. We were not able to identify relevant human studies. Nevertheless, animal data send the clear message that determining how prenatal high calorie malnutrition exaggerates the detrimental effects of a poor postnatal diet in at-risk human populations is urgent.

Cross-Domain Exposure Effects

Humans live in environments characterized by simultaneous exposures from multiple domains, each of which are socially structured and would likely exacerbate existing health disparities. The animal literature has begun to replicate this exposure complexity by conducting cross-exposure experiments (Table 4). Recent research by Wei and colleagues exemplifies this work. Wei [57] found that prenatal exposure to bisphenol A, a synthetic compound with hormone-like properties commonly employed in plastics production, resulted in defective insulin signaling among offspring exposed to standard diets. However, compared to unexposed controls, the prenatally exposed BPA pups developed more extensive liver impairment (e.g., increased inflammation and mild fibrosis) following exposure to high-fat diets. The authors concluded that prenatal BPA exposure predisposed offspring who later consumed high-fat diets to develop metabolic syndrome manifestations.

Table 4 Cross-exposure papers: studies illustrating gestational contaminant priming followed by nutrition second hit
Full size table

In 2014, Bolton and colleagues [58••] reported that prenatal diesel exhaust exposure, followed by a postnatal high fat (compared to a low fat) dietary second hit, programmed offspring for diet-induced metabolic, behavioral, and neuro-inflammatory changes in adulthood and that these changes were greatest among males. This work followed their prior report [59] that prenatal exposure to diesel exhaust followed postnatally by high-fat diet programmed offspring for weight gain and neuro-inflammation in adulthood. These findings, too, differed by sex: postnatal high-fat diet induced 340 % more weight gain in females with prenatal diesel exposure than females with prenatal filtered air exposure, while postnatal high-fat diet induced 450 % higher insulin in prenatal diesel-exposed males, compared to prenatal filtered air-exposed males.

Another example illustrates the potential cumulative effects of multiple, cross-domain prenatal insults. Tamashiro and colleagues conducted a unique experiment in rats, in which dams were exposed to high-fat diet, an 8-day variable stress protocol, a combination of the two, or control conditions during pregnancy [60•]. Offspring were fed either a control or high-fat diet upon weaning. The postweaning high-fat diet induced greater weight gain and, among male offspring, lower glucose tolerance than in controls.

Humans living on the disadvantaged trajectory are most likely simultaneously exposed to poor air quality, high stress, and nutrition-poor food environments. Social epidemiologic research suggests that barring substantial intervention, residents in disadvantaged areas are likely to remain there. Therefore, these cross-domain exposure effects have profound implications for social epidemiologic research and the construction of persistent health disparities.

Discussion

The mechanisms by which health disparities propagate through generations are not well understood. In this article, we considered developmental programming as a possible contributor to persistent health disparities. We reviewed published research in three prenatal and postnatal exposure domains: air pollution, psychosocial stress, and overnutrition. A growing body of evidence supports the hypothesis that prenatal exposures to environmental stressors prime offspring for exaggerated response to childhood or adulthood exposures. While early in its development, these findings are similar across species, including a small number of observational human studies, as well as across domains. Taken together, this research justifies further consideration of how developmental programming may contribute to constructing and exacerbating health disparities across generations.

Effect Modification of Established Social Processes by Developmental Priming Status

Health is propagated across generations through multiple pathways, including genetics, behavioral modeling and literacy, and environmental continuity. We propose developmental programming not as a competing or alternate theory of intergenerational transmission but rather as a previously unappreciated process that can magnify the effects of social epidemiologic intergenerational processes. While genetic susceptibility puts some individuals at higher risk of specific health conditions, it rarely varies by race [61, 62] and is not modifiable, so will not be discussed further in this review.

In our conceptual model (Fig. 1), health disparities are exacerbated from one generation to the next as a result of (a) the effects of suboptimal gestational environments on exaggerated offspring response to subsequent environmental and social exposures (β1 versus β1* and β2 versus β2*) and (b) intergenerational continuity of adverse exposures (prevalence of Ag0, Ag1, Dg0, Dg1 scenarios). Vast social epidemiologic evidence supports the continuity of adverse exposures, perhaps most thoroughly for residential conditions. Racial residential segregation results in the geographic clustering of non-white or non-wealthy individuals in the USA [63]. Environmental disamenities are also non-randomly distributed [64] and cluster geographically, particularly in low-wealth, non-white neighborhoods. This is true for sources of air pollution [65, 66], crime, environmental sources of psychosocial stress such as graffiti and social disorder [67–70], liquor stores [71, 72], unhealthy food environments [73–76], and inadequate physical activity resources [77–79]. Social epidemiologic research consistently indicates that residence in close proximity to negative environmental conditions is associated with increased risk of a variety of poor health outcomes [80–82].

Other mechanisms that contribute to the inequitable allocation of chronic, adverse exposures have been proposed; their effects could also be modified by an individual’s developmental priming status. Contemporary social conditions deriving from historic injustices, including subjugation (slavery for African Americans and displacement of Native Americans) [83, 84] and ongoing racial/ethnic discrimination [85–87] contribute to chronic social stress and have been considered in the social epidemiologic literature. Intergenerational poverty [88, 89] and parental modeling of health-related behavior [90] have also been implicated in poor offspring health outcomes and are among the possible pathways through which health disparities are propagated.

The implications of the combined gestational priming plus differential distribution of adverse exposures is that residents of suboptimal environments will have both poorer health at baseline and a propensity for an exaggerated negative health response to subsequent adverse exposures, which are differentially distributed throughout the next generation’s life course. It suggests that developmental priming processes and social epidemiologic exposures interact to socially structure health states.

Social Epidemiologic Research Gaps

Our proposed theory of the contribution of developmental programming to intergenerational health disparities is based upon empirical evidence in our literature review and decades of social epidemiologic research. Yet, research on the intersection between developmental programming and health disparities is scant.

A handful of studies suggest differences in methylation patterns according to social factors [91, 92], but this area of study is in its infancy. Another critical component in testing our theory is investigation of the extent to which developmental programming biomarkers explain differences in health and responsivity to exposures in advantaged versus disadvantaged groups. Such differences likely vary according to disease outcome, specific social groups, and life stage.

Inclusion of developmental programming biomarkers would help to overcome limitations stemming from the vast number of factors that are entangled with social disadvantage. However, key barriers to this line of inquiry are the still growing knowledge about methylation sites and other biomarkers relevant to specific disease processes. Such knowledge is critical for linking intergenerational health disparities with mechanisms identified in biological developmental programming research.

Social Epidemiologic Research Challenges

Investigation of developmental programming as a contributor to health disparities requires observational research on socially patterned exposures and outcomes in human study populations. Such epidemiologic research poses several methodological challenges.

Resilience

Not every person whose mother is exposed to poor air quality goes on to develop airway hyperreactivity; nor does every person whose mother was obese during pregnancy become obese him/herself. The presence of resilience, or the ability to properly adapt to stress and adversity, is either a robust homeostasis-supporting biologic phenomenon or it constitutes counter-evidence for developmental programming effects. Research that carefully assesses prenatal and postnatal exposure from a developmental priming perspective, with special attention to resilience and the factors that seem to support healthy coping, will help quantify what portion of health outcomes, if any, might be attributable to programming effects.

Long-Term, Intergenerational Follow-Up

Our most robust understanding of developmental programming examines intergenerational effects of experimentally assigned maternal exposures. Adverse offspring later-life effects are often observed even when the gestationally exposed and non-exposed are similar (in their size, other biomarkers) at birth. Therefore, the indicators of adverse fetal development (e.g., birth weight) typically used in epidemiologic studies are inadequate as measures of adverse developmental programming. Rather, measurement of developmental programming requires data collection on both mothers and offspring, in study populations large enough to test prenatal and postnatal interactions. Research needs will include the use of data linkage, improved retrospective maternal health status assessment, and intergenerational follow-up. Studying priming effects or second hits in humans, even observationally, will require long-term cohort construction and funding programs that endure across generations.

Within-Person, Within-Family Variability

Many of the exposures that we know contribute to fetal priming are ubiquitous (e.g., social stress) and have become part of our daily environment (e.g., diesel exposure). As a result, constructing unexposed groups for causal comparisons is challenging. In addition, it is difficult to disentangle maternal effects (e.g., maternal high-fat diet) from offspring effects (e.g., ongoing offspring high-fat diet). Existing human research studying prenatal/postnatal interactions typically rely on short-term experimental interventions such as a short period of over-feeding or inactivity [93] to attain variability in postnatal exposures. However, such effects are unlikely to be comparable with the chronic exposures humans experience. Social epidemiologic research of developmental priming may require large, cross-national studies to construct adequate exposure variability to observe intergenerational effects.

Complexity

Our review suggests that multiple, interrelated and socially patterned prenatal and postnatal exposures contribute to intergenerational disparities. The inclusion of developmental programming into social epidemiologic models introduces additional complexity to etiologic pathways for understanding health disparities. For example, maternal overnutrition leads to neurological rewiring of appetite circuits and the propensity to gain weight among offspring; thus elevated appetite is a partial mediator in the pathway from developmental programming to obesity. The investigation of developmental programming as both a mediator and a moderator of typical social epidemiologic exposures requires innovative analytic approaches such as emerging decomposition methods [94]. Furthermore, these second hits may have lagged effects within the life cycle or across generations and may induce differential effects according to life stage. Examination of these types of complexities requires approaches such as structural equation modeling (SEM) [95] or complex systems modeling [96].

Translation for Public Health Policy

The reduction or elimination of health disparities is a national priority named by many health-related agencies, including the Centers for Disease Control and Prevention, Healthy People 2020 and Agency for Healthcare Research and Quality. Despite resources devoted to this endeavor, health disparities persist. The evidence reported here, which developmental programming primes subsequent generational responses to ubiquitous social-epidemiologic exposures, introduces important considerations for efforts aimed at reducing health disparities.

Targeted Exposure and Intervention Effects

In this review, we provide evidence that postnatal exposures in three domains—air pollution, psychosocial stress, and overnutrition—result in stronger negative effects among prenatally exposed individuals (Fig. 1: β versus β*). Left unaddressed, poor health outcomes associated with these exposures are likely to exacerbate existing health disparities through developmental programming and social patterning mechanisms. Conversely, interventions targeting these exposures among the most vulnerable populations have the potential to exert powerful, positive effects (Fig. 2: α1* versus α2*).

Fig. 2
figure 2

Theoretical role of the relationship between environmental enrichment and developmental programming on reducing health disparities over three generations. A gx and D gx denote health status Y gx for generation g and exposure status x. α g denotes the magnitude of disparity in any in generation g

Full size image

Environmental Remediation and Enrichment

The research reviewed here suggests that biological predispositions for adverse health effects may be established during gestation through developmental programming or priming. For priming to exert an adverse effect, however, it requires a second adverse environmental exposure for a predisposition to become actualized. Therefore, environmental enrichment represents a potent tool for improving health trajectories (Fig. 2: α1* versus α2*). While we cannot undo the priming that has already occurred, we can reduce its ultimate effects by restricting the number and nature of second insults a given primed individual receives. Public policy and industrial regulation can be used to improve air quality while economic and cultural reform could reduce social stressors by socially and financially supporting disadvantaged populations. Changing social environments is one of the most effective routes to affecting population prevalence of both risk factors and disease outcomes [97]. Likewise, long-term improvements in dietary practices will likely require policy changes targeting food availability, affordability, and culture [98]. Addressing these contextual exposures is undoubtedly challenging; yet evidence suggests that altering environmental exposures is one of the few interventions that can be used to improve health for the most vulnerable groups. For instance, environmental enrichment following lead exposure among children has been associated with cognitive improvements [99]. The research presented here also suggests that environmental remediation will not only reduce the magnitude and/or likelihood of the second hit, it will also reduce the priming effect for the next generation.

Expanded Time Frame

While national public health objectives seek to eliminate health disparities by 2020, our conceptual model illustrates the long period of time that may be required to ameliorate them, even with dramatic and sustained reduction in exposure prevalence. The magnitude of risk transmission from one generation to the next is unknown. Quantification of the intergenerational transmission rate, as called for under Resilience above, would improve our ability to redefine the time frame in which elimination of health disparities is possible.

With multicomponent interventions that target both environmental and behavioral objectives thought to most potently reduce adverse developmental programming, we should seek to attain incremental improvements in health from one generation to the next (e.g., the health improvement from D0 to D10, and D10 to D20 in Fig. 2). Identification of such targets is needed and will require interdisciplinary research and programs of broad scope.

Conclusion

Public health scholars, practitioners, and policymakers have called for an explicit examination of drivers of disparities. Our conceptual framework suggests that this is even more important due effect magnification of social epidemiologic exposures by developmental programming. Awareness of key developmental programming processes provides an opportunity to identify high-leverage factors for reducing persistent disparities. It also affirms that environmental modification—providing healthy and safe environments in which to live, work, and play—is likely the single best intervention to ensure health for all.