This is the old United Nations University website. Visit the new site at http://unu.edu
Abstract
Introduction
Scope of the problem
Minorities in poverty
Effects of poverty on development
Epidemiological evidence
Epidemiological measures of effect
Single-component models
Nutrition
Health
Education
Multicomponent models
Developing countries
Practical applications
Integrative services model
Discussion
References
Keith G. Scott, Rachel Nonkin Avchen, and Holly A. Hollomon
The authors are affiliated with the Department of Psychology in the University of Miami in Coral Gables, Florida, USA.
The causes of negative child outcomes need to be reconceptualized in terms of the effects of multiple risks. This paper reviews the extent of the problems associated with adverse development in children from deprived backgrounds using two parallel lines of research: epidemic-logical evidence and early intervention. Epidemiological studies suggest that multiple risk factors interact, magnifying the chances of non-optimal development for at-risk populations. Furthermore, evidence from intervention studies suggests that full-service intervention models are the most effective format for reducing poor developmental outcomes for at-risk children.
Children from impoverished backgrounds face obstacles from the moment of conception, and some speculate that risk may be transmitted across generations. Barriers that maybe endured by such children include poor prenatal care, inadequate nutrition, deficient medical care, and insufficient education. Consequently, many factors have been indicated as influential in fostering poor developmental outcomes.
The causes of negative child outcomes need to be reconceptualized in terms of the effects of multiple risks. In particular, accumulating epidemiological evidence suggests that risks strongly interact, creating a need for multifaceted and comprehensive intervention services for at-risk children [1]. In order to change environmental conditions that directly affect child development, efforts must be directed at different combinations of risk. Risk can be understood as a “proxy for need” of services [1].
This paper reviews the extent of the problems associated with non-optimal development in children from deprived backgrounds. Two parallel lines of research will guide the paper: epidemiological evidence and early intervention. Such a review is intended to inform the discussion concerning the development of an integrative programme to promote maximum nutritional, medical, and educational opportunities for child development. In particular, this paper examines the effects of poverty on development. When evidence from developing countries is not accessible, examples from research in industrialized nations will be used.
Conditions of poverty as measured in industrialized nations are relevant when studying the effects of poverty in developing countries due to the “epidemiological transition.” The epidemiological transition suggests that ambient factors commonplace in industrialized countries are inherited by developing countries as they advance [2]. Thus, developmental outcomes of poor children in the United States maybe predictive of outcomes of children in developing nations. In addition, this paper will distinguish between outcomes based on single-component interventions, multicomponent interventions, and full-service interventions.
Currently poverty may be the most pervasive risk factor affecting child development. In the world today, there are approximately 1.3 billion people living in poverty [3]. In the developing world, estimates suggest there are over 500 million children under the age of five years, and approximately 97 million of them live in the least developed countries [4]. Although exact estimates of child poverty in developing countries are difficult to make, estimates from the United States suggest that approximately 14 million children, including 5 million pre-schoolers, were living in poverty in 1995 [5].
Such figures may seem astonishing, since the United States is one of the world’s wealthiest nations. Even so, children in the United States are 1.6 times more likely to live in poverty than children living in Canada, three times more likely than those living in France or Germany, and two times more likely than those in Britain [6].
Besides general mortality indicators, it is necessary to consider ethnicity as a risk factor for children from deprived backgrounds. There have been persistent differences between ethnic groups in infant mortality rates (IMR). Specifically, in 1994 the IMR (per 1000 live births) for black infants in the United States was 17.1, which was 9.2 times higher than that for white infants [6]. Analysis of the 1989-1991 birth cohort in the United States showed that American Indian/Alaskan natives had the second highest IMR (12.6), followed by infants of Hispanic origin (7.6). White infants and infants with only one parent from Central or South America and Asia or Pacific Islands all had IMRs of 6.6 [7].
It is obvious from the above statistics that mortality, specifically, and poverty, generally, are not distributed equally across different demographic groups. African-American children, Hispanic-American children, and children from single-parent homes are disproportionately impoverished [8]. Moreover, the effect of adverse environments on development is substantially impacted by the duration of time spent in poverty [9]. In the United States, 18.3% of children under the age of 18 were living in poverty, based upon 1989 family income [6]. A breakdown by ethnicity exemplifies the nonparallel representation of minorities in poverty: 12.5% of white children were living in poverty compared with 39.8% of black children and 32.2% of Hispanic children [6]. Further, estimates from 1992 suggest African-American children are more likely to experience long-term poverty than are white children. Twenty-nine percent of African-American children were poor for 10 or more years, as compared with less than 1% of white children [8].
Regardless of ethnicity or geographical region, the impact of poverty on development is associated with a disproportionate number of adverse outcomes. Factors associated with the duration, severity, and timing of poverty are important to consider when assessing outcome for children. Research has found tangible differences between developmental deficits in children living in long-term poverty relative to those who experience short-term poverty [9]. Differences between the extremely poor and the moderately poor have also been identified. For example, one study found that the incidence of low birthweight (less than 2,500 g) was greater among children experiencing severe rather than moderate levels of poverty [9]. Studies have also shown that toddlers and young children in poverty have a lower rate of school completion than children and adolescents who endure poverty in later years [10].
Given the many influences associated with poverty, the diversity in developmental outcomes for children from deprived backgrounds is understandable. Poverty is fundamentally linked with a reduced opportunity for optimal development. Specifically, poverty is known to impact major areas influential in child development, such as access to appropriate nutrition, medical care, and education. A detailed discussion of these issues can be found in this volume [11].
Traditional regression models and mean difference models have historically been used to study the impact of poverty on child development. These methods may not be the most effective to analyze the problem. The effects of poverty on development can most effectively be understood in terms of an epidemiological multiple-risk-factor model. Such a model allows a differentiation between the impact of the risks associated with poverty on the individual and the consequences of poverty to the population. This distinction cannot be achieved using regression.
To distinguish between risk associated with the individual and risk to the population is important for making recommendations to clinicians and policy makers. The prevalence of a risk factor will govern its impact on the population. It is possible to have a rare but serious risk factor that is of great clinical importance when present in the individual but is of minor importance to the population. On the other hand, exposure of a large segment of a community to a risk factor can have a great impact on the occurrence of a disorder in the population. Such a relationship can be evident even when the association of risk with an individual appears to be relatively weak when examined in terms of regression coefficients or mean differences. Small mean differences or very modest correlations (even in the range generally considered negligible by researchers in child development) can have large effects on populations.
Much of the importance of population effects is associated with changes in the shape of the normal distribution. For instance, a large increase in the number of cases in the lower tail of the normal distribution will show only a small effect when expressed in terms of product-moment correlations or mean differences [12]. A reader who finds this surprising might reflect on the observation [13] that the correlation between smoking and lung cancer results in a product-moment correlation of approximately r = 0.10. The difference between smokers and nonsmokers, expressed in terms of risk ratios from the same data, shows that smokers are approximately 11 times more likely to contract lung cancer than nonsmokers.
Epidemiological statistics focus on differences in proportions, whereas regression focuses on prediction of means and variances. The risk ratio (RR) is an important measure of individual risk. It measures the rate of disability among those who are exposed to a given risk factor relative to the rate of disability among those who are not exposed [14]. The population-attributable risk fraction (PAF) is used to estimate the effect a particular risk has on the population. Particularly, the PAF estimates the proportion of cases in the population that would be prevented if the risk factor was eliminated. The PAF takes into account the prevalence of the risk factor in the community, the rate of the disorder among those who are exposed, and the rate of the disorder among those who are not exposed.
The effect of two risk factors, low maternal education (<12 years) and very low birthweight (<1,500 g), on determining special education needs provides a good illustration of RR and PAF [15]. A child who has a mother with low education and who was born with a very low birthweight is 3.10 times more likely to be identified as needing special education services at age 10 than a child whose mother has at least 12 years of education and who was not born with a very low birthweight. On the other hand, only 0.1% of the special education cases were attributable to the joint occurrence of low maternal education and very low birthweight. Thus, although the increase in risk associated with the combined occurrence of these risk factors on the individual is substantial, they rarely occurred in the study population. In a population in which the joint occurrence of these risk factors is more prevalent, as it is in some developing countries, the fraction of cases attributable to the risk factors could be much larger.
To summarize, a risk factor that may be of high relevance to the management of an individual case may not necessarily indicate its effect on the population. In order to address the importance of the risk in the population, data about the prevalence of the risk in the population must be examined. From a public health planning perspective, effects measured by regression are of limited use, since this method combines information about the strength and prevalence of a risk factor in the study population into a single coefficient.
Another important finding from epidemiological studies is that risk factors are not randomly distributed in populations. Using regression models it is possible to estimate the effects of one variable or risk factor while holding the others constant. This practice is sometimes referred to as using statistical controls or partialing out effects. Although such a method may be scientifically useful and interesting, it may lead to serious misunderstandings from a public health perspective.
Statistical control of risk factors can be problematic because they often co-occur or cluster together. In a recent study, Hollomon [16] illustrated this problem by analyzing a data set using several different methods. Children of teenage mothers were identified as the group at lowest risk for later cognitive problems when logistic regression was used to statistically control for educational level, birthweight, sex, gestational age, prenatal care, marital status, and maternal age. The cluster of risks controlled for in this analysis consists of those characteristic of births to teenage mothers in the United States. A cohort of births to teenage mothers without these risks is almost, if not totally, nonexistent. Not only do these risk factors occur in clusters in teenagers, but they also interact with each other. The public health importance and population impact cannot be understood by isolating effects statistically. In understanding the impact of poverty on children, an emphasis must be placed on understanding the clusters of risks that are present in a society and then addressing the problem with a multicomponent approach.
Single-component models approach intervention by targeting specific areas known to impact child development. Interventions to be discussed include research designs that focus on modifying areas of nutrition, health, or education. Nutritional interventions typically support dietary change by providing deficient nutrients or supplementary feeding. Medical interventions have focused on increasing access to primary, preventive child care, and educational interventions have emphasized an early awareness of skills important for successful school achievement. The population targeted for these interventions includes children at risk for adverse developmental outcomes. Outcome measures primarily focus on cognitive and physical domains of child development.
Short-term interventions
Long-term interventions
Early supplementary feeding is known to combat later developmental delays. Yet the duration of intervention may influence the impact on development. A short-term supplementary feeding programme was administered in West Java, Indonesia. The purpose of the investigation was to study the impact supplementary feeding had on weight, height, motor development, and mental development [17]. Twenty day- care centres with an enrollment of 15 or more children aged 6 to 20 months were selected for study. Nine of the centres received a 90-day supplementation, and 11 were used as controls. All of the children showed a marked deficit in weight and growth stunting at inception of the study, and no group differences in motor or mental development existed between the groups before supplementation.
The results indicated that the supplemented group experienced a large, positive weight change as compared with the control group. No differences in height were found between the groups. The supplemented group increased 14.1 points on the Bayley (psychomotor development index) as compared with the non-supplemented group. No differences between groups were found on the mental development index. Thus, short-term supplementation was effective at altering weight and motor development, but not height or cognitive development. Supplementing women during pregnancy and lactation but not supplementing the offspring has also been found to have a beneficial effect on the motor development but not the mental development of children [18].
There is evidence to suggest that long-term early supplementary feeding does affect cognitive development. Supplementary feeding during pregnancy and the first months of postnatal life has enhanced mental development among toddlers [19].
In a longitudinal study in Guatemala, 2,000 children aged seven years and younger from four villages participated in a nutritional intervention between 1969 and 1977 [20]. The children in two of the villages received a high-protein supplement (atole), and those in the other two villages received a lesser supplement containing one-third of the calories and no protein (fresco). In 1988, 70% of the children (1411) were contacted; at that time they ranged in age from 11 to 26 years.
Cognitive development was measured using school performance variables. Subjects who received at least two years of postnatal supplementation (n = 611) were examined. The analyses were more favourable for subjects who received atole. The results indicated that as children aged, the developmental benefits of early supplementation increased.
Specific analyses indicated that children in the atole villages performed better on tests of arithmetic, reading, and general knowledge. Treatment by maximum grade attainment was also observed. Children receiving atole scored significantly higher on tests of reading than those receiving fresco and were consistently at the upper end of the grade distribution. With regard to socio-economic factors, analyses revealed that the atole villages had significantly lower maternal education and significantly higher father employment than the fresco villages. Yet, subjects receiving atole performed significantly better than subjects of similar socio-economic levels receiving fresco in tests of literacy, standardized reading and vocabulary, and general knowledge. The authors concluded that differences in performance on tests of complex mental abilities in adolescence can be attributed to differences in individual transactions with the environment, and early supplementary feeding directly impacts on these transactions [20].
Intervention efforts have been directed at providing primary and preventive care as a means of improving general health for children in adverse situations. Several studies have measured the effects of sustained contact with a primary physician or nurse on infant health. A study of infants up to eight months of age in low-income African-American families found that there were no differences between intervention and control subjects on measures of general health, morbidity, incidence of accidents, or immunization rates [21]. However, the intervention group performed significantly better on measures of gross motor skills and had significantly fewer upper respiratory symptoms than did the control group.
When public health nurses provide sustained services, the outcomes are not noticeably different. Ninety-eight infants up to nine months of age from low-income families participated in an intervention aimed at providing case management to facilitate child health clinic and immunization services [22]. Differences were found between the groups. Infants in the intervention group had significantly more adequate child-health clinic visits than the control group, who received segregated case-management services. Although there were no differences between the groups in rates of adequate immunization, differences in cost-effectiveness were identified between services provided by public health nurses and fragmented services.
Medical interventions have also been directed at infants with particular conditions. For example, interventions with low-birthweight (LBW) infants have attempted to reduce harmful stimulus in neonatal intensive care units and newborn nurseries [23]. Particularly, LBW infants are placed in soothing environments that promote behavioural and central nervous system organization. Short-term effects such as weight gain, decreased apnea, and positive changes in state organization have been found. However, long-term changes have not been detected [24].
There is a long history of early educational intervention designed to meet the needs of children at risk for poor educational outcomes due to poverty. Some programmes focus primarily on the educational component. A follow-up to a three-year-long intervention was conducted, providing an educational curriculum to severely malnourished children through a home-visiting programme in Jamaica [25]. At age 15, children who received intervention scored higher on an IQ test and measures of school achievement than malnourished children who did not receive intervention.
Model programmes
Programme initiators motivated by different interests have provided many and varied intervention programmes to children at risk for developmental problems in both the United States and developing countries. There has been a shift away from isolated intervention efforts to more comprehensive models. Multiple intervention services are provided in order to address the multiple needs of at-risk children. These programmes can be discussed in terms of model programmes that test the best practices under optimal conditions in a small group of children. Other programmes represent the widespread practical application of model programmes.
Perry Preschool
The group most frequently targeted for intervention in the United States is children who are considered at risk due to chronic exposure to severe poverty. The Consortium for Longitudinal Studies provides the best evidence for the efficacy of early intervention with these children [26]. One programme evaluated by the Consortium, the Perry Preschool Program, provided a high-quality, active-learning pre-school experience to children living in poverty. Researchers found that children who participated in the intervention had fewer special education placements, fewer grade retentions, higher rates of high school graduation, and more optimal post-secondary employment [27-29].
Abecedarian
The Carolina Abecedarian Project varied the duration and intensity of early intervention to an impoverished population. Children received no educational intervention or intervention from infancy through grade 3, infancy through pre-school, or entry into school through grade 3. Services in pre-school included centre-based care, educational curriculum, nutrition, on-site medical care, and supportive social services. Children who received pre-school intervention had higher IQ and achievement scores through age 12 [30]. There was less support for the effectiveness of intervention only in children of school age. The researchers concluded that more intense intervention services resulted in better long-term outcomes [30, 31]
The Infant Health and Development Program
The Infant Health and Development Program (IHDP) was a multisite, randomized clinical trial designed to evaluate the efficacy of early intervention. The IHDP targeted children who were at risk because they were born with low birthweight (less than 2,500 g) and were premature (less than 37 gestational weeks). It combined early child development and family support with paediatric follow-up in an attempt to reduce developmental, behavioural, and health problems experienced by an at-risk population [32]. From birth to age three, the intervention group received paediatric follow-up, home visits, parent support groups, a developmental curriculum encompassing cognitive, social, motor, and linguistic skills, and 25 hours of centre-based care a week. The paediatric follow-up group received paediatric and referral services only [32, 33].
At age three immediately following intervention, the IQ scores of the intervention group were significantly higher than those of the follow-up group. The follow-up group was at increased risk for behaviour problems compared with intervention children. Finally, there were no differences between the groups in the number of serious health conditions, but mothers of children in the intervention group reported more minor illnesses. This result may reflect closer monitoring of mild symptoms by these mothers [32].
By age five, no intellectual difference was observed between the two groups for the lighter infants (<2,001 g), but the heavier children receiving intervention (2,001-2,500 g) continued to exhibit higher IQ scores than the corresponding follow-up children. No differences in behaviour or health status remained at age five [34]. A follow-up of the IHDP completed at eight years obtained measures of school achievement in addition to cognitive measures [35]. Among the heavier LEW children, those who had received intervention continued to show a significant advantage in IQ (4.4 points) and had higher scores on a mathematics achievement test than control children. There were no significant differences between any groups in performance on a reading achievement test, number of special education placements, behavioural measures, or health status [35].
In some other countries, exposure to conditions of poverty and malnutrition is more severe and more likely to interfere with the normal course of development than in the United States, where compensatory services are more readily available. Unfortunately, the efficacy of multicomponent early interventions has been less well established under these conditions. McKay conducted a multicomponent intervention trial on an impoverished sample in Colombia that included health care, nutritional supplementation, and pre-school education [17]. They found that the earlier the intervention services began, the better the outcome for the child.
The Jamaican Study was a two-year intervention that varied levels of intervention services [36]. All children received free medical care. Groups who received one of four levels of intervention were compared: control group, nutritional supplementation, psychosocial stimulation, and a combined group. Results indicated that the control group had the lowest developmental quotient, the single-component interventions scored in the middle, and the combined intervention group scored the highest. Nutritional supplementation benefited motor and performance subscales, whereas psychosocial stimulation improved all subscales. Supplementation appeared to have a gradual and more long-term effect, whereas stimulation had an immediate effect that lessened with time [36].
Head start
In addition to determining the efficacy of highly controlled early intervention programmes, it is essential to determine if these services can be implemented on a population basis. The Head Start programme was the first nationally implemented intervention programme with educational outcomes as the primary focus. A comprehensive services model was adopted to achieve the goal that all children be ready to learn at school entry [37].
The primary components deemed influential on child development were education, health, parent involvement, and social services [37]. The emphasis on parental involvement makes Head Start one of the first two-generation intervention programmes. The goal is to influence parental self-sufficiency in addition to child development [38].
In theory, the provision of these services improves a child’s ability to learn and be successful in an academic realm. Evaluations of Head Start’s effectiveness have yielded mixed results. The nationally mandated Westinghouse Report investigated the impact of participation in Head Start on later school achievement and found little effect of early intervention on school achievement [39]. This report has been criticized for a number of reasons [39, 40], and recent studies show more encouraging results [41, 42]. Other large-scale implementations of multicomponent interventions have been more clearly successful in improving the long-term educational outcomes of children from impoverished environments [43].
Linda ray intervention program
Cognition and language
Behaviour
Home environment
It is becoming increasingly recognized that the problems experienced by children and their families do not occur in isolation; therefore, the services that address them should not occur in isolation. The goal of full-service integration is to increase the availability, access, and utilization of services to children and families living in poverty, while reducing the cost of implementation [44]. Models of continuous care have been implemented successfully in the area of health care [22] and child development services [45]. The following is an example of a model intervention program based on an integrative services model that goes beyond nutrition, health care, and education to incorporate all aspects of service delivery.
Programme description
The Linda Ray Intervention Program (LRIP) was designed to provide a total service intervention for infants from low socio-economic status inner-city neighborhoods who had been exposed to cocaine in utero. The LRIP is a model intervention programme that serves children from birth to age three. Children are randomly assigned to home-based or centre-based groups until full-capacity enrollment is reached. Subsequently, from the same referral sources, a primary-care group is enrolled that receives primary medical care, social work services, developmental assessments, and transportation to these appointments.
In addition to the services provided in the primary-care group, a teacher visited the home-based children twice a week for 1.5 hours per visit to demonstrate child-care activities based on the intervention curriculum (described under the educational component). Children in the centre-based group received curriculum instruction at the Linda Ray Intervention Center five days a week for five hours per day. Children in the latter group had the same teacher and peer group throughout the three-year intervention.
The LRIP was conceptualized as a model intervention based on a public health approach that is empirically driven with correspondence rules relating developmental outcomes to sources of risk to intervention strategies. The tradition of public health research and programme planning has focused on preventing specific undesirable outcomes, such as disease or risk factors for a disease. A risk factor is any characteristic or circumstance of a person or group that is associated with the development of an undesirable outcome. It may be medical, social, economic, cultural, or some intersection of these variables. The public health model uses a retrospective paradigm working backward from an adverse outcome to risk factors associated with it that will be targeted for prevention [1].
There are four major steps in the process of developing an intervention based on the risk factor model. First, researchers select an adverse outcome to be targeted that has been identified through surveillance data. In public health, surveillance data are defined as the “ongoing, systematic collection, analysis, and interpretation of outcome-specific data for use in planning, implementation, and evaluation of public health practice” [46, p. 3]. Second, a source of risk is identified that is associated with an increase in the targeted outcome. Third, a service or strategy is designed that will prevent the occurrence of the risk factor. Finally, an intervention component is developed or a pre-existing agency is identified that will provide the prevention service.
Intervention components
On the basis of the surveillance data, several areas of risk were identified to be targeted by the intervention programme. Risk was conceptualized in terms of risk factors associated with the pre-school child, infant/toddler, birth characteristics, child care, maternal/family characteristics, and the overarching health and social context. Specific sources of risk were repeatedly identified within these broad areas of risk, including education, medical care, nutrition, family education and support, social services (crisis management, family planning, emergency assistance, etc.), and transportation. The repetition of risk sources and the services needed to address them across domains suggested an integrative services model that is qualitatively different from previous multicomponent interventions.
Education
First and foremost, the goal of the intervention programme was to prepare these at-risk children with the readiness skills necessary for school entry, especially literacy skills. A comprehensive educational programme was developed to address many areas of development affecting academic outcomes. The major intervention service developed to address educational risk was the outcome curriculum.
Children exposed to cocaine prenatally have heterogeneous outcomes, ranging from no documented problems to deficits in language development [47], fine and gross motor skills [48,49], and social-emotional functioning [47,48, 50]. Although early researchers attributed these delays to in utero substance exposure, recent studies focus on the role of the environment in exacerbating the outcomes of these children [51]. On the basis of these findings, it was determined that a curriculum strong in all developmental areas was needed.
Thus, a broad, developmentally sequenced model was adopted, rather than one focusing on specific developmental deficits.
Medical care
The LRIP participants have daily access to nursing and medical care and 24-hour emergency medical assistance. They receive on-site primary well-child care, including immunizations. Parents are educated about appropriate hygiene and prevention of chronic minor health problems associated with this population. In addition, the programme coordinates family planning and prenatal care services for subsequent pregnancies.
Nutrition
The children are provided meals and snacks at the centre, which constitute two-thirds of the recommended daily allowances. A certified nutrition specialist serves as a consultant. Parents are educated as to developmentally appropriate feeding practices within cultural parameters. Medical staff monitor weight gain and review diets during routine visits. Teachers and social workers coordinate access to food-supplementation programmes.
Family education and support
Two-generation programmes that promote both the child’s development and the self-sufficiency of the family (parent education, employment, etc.) are viewed as the most efficient method of maintaining intervention effects [38]. Providing parent education and support is one of the central components of the LRIP. Weekly parent education classes are held at the centre, where topics such as drug education, proper hygiene, making toys, and responsive caregiving are taught. Additional instruction is provided to alternative primary caregivers through the initiation of Father and Grandmother Groups that provide a social support network for alternative caregivers.
Social services
In addition to the services administered directly through the LRIP, social workers help inform the families of the resources available in the community and coordinate their access to those resources. They work closely with the public health nurses who visit families of the LRIP and with existing public service agencies. The social workers as well as the home-based teachers act as advocates for their clients with these agencies.
Transportation
The families of the LRIP need to be able to access the services available to them to ensure success. Therefore, transportation is provided daily to all centre-based children through LRIP-operated buses. Transportation to assessment and medical appointments is provided for all children. Mass transit passes are provided to families to facilitate access to resources. Without adequate transportation, many families would not be able to take advantage of the intervention programme or services within the community.
Preliminary analyses
Although complete data on the first cohort of children will not be available until the summer of 1998, preliminary findings are being used to modify the curriculum for use with the second cohort currently being enrolled. Preliminary evaluations have been completed through 24 months for 50 centre-based, 45 home-based, and 30 primary-care children. The scores for the home and centre groups in the areas of early social communication, language, and emergent literacy skills are generally in the low normal range. The curriculum and the programme have been enhanced in these areas for the second cohort now being enrolled. Eventually, a study will examine the cost-effectiveness of each intervention strategy in terms of its influence on child outcome.
Preliminary analyses suggest that intervention beyond primary care had a significant effect on cognitive and language scores at 24 months. Children who received intervention scored higher on the mental scale of the Bayley Test of Mental Development (II) and on the Reynell Scales of Language Development (receptive and expressive subscales) than those in the primary-care group.
The groups maintained a consistent order in terms of outcomes, with the centre-based group performing highest, followed by the home-based group, and the primary-care group scoring lowest. A comparison between centre-based and home-based intervention provided some evidence that the more intensive, centre-based intervention yielded better outcomes. There was a marginally significant advantage for centre-based over home-based intervention on the Bayley Test and a significant advantage on the expressive subscale of the Reynell Scale.
A comparison of the children receiving centre- and home-based intervention on the child behaviour checklist revealed differences between groups in teachers’ reports of behaviour problems. Children who attended the centre-based intervention were rated as having fewer behaviour problems than those receiving home-based intervention. This difference was especially apparent in the types of externalizing problems, such as destructive and aggressive behaviour, which are associated with subsequent substance abuse.
Children receiving centre-based intervention had more routines in their day-to-day transactions and had a higher quality of home environment than those receiving home-based intervention. Although the intervention services of the LRIP are primarily child-focused, they appear to have positive secondary effects on aspects of the child’s caregiving environment. This advantage exists primarily for children who receive intervention at the centre rather than within the home.
Developmental epidemiology
Implications for public policy
In addition to a direct comparison of three levels of early intervention on child outcome, a series of longitudinal studies of risks affecting long-term educational outcomes is being conducted in conjunction with the Florida Department of Education and the Department of Health. These two areas represent parallel lines of research that provide independent information about the interaction of risk factors and their influence on child development. Along the epidemiological line, a computerized linkage of the birth records of all children in the state of Florida with public school records yields an instantaneous longitudinal data set capable of identifying risk factors present at birth that serve as markers for poor educational outcomes. The data linkage procedure has been validated previously and yields 97% sensitivity and specificity estimates as compared with a gold standard linked by hand [52].
Scott et al. [53] provided the groundwork for the application of the public health approach to child development by describing a method of investigation termed developmental epidemiology. Developmental epidemiology is defined as “the study of the distribution of behavioral outcomes in infancy and childhood and the indicators of their occurrence” [53, p. 352]. The major goal of this field of study is to estimate the magnitude of risk of a poor developmental outcome due to antecedent exposure to one or more risk factors in a manner that will guide prevention and intervention efforts.
Studies from the epidemiological data set have identified low birthweight, low maternal education at birth, and iron-deficiency anaemia as risk factors for special education placement (as a proxy for disability) at age 10 [15, 54]. A dose-response relationship was found between birthweight and maternal education. Children who were born with low birthweight or to a mother who had not completed high school were more likely to be placed in special education. These risk factors also interacted, putting children who had both risk factors at the highest level of individual risk. Although clinically important, such children accounted for a small number of the overall cases of special education. Children born to mothers with low levels of education are an important group to target for early intervention from a public policy perspective, because they comprise a large percentage of children receiving special education services.
Within the general category of special education, low maternal education and male sex were factors strongly associated with risk for behaviour disorders [55]. Low maternal education, low birthweight, having an unmarried mother, and male sex were linked to mild and moderate mental retardation, whereas birthweight and delivery complications were risk factors for profound mental retardation [56; Chapman DA, Scott KG, Blair C, Krieger-Hurtado E, Urbano RC, personal communication, 1998). The risk factors associated with learning disabilities include male sex, low maternal education, unmarried mother, young maternal age, low birthweight, late prenatal care, and belonging to a minority group [Blair CB, personal communication, 1998]. Maternal education below high school completion is a risk factor common across educational outcomes. It may serve as a mediator between low socio-economic status and non-optimal development through the mechanism of inadequate provision of stimulation to the child.
A risk factor can have an independent effect, or combinations of risk factors can have joint effects that lead to an adverse outcome. Exposure to multiple risk factors can lead to a variety of poor developmental and educational outcomes. There can be diverse risk factors for similar outcomes (equifinality) or a common risk factor for diverse outcomes (multifinality) [23]. Therefore, the most effective method of intervention is an integrated service model that targets all identified risk factors.
Studies based on early intervention and developmental epidemiology have implications for public policy. First, families at risk experience multiple risk factors that require multiple services to address them. Second, studies in other fields (e.g., health care) have indicated that integrated services increase utilization of services and yield a positive cost-benefit analysis [22]. Third, multiple-component intervention models such as Perry Preschool, Abecedarian, and IHDP have resulted in positive long-term developmental and educational outcomes for children at risk for problems and delays. Cost analyses have indicated substantial benefits for early intervention in terms of reduced special education placement, reduced grade retention, higher graduation rates, more optimal post-secondary employment, reduced welfare costs, and decreased justice system contact [57, 58]. It is not uncommon in intervention studies to see delayed benefits of early intervention in areas such as these, known as “sleeper effects,” in the absence of short-term benefits.
The conclusion from these findings is that a full-service early intervention is the most effective and cost-efficient way to meet the needs of children at risk because of chronic exposure to poverty. However, these programmes require substantial initial funding before long-term benefits can be observed. Integrated service models may provide the answer for reducing spending in the long run; however, the initial outlay of funding must be supported by the government.
A caveat to this conclusion is that the situation will not be resolved immediately. There is evidence for intergenerational effects of poverty on child development [59, 60]. For example, mothers who were born with low birthweight have an elevated risk of having a LBW child, regardless of nutritional, medical, and socio-economic factors during pregnancy [60]. There appears to be a multigenerational process in which the conditions of a mother’s birth and childhood growth contribute to her reproductive success [59]. The implication is that it may take more than one generation of intervention to achieve optimal child development.
On the basis of the evidence from epidemiological data and early intervention evaluation, it is recommended that funding for a full-service model with integrated services be made available to at-risk children. In the United States, children living in poverty have benefited from interventions based on this format. It is speculated that the best method to prevent developmental delays and suboptimal educational outcomes in children exposed to chronic and severe poverty conditions in other countries is through a fully integrated, multigeneration intervention programme.
1. Backett EM, Davies AM, Petros-Barvazian A. The risk approach in health care: with special reference to maternal and child health, including family planning. Geneva: World Health Organization, 1984.
2. World Health Organization. The world health report 1997: conquering suffering, enriching humanity. [On-line] http://www.who.org/whr/1997/exsum97e.htm, 1997.
3. World Health Organization. Fact Sheet N 91: Intensified cooperation with countries. [On-line] http://www.who.org/inf/fs/fact091.html, 1995.
4. United Nation’s Children Fund. The state of the world’s children. Oxford: Oxford University Press, 1998.
5. Lewit EM, Terman DL, Behrman RE. Children and poverty: analysis and recommendations. Future Child 1997;7:4-24.
6. Children’s Defense Fund. The state of America’s children. Washington, DC: Children’s Defense Fund, 1997.
7. Centers for Disease Control. Health, United States, 1995. Hyattsville, Md, USA: National Center for Health Statistics, Public Health Service. [On-line], http://www.cdc.gov/nchswww/datawh/statab/pubd/hust20htm, 1996.
8. Corcoran ME, Chaudry A. The dynamics of childhood poverty. Future Child 1997;7:40-54.
9. Korenman S, Miller JE, Sjaastad JE. Long-term poverty and child development in the United States: results from the NLSY. Children and Youth Services Review 1995; 17:127-55.
10. Brooks-Gunn J, Duncan GJ. The effects of poverty on children. Future Child 1997;7:55-71.
11. Grantham-McGregor SM, Fernald LC, Sethuraman K. The effects of health and nutrition on cognitive and behavioural development in children in the first three years of life. Part 1. Low birthweight, breastfeeding, and protein-energy malnutrition. Food Nutr Bull 1999:20:53-75.
12. Scott KG, Masi W. The outcome from and utility of or register of risk. In: Field TM, Sostek AM, Goldberg S, Shuman HH, eds. Infants born at risk: behavior and development. New York: Spectrum, 1979:485-96.
13. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ, USA: Lawrence Erlbaum Associates, 1983.
14. Hennekens CH, Buring JE. Epidemiology in medicine. Boston, Mass, USA: Little, Brown, and Company, 1987.
15. Hollomon HA, Dobbins DR, Scott KG. The effects of biological and social risk factors on special education placement: birth weight and maternal education as an example. Res Dev Disabil 1998;19:281-94.
16. Hollomon HA. Risk factors associated with mild mental retardation, learning disabilities, and low achievement: an epidemiological approach. Psychology. Coral Gables, Fla, USA: University of Miami, 1998.
17. McKay H, Sinisterra L, McKay A, Gomez H, Lloreda P. Improving cognitive abilities in chronically deprived children. Science 1978;200:270-8.
18. Husaini MA, Karyadi L, Husaini YK, Sandjaja, Karyadi D, Pollitt E. Developmental effects of short-term supplementary feeding in nutritionally-at-risk Indonesian infants. Am J Clin Nutr 1991;54:799-804.
19. Pollitt E, Oh S-Y. Early supplementary feeding, child development, and health policy. Food Nutr Bull 1994;15:208-14.
20. Joos SK, Pollitt E, Mueller WH. The Bacon Chow Study: effects of maternal nutritional supplementation on infant mental and motor development. Food Nutr Bull 1982;4:1-4.
21. Pollitt E, Gorman KS, Engle PL, Martorell R, Rivera J. Early supplementary feeding and cognition: effects over two decades. Monogr Soc Res Child Dev 1993;58:v-99.
22. Barnes-Boyd C. Effects of sustained nurse/mother contact on infant outcomes among low-income African-American families. Public Health Nurs 1995;12:378-85.
23. Erkel EA, Morgan EP, Staples MA, Assey VH, Michel Y. Case management and preventive services among infants from low-income families. Public Health Nurs 1994; 11:352-60.
24. Beckwith L, Sigman MD. Preventive interventions in infancy. Child Adolesc Psychiatric Clin North Am 1995; 4:683-700.
25. Grantham-McGregor S, Powell C, Walker S, Chang S, Fletcher P. The long-term follow-up of severely malnourished children who participated in an intervention program. Child Dev 1994;65:428-39.
26. Royce IM, Darlington RB, Murray HW. Pooled analyses: findings across studies. In: Consortium for Longitudinal Studies, ed. As the twig is bent.. .lasting effects of preschool programs. Hillsdale, N): Lawrence Erlbaum Associates, 1983:411-59.
27. Schweinhart LJ, Weikart DP. Significant benefits: the High/ Scope Perry Preschool Study through age 27. Ypsilanti, Mich, USA: High/Scope Press, 1993.
28. Schweinhart LJ, Weikart DP, Lamer MB. Consequences of three preschool curriculum models through age 15. Early Child Res Q 1986;1:15-45.
29. Weikart D, Bond J, McNeil J. The Ypsilanti Perry Preschool Project. Preschool years and longitudinal results through fourth grade. Ypsilanti, Mich, USA: High/Scope Educational Research Foundation, 1978.
30. Campbell FA, Ramey CT. Effects of early intervention on intellectual and academic achievement: a follow-up study of children from low-income families. Child Dev 1994;65:684-98.
31. Ramey CT, Ramey SL. Effective early intervention. Ment Retard 1992,30:337-45.
32. Infant Health and Development Program. Enhancing the outcomes of low-birth-weight, premature infants: a multisite, randomized trial. JAMA 1990;263:3035-42.
33. Ramey CT, Bryant DM, Wasik BH, Sparling JJ, Fendt KH, LaVange LM. Infant health and development program for low birth weight, premature infants: program elements, family participation, and child intelligence. Pediatrics 1992;89:454-65.
34. Brooks-Gunn J, McCarton CM, Casey PH, McCormick MC, Bauer CR, Bernbaum JC, Tyson J, Swanson M, Bennett FC. Early intervention in low-birth-weight premature infants: results through age 5 years from the Infant Health and Development Program. JAMA 1994;272: 1257-62.
35. McCarton CM, Brooks-Gunn J, Wallace IF, Bauer CR, Bennett FC, Bernbaum JC, Broyles RS, Casey PH, McCormick MC, Scott DT, Tyson J, Tonascia J, Meinart CL. Results at age 8 years of early intervention for low-birth-weight premature infants. JAMA 1997;277:126-32.
36. Grantham-McGregor SM, Powell CA, Walker SP, Himes JH. Nutritional supplementation, psychosocial stimulation, and mental development of stunted children: the Jamaican study. Lancet 1991;338:382.
37. Head Start. Head Start bureau home page. [On-line] http://www.acf.dhhs.gov/programs/hsb/, 1998.
38. Parker FL, Piotrkowski CS, Horn WF, Greene SM. The challenge for Head Start: realizing its vision as a two-generation program. In: Smith S, ed. Two generation programs for families in poverty. A new intervention strategy. Norwood, NJ, USA: Ablex, 1995:135-59.
39. Brown B. Head Start: how research changed public policy. Young Children 1985;49:9-13.
40. Cole OJ, Washington V. A critical analysis of the assessment of the effects of Head Start on minority children. J Negro Educ 1986;55.
41. Lee VE, Brooks-Gunn J, Schnur E, Liaw F-R. Are Head Start effects sustained? A longitudinal follow-up comparison of disadvantaged children attending Head Start, no preschool, no other programs. Child Dev 1990;61:495-507.
42. McKey RH, Condelli L, Ganson H, Barrett BJ, McConkey C, Plantz M. Executive summary: the impact of Head Start on children, families, and communities. Washington, DC: CSR, 1985.
43. Reynolds AJ. One year of preschool intervention or two: does it matter? Early Child Res Q 1995;10:1-31.
44. Illback RJ. Poverty and crisis in children’s services: the need for services integration. J Clin Child Psychol 1994,23:413-24.
45. Kaul V. Integrated child development services in India. Childhood 1993;1:243-5.
46. Thacker SB. Historical development. In: Teutsch SM, Churchill RE, eds. Principles and practice of public health surveillance. New York: Oxford University Press, 1994:3-17.
47. Gregorchik LA. The cocaine exposed children are here. Phi Delta Kappan 1992;173:709-11.
48. Schneider J, Griffith D, Chasnoff I. Infants exposed to cocaine in utero: implications for developmental assessment and intervention. Infants and Young Children 1989:2:25-36.
49. Southern Association of Children Under Six. Prenatal exposure: the South looks for answers. Little Rock, Ark, USA: Elizabeth F. Shores, 1991.
50. Chasnoff IJ, Burns WJ, Schnoll SH, Burns KA. Cocaine use in pregnancy. N Engl J Med 1985;313:666-9.
51. Scherling D. Prenatal cocaine exposure and childhood psychopathology: a developmental analysis. Am Orthopsychiatr Assoc 1994:64:9-19.
52. Boussy CA. A comparison of hand and computer-linked records. Doctoral dissertation. University of Miami, Miami, Fla, USA, 1993.
53. Scott KJ, Shaw K, Urbano JC. Developmental epidemiology. In: Friedman S, Haywood C, eds. Developmental follow-up. New York: Academic Press, 1994:351-73.
54. Kreiger-Hurtado E, Claussen AH, Scott KG. Early childhood anemia and mild/moderate mental retardation. Am J Clin Nutr, in press.
55. Mason CA, Chapman DA, Scott KG. Risk factors for severe emotional disabilities and emotional handicaps [SED/EH]: an epidemiological perspective. Am J Community Psychol, in press.
56. Chapman DA. The epidemiology of mild, moderate/severe, and profound mental retardation: a multiple risk factor approach. Psychology. Coral Gables, Fla, USA: University of Miami, 1998.
57. Barnett SW. Lives in the balance: age-27 benefit-cost analysis of the High/Scope Perry Preschool Program. Ypsilanti, Mich, USA: High/Scope Foundation, 1996.
58. Fewell RR, Scott KG. The cost of implementing the intervention. In: Gross RT, Spiker D, Haynes CW, eds. Helping low birth weight, premature babies: the infant health and development program. Stanford, Calif, USA: Stanford University Press, 1997:479-502.
59. Emanuel I. Invited commentary. An assessment of maternal intergenerational factors in pregnancy outcome. Am J Epidemiol 1997;146:820-5.
60. Starfield B, Shapiro S, Weiss J, Liang K, Ra K, Paige D, Wang X. Race, family income, and low birth weight. Am J Epidemiol 1991;134:1167-74.
