Overview of Risk Factors and Prevention

Current approaches to understanding the etiology of mental disorders in childhood are driven by empirical advances in neuroscience and behavioral research rather than by theories. Epidemiological research on the factors that make children vulnerable to mental illness is important for several reasons: delineating the range of risk factors for particular mental disorders helps to understand their etiology; the populations most at risk can be identified; understanding the relative strength of different risk factors allows for the design of appropriate prevention programs for children in different contexts; and resources can be better allocated to intervene so as to maximize their effectiveness.

Risk Factors

There is now good evidence that both biological factors and adverse psychosocial experiences during childhood influence—but not necessarily“cause”—the mental disorders of childhood. Adverse experiences may occur at home, at school, or in the community. A stressor or risk factor may have no, little, or a profound impact, depending on individual differences among children and the age at which the child is exposed to it, as well as whether it occurs alone or in association with other risk factors. Although children are influenced by their psychosocial environment, most are inherently resilient and can deal with some degree of adversity. However, some children, possibly those with an inherent biological vulnerability (e.g., genes that convey susceptibility to an illness), are more likely to be harmed by an adverse environment, and there are some environmental adversities, especially those that are long-standing or repeated, that seem likely to induce a mental disorder in all but the hardiest of children. A recent analysis of risk factors by Kraemer and colleagues (1997) has provided a useful framework for differentiating among categories of risk and may help point this work in a more productive direction.

Risk factors for developing a mental disorder or experiencing problems in social-emotional development include prenatal damage from exposure to alcohol, illegal drugs, and tobacco; low birth weight; difficult temperament or an inherited predisposition to a mental disorder; external risk factors such as poverty, deprivation, abuse and neglect; unsatisfactory relationships; parental mental health disorder; or exposure to traumatic events.

Biological Influences on Mental Disorders
It seems likely that the roots of most mental disorders lie in some combination of genetic and environmental factors—the latter may be biological or psychosocial (Rutter et al., 1999). However, increasing consensus has emerged that biologic factors exert especially pronounced influences on several disorders in particular, including pervasive developmental disorder (Piven & O’Leary, 1997), autism (Piven & O’Leary, 1997), and early-onset schizophrenia (McClellan & Werry, in press). It is also likely that biological factors play a large part in the etiology of social phobia (Pine, 1997), obsessive-compulsive disorder (Leonard et al., 1997), and other disorders such as Tourette’s disorder (Leckman et al., 1997).

Two important points about biological factors should be borne in mind. The first is that biological influences are not necessarily synonymous with those of genetics or inheritance. Biological abnormalities of the central nervous system that influence behavior, thinking, or feeling can be caused by injury, infection, poor nutrition, or exposure to toxins, such as lead in the environment. These abnormalities are not inherited. Mental disorders that are most likely to have genetic components include autism, bipolar disorder, schizophrenia, and attention-deficit/hyperactivity disorder (ADHD) (National Institute of Mental Health [NIMH], 1998). Second, it is erroneous to assume that biological and environmental factors are independent of each other, when in fact they interact. For example, traumatic experiences may induce biological changes that persist. Conversely, children with a biologically based behavior may modify their environment. For example, low-birth-weight infants who have sustained brain damage, and thereby become excessively irritable, may change the behavior of caretakers in a way that adversely affects the caretaker’s ability to provide good care. Thus, it is now well documented that a number of biologic risk factors exert important effects on brain structure and function and increase the likelihood of subsequently developing mental disorders. These well-established factors include intrauterine exposure to alcohol or cigarette smoke (Nichols & Chen, 1981), perinatal trauma (Whitaker et al., 1997), environmental exposure to lead (Needleman et al., 1990), malnutrition of pregnancy, traumatic brain injury, nonspecific forms of mental retardation, and specific chromosomal syndromes.

Psychosocial Risk Factors
A landmark study on risks from the environment (Rutter & Quinton, 1977) showed that several factors can endanger a child’s mental health. Dysfunctional aspects of family life such as severe parental discord, a parent’s psychopathology or criminality, overcrowding, or large family size can predispose to conduct disorders and antisocial personality disorders, especially if the child does not have a loving relationship with at least one of the parents (Rutter, 1979). Economic hardship can indirectly increase a child’s risk of developing a behavioral disorder because it may cause behavioral problems in the parents or increase the risk of child abuse (Dutton, 1986; Link et al., 1986; Wilson, 1987; Schorr, 1988). Exposure to acts of violence also is identified as a possible cause of stress-related mental health problems (Jenkins & Bell, 1997). Studies point to poor caregiving practices as being a risk factor for children of depressed parents (Zahn-Waxler et al., 1990).

The quality of the relationship between infants or children and their primary caregiver, as manifested by the security of attachment, has long been felt to be of paramount importance to mental health across the life span. In this regard, the relationship between maternal problems and those factors in children that predispose them to form insecure attachments, particularly young infants’ and toddlers’ security of attachment and temperament style and their impact on the development of mood and conduct disorders, is of great interest to researchers. Many investigators have taken the view that the nature and the outcome of the attachment process are related to later depression, especially when the child is raised in an abusive environment (Toth & Cicchetti, 1996), and to later conduct disorder (Sampson & Laub, 1993). The relationship of attachment to mental disorders has been the subject of several important review articles (Rutter, 1995; van IJzendoorn et al., 1995).

There is controversy as to whether the key determinant of “insecure” responses to strange situations stems from maternal behavior or from an inborn predisposition to respond to an unfamiliar stranger with avoidant behaviors, such as is found in socially phobic children (Belsky & Rovine, 1987; Kagan et al., 1988; Thompson et al., 1988; Kagan, 1994, 1995). Kagan demonstrated that infants who were more prone to being active, agitated, and tearful at 4 months of age were less spontaneous and sociable and more likely to show anxiety symptoms at age 4 (Snidman et al., 1995; Kagan et al., 1998). These findings are of considerable significance, because long-term study of such highly reactive, behaviorally inhibited infants and toddlers has shown that they are excessively shy and avoidant in early childhood and that this behavior persists and predisposes to later anxiety (Biederman et al., 1993). There is also some controversy as to whether “difficult” temperament in an infant is an early manifestation of a behavior problem, particularly in children who go on to demonstrate such problems as conduct disorder (Olds et al., 1999). One analysis of the attachment literature suggests that abnormal or insecure forms of attachment are largely the product of maternal problems, such as depression and substance abuse, rather than of individual differences in the child (van IJzendoorn et al., 1992).

The relationship between a child’s temperament and parenting style is complex (Thomas et al., 1968); it may be either protective if it is good or a risk factor if it is poor. Thus, a difficult child’s chances of developing mental health problems are much reduced if he or she grows up in a family in which there are clear rules and consistent enforcement (Maziade et al., 1985), while a child exposed to inconsistent discipline is at greater risk for later behavior problems (Werner & Smith, 1992).

Family and Genetic Risk Factors
As noted above in the relationships between temperament and attachment, in some instances the relative contributions of biologic influences and environmental influences are difficult to tease apart, a problem that particularly affects studies investigating the impact of family and genetic influences on risk for childhood mental disorder. For example, research has shown that between 20 and 50 percent of depressed children and adolescents have a family history of depression (Puig-Antich et al., 1989; Todd et al., 1993; Williamson et al., 1995; Kovacs, 1997b). The exact reasons for this increased risk have not been fully clarified, but experts tend to agree that both factors interact to result in this increased risk (Weissman et al., 1997). Family research has found that children of depressed parents are more than three times as likely as children of nondepressed parents to experience a depressive disorder (see Birmaher et al., 1996a and 1996b for review). Parental depression also increases the risk of anxiety disorders, conduct disorder, and alcohol dependence (Downey & Coyne, 1990; Weissman et al., 1997; Wickramaratne & Weissman, 1998). The risk is greater if both parents have had a depressive illness, if the parents were depressed when they were young, or if a parent had several episodes of depression (Merikangas et al., 1988; Downey & Coyne, 1990; McCracken, 1992a, 1992b; Mufson et al., 1992; Warner et al., 1995; Wickramaratne & Weissman, 1998).

Effects of Parental Depression
Depressed parents may be withdrawn and lack energy and consequently pay little attention to, or provide inadequate supervision of, their children. Alternatively, such parents may be excessively irritable and overcritical, thereby upsetting children, demoralizing them, and distancing them (Cohn et al., 1986; Field et al., 1990). At a more subtle level, parents’ distress— being pessimistic, tearful, or threatening suicide—is sometimes seen or heard by the child, thereby inducing anxiety. Depressed parents may not model effective coping strategies for stress; instead of “moving on,” some provide an example of“giving up” (Garber & Hilsman, 1992). Depression is also often associated with marital discord, which may have its own adverse effect on children and adolescents. Conversely, the behavior of the depressed child or teenager may contribute to family stress as much as being a product of it. The poor academic performance, withdrawal from normal peer activities, and lack of energy or motivation of a depressed teenager may lead to intrusive or reprimanding reactions from parents that may further reduce the youngster’s self-esteem and optimism.

The consequences of maternal depression vary with the state of development of the child, and some of the effects are quite subtle (Cicchetti & Toth, 1998). For example, in infancy, a withdrawn or unresponsive depressed mother may increase an infant’s distress, and an intrusive or hostile depressed mother may lead the infant to avoid looking at and communicating with her (Cohn et al., 1986). Other studies have shown that if infants’ smiles are met with a somber or gloomy face, they respond by showing a similarly somber expression and then by averting their eyes (Murray et al., 1993).

During the toddler stage of development, research shows that the playful interactions of a toddler with a depressed mother are often briefer and more likely to be interrupted (by either the mother or the child) than those with a nondepressed parent (Jameson et al., 1997). Research has shown that some depressed mothers are less able to provide structure or to modify the behavior of excited toddlers, increasing the risk of out-of-control behavior, the development of a later conduct disorder, or later aggressive dealings with peers (Zahn-Waxler et al., 1990; Hay et al., 1992). A depressed mother’s inability to control a young child’s behavior may result in the child failing to learn appropriate skills for settling disputes without reliance on aggression.

Stressful Life Events
The relationship between stressful life events and risk for child mental disorders is well established (e.g., Garmezy, 1983; Hammen, 1988; Jensen et al., 1991; Garber & Hilsman, 1992), although this relationship in children and adolescents is complicated, perhaps reflecting the impact of individual differences and developmental changes. For example, there is a relationship between stressful life events, such as parental death or divorce, and the onset of major depression in young children, especially if they occur in early childhood and lead to a permanent and negative change in the child’s circumstances. Yet findings are mixed as to whether the same relationship is true for depression in midchildhood or in adolescence (Birmaher et al., 1996a and 1996b; Garrison et al., 1997).

Childhood Maltreatment
Child abuse is a very widespread problem; it is estimated that over 3 million children are maltreated every year in the United States (National Committee to Prevent Child Abuse, 1995). Physical abuse is associated with insecure attachment (Main & Solomon, 1990), psychiatric disorders such as post-traumatic stress disorder, conduct disorder, ADHD (Famularo et al., 1992), depression (Kaufman, 1991), and impaired social functioning with peers (Salzinger et al., 1993). Psychological maltreatment is believed to occur more frequently than physical maltreatment (Cicchetti & Carlson, 1989); it is associated with depression, conduct disorder, and delinquency (Kazdin et al., 1985) and can impair social and cognitive functioning in children (Smetana & Kelly, 1989).

Peer and Sibling Influences
The influence of maladaptive peers can be very damaging to a child and greatly increases the likelihood of adverse outcomes such as delinquency, particularly if the child comes from a family beset by many stressors (Friday & Hage, 1976; Loeber & Farrington, 1998). One way to reduce antisocial behavior in adolescents is to encourage such youths to interact with better adapted youths under the supervision of a mental health worker (Feldman et al., 1983). Sibling rivalry is a common component of family life and, especially in the presence of other risk factors, may contribute to family stresses (Patterson & Dishion, 1988). Although almost universal, in the presence of other risk factors it may be the origin of aggressive behavior that eventually extends beyond the family (Patterson & Dishion, 1988). In stressed or large families, parents have many demands placed on their time and find it difficult to oversee, or place limits on, their young children’s behavior. When parental attention is in short supply, young siblings squabbling with each other attract available attention. In such situations, parents rarely comment on good or neutral behavior but do pay attention, even if in a highly critical and negative way, when their children start to fight; as a result, the act of fighting may be inadvertently rewarded. Thus, any attention, whether it be praise or physical punishment, increases the likelihood that the behavior is repeated.

Correlations and Interactions Among Risk Factors
Recent evidence suggests that social/environmental risk factors may combine with physical risk factors of the child, such as neurological damage caused by birth complications or low birthweight, fearlessness and stimulation-seeking behavior, learning impairments, autonomic underarousal, and insensitivity to physical pain and punishment (Raine et al., 1996, 1997, 1998). However, testing models of the impact of risk factor interactions for the development of mental disorders is difficult, because some of the risk factors are difficult to measure. Thus, the trend these days is to move away from the consideration of individual risk factors toward identifying measurable risk factors and their combinations and incorporating all of them into a single model that can be tested (Patterson, 1996).

The next section describes a series of preventive interventions directed against the environmental risk factors described above.

Prevention

Childhood is an important time to prevent mental disorders and to promote healthy development, because many adult mental disorders have related antecedent problems in childhood. Thus, it is logical to try to intervene early in children’s lives before problems are established and become more refractory. The field of prevention has now developed to the point that reduction of risk, prevention of onset, and early intervention are realistic possibilities. Scientific methodologies in prevention are increasingly sophisticated, and the results from high-quality research trials are as credible as those in other areas of biomedical and psychosocial science. There is a growing recognition that prevention does work; for example, improving parenting skills through training can substantially reduce antisocial behavior in children (Patterson et al., 1993).

The wider human services and law enforcement communities, not just the mental health community, have made prevention a priority. Policymakers and service providers in health, education, social services, and juvenile justice have become invested in intervening early in children’s lives: they have come to appreciate that mental health is inexorably linked with general health, child care, and success in the classroom and inversely related to involvement in the juvenile justice system. It is also perceived that investment in prevention may be cost-effective. Although much research still needs to be done, communities and managed health care organizations eager to develop, maintain, and measure empirically supported preventive interventions are encouraged to use a risk and evidence-based framework developed by the National Mental Health Association (Mrazek, 1998).

Some forms of primary prevention are so familiar that they are no longer thought of as mental health prevention activities, when, in fact, they are. For example, vaccination against measles prevents its neurobehavioral complications; safe sex practices and maternal screening prevent newborn infections such as syphilis and HIV, which also have neurobehavioral manifestations. Efforts to control alcohol use during pregnancy help prevent fetal alcohol syndrome (Stratton et al., 1996). All these conditions may produce mental disorders in children.

This section describes several exemplary interventions that focus on enhancing mental health and primary prevention of behavior problems and mental health disorders. Prevention of a disorder or its recurrence or exacerbation is discussed together with that disorder in other sections of this chapter. Prevention strategies usually target high-risk infants, young children, adolescents, and/or their caregivers, addressing the risk factors described above.

Project Head Start
Project Head Start, though generally conceived of as an early childhood intervention program, is probably this country’s best known prevention program. In 1965, when it was designed and first implemented in 2,500 communities, Head Start’s target population was economically disadvantaged preschool children. Its goal was to improve the social competence of these children through an 8-week comprehensive intervention that included a center-based component and a home visit by community aides, focusing on social, health, and education services (Karoly et al., 1998). A number of psychologists, most notably Jerome Bruner (1971), argued that children can be trained to think in a more logical way and that the development of logic is not entirely predetermined. Bruner’s views were very influential in launching early intervention programs such as Head Start. There is now ample evidence that, by providing an appropriately stimulating environment, significant advances in knowledge and reasoning ability can be achieved.

The program has served over 15 million children and has cost $31 billion since its inception (General Accounting Office, 1997). It has changed in many ways in the intervening years, and there now is considerable program variation across localities (Zigler & Styfco, 1993). Early evaluations of Head Start showed promising results in terms of higher IQ scores, but over the years many of the findings have met with criticism and skepticism. The reason is that there has been no national randomized controlled trial to evaluate the program as originally designed (Karoly et al., 1998).

Repeated evaluations of Head Start programs that did not employ such a rigorous design (Berrento-Clement et al., 1984; Seitz et al., 1985; Lee et al., 1990; Yoshikawa, 1995) have shown that, although focused early education can improve test scores, the advantage is short-lived. The test scores of children of comparable ability who do not receive early childhood education quickly catch up with those who have been in Head Start programs (Lee et al., 1990). Yet there appear to be more enduring academic outcomes. A review of 36 studies of Head Start and other early childhood programs found them to lower enrollment in special education and to enhance rates of high school graduation and promotion to the next grade level (Barnett, 1995). Head Start and other forms of early education offer arguably even more important benefits, which do not become apparent until children are older. The advantages are mainly social, rather than cognitive, and include better peer relations, less truancy, and less antisocial behavior (Berrento-Clement et al., 1984; Provence, 1985; Seitz et al., 1985; Webster-Stratton, 1998; Weikart, 1998). Although important from a societal perspective, it is not known whether these very significant benefits are due to direct effects on the child or to the parent education programs that often accompany Head Start programs (Zigler & Styfco, 1993).

Carolina Abecedarian Project
The Carolina Abecedarian Project is an example of an early educational intervention for high-risk children that has been tested more rigorously than Head Start in well-designed, randomized, and controlled trials. It addresses the issue of the timing of the intervention, that is, when an intervention should begin and how long it should continue. Unlike Head Start, children were enrolled in this program at birth and remained in it for several years.

In the Carolina Abecedarian Project, children who had been identified at birth as being at high risk for school failure on the basis of social and economic variables were enrolled in a child-centered prevention-oriented intervention program delivered in a day care setting from infancy to age 5 (Campbell & Ramey, 1994 ¹). The preschool intervention operated 8 hours a day for 50 weeks a year and included an infant curriculum to enhance development and parent activities. At elementary school age, a second intervention was provided: the children, who were then in kindergarten, received 15 home visits a year for 3 years from a teacher who prepared a home program to supplement the school’s basic curriculum. There were significant positive effects from the two-phase intervention on intellectual development and academic achievement, and these effects were maintained through age 12, which was 4 years after the intervention ended.

Infant Health and Development Program
The Infant Health and Development Program (IHDP) also began at birth and continued for several years and was also designed for low-birth-weight and premature infants (McCarton et al., 1997²). The intervention was provided until the children reached 3 years of age. It included pediatric care, home visits, parent group meetings, and center-based schooling 5 days a week from 12 months of age to 3 years. At the end of the intervention, the group receiving it had significantly higher mean IQ scores than did the control group. Of note, although children’s behavior problems were not targeted by the intervention, mothers of children in the intervention group reported significantly fewer behavior problems than those in the control group.

Elmira Prenatal/Early Infancy Project
The Elmira Prenatal/Early Infancy Project is an excellent example of a preventive intervention that targeted an at-risk population to prevent the onset of a series of health, social, and mental health problems in children and in their mothers (Olds et al., 1998 and previous years³ ). This study warrants special attention because of its positive and enduring findings, randomized, controlled design, cost-benefit analysis, and unusually long-term follow up of 15 years. The study began by focusing on pregnant women bearing their first child in a small, semirural county in upstate New York. The children of these women were considered high risk because of their mother’s young maternal age, single-parent status, or low socioeconomic level. There were four study groups to which random assignment was made. The first group received developmental screening at ages 1 and 2; the second group received screening and free transportation to health care; the third group received screening, transportation, and nurse home visits once every 2 weeks during pregnancy; and the fourth group received all of the above plus continued home visits by a nurse on a diminishing schedule until the infants were 24 months of age. The intervention focused on parent education, enhancement of the women’s informal support systems, and linkage with community services.

Women in both groups receiving home visits from nurses had many positive behavioral outcomes compared with groups that received screening only or screening plus transportation. Among the women at highest risk for caregiver dysfunction, those who were visited by a nurse had fewer instances of verified child abuse and neglect during the first 2 years of their children’s lives. They were observed in their homes to restrict and punish their children less frequently, and they provided more appropriate play materials. There were no differences between groups in the rates of new cases of child abuse and neglect or in the children’s intellectual functioning in the period when the children were 25 to 48 months of age. However, nurse-visited children had fewer behavioral and parental coping problems (as noted in the physician record). Nurse-visited mothers were observed to be more involved with their children than were mothers in the comparison groups.

A cost-benefit analysis estimated program costs (direct costs of nurse visitation, costs of services to which nurses linked families, and costs of transportation) and benefits (cost outcomes presumed to be affected by the program through improved maternal and child functioning, such as less use of Aid to Families With Dependent Children, Medicaid, food stamps, child protective services, and greater tax revenues generated by women’s working). Taking a time point of 2 years after the program ended, the net cost of the program for the sample as a whole was $1,582 per family, but for low-income families, the cost of the program was recovered with a dividend of $180 per family.

Fifteen years after the birth of the index child (13 years after termination of the intervention), women who were visited by nurses during pregnancy and infancy had significantly fewer subsequent pregnancies, less use of welfare, fewer verified reports of abuse and neglect, fewer behavioral impairments due to use of alcohol and other drugs, and fewer arrests. Their children, now adolescents, reported fewer instances of running away, fewer arrests, fewer convictions and violations of probation, fewer lifetime sex partners, fewer cigarettes smoked per day, and fewer days having consumed alcohol in the last 6 months. The parents of these adolescents reported that their children had fewer behavioral problems related to use of alcohol and other drugs.

Primary Mental Health Project
The Primary Mental Health Project (PMHP) is a 42-year-old program for early detection and prevention of young children’s school adjustment problems. PMHP currently operates in approximately 2,000 schools in 700 school districts nationally and internationally. Seven states in the United States are implementing the program systematically, based on authorizing legislation and state appropriations.

PMHP has four key elements: (1) a focus on primary grade children; (2) systematic use of brief objective screening measures for early identification of children in need; (3) use of carefully selected, trained, closely supervised nonprofessionals (called child associates) to establish a caring and trusting relationship with children; and (4) a changing role for the school professionals that features selection, training, and supervision of child associates, early systematic screening, and functioning as program coordinator, liaison, and consultant to parents, teachers and other school personnel.

The PMHP model has been applied flexibly to diverse ethnic and sociodemographic groups in settings where help is most needed. Over 30 program evaluation studies, including several at the state level, underscore the program’s efficacy (Cowen et al., 1996). Significant improvements were detected in children’s grades, achievement test scores, and adjustment ratings by teachers and child associates. PMHP represents a successful mental health intervention that does not require highly trained and skilled mental health professionals.

Other Prevention Programs and Strategies
These and other prevention trials demonstrate that positive adaptation and social-emotional well-being in children and youth can be enhanced, and that risk factors for behavioral and emotional disorders can be reduced, by intervening in home, school, day care, and other settings. Programs have focused not only on mental health but also on other problem behaviors. (Botvin et al., 1995; St. Lawrence et al., 1995;Kellam & Anthony, 1998).

Other prevention trials are showing similar benefits. For example, a large-scale, four-site school- and home-based prevention trial, known as FastTrack, has shown clear benefits in reducing behavior problems among high-risk children, as well as in reducing needs for and use of special education, which has substantial cost-effectiveness implications (Conduct Problems Prevention Research Group, 1999a, 1999b). Another trial is now under way to test the efficacy of a preventive intervention provided to adolescents whose parents are currently being treated for depression within a health maintenance organization (Clark et al., 1998). Treatment of mood disorders also has potential effectiveness for the primary prevention of suicide, as explained in the later section on Depression and Suicide in Children and Adolescents.

² Also see IHDP, 1990; Ramey et al., 1992; Brooks-Gunn et al., 1994a, 1994b; Casey et al., 1994.

³ Also see Olds et al., 1986a, 1986b, 1988, 1993, 1994a, 1994b, 1995, and 1997.