Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
338 FROM NEURONS TO NEIGHBORHOODS development in the preschool years (Shonkoff et al., 2000; Zigler et al., 1996). These highly diverse initiatives have been included under the broad umbrella of what is called âearly intervention.â CONCEPTS OF INTERVENTION AND THEORIES OF CHANGE Disentangling the Concept Early childhood intervention is more a concept than a specific program (Guralnick, 1998; Shonkoff and Meisels, 2000). Much of its diversity is related to differences in target groupsâfrom the broad-based agendas of health promotion and disease prevention, early child care, and preschool education to the highly specialized challenges presented by developmental disabilities, economic hardship, family violence, and serious mental health problems, including child psychopathology, maternal depression, and pa- rental substance abuse. Within this context, the diversity among and within subgroups is as great as that across the general population. Generalizations about children with developmental disabilities are par- ticularly problematic. As a distinct population, they represent a markedly heterogeneous group of individuals with a wide variety of impairments that differ in both their defining features and level of severity. These impair- ments may include various combinations of delayed or atypical skills in cognition, communication, motor performance, emotional reactivity, and social relatedness, among others. Specific disorders range from commonly recognized conditions (e.g., Down syndrome, cerebral palsy, spina bifida, and autism) to relatively rare and less known disorders (e.g., Rett syn- drome, trisomy 13, and metachromatic leukodystrophy), with a large pro- portion of children whose conditions elude both a definitive diagnosis and a known cause (Guralnick, 1997; Shonkoff and Marshall, 2000). Children with developmental problems that are presumed to be second- ary to the influences of an adverse caregiving environment (e.g., poverty, family violence, parental mental illness) comprise a similarly heterogeneous population. In both circumstances (i.e., whether the vulnerability origi- nates primarily in the biology of the child or the stresses in the environ- ment), the cultural values of the family create a distinctive childrearing context that can present yet another set of challenges to the delivery of professional services in a highly pluralistic society (GarcÃa Coll and Magnuson, 2000; Lewis, 2000). Nevertheless, all children deemed eligible for early intervention programs share a common characteristicâconcern about their development or behavior, regardless of the cause, and a belief that formalized services can increase the probability of a more positive outcome. Another major source of diversity among early childhood interventions
HEALTHY DEVELOPMENT THROUGH INTERVENTION 339 is the marked heterogeneity of service formats. These include multiple variations and combinations of center-based and home-based models, guided by different blends of child-focused and family-focused philoso- phies. Significant differences in staffing configurations contribute addi- tional variability, ranging from the highly professionalized services deliv- ered by educators, developmental therapists, social workers, and nurses with advanced degrees to the highly personalized supports provided by community workers with limited formal education or training. Widely differing views on the definition of âearlyâ provide yet another element of variability, ranging from preschool programs targeting 4-year-olds to pre- natal services focused on expectant mothers. The breadth and the depth of these differences illustrate the diversity of the field of early childhood inter- vention (Guralnick, 1997; Shonkoff and Meisels, 2000; Zeanah, 2000). The extent of this heterogeneity underscores the challenges confronting policy development, service coordination, and evaluation research. Closely related to the diversity of early childhood programs is the extent to which interventions are defined differently depending on the dis- ciplinary lens through which they are viewed. Early intervention is a collec- tion of service systems whose roots extend deeply into a variety of profes- sional domains, including health, education, and social services (Meisels and Shonkoff, 2000). It is a field whose knowledge base has been shaped by a diversity of theoretical frameworks and scientific traditions, from the instruction-oriented approach of education (Bailey, 1997; Bruder, 1997; Wolery, 2000) to the psychodynamic approach of mental health services (Emde and Robinson, 2000; Greenspan, 1990; Lieberman et al., 2000; Osofsky and Fitzgerald, 2000), and from the conceptual models of develop- mental therapies (Harris, 1997; McLean and Cripe, 1997; Warren et al., 1993) to the randomized control trials of clinical medicine (Infant Health and Development Program, 1990; Palmer et al., 1988). At its best, early intervention embodies a rich and dynamic example of multidisciplinary collaboration. Less constructively, it can reflect narrow parochial interests that invest more energy in the protection of professional turf than in serving the best interests of children and families. As its knowledge base has matured, the field of early childhood inter- vention has evolved from its original focus on children to a growing appre- ciation of the extent to which family, community, and broader societal factors affect child health and development. A natural outgrowth of this evolution is a recognition that individual programs are always delivered within a multilayered context, and that their effects are always moderated by the influences of more pervasive social, economic, and political forces. Thus, successful policies for children who live in adverse circumstances may have less to do with the impact of specific services and be more a matter of changing the larger environment in which the children are reared. This
340 FROM NEURONS TO NEIGHBORHOODS growing awareness is likely to lead to further expansion of the concept of early childhood intervention to include such wide-ranging policy concerns as housing, employment practices, community policing, and taxation, among many others (Garbarino and Ganzel, 2000; Sameroff and Fiese, 2000). Theories of Change All successful interventions are guided by a theoretical model that speci- fies the relation between their stated goals and the strategies employed to achieve them (Weiss, 1995). Sometimes these frameworks are articulated explicitly; other times, they are implicit but not clearly formulated. After more than a quarter century of remarkable growth and continuing matura- tion, the basic sciences of child development and neurobiology have con- verged with the learned experiences of a broad array of intervention poli- cies and programs to generate sufficient knowledge to build an intellectually rigorous, common theory of change for the field. The essential characteristics of this framework are drawn from the core concepts outlined in Chapter 1 and buttressed by the wealth of information contained in this report. They stand on the shoulders of decades of creative theoretical formulations about the process of human development. Most prominent among these are the transactional model first formulated by Sameroff and Chandler (1975) and later adapted to the challenges of early intervention by Sameroff and Fiese (1990, 2000); the ecological model articulated by Bronfenbrenner (1979) and subsequently expanded to a bio- ecological model by Bronfenbrenner and Ceci (1994); the concepts of vul- nerability and resilience applied to a wide variety of biological and environ- mental conditions by Werner and Smith (1982), Garmezy and Rutter (1983), and Rutter (2000); the process model of parenting developed by Belsky (1984); the social support model for families of children with dis- abilities popularized by Dunst (1985); the developmental contextual per- spective proposed by Lerner (1991); the biosocial model adopted for an intervention targeting low-birthweight, premature infants by Ramey and colleagues (1992); the principles of developmental psychopathology formu- lated by Cicchetti and Cohen (1995); the social context model constructed by the MacArthur Foundation Research Network on Psychopathology and Development (Boyce et al., 1998); and the developmental framework for early intervention for both biologically and environmentally vulnerable children presented by Guralnick (1998). Taken together, the substance of these models converges to a remark- able degree and applies equally well across the diverse mixture of policies and programs that characterize early childhood intervention in the United States today. This shared theory of change has several central features:
HEALTHY DEVELOPMENT THROUGH INTERVENTION 341 â¢ All strategies of intervention, regardless of the target group or the desired outcomes, can be derived from normative theories of child develop- ment. That is to say, the general principles of development apply to all children, independent of their biological variability or the range of environ- ments in which they live. â¢ All domains of development unfold under the interactive influences of genetic predisposition and individual experience. The trajectories of experience-expectant skills (e.g., motor development) are relatively less sus- ceptible to intervention effects and those of experience-dependent skills (e.g., literacy) are affected more significantly, but no area of human compe- tence is completely predetermined by intrinsic factors. â¢ Young childrenâs relationships with their primary caregivers have a major impact on their cognitive, linguistic, emotional, social, and moral development. These relationships are most growth-promoting when they are warm, nurturing, individualized, responsive in a contingent and recip- rocal manner, and characterized by a high level of âgoodness of fit.â â¢ A young childâs environment is both physical and social. Its impact on development is mediated through the nature and quality of the experi- ences that it offers and the daily transactions that transpire among people inside and outside the home. â¢ The ability of caregivers to attend to the individualized needs of young children is influenced by both their internal resources (e.g., emo- tional health, social competence, intelligence, educational attainment, per- sonal family history) and the external circumstances of their lives (e.g., family environment, social networks, employment status, economic secu- rity, experience with discrimination). The cumulative burden of multiple risk factors and sources of stress compromises the capacity of a caregiver to promote sound health and development. The buffering function of protec- tive factors and sources of support enhances it. â¢ Early intervention programs are designed to affect children directly (through the provision of structured experiences) and indirectly (through their impact on the caregiving environment). Child-focused interventions involve developmentally guided educational opportunities or specifically prescribed therapies or both. Caregiver-focused interventions include vary- ing combinations of information, instruction, emotional support, and assis- tance in securing needed resources and related services. â¢ The determination of appropriate child and family outcomes, and their assessment, require an appreciation of the importance of individual differences among children, an understanding of the extent to which the caregiving environment is changeable, and a realistic appraisal of the match between the resources of the service program and the goals of the interven- tion.
342 FROM NEURONS TO NEIGHBORHOODS â¢ The success of an intervention is determined by the soundness of the strategy, its acceptability to the intended recipients, and the quality of its implementation. EMPIRICAL FINDINGS, PROFESSIONAL EXPERIENCE, AND CURRENT PRACTICE A comprehensive review and synthesis of the full corpus of early inter- vention research was beyond the resource capacity of the committee. The literature analyzed for this report was therefore culled largely from an extensive number of published reviews (Barnett, 1995; Benasich et al., 1992; Berlin et al., 1998; Brooks-Gunn et al., 2000; Casto and Mastropieri, 1986; Currie, 2000; Farran, 1990, 2000; Gomby et al., 1995, 1999; Guralnick, 1997, 1998; Halpern, 2000; Karoly et al., 1998; Lazar et al., 1982; Shonkoff and Hauser-Cram, 1987; St. Pierre et al., 1995b; Yoshi- kawa, 1995) and supplemented by original publications for a selected num- ber of flagship studies. In contrast to its rich and widely endorsed conceptual foundation, the empirical knowledge base on the efficacy of early childhood intervention is relatively uneven. The diversity of target populations and service models that have been studied, and the methodological deficiencies of much of the available literature, contribute to this lack of consistency in the existing database. Most important in this regard is the extent to which a large proportion of studies that address questions of causality have suffered from inappropriate research designs, inadequate analytic approaches, or both, as described in Chapter 4. Notwithstanding these limitations, more than three decades of developmental research and program evaluation have generated the following core of replicated findings, whose convergence strengthens their presumed validity: â¢ In the absence of formal intervention, social class differences in scores on standardized developmental measures that favor children in better edu- cated, higher-income families begin to emerge between 18 and 24 months of age and increase over time (Golden and Birns, 1976; McCall, 1979). â¢ In the absence of formal intervention, there is a general decline in performance on standardized developmental measures for children with established cognitive disabilities, documented most clearly in toddlers and preschoolers with Down syndrome, across the first five years of life (Guralnick, 1998; Guralnick and Bricker, 1987). â¢ Well-designed and successfully implemented interventions can en- hance the short-term performance of children living in poverty, with re- ported effect sizes ranging up to 1.0 standard deviation in the preschool
HEALTHY DEVELOPMENT THROUGH INTERVENTION 343 years (Farran, 1990, 2000; Guralnick, 1998; Karoly et al., 1998; Ramey and Campbell, 1984; Schweinhart et al., 1993). â¢ Well-designed and successfully implemented interventions can pro- mote significant short-term gains on standardized cognitive and social mea- sures for young children with documented developmental delays or disabili- ties, with reported effect sizes ranging from 0.5 to 0.75 standard deviation (Casto and Mastropieri, 1986; Farran, 1990, 2000; Guralnick, 1998; Shonkoff and Hauser-Cram, 1987). â¢ Short-term impacts on the cognitive development of young children living in high-risk environments are greater when the intervention is goal- directed and child-focused in comparison to generic family support pro- grams (Farran, 2000; Guralnick, 1998). â¢ Measured, short-term impacts on the cognitive and social develop- ment of young children with developmental disabilities are greater when the intervention is more structured and focused on the child-caregiver rela- tionship, although the effects are highly variable in view of the marked diversity of child impairments and their severity (Farran, 2000; Guralnick, 1988, 1998; Shonkoff and Hauser-Cram, 1987). â¢ Short-term IQ gains associated with high-quality preschool interven- tions for children living in poverty typically fade out during middle child- hood, after the intervention has been completed; however, long-term ben- efits in higher academic achievement, lower rates of grade retention, and decreased referral for special education services have been replicated (Barnett, 1995; Karoly et al., 1998; Lazar et al., 1982), with reported long- term effect sizes ranging from 0.1 to 0.4 standard deviation (Ramey and Campbell, 1984; Schweinhart et al., 1993). â¢ Extended longitudinal investigations into the adolescent and adult years are relatively uncommon but provide documentation of differences between the intervention and control groups for economically disadvan- taged children in high school graduation, income, welfare dependence, and criminal behavior (Karoly et al., 1998; Schweinhart et al., 1993; Yoshikawa, 1995). â¢ Long-term follow-up data on children with disabilities are scarce, although follow-up studies of children with autism demonstrate persistent benefits of intensive preschool interventions that are followed by continu- ing specialized services during middle childhood (McEachin et al., 1993; Lovaas, 1987). â¢ Analyses of the economic costs and benefits of early childhood inter- ventions for low-income children have demonstrated medium- and long- term benefits to families as well as savings in public expenditures for special education, welfare assistance, and criminal justice (Barnett, 2000; Barnett and Escobar, 1990; Karoly et al., 1998).
344 FROM NEURONS TO NEIGHBORHOODS Successful child-focused intervention programs for economically disad- vantaged groups are designed to provide children with cognitively stimulat- ing environments that they are presumed to be less likely to experience at home. Such programs typically offer rich, school-based learning curricula, often in combination with a wide variety of developmentally enhancing activities in a classroom setting. Several recent comprehensive reviews of such interventions have attempted to discern patterns of impact across programs (Bryant and Maxwell, 1997; Farran, 1990, 2000; Yoshikawa, 1994, 1995). Unfortunately, despite a plethora of investigations, most conclude that it is difficult to draw clear conclusions about the effectiveness of any of a variety of specific intervention approaches. The limitations of this literature are due largely to basic problems in research design (e.g., lack of random assignment, lack of comparable com- parison groups) that make the findings of individual studies less reliable and difficult to compare with each other. A more fundamental barrier to comparisons across studies, however, is the considerable variability among intervention programs on a number of important dimensions, such as the age of the children at time of entry, the characteristics of the target popula- tion, the nature of the program components, the intensity and duration of service delivery, issues regarding comparison or control conditions, and the nature of the staff and their training. Consequently, it is not possible to be certain that differences in outcomes, when they are found, are due to any one (or a combination) of these factors. Generally speaking, programs that have demonstrated the largest and longest-lasting cognitive gains have been administered to children with multiple risks and have offered the most intensive and longest-lasting services. For example, the largest initial IQ gains were documented in the Milwaukee Project, which targeted low- income, black mothers with intellectual limitations and offered full-day infant and preschool child care for the first five years of life, as well as parent education and job training (Garber, 1988). The association between the intensity or duration of service and child outcomes, however, has not been a consistent finding in other studies. In contrast to the extensive attention paid to cognitive performance, relatively few evaluations of child-focused interventions for low-income children have provided short-term outcome data on social adjustment. Those studies that have reported such information generally have not found much evidence of either reduced problems or increased positive behavior. Nevertheless, some researchers have argued that the subsequent documen- tation of differences in progress through school and into adulthood (as illustrated by differential rates of welfare dependence and criminal behav- ior) reflect a social rather than a cognitive impact (Barnett, 1995; Yoshi- kawa, 1995). In addition to the broad array of child-focused programs that have
HEALTHY DEVELOPMENT THROUGH INTERVENTION 345 undergone extensive evaluation, many interventions for low-income chil- dren have focused primarily on parents and parenting, employing various combinations of home visits, group supports, and informational sessions (Brooks-Gunn et al., 2000; Seitz and Provence, 1990). Some of these programs have combined parent-focused components with center-based child care. Despite the considerable diversity of designs, most services are based on the common assumption that parents play a central role in their childrenâs development and that interventions for low-income children are most efficient when they target parentsâ behavior directly. Such services typically provide some form of social support, both instrumental and emo- tional, as well as instruction about childrenâs development. The expecta- tion is that reliable support will reduce parental stress and consequently enhance parental mental health and caregiving capacity, whereas instruc- tional materials on childrenâs development will improve parenting behavior by addressing parentsâ presumed lack of information about what is appro- priate and developmentally enhancing for their children (Seitz and Provence, 1990). A recent review by Olds and colleagues (1999) suggests that inter- ventions for socioeconomically disadvantaged families that are largely par- ent-focused work best when the parents perceive that they or their children need help. Generally speaking, programs that offer both a parent and a child component appear to be the most successful in promoting long-term devel- opmental gains for children from low-income families. Most of the docu- mented benefits have clustered in the realm of social development, perhaps because of early program impacts on risk factors for antisocial behavior. A review by Yoshikawa (1995) of the effects of early childhood intervention programs found that all four of the programs that showed a long-term impact on chronic delinquency had influenced multiple family risk factors in early childhood, including parent-child interaction. Nevertheless, these findings are suggestive, not conclusive, and comparable information on a broader range of interventions is necessary before meaningful conclusions can be drawn about which program components, and in which combina- tion, are successful in promoting positive long-term social outcomes. In recent years, a growing number of interventions have focused on family literacy as a key strategy for improving the home learning environ- ment for young children. Some programs (e.g., Even Start) offer inter- generational literacy activities that include child and adult instruction, as well as parenting education (St. Pierre and Swartz, 1995). Others (e.g., the Home Instruction Program for Preschool Youngsters, or HIPPY) emphasize instruction for parents on how to create a stimulating environment for their children, as well as offering model activities and complementary materials such as books (Baker et al., 1999). To date, evaluation results for both types of programs have been modest and inconsistent. In the future, much
346 FROM NEURONS TO NEIGHBORHOODS more rigorous evaluations with randomized assignment will be needed to assess their effectiveness. In contrast to the marked heterogeneity of program models for children living in socioeconomically disadvantaged circumstances, early interven- tion services for young children with developmental disabilities operate within a more circumscribed arena, guided by a federal entitlement to services for all children with a diagnosed impairment or a documented developmental delay (with the additional option for states to serve infants at risk for subsequent developmental problems). This entitlement was first established in 1986 under Part H of Public Law 99-457 and reauthorized in 1997 under Part C of the Individuals With Disabilities Education Act (Pub- lic Law 105-17). Although the mandate for individualized family service plans provides room for considerable variability, virtually all programs for children with special health or developmental needs employ a family-cen- tered model that combines individual child therapies and educational expe- riences with an array of parent services, such as support groups, individual counseling, and instrumental assistance in securing materials and related services specific to the childâs disability. Finally, unlike interventions for low-income children, programs for children with special needs are required to provide access to a designated array of professional services in natural environments, including those offered by educators, physical and occupa- tional therapists, and speech and language pathologists (Harbin et al., 2000). Beyond both the prescriptions of the law and the evolving conceptual and empirical foundations of the field, much of the knowledge base that shapes the current practice of early childhood intervention is based on professional experience. Central to this perspective is a firm belief in the benefits of family-centered services, the importance of cultural competence, and the impact of relationships on relationships. In this context, a broad spectrum of policies and programs are implemented by a wide variety of service providers, guided by a clear conviction that the impact of their efforts is determined by the extent to which their relationships with families affect the relationships between parents and their children, which, in turn, have a significant impact on child health and development (Barnard, 1998; Berlin et al., 1998; Gilkerson and Stott, 2000; McDonough, 2000). ASSESSING DEVELOPMENTAL OUTCOMES AND MEDIATORS Assessing Child Abilities The evaluation of development in young children is a complex task. The growing cultural diversity of the early childhood population in the United States intensifies that complexity. Superimposed on this formidable
HEALTHY DEVELOPMENT THROUGH INTERVENTION 347 challenge, the high-stakes assessment of competence in children who are adapting to a wide variety of biological vulnerabilities and environmental stressors remains one of the thorniest issues facing the early intervention field. Thus, for more than three decades, researchers and service providers have struggled with both the identification of significant child outcomes and their valid and reliable measurement (Brooks-Gunn and Weinraub, 1983; Cicchetti and Wagner, 1990; Gilliam and Mayes, 2000; Honzik, 1983; Meisels, 1994, 1996; Zigler and Trickett, 1978). Traditional Emphasis on IQ and Early Skill Acquisition From its earliest beginnings, the field of early childhood intervention has focused considerable attention on the promotion of intelligence. Al- though there is still widespread interest in this objective, there is also a great deal of concern about the way in which this elusive construct is conceptual- ized and measured. Debate on this issue has been lively in both academic and policy circles. It centers on both the general challenges inherent in developmental assessment during the early childhood period and the spe- cific value and limitations of an IQ test as an appropriate measure of program effects (McCall et al., 1972; Meisels and Atkins-Burnett, 2000). Perhaps the most important limitation of an IQ score in the context of evaluating the performance of children in an early intervention program is the fact that it is standardized for age and therefore is not useful as a measure of growth or developmental change (see Chapter 4). Nevertheless, its popularity as a measure of intervention impact has been remarkably robust. Moreover, although the evaluation literature is vast and diverse in its focus, highly variable in its methodological rigor, and often inconsistent in its findings, there is a clear pattern regarding short-term impacts on standardized test performance. Specifically, a wide variety of services, both for children living in poverty and for those with biological vulnerabilities, have demonstrated significant gains in IQ during the first five years, fol- lowed by a subsequent fade-out of effects during middle childhood (Campbell and Ramey, 1994; Lally et al., 1988; McCarton et al., 1997; Schweinhart et al., 1993; Walker and Johnson, 1988). The magnitude of these initial treatment-control differences has been moderately high, rang- ing from effect sizes of 0.5 to 0.75 standard deviation (Casto and Mastropieri, 1986; Farran, 1990, 2000; Guralnick, 1998; Shonkoff and Hauser-Cram, 1987). The most striking exception to the fade-out phenomenon has been demonstrated in an intensive intervention program for children with au- tism, which produced sustained treatment-control differences in IQ scores well into the middle childhood years, while the children continued to re- ceive special services as needed (McEachin et al., 1993; Lovaas, 1987).
348 FROM NEURONS TO NEIGHBORHOODS Despite the strength of this study, in the face of significant ethical and practical challenges (Rogers, 1998), legitimate methodological concerns have been raised, including lack of random assignment and questions about the actual intensity and duration of the intervention (Gresham and MacMillan, 1998). A major multisite replication1 is now in progress and is likely to clarify these and related issues and begin to identify the character- istics of subgroups of children who vary in their response to the program. Growing Interest in Underlying Functional Capacities Increasing numbers of early childhood investigators and service provid- ers criticize conventional intelligence testing that relies on the administra- tion of single instruments in standardized settings. Central to this concern is a belief that traditional cognitive measures are unrelated to the everyday context of childrenâs lives, that they impose a linear orientation on a pro- cess that is typically characterized by spurts, plateaus, and extensive vari- ability, and that they are particularly inappropriate when used to evaluate the competence of children with disabilities or youngsters who are reared in families that reflect nonmajority cultures (Meisels, 1996). As an alternative, critics have suggested greater focus on assessing the processes of social and emotional development, as well as the underlying functional capacities that lead to cognitive gains, rather than simply mea- suring the achievement of concrete milestones (Cicchetti and Wagner, 1990; Hauser-Cram and Shonkoff, 1988, 1995; McCune et al., 1990). Closely linked to this emerging perspective is the call for an approach to evaluation and intervention that is embedded within the childâs natural environment and conducted in an ongoing information-gathering manner rather than as a series of disconnected snapshots of competence (Meisels, 1996). This reorientation is particularly important for the evaluation of children with significant motor and sensory impairments, whose progress is often not reflected in standardized test scores (Brooks-Gunn and Lewis, 1983; Shonkoff, 1983). Although the program evaluation literature in these new domains of interest is extremely limited, the underlying developmental science has grown considerably, as described in Chapters 5 and 6. Among the potential target areas for greater attention in measuring program effects, three are particularly noteworthy: self-regulation, interpersonal skills and relation- ships, and knowledge acquisition skills and problem-solving abilities. Self-regulation. As critical early mediators of successful development, 1Although no data from the multisite replication have been published yet, a description of the project appears in Smith et al. (2000).
HEALTHY DEVELOPMENT THROUGH INTERVENTION 349 self-regulatory behaviors offer an attractive focus for early intervention services (see Chapter 5). Dimensions that appear particularly promising include emotional reactivity, attention and activity level, and other behav- ioral aspects of school readiness, such as taking turns and following direc- tions. Facilitating the capacity for self-regulation can provide a construc- tive framework for addressing temperamental differences in all young children, as well as a useful strategy for promoting mastery in those with disabilities (Barton and Robins, 2000). Infants with very low birthweight are particularly vulnerable with respect to regulatory difficulties, most no- tably in their ability to handle different levels of intensity of interaction (Field, 1979; Goldberg et al., 1980). The hypothesized relation between early disorganization and later attention deficit hyperactivity disorder pre- sents a rich area for investigation as a potential opportunity for preventive intervention in the early childhood years. Interpersonal skills and relationships. Extensive research has demon- strated that the establishment of stable and secure relationships is a central feature of healthy human development, and therefore a critical goal of developmental promotion and early childhood intervention (see Chapter 6). Beginning with the infantâs attachment to his or her primary caregivers and extending to the bonds that young children develop with other adults, siblings, and peers, early relationships are viewed as both the foundation and the scaffold on which cognitive, linguistic, emotional, social, and moral development unfold. Early social interactions serve as an essential vehicle for children to learn about how their actions elicit responses from others, how to explore their environment with confidence, and how to experience and deal with thoughts and feelings. Consequently, increasing numbers of program evaluators are measuring aspects of the parent-child relationship as both mediator and outcome variables (Brooks-Gunn et al., 2000; Kelly and Barnard, 2000; Zeanah et al., 2000). Knowledge acquisition skills and problem-solving abilities. As an alter- native to relying exclusively on standardized cognitive assessments, consid- erable value lies in an evaluation of the underlying capacities that make it possible for children to learn. Among those that are of greatest potential interest are new methods of measuring mastery motivation, problem-solv- ing strategies, and the ability to generalize learning from one situation to another (see Chapter 5). Medium-Term Impacts on Subsequent School Achievement The War on Poverty in the 1960s and the establishment of a federal entitlement to early intervention services for infants and toddlers with de- velopmental disabilities in the 1980s were both motivated by a belief that preschool programs for vulnerable children in the early years could enhance
350 FROM NEURONS TO NEIGHBORHOODS later academic achievement and reduce the subsequent need for special education services. After more than 30 years of empirical study, the re- search literature on this issue is uneven but promising. Beginning with the data syntheses of the Consortium for Longitudinal Studies (Lazar et al., 1982), early childhood researchers in growing num- bers have looked beyond the disappointing fade-out of early IQ effects after the intervention is completed, focusing increasingly on intervention-control group differences in school performance during middle childhood and ado- lescence (i.e., differences in later performance between children who re- ceived the intervention during the preschool years and those who did not). This approach began with aggregated findings from 11 program evalua- tions reported by Lazar and his colleagues (1982), which revealed signifi- cant impacts of early intervention on both grade retention (i.e., repeating a grade) and the need for special education services, with greater differences found for those studies that had more nearly randomized research designs. More recently, the Abecedarian Project demonstrated a statistically nonsig- nificant trend toward less grade retention and special education at age 12, which reached significance at age 15 (Campbell and Ramey, 1994, 1995). Notwithstanding their statistical significance, however, the small magni- tude of the intervention-control differences in many of these studies have led some critics to question their value (e.g., Locurto, 1991). However, since the one-time costs of repeating a grade are roughly $6,000 per year and the continuing costs of special education are approximately $8,000 per year, relatively small impacts on grade retention and especially the use of special education services can produce substantial financial benefits (Currie, 2000). The frequently replicated finding of positive impacts of early interven- tion services on school performance, however, has not been universal. For example, no differences in either special education or grade retention were found in follow-up investigations of the Houston Parent-Child Develop- ment Center to age 11 (Johnson and Walker, 1991) or for the Syracuse Family Development Research Program up to age 15 (Lally et al., 1988). Researchers in the Infant Health and Development Program also found no differences in either retention or special education at age 8 (McCarton et al., 1997). However, the sample children were only in first and second grade at the time of the follow-up assessments, and the intervention-control group differences in special education placement for the Perry Preschool sample did not appear until the third grade (Weikart et al., 1978). The interpretation of these discrepant findings is not entirely clear. Beyond obvious differences in the nature of the preschool intervention and the program participants, it is difficult to determine how much these find- ings are related to differences in criteria for repeating a grade or for special
HEALTHY DEVELOPMENT THROUGH INTERVENTION 351 education assignment among the study sites and across time. Nevertheless, there are sufficient data to conclude that early intervention services for children living in poverty that are provided during the first five years of life can reduce subsequent rates of grade retention and use of special education services in middle childhood. The important research question is to deter- mine why some programs are more successful than others. Comparable longitudinal studies have not been conducted on children with diagnosed developmental disabilities. Assessments of school achievement provide another set of criteria by which the impact of early intervention services may be measured. Once again, the literature demonstrates positive program effects but the patterns of impact are variable and not detected universally. Graduates of the Abecedarian Project scored significantly higher than controls in reading and knowledge on the Woodcock-Johnson Test of Achievement at age 12 and in mathematics and reading at age 15 (Campbell and Ramey, 1994, 1995). Perry Preschool participants achieved significantly higher scores in reading, arithmetic, and language on the California Achievement Test (Schweinhart et al., 1993). Follow-up studies of children served by Parent- Child Development Centers indicate positive trends in reading, vocabulary, and language on the Iowa Tests of Basic Skills, but the differences did not reach statistical significance (Johnson and Walker, 1991). At age 8, there were no overall differences on the Woodcock-Johnson Test between the intervention and follow-up groups from the Infant Health and Develop- ment Program, but the heavier of the low-birthweight intervention group had significantly higher mathematics scores than a matching subset of the control group (McCarton et al., 1997). Taken together, the follow-up literature provides abundant evidence of intervention-control group differences in academic achievement during middle childhood, but no consistent or distinctive pattern of advantage associated with a particular type of preschool curriculum or program for- mat. Moreover, the nature of the outcomes (i.e., grade retention, special education placements, and academic achievement scores) do not lend them- selves to analyses that address questions regarding growth, as described in Chapter 4. Perhaps of greater concern is the possibility that the absence of reproducible patterns of outcomes across studies is a reflection of the extent to which published reports focus primarily on those variables for which statistically significant differences are found, with little attention given to the much larger number of measured outcomes that demonstrate no pro- gram-control differences. This criticism was raised by Locurto (1991) in an analysis of data from the Perry Preschool Project and the Milwaukee Project, which noted their mutually inconsistent and counterintuitive findings re- garding the relation between IQ scores and academic achievement.
352 FROM NEURONS TO NEIGHBORHOODS Long-Term Influences on Productive Adult Citizenship Measuring the relation between participation in a preschool interven- tion program and long-term outcomes through the adolescent and adult years is a complex and highly speculative venture. On one hand, a hypoth- esized impact fuels public interest in the potential return on investment in the early childhood period. On the other hand, it raises expectations that may be unrealistically ambitious, tends to downplay the value of the en- hanced well-being of children during the intervention itself, and fails to account for the significant impacts of intervening influences on develop- ment in middle childhood and early adolescence. Central to the concept of long-term intervention effects is the notion of shifting developmental momentum prior to school entry in a manner that increases the likelihood that an otherwise vulnerable child will embark on a more positive pathway into middle childhood. Whether this favorable trajectory is sustained into the adolescent and later adult years obviously will depend on subsequent influences at multiple points along the life course. That is to say, significant medium- and long-term benefits of early child- hood intervention may be viewed as a continuing developmental pathway that is contingent on a chain of positive effects that increase the probability of remaining on track. Very few early childhood intervention programs have followed their sample into the adolescent and adult years. The most extensive data have been collected for graduates of the High/Scope Perry Preschool Program, which reveal statistically significant differences at age 27 favoring the inter- vention group over the controls in income and in rates of high school graduation, criminal arrests, and welfare participation, but no differences in teen pregnancy (Schweinhart et al., 1993). Intervention-control group differences in criminal behavior also were reported for the Syracuse Family Development Research Program (Lally et al., 1988) but were not found in a follow-up of graduates of the Elmira Prenatal/Early Infancy Project (Olds et al., 1997). ASSESSING FAMILY MEDIATORS OF CHILD WELL-BEING A variety of family-focused intervention models have been designed to improve the developmental trajectories of children at risk for problems as a result of environmental or biological vulnerability, as well as for those with diagnosed disabilities. The theory of change that guides such programs is grounded in the assumption that strengthened parent-child relationships and enhanced home environments promote positive outcomes for all young children across a broad range of functional domains (Guralnick, 1998; Sameroff and Fiese, 2000). Professional experience indicates that sensitiv-
HEALTHY DEVELOPMENT THROUGH INTERVENTION 353 ity to cultural differences is also essential to service effectiveness, although the empirical knowledge base in this area is limited (GarcÃa Coll and Magnuson, 2000). Caregiver-Child Relationships and Interactive Behaviors Extensive research conducted over the past several decades has pro- vided rich documentation of the mutual influences that caregivers and young children have on each other (see Chapter 6). Caregiver characteristics that promote healthy child development include warmth, nurturance, stability, predictability, and contingent responsiveness. Childrenâs characteristics that influence the nature of their interactions with their caregivers include predictability of behavior, social responsiveness, readability of cues, activ- ity level, and mood. Caregiver behavior may be affected adversely by immaturity or inexperience, low educational attainment, or mental health problems (e.g., depression, anxiety) related to family violence, substance abuse, economic stress, or constitutional illness. Child behavior may be affected adversely by prematurity, poor nutrition, illness, disability, or tem- peramental difficulties. Beyond the significance of any particular attribute (either positive or negative), the quality of the caregiver-child relationship is influenced most often by the goodness of fit between the styles of both contributors. Consequently, helping parents understand their childâs unique characteristics and providing guidance on how to build a mutually reward- ing relationship that facilitates the childâs development and promotes a sense of parental well-being are common goals shared by a wide variety of early childhood programs. Despite the marked heterogeneity of children, families, and service models that characterize the early childhood field, there is strong consensus on the central importance of child-caregiver relationships. Low income creates a particularly stressful context in which positive interactions with children are threatened, and punitive or otherwise negative relationships may result. The high prevalence of depression, attachment difficulties, and posttraumatic stress among mothers living in poverty serves to undermine their development of empathy, sensitivity, and responsiveness to their chil- dren, which can lead to diminished parenting behaviors and thus decreased learning opportunities and poorer developmental outcomes (McLeod and Shanahan, 1993; McLoyd, 1990; Pianta and Egeland, 1990). Research evidence supporting the potential positive impacts of early childhood programs on parent-child interaction is encouraging. Brooks- Gunn and her colleagues (2000) conducted a recent review of 24 parent- focused programs, 17 of which were home based and 7 of which combined home and center components. Of the 17 home-based programs, 13 as- sessed parent-child interactions or relationships, and 11 of the 13 docu-
354 FROM NEURONS TO NEIGHBORHOODS mented significant intervention effects. Six of the seven home-center com- binations reported similar findings. The majority of the effects reflected increased rates of sensitive parenting behaviors, although these gains were generally not associated with significant differences in child outcomes. Several investigators have noted the extent to which parentsâ beliefs influence a wide range of caregiving behaviors, including specific child- rearing practices (e.g., discipline and limit setting), and how cultural differ- ences influence the way in which the home environment is structured to create a variety of learning opportunities (GarcÃa Coll, 1990; GarcÃa Coll et al., 1996; Harrison et al., 1990; Thompson et al., 1999). Home Environment and Family Experiences The most widely cited and well-documented finding in the early child- hood intervention literature is the strong correlation between family socio- economic status and child health and development (see Chapter 10). Specifically, children in families with lower incomes and lower maternal educational attainment are at greater risk for a variety of poorer outcomes, including school failure, learning disabilities, behavior problems, mental retardation, developmental delay, and health impairments (Aber et al., 1997; Chase-Lansdale and Brooks-Gunn, 1995; Duncan and Brooks-Gunn, 1997; Huston, 1991; McLoyd, 1998). Poor children who are members of racial or ethnic minority groups are particularly vulnerable (McLoyd, 1990; Shonkoff, 1982). Less well appreciated is the disproportionate prevalence of children with biologically based developmental disabilities in low-in- come and less-educated families. In fact, Bowe (1995) reported that at least one-third of the families of children with a developmental disability are living at or below the poverty line. Notwithstanding the strong predictive validity of demographic mark- ers, they have relatively limited utility as guides for designing effective interventions because they tell us relatively little about the causal mecha- nisms that explain their impacts on child development. Thus, researchers and service providers are focusing increasingly on the importance of within- group variability and individual differences among children and families (Berlin et al., 1998; Brooks-Gunn and Duncan, 1997). Closely related to the salience of such variability is the importance of the home environment as a marker of either vulnerability or protection for young children (Brad- ley, 1995; Bradley et al., 1989). As a source of risk, the home may reflect an atmosphere of disorganiza- tion, neglect, or frank abuse. As a source of resilience and growth promo- tion, it is characterized by regularized daily routines and both a physical and a psychological milieu that supports healthy child-caregiver interac- tions and rich opportunities for learning. In a literature review cited earlier
HEALTHY DEVELOPMENT THROUGH INTERVENTION 355 (Brooks-Gunn et al., 2000), 11 of 17 evaluations of parent-focused home- based programs used the HOME inventory (Home Observation for Mea- surement of the Environment; Caldwell and Bradley, 1984) as a measure of the caregiving milieu, and 8 of the 11 demonstrated at least some positive program influence. In addition, impacts on the quality of the home envi- ronment were assessed in four programs that combined home-based and center-based components, two of which (the Infant Health and Develop- ment Program and the Houston Parent-Child Development Center) docu- mented modest positive effects (Andrews et al., 1982; Bradley et al., 1989) and two of which (Project CARE and the Teenage Pregnancy Intervention Program) found no intervention-control group differences (Field et al., 1982; Wasik et al., 1990). The quality of daily family life (e.g., emotional well-being, level of personal control, life satisfaction, and interpersonal relationships) serves as another important protective or risk factor for both child and family out- comes (Crnic et al., 1983; Sameroff et al., 1987). In this context, the protective influences of family cohesion, as well as the adverse impacts of family violence and parental mental illness, are particularly significant. Maternal depression or substance abuse, for example, presents a major threat to child health and development (Bauman and Dougherty, 1983; Downey and Coyne, 1990; Field, 1995; Lester et al., 2000; Mayes, 1995; Seifer and Dickstein, 2000). Similarly, children who witness family vio- lence or who are the victims of physical abuse directly experience signifi- cant consequences, such as psychosomatic disorders, anxiety, fears, sleep disruption, excessive crying, and school problems (Cicchetti and Toth, 1995; Osofsky, 1995; Pynoos et al., 1995; Scheeringa et al., 1995). Few early childhood intervention programs include sufficient profes- sional expertise to treat serious parent or family psychopathology, which can overwhelm the most valiant efforts of a conventional education and support approach. Limited data suggest, however, that attention to such needs may be fruitful. In one example, a home visiting program for socially isolated, pregnant women employed two service modelsâone focused on providing information and resources and the other on developing a thera- peutic relationship between the home visitor and the expectant mother. Follow-up study revealed that women who received the mental health pro- gram approach reported fewer depressive symptoms, and the impact was particularly significant for those who experienced multiple risks (Barnard et al., 1988; Booth et al., 1989). Assessing Community Mediators of Child Well-Being The concept of community can be defined in multiple waysâas a net- work of social connections, a target for resource allocation, and simply a
356 FROM NEURONS TO NEIGHBORHOODS physical space. The hypothesized impacts of community factors on child health and development range from the positive effects of an environment rich in social capital and collective efficacy to the adverse influences of one that is burdened by poverty, violence, and other social drains. Although the potential effects of community-level variables on child health and develop- ment have been well described, their explicit measurement in early interven- tion impact studies has been limited, and the extent to which they are amenable to change is unclear (Duncan and Raudenbush, 1999; Earls and Buka, 2000; Manski, 1993). Potential domains of influence include both threats and facilitators and are described in greater detail in Chapter 12. The relative absence of significant attention to community-level interven- tions, in contrast to the dominant focus on child- and family-oriented strat- egies, is another indication of the relatively limited scope of early childhood programs in the United States, which are conceptualized within a more individualistic and less interdependent framework (see Chapter 3). Threats to Physical Health and Safety Potential threats to the physical health and well-being of young chil- dren include poor housing, with its associated risk of increased exposure to infectious diseases and higher incidence of injuries; environmental toxins, such as lead (which can adversely affect brain development); and endemic substance abuse and violence, with their associated risk of child maltreat- ment (Klerman, 1991; Korenman and Miller, 1997). When safety concerns limit the extent to which children are allowed to play outside their homes, learning opportunities are restricted and development may be compro- mised. Significant interactions between the adverse physical features of a poor neighborhood and the associated social context of a dangerous envi- ronment present a particularly serious threat to childrenâs well-being. Em- pirical data linked explicitly to early intervention program effects, however, are unavailable. Threats to Social and Educational Opportunity Beyond their threats to childrenâs physical health and safety, certain characteristics of communities add further disadvantage by undermining a sense of opportunity or individual possibility, beginning in infancy and extending throughout childhood. Aspects of this burden include the ad- verse consequences of limited recreational facilities, inadequate child care, and substandard schools. Racism or other forms of discrimination based on ethnic status, social class, or the presence of a developmental disability lead to both overt and subtle messages of social exclusion that can have significant debilitating effects on a young childâs emerging sense of self
HEALTHY DEVELOPMENT THROUGH INTERVENTION 357 (GarcÃa Coll and Magnuson, 2000; Stoneman, in press). Once again, these concepts have been well described, but their empirical documentation has been limited. Some researchers have hypothesized that the negative impacts of com- munity factors on child well-being may be significant only in the most impoverished environments, and that modest community-level interven- tions in such circumstances may be of limited benefit (see Chapter 12). Notwithstanding the modest science base in this area, family relocation has been demonstrated to result in positive child outcomes for some children, but results suggest that a large (i.e., 1+ standard deviation) change in neigh- borhood conditions, as might be reflected in a move from an inner-city housing project to a neighborhood with only half as many poor families, is necessary to produce significant effects (Katz et al., 1999; Ludwig et al., in press). Severe Deprivation The concept of environmental deprivation in the early childhood years is complex and highly charged. Several observers have raised concerns about the inappropriate labeling and associated stigmatization of low-in- come families, many of whom are members of minority groups, who are unfairly and inaccurately categorized as neglectful. This is particularly problematic in circumstances in which children are developing normally in caring and nurturing environments but are not mastering the social behav- iors or cognitive skills that are expected by teachers and required by schools (GarcÃa Coll and Magnuson, 2000). Nevertheless, some children do indeed grow up in environments that are characterized appropriately as deprived, inadequate, or destructive. Historical examples include institutions for young children with Down syndrome and cerebral palsy; contemporary models are best exemplified by Romanian orphanages and children living in extremely abusive homes domi- nated by severe mental illness and substance abuse. In both circumstances, research has demonstrated the devastating impacts of early and severe dep- rivation, as well as the remarkable capacity of children, both with and without biologically based disabilities, to recover from extraordinary devel- opmental assaults if an alternative environment is provided as early as possible (Benoit et al., 1996; Provence and Lipton, 1962; Spitz, 1945) (see Chapter 9). Facilitators of Growth-Promoting Opportunities In contrast to strong evidence documenting the adverse impacts of high-risk environments, a number of enhancing community characteristics
358 FROM NEURONS TO NEIGHBORHOODS have been postulated to increase the probability of more positive child outcomes. These include supportive social networks for families, particu- larly for mothers; inclusive community settings, such as organized pro- grams that offer a welcoming environment for children of diverse back- grounds and make appropriate accommodations for children with special medical or developmental needs; and other manifestations of social capital or collective efficacy that are accessible to children and families (Sampson et al., 1997). These facilitators may be particularly important for victims of systematic discrimination or social isolation. Empirical evidence for such associations, however, has not yet been produced. The extent to which community resources can promote developmental opportunities for young children is presumed to be determined by both the nature of the offerings and the commitment of the community to ensure their availability. Common examples include accessible and affordable child care and preschool programs of high quality and a diverse selection of recreational activities. As important as the programs themselves are the intangible sense of community and the message of social inclusion, which communicates to all children and families that opportunities are available to them and that expectations for their healthy development and later achievement are high. Such messages are likely to be particularly critical for children with disabilities, children who are poor, and children of racial or ethnic minority status. Promoting such social capital and increasing its accessibility for vulnerable families is an important component of the early childhood intervention agenda. Systematic research in this area has not yet been done. Social Policies that Affect Families with Young Children Social policies often have considerable impact on the well-being of young children and their families, directly or indirectly, and by either com- mission or omission (Shonkoff et al., 2000). Some, such as federally man- dated early intervention and special education services under the Individu- als with Disabilities Education Act and state-mandated child care regulations, are recognizable components of the early intervention arena. Others, such as the provision of unpaid job leave for parents of newborns under the Family and Medical Leave Act and the time limits and mandated work requirements of Temporary Assistance to Needy Families under the Personal Responsibility and Work Opportunity Reconciliation Act, are not linked directly to the field of early childhood intervention but have a signifi- cant impact on its agenda. By the same token, many important social policies do not have their origins in direct government action. Prominent examples include the shift in pediatric health services toward a managed
HEALTHY DEVELOPMENT THROUGH INTERVENTION 359 care model and corporate policies and practices that affect working hours, fringe benefits, and other supports for employees with young children. Some policies (e.g., minimum wage laws, the earned income tax credit) have an impact on child health and development by affecting the availabil- ity of material resources and therefore the quality of family life and parentsâ ability to provide learning experiences for their children. Others (e.g., mandated child safety caps for medicine containers, legal limits on hot water heater temperature settings, the fortification of foods with iron or folic acid) are designed to reduce reliance on individual caregiver behavior by controlling external environmental threats to health and safety. Taken collectively, the range of potential policies that can influence the well-being of young children is considerable. This indicates a need to consider a much broader scope and definition of the concept of early childhood intervention. LESSONS LEARNED AND FUTURE CHALLENGES Essential Features of Effective Interventions Despite the methodological limitations of the existing science base and the marked diversity of disciplinary perspectives and program models that are represented in the research literature, a common set of essential features has emerged across a broad spectrum of early childhood intervention sys- tems. These include a mix of both well-documented empirical findings and state-of-the-art guidelines based on professional consensus. Before examining the characteristics that are associated with effective interventions, it is necessary to acknowledge the specific problems inherent in the available data. Notwithstanding important exceptions, much of the empirical knowledge base is compromised by incomplete information on sample children and families, inadequate documentation of the services planned or delivered, and substantial methodological limitations in study design and data analysis. These limit their utility for addressing causal questions (see Chapter 4). Moreover, except for selected reports, most intervention studies focus on the quantification of aggregate program ef- fects rather than the more useful analysis of differential program impacts based on complex interactions among child, family, and service variables. In this context, promising new studies of early childhood intervention are beginning to employ a variety of quantitative and qualitative research methods to address a more focused set of questions. For example, what can be learned about tailoring specific services to children and families in differ- ent circumstances and with different needs? How does a policy or program decide when to focus on the child, the family, the community, or other significant influences in the childâs life, and in what mix? What can we learn about thresholds of program intensity and levels of parent engagement that
360 FROM NEURONS TO NEIGHBORHOODS are necessary for measurable impact, particularly as they may vary for different populations? What is known about the developmental timing and duration of different interventions? What is required to sustain positive change, both in terms of the processes that must be set in motion and the ongoing services, if any, that must be continued? What are the major barri- ers and constraints that limit the possibilities for positive change? Future research will undoubtedly provide answers to these and other compelling questions about the differential impacts of early childhood ser- vices. The clear determination of causal connections between specific inter- ventions and specific outcomes, however, will depend on the extent to which investigators adhere to the principles discussed in Chapter 4. Cur- rent knowledge points to the need for greater empirical attention to the following essential features of effective interventions: (1) individualization of service delivery; (2) quality of program implementation; (3) timing, in- tensity, and duration of intervention; (4) provider knowledge, skills, and relationship with the family; and (5) a family-centered, community-based, coordinated orientation. Individualization of Service Delivery Extensive research from a variety of service system perspectives con- verges on the principle that effective intervention demands an individual- ized approach that matches well-defined goals to the specific needs and resources of the children and families who are served. Thus, there is scant support for a one-size-fits-all model of early childhood intervention. Con- sequently, there is little justification for an approach to program evaluation that asks generic questions about whether services are effective, in contrast to an assessment strategy that investigates the extent to which specific kinds of interventions have differential impacts on specific kinds of children in specific types of families. Central to this fundamental principle of effective services is the importance of understanding the diverse cultural contexts within which young children grow up, and the need for individualized functional child assessments that measure important capacities that are linked to the intervention in an ongoing, reciprocal fashion. For young children whose development may be compromised by an impoverished, disorganized, or abusive environment, as well as for those with a documented disability (who themselves represent a remarkably het- erogeneous population), interventions that are tailored to specific needs have been shown to be more effective in producing desired child and family outcomes than services that provide generic advice and support (Brooks- Gunn et al., 2000; Farran, 1990, 2000; Guralnick, 1998). Furthermore, programs that directly target the everyday experiences of children appear to be more effective in improving their acquisition of skills than those that
HEALTHY DEVELOPMENT THROUGH INTERVENTION 361 seek to promote child development indirectly by enhancing the general quality of the caregiving environment (Farran, 2000). Similarly, services that are focused explicitly on parenting behaviors have greater impact on parent-child interactions than do generic parent education efforts (Brooks- Gunn et al., 2000). These patterns are reflected in the relatively greater child-focused impacts of center-based interventions (Farran, 2000) and greater parent-focused effects of home-based programs (Brooks-Gunn et al., 2000). A confirmatory review of 27 early intervention programs by Benasich and colleagues (1992) found short-term child cognitive benefits in 90 percent of center-based services, in contrast to 64 percent of home-based interventions. One year after program termination, child developmental gains persisted for 67 percent of the center-based programs and 44 percent of the home-based interventions. Research demonstrating differential effectiveness for specific subgroups of children and families further supports the need for individualization of services to ensure maximum impact. For example, children whose mothers had the lowest IQ gained the most from the Abecedarian Project (Campbell and Ramey, 1994, 1995). Similarly, children whose mothers had less edu- cation demonstrated greater benefits from the Infant Health and Develop- ment Program, although greater child impacts were also documented for children at lower biological risk as measured by birthweight (Brooks-Gunn et al., 1994; Liaw and Brooks-Gunn, 1993; McCarton et al., 1997). For children with defined disabilities, both the nature of the impairment and its level of severity demand a highly differentiated approach to service plan- ning and delivery. Generally speaking, for both biologically and environ- mentally vulnerable populations, program impacts are generally greater for more disadvantaged families and for children with less severe disabilities (although the latter may be a function of the developmental measures that are used). More definitive understanding of the causal relations between specific interventions and specific outcomes for specific target populations will require further randomized experimental studies. Linked to the need for individualized intervention strategies, current practice (and, in fact, federal law for children with disabilities) mandates that service outcomes be tailored to the particular interests of each indi- vidual family (Meisels and Shonkoff, 2000). In this context, parents of children with the same developmental disability may have very different goals and aspirations. Similarly, families experiencing comparable levels of economic hardship may have different needs and desires for assistance. Quality of Program Implementation The extent to which model demonstration programs are endowed with abundant resources and highly trained staff, evaluated successfully, and
362 FROM NEURONS TO NEIGHBORHOODS subsequently replicated with inadequate budgets and less skilled personnel is a highly problematic burden for the early intervention field. Thus, a second feature of early childhood services that is endorsed widely across all service systems is the fundamental importance of the quality of the inter- vention that is actually delivered and received by target children and fami- lies. The research literature on child care provides abundant evidence of the positive correlation between quality of care and developmental outcomes for children (see reviews by Lamb, 1998; Love et al., 1996; Scarr and Eisenberg, 1993; and Smith, 1998). Moreover, in the absence of subsidies, children from low-income families who are at greater developmental risk are more likely to receive lower-quality care (NICHD Early Child Care Research Network, 1997c; Phillips et al., 1987b). The impact of quality has been shown to be particularly important for children from families who bear the burden of multiple risk factors, who are also the children with the greatest probability of being enrolled in poor-quality programs (Currie, 2000; Peisner-Feinberg and Burchinal, 1997). Variations in quality among intervention programs designed to address the problems of economic disadvantage are widespread. One evaluation of a sample of Head Start programs, for example, generated developmentally appropriate ratings for only 3 of the 32 classrooms studied (Bryant et al., 1994). More promising results from the recent FACES data collection reflect greater attention to quality standards in Head Start centers (Admin- istration on Children Youth and Families, 1998, 2000). Generally speak- ing, concerns about the quality of program implementation have received much less explicit attention in the literature on services for young children with developmental disabilities. The critical importance of the quality of program implementation is also a key issue with respect to the future of the evaluation enterprise. As noted in Chapter 4, the premature assessment of an intervention impact before one is confident that it can be faithfully implemented is likely to be both a waste of money and a demoralizing influence on those who are trying to develop promising new programs. Timing, Intensity, and Duration of Intervention The research literature on service intensity, duration, and age of initia- tion is perhaps the most complex and inconclusive aspect of the early childhood intervention knowledge base. Many investigators have reported findings that support the value of âearlierâ and âmore.â Others have chal- lenged such conclusions as advocacy-driven research. The concept of intensity is defined operationally in many ways. Most typically it has been measured by the amount of professional time (e.g.,
HEALTHY DEVELOPMENT THROUGH INTERVENTION 363 hours per day, days per week, or weeks per year) spent with families or children. Unfortunately, however, relatively few effectiveness studies have collected sufficient data to assess this important variable. One important exception is the Infant Health and Development Program, which has gener- ated a rich database on services received by individual sample members and has documented a positive association between intensity of participation and child cognitive gains (Ramey et al., 1992). The nonrandom nature of the differences in program participation, however, precludes definitive in- terpretation of intensity effects. In two studies of a home visiting program for poor families with infants in Jamaica, one of which used a random assignment design, weekly visits were associated with higher child develop- mental test scores than biweekly visits, and children who were visited bi- weekly scored higher than those who received services at monthly intervals (Powell and Grantham-McGregor, 1989). Intensity effects have also been noted for children with autism, as increased program intensity is associated with more substantial short- and long-term outcomes (McEachin et al., 1993; Lovaas, 1987). Duration of intervention has also been studied and found to be associ- ated with measurable family impacts. For example, mothers who partici- pated in the Prenatal/Early Infancy Project for 2 years were less likely to maltreat their children than mothers who received 9 months of service. Moreover, a 15-year follow-up revealed an inverse relation between the amount of service received and a number of negative maternal outcomes, including child maltreatment, repeat pregnancy, welfare dependence, sub- stance abuse, and interactions with the criminal justice system (Kitzman et al., 1997; Olds et al., 1997). The Abecedarian intervention plus a follow- up program into the elementary school years was more effective than the preschool intervention alone (Campbell and Ramey, 1994). For families of young children with developmental disabilities, the vari- ability in service intensity is considerable. Differences in both amount and duration of intervention may be related to the age of referral, the nature and severity of the childâs impairment, or the familyâs resources and needs. In a systematic investigation of services received by families of children with Down syndrome, motor impairment, and developmental delays of uncer- tain etiology, the Early Intervention Collaborative Study documented an average of 7 hours of service per month over the first 12 months of program participation, with a range from less than 1 hour to over 20 hours monthly (Shonkoff et al., 1992) In this sample of 190 children enrolled in 29 com- munity-based programs, the strongest predictor of service intensity in the first year was the childâs score on the Bayley Scales at the time of program entry (i.e., children with more severe impairments received more hours of service). The measurement of program intensity for children with disabilities
364 FROM NEURONS TO NEIGHBORHOODS also includes parent time spent on developing relationships and acquiring knowledge, as well as the extent to which structured learning opportunities are embedded in typical family routines. Indeed, the ultimate impacts of such programs are presumed to be dependent on the degree to which fami- lies are able to incorporate specific intervention techniques into their every- day interactions with their children (Gilkerson and Stott, 2000). Finally, both empirical data and clinical experience indicate that earlier identification and intervention are more important for some conditions or circumstances than for others. For example, early diagnosis and treatment is clearly effective in reducing the adverse impacts of a hearing loss on functional communication and cognition (Brasel and Quigley, 1977). Simi- larly, early tactile/kinesthetic stimulation of premature newborns has been associated with greater weight gain, higher survival rates, and higher neurobehavioral scores (Field et al., 1986). Children who are adopted out of institutionalized orphanages before 12 months of age have better devel- opmental outcomes than those who are adopted at an older age (Benoit et al., 1996). The impacts of prenatal home visits have been correlated with enhanced health and safety outcomes and decreased parental interaction difficulties for some groups but have shown minimal effects for others (Olds and Kitzman, 1993). The persistent effects of the Abecedarian Project have been attributed by some investigators to the initiation of the interven- tion in early infancy and its extension over the first five years of life. In summary, earlier has been shown to be better (and defined differ- ently) for some conditions than for others. There is no basis, however, for concluding that individualized interventions provided after certain ages can have no positive impacts. Furthermore, notwithstanding the importance of preventing early developmental concerns from becoming more serious prob- lems later, the premature initiation of services may lead in some circum- stances to inappropriate labeling or the removal of children from typical experiences, thereby reducing the possibility of self-righting corrections or compensatory growth spurts. Finally, questions about intensity and dura- tion must always be considered in the context of assessing the ratio of costs to benefits. Modest benefits from shorter and less intense services may be small, but their cost is relatively low. In contrast, significantly higher benefits may be derived from longer and more intense services, but the cost of those greater gains may be quite high. Weighing the difference between costs and benefits in the determina- tion of appropriate program âdosagesâ is a critical policy challenge. Unfor- tunately, the data needed to assess this issue are quite limited. Moreover, it is most important to recognize that the only way to provide definitive answers to questions about the relative impacts of the timing, intensity, and duration of service delivery is to conduct randomized experimental studies on specific populations.
HEALTHY DEVELOPMENT THROUGH INTERVENTION 365 Provider Knowledge, Skills, and Relationship with the Family The extent to which service providers have the knowledge and skills necessary to address the needs of their target population is a fundamental challenge facing all human services. This challenge is particularly compel- ling for early childhood programs, in view of the broad array of conditions and circumstances with which they are confronted. Examples include in- fants with significant developmental disabilities with or without complex medical concerns, preschoolers with severe behavioral disorders, mothers with clinical depression, and families dealing with the stresses of poverty, marital discord, substance abuse, and recurrent domestic violence. Each of these types of problems typically requires a level of professional expertise that exceeds the generic skills of a child care provider, early childhood educator, child protective services worker, or nonprofessional home visitor. A substantial body of research in child care settings has clearly linked well-trained, qualified teachers and staff to better child outcomes, particu- larly for low-income children who are at risk for early developmental prob- lems and later educational underachievement (see Lamb, 1998). However, as greater numbers of children with disabilities are enrolled in programs, child care providers and early childhood educators are increasingly faced with the inadequacy of their professional training and the paucity of expert consultation available to help them address a wide variety of special needs. The massive shortage of mental health professionals to deal with very young children and the uneven level of skills and excessive rate of turnover among child care workers are particularly critical problems in this regard (Knitzer, 2000). Resource limitations and pressures to âdo more with lessâ present enormous challenges to programs that serve families who are coping with complex developmental and socioeconomic concerns. Marked disparities in the training and skills of home visiting program staff are prominent examples of this phenomenon (Gomby et al., 1999; Olds et al., 1999). In this context, the ultimate impact of any intervention is dependent on both staff expertise and the quality and continuity of the personal relationship established between the service provider and the family that is being served. For example, mothers and children who received high ratings for active engagement in the Infant Health and Development Program were found to have better home environments and higher child IQ scores at 36 months (Liaw et al., 1995), although the direction of effect could not be determined with assurance. The challenges of establishing relationships with individuals who face varying combinations of child disability and adverse environmental circum- stances are substantial. Families of children with special needs seek guid- ance in understanding how to promote their childâs atypical development,
366 FROM NEURONS TO NEIGHBORHOODS and service providers are trained to respect parentsâ knowledge about their childâs unique personal characteristics. Children living in impoverished or disorganized environments are presumed to need compensatory, enriching experiences, and their parents are generally presumed to need help in ad- dressing basic childrearing needs. This tension between intervention mod- els that view parents as the ultimate authority with respect to their childrenâs interests and those that view them as requiring significant assistance de- mands highly skilled staff and creates a complex challenge for the early childhood field. Family-Centered, Community-Based, Coordinated Orientation The concepts of family-centered, community-based, coordinated ser- vices are firmly embedded in the professional experience and philosophies that guide all early childhood programs, from the generic child care facility to the most highly specialized intervention for young children with complex developmental disabilities or severely compromised living arrangements. Thus, although the empirical evidence for these concepts is thin, the theo- retical and experiential support is strong. Central to the concept of family-centered care is the notion of empow- ering parents as the true experts with respect to their own childâs and familyâs needs and the goal of building a strong, mutually respectful, work- ing partnership in which parents and professionals relate comfortably in a collaborative effort to achieve family-driven objectives (Turnbull et al., 2000). The essential characteristics of a community-based model are re- flected in the extent to which services are delivered in a nonstigmatizing, normative environment that has both physical and psychological proximity to where young children and their families live. The essence of coordinated services is embedded in the synergistic organization of a variety of program- matic resources in a rational, efficient, and cost-effective manner that mini- mizes bureaucratic complexity and avoids unnecessary burdens on families. The essential features of a family-centered approach to early childhood services include: (1) treating families with dignity and respect, particularly with respect to their cultural and socioeconomic characteristics; (2) provid- ing choices that address family priorities and concerns; (3) fully disclosing information so that families can make informed decisions; and (4) provid- ing support in a manner that is empowering and that enhances parental competence. The extent to which a program is viewed as family-centered is generally determined by measures of parent satisfaction, service utilization, and level of participant attrition. Providing developmental promotion and early intervention services in a community-based context facilitates access and reduces the stigma associ- ated with service provision in a segregated setting. For children with devel-
HEALTHY DEVELOPMENT THROUGH INTERVENTION 367 opmental disabilities, the promotion of competence in normative commu- nity contexts is particularly important as a vehicle for both acquiring func- tional skills and gaining social acceptance (Guralnick, in press). The evolving nature and imprecise measurement of the concepts of âcoordinated,â âcommunity-based,â and âfamily-centeredâ underscore the critical need for more descriptive, exploratory investigations in this area, including both qualitative and quantitative research. Indeed, as de- scribed in Chapter 4, the level of maturation of the knowledge in this area indicates that experimental, randomized studies would be premature and of less value at the current time. Opportunities, Constraints, and Challenges As the concept of early childhood intervention continues to evolve, it faces a multitude of ongoing challenges. Some must await the generation of new knowledge; others will depend on the resolution of old political con- flicts. In the final analysis, the future vitality of the field will be served best by a creative blend of critical self-evaluation and openness to fresh think- ing. The following seven challenges are particularly important at this point in time: (1) increasing access and participation, (2) ensuring greater quality control, (3) defining and achieving cultural competence, (4) identifying and responding to the special needs of distinctive subgroups, (5) influencing and evaluating the impacts of postintervention environments, (6) strengthening the service infrastructure, and (7) assessing the costs of early childhood investments. Increasing Access and Participation Marked inequalities in access to state-of-the-art early childhood ser- vices are a serious problem. Diminished accessibility is related to a variety of potential barriers, including cost, language, culture, citizenship status, transportation, eligibility standards, program scheduling, and stigma asso- ciated with labeling, among others. Beyond the failure of existing policies and programs to ensure the identification and enrollment of all children and families who could benefit from available services, many early childhood intervention efforts experi- ence significant participant attrition. For example, in one study of Parent- Child Development Centers, 47 percent of the treatment group dropped out in the first year of the program (Walker et al., 1995). Of the 985 children enrolled in the Infant Health and Development Program, 81 received no services whatsoever (Liaw et al., 1995). Average attendance in the High/ Scope Perry Preschool Program was 69 percent in the center-based compo- nent (Weikart and Schweinhart, 1992), and only 56 percent of the families
368 FROM NEURONS TO NEIGHBORHOODS enrolled in the Comprehensive Child Development Programs were actively engaged after three years of participation (St. Pierre et al., 1994). A recent review of several model home visiting programs characterized the enroll- ment, involvement, and retention of families as a common struggle. For example, data from the Hawaii Healthy Start Program and the Nurse Home Visitation Program estimated that 10-25 percent of families who were in- vited to enroll in these programs chose not to participate. Once the families were enrolled, they received an average of about half of the scheduled visits, regardless of the intended frequency. Between 20 and 67 percent of all the families who enrolled in the home visiting programs reviewed left the pro- gram before it was scheduled to end (Gomby et al., 1999). Significant dropout rates present problems for both service delivery and for the evaluation of intervention impacts. On one hand, less than univer- sal âtake-upâ is a fact of life and may reflect rational responses by parents who do not perceive the potential benefits of a program to be worth the costs of the time and commitment required of them. In such circumstances, the failure of families to continue to participate in an early childhood program may indicate the need to reevaluate the goals of the intervention, the nature of the services that are provided, and the goodness-of-fit between what the program offers and what the target families perceive as their needs. Thus, assessing the impact of an offer of service could be of great value. On the other hand, an evaluation of the effectiveness of a program that experiences significant sample attrition must be interpreted with par- ticular caution. Although such an assessment may produce interesting and important findings, particularly in the context of its potential impact in the real world, it says very little about how effective the service model would have been if it had been received more favorably by the intended recipients. In summary, the problem of sample attrition is a function of both the questions that are asked and how the findings are interpreted. Ensuring Greater Quality Control The importance of quality control is described earlier in this chapter as an essential feature of effective intervention. When addressed honestly, it represents a powerful strategy for enhancing the early childhood agenda. When thwarted by inadequate resources, professional inertia, or the fear of critical findings, it represents a serious threat to the field and to the children and families it is designed to serve. In view of the extensive literature that has been accumulated based on descriptive and quasi-experimental research, the need for an open and honest commitment to true experimentation with randomization represents one of the most significant challenges facing early childhood policy makers, service providers, and program evaluators. In the final analysis, the future
HEALTHY DEVELOPMENT THROUGH INTERVENTION 369 vitality of the field will depend on the extent to which well-designed experi- ments can be conducted in a nonthreatening atmosphere in an effort to promote continuous quality improvement based on continually expanding knowledge. Defining and Achieving Cultural Competence The development of all young children and the functioning of all fami- lies unfold within a distinctive cultural context. This fundamental concept, which is discussed in greater detail in Chapter 3, is reflected in the values and beliefs that shape parenting practices and the expectations that families (as well as societies) have for their children, beginning from the moment of birth (GarcÃa Coll and Magnuson, 2000; Greenfield and Suzuki, 1998; Super and Harkness, 1986). The importance of examining the design and implementation of early childhood policies and practices through a cultural lens cannot be overstated. All early childhood intervention initiatives, as described earlier in this chapter, are generally predicated on both a presumption of vulnerability (in the child or the family or both) and a belief that specific services can alter the childâs daily experiences in a way that will increase the odds of a more favorable developmental outcome. Implicit in this model are two assump- tionsâfirst, that the caregiving environment needs modification and, sec- ond, that there is clear agreement between the family and the service pro- gram on the desired outcomes. In both cases, the potential for biased, ethnocentric value judgments on the part of the service provider is high, and therefore the need to guard against inappropriate or intrusive interventions is real. As described in Chapter 3, there is considerable variability in the cul- tural practices that characterize families with young children. This diver- sity is manifested in a wide variety of scripts that reflect routine approaches to daily childrearing tasks, in conjunction with significant differences in the kinds of emerging skills that caregivers value and nurture in young children. For example, the cultural practices of many ethnic minority families in the United States, including but not restricted to recent immigrants, differ sub- stantially from those of white, middle-class families. Consequently, the children in such families may exhibit developmental patterns or specific skills that differ from those required to succeed in school. In such circum- stances, it would be most inappropriate to label the child as developmen- tally delayed or disabled, even though he or she would be at greater risk of failing in school and could benefit from services designed to enhance school readiness. Differences in parental behaviors are also vulnerable to being judged as inappropriate or misclassified as abnormal. For example, different cultural
370 FROM NEURONS TO NEIGHBORHOODS beliefs and practices related to leaving young children unsupervised, or different patterns of discipline in association with disagreement about the criteria for defining maltreatment, can result in highly sensitive challenges for a child welfare system (Korbin, 1994; Rose and Meezan, 1996). Fami- lies from different cultures also have different beliefs and understanding about developmental disabilities and mental health problems, which may influence their reactions to diagnoses, adaptation to the challenges of caring for a child with special needs, and preferences among service delivery op- tions (Bernheimer et al., 1990; Coates and Vietze, 1996; Lynch and Hanson, 1998; Seligman and Darling, 1997). Recognizing the critical importance of cultural differences, as described in Chapter 3, each of the diverse service streams that constitute the early childhood intervention landscape endorses the central importance of cul- tural competence as a cornerstone of state-of-the-art practice. Conse- quently, an increasing number and variety of resources to provide guidance in this area have proliferated in recent years. Most of the available mate- rial, however, is conceptual rather than empirical (Johnson-Powell, 1997; Lewis, 2000; Lynch and Hanson, 1998), Thus, despite widespread consen- sus about its importance, the underlying science of cultural competence remains to be developed. Identifying and Responding to the Special Needs of Distinctive Subgroups Notwithstanding the common knowledge base that transcends the com- partmentalized world of early childhood intervention, specific population subgroups confront unique challenges that require specialized expertise. For children, the presence of a biologically based disability, such as cerebral palsy or a sensory loss, requires an intervention strategy that incorporates knowledge about both normative child development and adaptation to a specific physical impairment. For mothers, the diagnosis of depression or a substance abuse problem adds an enormous burden to the normative stresses of parenting, and necessitates services that go beyond the provision of simple advice and support. And for families that confront severe economic hardship and ongoing domestic violence, the needs of their young children extend beyond the addition of educational enrichment activities. The overarching challenge for both policy makers and service providers is to integrate specialized services (when they are required) in a comprehensive framework that addresses the generic needs of all children and families, while recognizing the importance of individual differences and the necessity of cultural competence in an increasingly pluralistic society. Another dimension of special needs requiring sensitive attention is the extent to which early childhood interventions might be beneficial for some but have unintended negative consequences for others. From the perspec-
HEALTHY DEVELOPMENT THROUGH INTERVENTION 371 tive of the family, programs that focus explicitly on parent training may send a message of presumed incompetence, which might undermine a motherâs or fatherâs self-confidence and contribute inadvertently to less effective performance. Similarly, parenting interventions that respond to cultural differences in a dismissive or pejorative manner are likely to pre- cipitate significant conflict or be rejected as unacceptable. Related to these concerns, some observers have cautioned that the provision of formally organized support services may sometimes interfere with the natural devel- opment of the informal social networks needed by all families (Affleck et al., 1989). Indeed, one study found that professionally directed support groups may actually serve as additional stressors for some mothers (Krauss et al., 1993). Some services may also have the unintended consequence of limiting child opportunities. For example, inappropriate interventions may cause some parents to interact with their child in an unnatural, therapeutic man- ner rather than through a natural and comfortable parent-child relation- ship. From the perspective of the child, a tightly structured intervention that is delivered in a highly prescriptive style may interfere with the normal adaptive and self-righting mechanisms that are inherent in the developmen- tal process. In a comparable fashion, isolation from normative settings for a child with a disability results in a distorted social world that provides limited opportunities for healthy adaptation. Inevitable tensions between the generic and idiosyncratic characteris- tics and needs of children and families create a complex agenda for the early childhood field. All children, with or without biological or environmental vulnerabilities, do best when they are reared in a nurturing environment that responds to their individuality and invests in their well-being. All families, regardless of their material resources, depend on informal social supports and varying levels of professional service. Thus, despite the chal- lenges of special needs, the general principles of development apply to all children and families across the broad array of early childhood service systems. Influencing and Assessing the Impacts of Postintervention Environments The demands of policy makers for evidence of long-term impacts as a result of investments in early childhood programs have put service provid- ers and program evaluators in a difficult bind. Central to this dilemma is the widely endorsed assertion that effective early intervention services do not serve as inoculations that confer a lifetime of immunity against the adverse effects of later experiences. Indeed, no intervention prior to school entry can ever be powerful enough to fully buffer a vulnerable child from
372 FROM NEURONS TO NEIGHBORHOODS the negative effects of attending a demoralizing school or living in a danger- ous neighborhood. The extensive documentation of IQ fade-out, particularly for children who live in impoverished environments and receive a mediocre public edu- cation, was described earlier. Nevertheless, the few studies that have fol- lowed early childhood program graduates through the school years and into adult life have demonstrated variable patterns of so-called sleeper effects in such areas as high school graduation, welfare dependence, in- come, and criminal behavior (Lally et al., 1988; Schweinhart et al., 1993). The need for more longitudinal data to further elucidate this phenomenon is clear, but follow-up studies must pay greater attention to the continuing influence of the childâs environment throughout the life span. The key challenge facing early childhood intervention professionals is the need to establish the standard of proof that must be met in order to endorse a program as effective. The immediate and short-term benefits for both families and taxpayers are real, and their value should not be dimin- ished. Moreover, the medium-term benefits of reduced grade retention and special education referrals can be quite large economically and could justify the initial costs of early intervention, even in the absence of longer-term impacts. Strengthening the Service Infrastructure Services to promote the health and well-being of all young children, as well as early intervention efforts for those who are developmentally vulner- able, cover a diverse and highly fragmented array of policies, programs, and funding sources. This fragmentation has been the object of considerable criticism for decades. The extensive knowledge base presented in this report provides a powerful tool to guide the design of a more rational and efficient infrastructure for early childhood services that incorporates the multiple streams that have evolved independently over the years. Beyond the general challenges of excessive service fragmentation and redundancy, the limited availability of mental health assistance for children under age 6 represents a massive gap in the current early childhood infra- structure (Knitzer, 2000). This shortcoming is particularly problematic in view of the high prevalence of emotional and behavioral problems in young children and the inextricable interrelation among cognitive, social, and emotional development, as elaborated in Chapters 5 and 6. Two striking examples illustrate the magnitude of this problem. First, the federal mandate to provide family-centered services for infants and toddlers with developmental disabilities or delays, under the provisions of Part C of the Individuals with Disabilities Education Act, focuses primarily on cognitive, language, and motor impairments and does not accord a
HEALTHY DEVELOPMENT THROUGH INTERVENTION 373 comparable entitlement to services for children whose difficulties lie in the domains of emotional and social development. In a similar fashion, mul- tiple federal programs address the problem of child maltreatment, and all states require mandatory reporting of suspected child abuse or neglect, yet large numbers of maltreated young children are managed in child welfare systems that have limited professional expertise in normative child develop- ment, developmental disabilities, and early childhood mental health. Both the failure to incorporate state-of-the-art mental health expertise into poli- cies and programs designed to address the needs of children with disabilities and the absence of sophisticated developmental services for young children who have been maltreated are dramatic examples of the significant gap between current knowledge and practice. Assessing Costs and Making Choices Among Early Childhood Investments The early era of early childhood intervention in the United States fo- cused relatively little attention on the question of cost. Whether the target population was dealing with the stresses of poverty or the challenges of developmental disability, public funds were appropriated on the basis of assumed need and the return on investment was rarely quantified. Beginning in the 1980s and continuing to the present, all health and human services have been faced with increasing demands for cost-effective- ness and demonstrated cost-benefit. This shift has been embedded in a changing political climate characterized by reductions in taxes and appro- priations for government social programs, devolution of authority from the federal to state and local levels, and an increasing reliance on market solu- tions to address health and human services needs. In this context, early childhood intervention programs face a less forgiving environment that demands evidence of both measurable impacts and more efficient service delivery. Although much of the impetus for greater accountability has been stereotyped as a lack of commitment to the well-being of vulnerable chil- dren, it is important to note that the rigorous assessment of costs and benefits is the best way to ensure that finite resources are used in the best interests of children and families. It is essential, however, to also recognize that the distribution of benefits matters. For example, savings to a school budget do not necessarily accrue to families and children in need. Thus, although public financial gains are generally desirable, a policy that re- sulted in the same net benefits, but proportionally more for the children and relatively less for the general taxpayer, might be preferred. The costs of early childhood services vary considerably. Averages per child range (in 1994 dollars) from Missouri Parents As Teachers (less than $1,000 per year) to the Avance Family Support and Education program
374 FROM NEURONS TO NEIGHBORHOODS ($1,600 per year for 1 or 2 years), Even Start Family Literacy Programs ($2,660 for 1 year), Child and Family Resource Programs ($3,220 per year for 5 years), Head Start Family Service Centers ($3,500 per year for 3 years), Head Start ($4,000 per year for 1 or 2 years), New Chance ($8,300 per year for 1.5 years), the Comprehensive Child Development Program ($8,600 per year for 5 years), and the Infant Health and Development Program ($10,000 per year for three years) (St. Pierre et al., 1995b). With such a wide discrepancy in costs, both annually and as a function of the number of years of program enrollment, the demand for evidence of inter- vention impacts is compelling and the need for more data on cost-effective- ness is clear. In their efforts to develop model early childhood programs that can have significant and long-lasting impacts, interventionists have periodically designed and evaluated elaborate service models with costs per child that have exceeded $8,000 to $10,000 per year. Yet legislators and service providers typically have budgets that are too small to offer such programs to more than a small fraction of the children who could profit from them. Early intervention researchers generally ask âWhat works?â and âHow does it work?â Budget-constrained policy makers and practitioners typi- cally seek the most cost-effective programs that serve the largest possible number of needy children. How significant an impact must an intervention have to be worth- while? Is a cheaper-by-half, scaled-back version of a proven program likely to provide at least half the benefits? If several interventions show docu- mented benefits but a decision maker cannot fund them all, how should he or she choose among the alternatives? Is the $4.66 billion the United States spent on Head Start programs in 1999 too much or too little? These are some of the difficult policy questions that arise. Cost-effectiveness and cost-benefit frameworks (Gramlich, 1990; Levin, 1983) are useful though imperfect means of addressing all of these issues. Each begins with a systematic accounting of the full costs of an intervention program. Dollar expenditures on staff salaries and benefits generally make up the bulk of such costs. However, a complete cost accounting must include subtler expenditures, such as the value of the time volunteers spend helping out with the program and the cost of the needed classrooms or other facilities. Thus, volunteer time is not free, since it could have been used productively for some other purpose; facilities are not free, because scaling up a program to provide services for tens of thousands of children often requires major expenditures to rent or build facilities. The benefits of early childhood programs are often difficult to quantify with a monetary value. Their accurate assessment requires a precise com- parison of developmental outcomes for children who received program services and otherwise similar children who did not. As described in Chap-
HEALTHY DEVELOPMENT THROUGH INTERVENTION 375 ter 4, randomized experimental designs often provide the strongest basis for estimating program impacts. But the measurement of child development is not always straightforward, and the assignment of a dollar value to a given outcome is a complex challenge. The most commonly measured dependent variable in the intervention literature is IQ. This report argues that impor- tant domains of child development and well-being also include emotional health and social behavior, and a full cost-benefit perspective underscores the importance of a broad conception of child outcomes. For example, IQ gains may benefit society by increasing the productivity of the nationâs workforce. However, as shown by data from the High/Scope Perry Pre- school Program, the societal value of improvements in childrenâs mental health and reductions in criminal activity can easily exceed the value of IQ- based productivity gains. Reactions to the assignment of dollar values to impacts on childrenâs development often range from skepticism to moral indignation. Advocates often ask how one can assign a dollar value to the lives that are saved by child care safety regulations, or to a boost in self-esteem for a child who avoids the stigma of assignment to a special education program. In contrast to the questions asked of cost-benefit analyses, some investigators conduct cost-effectiveness studies to avoid the problem of benefit valuation by sim- ply comparing the relative costs of programs designed to meet similar goals (Barnett, 2000). In view of their systematic attempt to account for the dollar value of all important program benefits, cost-benefit studies are more complex and ambitious than cost-effectiveness approaches. Program benefits, such as reduced use of special education services or decreased grade repetition, can be assigned specific dollar values based on what school systems spend to provide special and regular education services. However, these valuations of school-related benefits are likely to be conservative, in the sense that they omit the value to children and their families of avoiding stigmatizing educa- tion classifications and tracks. Quantifying the benefits of early childhood intervention programs for such long-term outcomes as criminal behavior and adult career success is exceedingly more difficult, since it requires the maintenance of contact with intervention and control group participants over a very long period of time. When the complete accounting of such benefits is not possible, a cost- benefit analysis evaluates all possible costs and benefits and then makes prudent judgments about whether the missing data would be likely to push the computed difference between benefits and costs in a positive or nega- tive direction. Table 13-1 summarizes results from a cost-benefit analysis of the High/ Scope Perry Preschool Program (Schweinhart et al., 1993). Random assign- ment of children to intervention and control groups, the systematic follow-
376 FROM NEURONS TO NEIGHBORHOODS up of both groups through age 27, and a careful accounting of both costs and benefits make this program a good example to illustrate the cost- benefit approach. On the minus side, the study sample was small (117 children in all were interviewed at age 27) and located in a single Midwest- ern city, which suggests a need for replication before much is made of the specific benefit estimates emerging from the study. Since program impacts on IQ test scores were no longer found by the time the children reached age 8, this study illustrates the potential value of an early childhood intervention that produces long-term impacts in impor- tant social domains despite the apparent transient nature of its short-term cognitive effects. Details on procedures for the cost and benefit accounting method are provided by Schweinhart and his colleagues (1993). Despite the value of the economic analyses that have been conducted on the Perry Preschool Program, several features of the intervention make it difficult to generalize from the findings. One remarkable factor is its high cost ($12,356 per child in 1992 dollars and $14,683 in 1998 dollars for a 1- to 2-year program), which is much higher than that of the typical Head Start program ($5,021 per child in 1998). This reflects its greater intensity of services compared with most early education programs, which makes it all but impossible to generalize to less intensive interventions. A second, and perhaps even more remarkable factor, is the size and nature of the reported benefits. At age 27, the full complement of measured benefits totaled $70,786, far in excess of the $12,356 costs of the program. Thus, despite the IQ fade-out by third grade, the answer to the question of whether the resources expended on the program were socially profitable is a resounding âyes.â Specifically, although the IQ impacts were short-lived, children who participated in the intervention spent significantly less time in special education programs and were less likely to repeat a grade, differ- ences that saved their schools some $6,872 when averaged across all chil- dren served. As large as these savings may seem, they are still not consid- TABLE 13-1 Costs and Benefits of the High/ Scope Perry Preschool Program Total cost per child $12,356 Total benefits per child $70,786 K-12 expenditures saved 6,872 Higher adult earnings 14,498 Crime saving 49,044 Other 372 NOTE: Adapted from Schweinhart et al. (1993:Table 43). Data are in 1992 dollars and discounted at 3 percent.
HEALTHY DEVELOPMENT THROUGH INTERVENTION 377 ered sufficient by some policy makers to outweigh program costs. Twice as large (some $14,498 per child) as the education benefits, however, were the higher earnings enjoyed by program participants in their young adult years. Beyond the savings linked to education costs and employment earnings, the greatest impact in the Perry Preschool benefit-cost calculus was the computed value of the favorable differences in crime victim and incarcera- tion costs for program participants relative to the control group. In fact, the costs to the criminal justice system and the costs of crime victimization are so large that the striking intervention-control differences in rates of criminal activity translated into a $49,044 program benefit per individual served. Relatively few of the total benefits were reaped by the children enrolled in the Perry Preschool program. In fact, of the $70,786 total benefits, only $8,815 accrued to the participating children, mostly in the form of higher earnings in their early adult years. The $49,044 crime-related benefits constitute savings to taxpayers and potential crime victims, as did a nearly $7,000 savings from lower enrollment in special education programs. Thus, while a complete cost/benefit accounting considers all sources of benefits, it is important to identify to what extent the participants themselves are receiving the benefits. The long-term analyses of the High/Scope Perry Preschool Program clearly demonstrate that a very intensive early intervention program can produce benefits far in excess of its costs. Despite the value of such a study, however, it does not begin to address a larger set of questions that are crucial for policy makers and practitioners. For example, do less expensive, less intensive programs also produce more benefits than costs? If resources are limited, is it better to offer high-quality programs to fewer children or more affordable programs to a larger number? Neither a single study nor a collection of evaluations of other high-cost, intensive studies can begin to answer these important questions. The only way to address this complex and critically important issue is to evaluate a range of high- and low-intensity programs. Such evaluations may well suggest that there are smaller but still positive benefits for smaller programs. Welfare-to-work experiments in the 1980s resulted in precisely that conclusion (Gueron and Pauly, 1991). Very expensive (e.g., $10,000 per participant), intensive programs produced more benefits than costs, but so too did less expensive (e.g., $1,000 per participant) training-based pro- grams, as well as extremely modest investments (e.g., $100 per participant) in job-search programs. Thus, cost-benefit information from a wide range of potential programs can help policy makers analyze trade-offs between intensity and coverage in the context of highly constrained budgets. Regrettably, the number of early intervention services for which cost and benefit data are available is exceedingly small and confined largely to
378 FROM NEURONS TO NEIGHBORHOODS highly intensive programs. Consequently, it is difficult to provide economi- cally guided answers to pressing questions about the optimal level of social investment that should be made in a wide range of early childhood efforts, such as Head Start, nutrition programs, and parenting education. Simi- larly, it is not possible to set safety and group size standards for child care settings based on reliable knowledge of their costs and benefits. The an- swers to these and many other important policy questions await further study. CONCLUSIONS Early childhood interventions are provided under the auspices of a wide variety of policies and programs. These include high-prevalence ser- vices, such as child care and early childhood education, as well as targeted interventions for a range of vulnerabilities, including economic hardship, childhood disabilities, parental substance abuse, and child maltreatment, among others. State-of-the-art early childhood programs are guided by a rich knowledge base that reflects a mixture of developmental theory, em- pirical research, and professional experience. A critical examination of this knowledge reveals considerable agreement on theoretical concepts, both replicable patterns and inconsistencies in the empirical data, and substan- tial gaps in potentially important areas of investigation. The research literature on the efficacy and effectiveness of early inter- vention programs encompasses thousands of peer-reviewed papers, mono- graphs, edited volumes, and project reports. Despite continuing debate about the nature of the underlying science and its methodological rigor, there is considerable agreement across all service streams about desired child outcomes, and about important family-based and community-based factors that influence child health and development. Generally speaking, well-designed early interventions that are child-focused produce immediate gains on standardized developmental measures, most commonly IQ scores. These findings have been replicated in multiple studies of children living in a variety of adverse circumstances and those with a wide range of diag- nosed disabilities, although the largest benefits are typically found in model demonstration projects that generally incur high costs per child. For children at risk because of low socioeconomic status, the short- term benefits of higher IQ scores typically fade out during the middle childhood years, but persistent intervention-control group differences have been documented, favoring those who received early services, in later aca- demic achievement, retention in grade, and referral for special education. Long-term follow-up data on low-income children are more limited but provide some evidence of intervention-control differences in high school graduation, employment, dependence on public assistance, and involve-
HEALTHY DEVELOPMENT THROUGH INTERVENTION 379 ment in the criminal justice system. Comparable longitudinal data on children with disabilities are not available. Complementary but distinct from child-focused interventions, a variety of early childhood services are delivered through family-focused models, many of which are home-based. Research on model programs reveals that well-designed services with explicitly defined goals can be effective in chang- ing parenting practices and influencing parent-child interactions. For fami- lies of children with cognitive, language, or sensory impairments, enhanced parental competencies linked to greater understanding of the childâs special needs are important mediators of improved child outcomes. The measur- able effects of parent-focused interventions on standardized child develop- ment scores in economically disadvantaged families, however, are less con- clusive, and there is little empirical documentation that nonspecific, general family support models for high-risk families, which typically are less expen- sive to deliver, have significant impacts on either parent behavior or as- sessed child performance. In the final analysis, there is considerable evidence to support the no- tion that model programs that deliver carefully designed interventions with well-defined goals can affect both parenting behavior and the developmen- tal trajectories of children whose life course is threatened by socioeconomic disadvantage, family disruption, or diagnosed disability. Programs that combine child-focused educational activities with explicit attention to par- ent-child interaction patterns and relationship building appear to have the greatest impacts. In contrast, services that are supported by more modest budgets and based on generic support, often without a clear delineation of intervention strategies matched directly to measurable objectives, appear to be less effective for families facing significant risk. The general question of whether early childhood programs can make a difference has been asked and answered in the affirmative innumerable times. This generic query is no longer worthy of further investigation. The central research priority for the early childhood field is to address more important sets of questions about how different types of interventions influ- ence specific outcomes for children and families who face differential op- portunities and vulnerabilities. To this end, program evaluators must as- sess the distinctive needs that must be met, the soundness of the intervention strategy, its acceptability to the intended recipients, the quality of its imple- mentation, and the extent to which less intensive, broader-based programs can be developed that are both beneficial and cost-effective. Programs with only transitory impacts on childrenâs IQ scores may still be socially profitable investments. The measurement of specific program effects on children must go beyond traditional cognitive evaluations (such as IQ) and include greater focus on a broad range of functional capacities, particularly in the social and emotional domains. Middle- and long-term
380 FROM NEURONS TO NEIGHBORHOODS follow-up studies also must pay greater attention to the assessment of subsequent and continuing environmental influences on development after the intervention has been completed. Although sometimes hard to quantify, program benefits and costs pro- vide vital information for budget-constrained policy makers and practitio- ners. Nevertheless, there currently are few systematic data on the costs and benefits of intensive early childhood interventions, and almost none on the less intensive, real-world services that are more likely to be implemented on a large scale. Practitioners and policy makers need careful evaluations of a broad portfolio of intervention programs, including both modest and inten- sive models, as programs with the largest impacts on children are not always the most practical to implement. Although not all decisions about allocating resources for early childhood programs need be based solely on considerations of financial costs and benefits, the need for better economic data is clear. The current agenda for early childhood policy and service delivery in the United States is embedded in four objectives: â¢ Full access to programs whose effectiveness has been demonstrated must be ensured for all eligible children and families. â¢ A culture of ongoing experimentation must be established to pro- mote the design, implementation, and evaluation of alternative ap- proaches for those circumstances in which existing interventions are found to have minimal impact. â¢ A strong commitment to rigorous quality control must be estab- lished and sustained, in order to ensure that all available resources are used in the most effective and efficient manner. â¢ It is essential that all early childhood policies and programs be de- signed and implemented within a culturally competent context and in a manner that respects the importance of individual differences among children and families. A fundamental challenge facing the nation is to find an appropriate balance between long-term investment in human capital development and the moral responsibility to ensure that the quality of life for young children does not fall below a minimum level of decency. Stated simply, certain services are deemed worthy of support because they generate significant long-term dividends. Other programs are essential not because they result in later financial benefits but because they reflect societyâs commitment to those who are most vulnerable and who cannot help themselves.
Knowledge into Action IV