Your independent source for Harvard news since 1898 | SUBSCRIBE

Your independent source for Harvard news since 1898

Helping Those Most in Need

March-April 2009

Theresa Betancourt with children in Sierra Leone

Theresa Betancourt with children in Sierra Leone

Photograph by Julia Rubin-Smith

Civil war ravaged Sierra Leone for more than a decade, destroying infrastructure and crippling the economy. A third of the country’s six million people were displaced; tens of thousands died.

As the country imploded, the lives of one particular group of children descended into hell on earth. During the course of the war, an estimated 7,000 to 10,000 children were forced into military service. Besides being forced to kill, these child soldiers were often forced to take drugs to reduce inhibitions against committing violent acts. Many of both genders were raped.

These former child soldiers in Sierra Leone constitute just one of the groups of disadvantaged children that projects supported by Harvard’s interdisciplinary Center on the Developing Child (CDC) are trying to help. Across the United States and around the world, Harvard scholars are working to find ways to buffer vulnerable children from the dire circumstances that surround them. Below, we offer snapshots of some of those projects.

 

Theresa Betancourt, then a doctoral student at the Harvard School of Public Health (HSPH), began tracking the Sierra Leonean child soldiers in 2002, as the war was ending and they were beginning to return home. Humanitarian-aid groups assisted with the reintegration process, which was far from easy. Some of the former child soldiers returned to find their families dead. Meanwhile, neighbors treated them with scorn, with blame for what they’d done during the war, or with mistrust, fearing that violent behavior would become a habit and the children would pose a threat to the community’s safety.

Studying a sample that includes children who went through formal reintegration programs; former soldiers who went home and “self-reintegrated” without such a program; and a control group of children who had not been forced to fight, Betancourt is measuring how mentally healthy or ill her subjects are, several years after the war’s end. In “the first longitudinal study of boys and girls affected in this way by war,” she is also looking at their subsequent experiences with employment, income, marriage, domestic violence, and “health risk behaviors” such as unprotected sex. And she is asking what factors explain these outcomes: besides considering whether they served as soldiers and went through reintegration programs, Betancourt has examined her subjects’ social-support networks; how much stigma they describe experiencing; how close-knit their home communities are; their family settings; whether they were raped during the war; and, in the case of former soldiers, to what degree their current social circles include other former child soldiers. She is trying to determine which matters more: earlier intense stress (child soldier or not) or more recent circumstances (experiencing stigma, having a family with enough money to afford secondary school, etc.)

Now an assistant professor of child health and human rights at HSPH, Betancourt hopes that her final analysis will provide not only clues for helping Sierra Leone heal, but a blueprint for picking up the pieces in other countries torn apart by violence: “We’re concerned with how we can contribute to an evidence base that can be used for advocacy and policy.”

The CDC funded analysis of the Sierra Leone data by doctoral student Ivelina Borisova, who is writing her dissertation about the effects of family settings on child soldiers’ reintegration and success in adolescence and adulthood. Borisova is pursuing her degree at the Harvard Graduate School of Education (HGSE), but began working with Betancourt after hearing about the latter’s work through her own interest in international human-rights issues. Their collaboration epitomizes the way CDC projects leap across traditional disciplinary lines and ignore the boundaries of Harvard faculties.

Photograph by Gabriel Amadeus Cooney

William Beardslee

In another project, Betancourt and William Beardslee, Gardner-Monks professor of child psychiatry at Harvard Medical School (HMS), are embarking on a joint endeavor in Rwanda to develop an approach for protecting families from ill effects when one member is HIV-positive. Treating the infection can prolong life and sometimes prevent the development of full-blown AIDS, but antiretroviral drugs alone don’t address emotional repercussions, says Betancourt.

Beardslee specializes in insulating families against the effects of parental depression; his book When a Parent Is Depressed outlines a plan—to be implemented by a mental-health professional or directly by a family with a depressed parent—for dealing with the illness. Among the plan’s components: children learn that depression is a biological illness that a parent can’t just snap out of with a bit of willpower, and that their parent’s depression is not their fault. Parents learn that they can be good parents in spite of their illness, and that specific actions—such as taking an interest in their children’s academic, extracurricular, and social activities—can lessen the likelihood that their illness will negatively affect their children. Simply starting the conversation is most important of all, says Beardslee: “It’s the demystification of the illness—making it discussable and getting treatment—that makes a difference.”

He has tested his approach, with favorable results, in multiple U.S. groups, as well as in Finland, Sweden, Norway, and Holland. And in the United States—where a staggering 48 percent of mothers in the Early Head Start program (for children age 3 or younger) are clinically depressed—the federal Head Start and Early Head Start programs commissioned Beardslee and Catherine Ayoub, an associate professor of psychology at HMS, in 2003 to incorporate the approach into program materials and training sessions.

 


 

Betancourt and Beardslee, who serve together on a CDC workgroup on global child mental health, are weighing how those strategies might translate to protecting families against HIV’s secondary effects, such as family discord and community stigmatization. They traveled to Rwanda in January to interview HIV-affected families and consider what makes for good parenting and resilience in the face of this disease.

 

At hsph, in the adjacent building to Betancourt’s office, assistant professor of society, human development, and health Karestan Koenen studies posttraumatic stress disorder (PTSD): specifically, analyzing whether genetic factors help to explain which hurricane victims develop PTSD and which do not.

Courtesy of Karestan Koenen

Karestan Koenen

Koenen’s subjects are survivors of the hurricanes that hit Florida in 2004—a series of four storms that, she says, caused “billions of dollars of damage,” but which “have been forgotten because Katrina came the next year and was so huge.”

The genetic polymorphisms she is examining include the serotonin transporter gene (serotonin is a neurotransmitter involved in mood). Studies have found that individuals with two short versions of the gene are much more susceptible to depression than those with two long versions. (People with one short and one long version fall somewhere in between.) “The question I’m interested in,” Koenen says, “is what makes some people vulnerable and some people resilient when bad things happen to them.”

Source: Caspi et al. “Influence of Life Stress on Depression: Moderation by a Polymorphism in the 5-HTT Gene.” Science 301:386-389 (2003). Reprinted with permission from AAAS.

Abused children face a higher-than-average risk of becoming clinically depressed—but only if their genetic makeup confers vulnerability, research indicates. This study examined the incidence of depression in children who had been abused and those who had not, sorting by serotonin transporter type. Children with two short versions of this gene—which affects the way the brain processes the neurotransmitter serotonin—had the highest risk of depression (top line). Two long alleles appears to be the most protective combination: for children with that genotype (bottom line), the risk of depression was essentially no greater if they’d been abused than if they had not.

Another line of Koenen’s research, involving children who sustain serious injuries, approaches this question from a different angle. In a sample of children admitted to Boston Medical Center with injuries (the majority of which resulted from car accidents and violent crime such as shootings and stabbings), she investigated what factors seemed to explain why some children developed PTSD while others with equally severe injuries did not. After examining parental mental health and other characteristics of the family environment, she concluded that the factors most predictive of a child developing PTSD were the level of stress the primary caregiver and other family members were under at the time of the injury and during healing; how much pain the child experienced; and the child’s age at the time of injury.

Treating children with PTSD, and preventing the disorder’s development, is a valid pursuit in and of itself—but there are secondary benefits as well, says Koenen. “We know that mental-health problems predict a lot of physical health problems later,” she says. “Yet mental health doesn’t receive the same attention as physical health. By treating mental-health problems early, you may prevent physical problems down the road.”

Koenen, a licensed clinical psychologist who previously worked for the Veterans’ Administration hospital in Boston and in a rape crisis center, first became interested in studying trauma and resilience as a Peace Corps volunteer in Niger. “I was struck by the conditions under which people lived—not just poverty, but also violence—and how they were able to lead happy lives under these conditions,” she says. “That really got me interested in what enables people to cope with bad situations.”

In a study supported by the CDC, she is studying PTSD’s physiological underpinnings: are children who already have stress etched into their bodies at the time they are injured more likely to develop the disorder? Do children who develop psychological problems show more physiological signs of stress than children who suffer a serious injury but are more resilient psychologically?

 


 

Colleagues in that study are examining the same biomarkers in different human samples—and in animals. (See main article for more on this study.) Animal studies typically carry the disclaimer that it’s unclear whether their results apply to humans, but getting this group of researchers together to ask questions simultaneously in human and animal models begins to bridge that gap. “It’s challenging,” says Koenen, “because we all come from different fields and talk different languages. Our projects have different timelines. But the center helps us work that out, in a way that in everyday academic life is difficult to find.”

 

 

Jon Chase/Harvard News Office

Matthew Nock

Loeb associate professor of the social sciences Matthew Nock studies a particularly troubled group of children: those who attempt suicide or harbor suicidal thoughts. Although little is known about what causes such thoughts or actions to develop, Nock has developed an innovative tool to try to identify at-risk children so they can be targeted for help.

Nock (another member of the CDC child mental-health workgroup) is investigating whether a new version of the Implicit Association Test developed by Cabot professor of social ethics Mahzarin Banaji might be used to assess suicide risk. The computerized test presents a series of images and asks subjects to label each one by choosing between two words presented at the same time. For example, to assess whether a subject has racist tendencies, the test presents a series of photographs of white and nonwhite faces, with one positive and one negative adjective alongside each photograph. In one round of testing, subjects are asked to associate the positive words with the white faces and the negative words with the black faces; in the other round, they are asked to do the opposite. The test measures how quickly (in fractions of a second) subjects press the computer keyboard to make the association; more rapid association of white faces with positive words and nonwhite faces with negative words is taken as evidence of prejudice. (Banaji has tweaked the test to correct for any bias resulting from the order in which the tasks are presented. Read more about her, and the test, here.)

Banaji’s team of researchers has developed versions of the test that assess prejudice against overweight people; people with disabilities; Arabs and Muslims; and the elderly, and that examine whether subjects more readily associate men than women with science careers. Nock’s version attempts to discern the presence in people’s minds of thoughts about suicide and death.

Courtesy of Matthew Nock

Above, a frame from the Implicit Association Test for suicidal thoughts. In one round of testing, the person taking the test would be instructed to identify images of suicide and death as "me"; in a separate round, the same test-taker would receive instructions to identify images of suicide and death as "not me." If the test is effective, people who actually harbor suicidal thoughts would more rapidly make the association between themselves and the suicidal images (by pressing a computer key) than people who are not suicidal. Nock is currently investigating whether his test reliably produces these results.

Testing it out in a group of children who had engaged in self-injurious behavior such as cutting themselves, Nock found “very large differences” in their performance on the test, compared to children with no history of self-injurious behavior. Over time, he found that the children with the most extreme scores were most likely to make a suicide attempt—but he cautions that the sample was not large enough to generalize that conclusion.

He is now trying the test in clinical settings—for example, with adults admitted to the psychiatric emergency department at Massachusetts General Hospital. Nock has already found significant differences in test performance between patients who have just made a suicide attempt and those who have not; tracking these subjects over time, he has also found that patients who make a second suicide attempt after leaving the hospital score differently on the test from those who don’t.

In a separate study funded by the CDC, he will test his tool in a larger sample of adolescents admitted to Children’s Hospital Boston. If the test proves reliable, Nock predicts it could be used to improve the accuracy of hospital admission and discharge decisions. “Decisions about how people get admitted to a hospital vary tremendously across clinicians, across sites,” he says, “and we’re bad at deciding when people are ready to leave. One of the highest-risk times for suicide death is within two weeks after people are discharged from the hospital. We may be able to use this test to improve our decisionmaking at these two points.”

 


 

In the future, the test could even be used for screening, to help teachers and school nurses differentiate between children contemplating suicide—who need help urgently—and those with depression or another mental illness, who still need help, but are not suicidal. “If we use mental disorders as a criterion, we can identify 90 percent of the kids who are going to make a suicide attempt,” he says, but the group of all children with mental disorders will be much larger. “How do you find the suicidal kids within that group? It’s like finding a needle in a haystack.”

Nock has also designed an experiment to address the concern—understandable, he notes—that administering a test with pictures and words related to death and suicide might cause children to consider suicide when they never had before. Based on the results, he is satisfied that the test doesn’t work this way. (Read more about Nock’s test in this article from the Boston Globe.)

 

 

Nock also studies children who commit self-injury—“direct deliberate tissue destruction where there is no intent to die.” He explains, “Kids engage in this behavior for other results, usually emotional regulation—kids who feel really upset or distressed may cut or burn themselves, and it decreases their negative emotions or stress. Or they do it to communicate to other people how distressed they’re feeling—kids often say, ‘When words don’t work, I use cutting or burning.’ ”

This behavior is more common among adolescents than among adults or younger children, he says: about 15 percent of high-school students, and a slightly higher percentage of college students, report having engaged in self-injury in the past year; among older adults, only 4 percent report ever having done it.

Meanwhile, between 2 and 5 percent of people report making a suicide attempt at some point during their lives, and between 10 and 15 percent of people report having had suicidal thoughts. Although not everyone who commits self-injury attempts suicide, the two behaviors are correlated: “In our work,” says Nock, “about 70 percent of kids who engage in self-injury have also made a suicide attempt.”

Nock and fellow Loeb associate professor of the social sciences Wendy Mendes are trying to find a physiological basis for the psychological relief that self-injurers describe. In one study, adolescents who had engaged in self-injury showed significantly higher skin conductance (a physical indicator of feeling stressed and agitated) in response to a frustrating situation than those who had not.

These findings are enlightening, but Nock and Mendes recognize their limits. For one thing, the study induced feelings of frustration by using a card game. In this common psychology-experiment scenario, subjects are told the experimenter wants them to match cards according to either color, number, or suit, and they must guess which dimension the experimenter desires through trial and error. As the researcher reveals whether the previous move was correct or incorrect, subjects quickly detect the pattern—but then the researcher mixes things up and begins telling subjects their responses were incorrect, even though they are following the same pattern deemed correct earlier. The task reliably “causes great stress and frustration” in subjects, says Nock—but “few people experience a stressful card game in real life, and that’s not what leads to self injury.”

Nock and Mendes are trying to get closer to the stresses of everyday life by using a state-of-the-art tool. With the LifeShirt (developed by a California company called VivoMetrics) they can track what happens in a person’s body before, during, and after incidents of self-injury brought on by true-life events such as a stressful day at work or a fight with a family member. (The simple mesh vest—similar to those used to designate teams in gym class—records respiration, heart rate and electrical activity, and other physiological measures as the wearer goes about daily activities. Worn underneath clothing, it is relatively inconspicuous; the only noticeable piece is an electronic belt pack—which looks not unlike an enormous BlackBerry—to record the information.)

Subjects (who were recruited based on a history of self-injury) wear the vest for 48 hours, and keep a journal so Nock and Mendes can correlate self-described emotion with the biological measures. The subjects are instructed to hit a button on the belt pack each time they engage in a coping behavior—be it taking deep breaths, going for a walk, smoking a cigarette, or self-injury. To date, a few dozen adults have gone through the study (which continues).

Those who injure themselves or attempt suicide are inherently difficult to study: one can’t invite subjects into the lab for self-mutilation, and people who contemplate, or have just attempted, suicide are acutely distressed, so asking them to consent to a research study may not be foremost in their caregivers’ minds. But it is precisely because these people experience such distress, says Nock, that it’s so crucial to understand their mental-health problems and develop effective interventions.

 

 

There is a sadness to working with disadvantaged children, who experience adversity early in life through no fault of their own. But work with abused children, those harmed by their own parents, is perhaps the saddest of all.

Abuse etches itself into children’s brains, profoundly affecting the way they see the world. For example, one study compared four- and five-year-old abuse victims to peers who had not been abused by presenting photographs of faces to both groups and asking them to name the emotions expressed in three different sets of images: one spanned a continuum from happy to fearful, one from happy to sad, and one from happy to angry.

The two groups of children were equally good at telling when a face was happy or not—but the abused children interpreted all the unhappy faces as angry, with no differentiation among anger, fear, and sadness. In their world view, anger predominated.

Catherine Ayoub, a forensic psychologist who has appointments at Massachusetts General Hospital and the Brazelton Touchpoints Center at Children’s Hospital as well as HMS, evaluates cases of alleged child abuse and adversarial divorce proceedings for judicial systems across the United States and Central and South America. She has seen children profoundly and permanently damaged by abuse; her recent research suggests that this is the rule, not the exception, and that the damage begins at an earlier age than many people realize.

Although it is well known that abuse causes older children to develop a negative world view “in which they tend to see other people as threatening, and there’s a sense of ‘I may not live till tomorrow,’ ” says Ayoub, “there was literature that said that if you were really young, it was like water off a duck’s back—you could make up for it” and recover completely.

But in a study of abuse victims that she conducted with Bigelow professor of education Kurt Fischer, “we found that at age 4, the maltreated children were not able to perform, cognitively, like the non-maltreated children,” says Ayoub, an early-childhood specialist who is directly involved with the CDC through her work on Un Buen Comienzo (see main article).

 


 

She and Fischer also found that maltreated children develop a pessimistic focus—“They take the positive and turn it into negatives”—and found significant changes in the levels of the stress hormone cortisol in severely maltreated children. “In mildly and moderately maltreated children, we didn’t find that,” says Ayoub. Now they’re exploring how severe the maltreatment needs to be to cause modified cortisol levels: another step toward developing effective support and therapy for abused children through gaining precise knowledge of abuse’s effects.

 

 

Courtesy of John Weisz

John Weisz with clients of the Judge Baker Children's Center

In other areas, we know plenty about how to protect children, but do not act on that knowledge, says John Weisz, president and CEO of the Judge Baker Children’s Center and chair of the CDC child mental-health workgroup.

“If you’re a typical parent in a typical town in the United States and your child has a mental disorder, chances are the clinics in your area are not doing the best practices for treating that disorder,” says Weisz, who also holds appointments at HMS and in the Faculty of Arts and Sciences psychology department. “We’ve had a difficult time making those treatments broadly available.”

The center is therefore supporting a wide-ranging look at the systemic issues that prevent children with mental-health problems from getting optimal support.

Ideas about how healthcare should function often do not align with stark reality, in part because “most organizations that provide mental-health care are barely surviving now,” says Weisz. “They only survive by having every hour be a billable hour. There are no excess funds to support training in new and effective practices, and there’s nothing in the reimbursement system that rewards such practices. If you do an hour of therapy, you’re paid the same amount for a therapy that’s been tested and doesn’t work as for using one has been tested and shown to work well.”

A higher reimbursement rate, or some sort of accreditation based on learning and using effective therapy techniques, could produce “a big difference in how treatment is provided,” echoes Nock.

Courtesy of Richard Frank

Richard Frank

Through the CDC, Nock and Weisz have joined efforts with Morris professor of healthcare policy Richard Frank, another member of the child mental-health workgroup who specializes in analyzing healthcare systems, treatments, and policies with an eye toward maximizing cost-effectiveness. (A number of studies have shown that the jurisdictions that spend most on Medicaid—or on a specific type of Medicaid expenditures, such as treatment of heart disease—don’t have the best outcomes, and may have the worst.)

In a CDC-supported project, Frank is extending this type of analysis to child mental-health care, developing quality indicators he will then use to compare money spent with quality of treatment—and to identify the most cost-effective treatments.

At the same time, also under the CDC umbrella, Weisz is working on a meta-analysis of research on anxiety, depression, attention deficit-hyperactivity disorder, and conduct problems. “You put these four together,” he says, “and you’re talking about most of the reasons children are referred for treatment.” The results of this enormous undertaking will inform Frank’s analysis and recommendations. Ultimately, they hope it will influence the widespread practice of treating children’s mental-health problems.

Because individual children may have multiple problems, Weisz and his colleagues are also working to synthesize the most effective treatments for each of the multiple disorders. The goal is a modular approach, with flexible components that fit together in combinations guided by a flow chart, to tailor treatment to the diverse needs of individual children. “Most mental-health treatments for children these days are very narrowly focused,” says Weisz. “There’s one manual for treating depression. There’s another manual for treating anxiety disorders. There’s a separate manual for conduct problems. What we’re trying to do is consolidate and create an integrated approach.”

The modular model also takes into account the demands on providers. “You can’t expect people who do clinical work all day every day to master 30 or 40 different treatment manuals,” says Weisz. “We need one source that allows them to do good, effective treatment without having to suspend their careers for five years to go learn how.”

Concerns of practicality are all the more pressing in the developing world, says Frank, whose work also involves creating models for incorporating mental-health coverage into fledgling health-insurance systems.

“In the United States, we’re always saying, ‘Let’s do the best practice,’ because we have a different idea of budget constraints,” he says. “In a developing country, you have to say, ‘What is really going to work here?’ The practical best thing may be something completely different from the theoretical best thing.”

The researchers hope this three-pronged package—the meta-analysis, the cost-effectiveness study, and the modular treatment scheme—will dovetail into a framework for radically improving mental-health services for children in the United States and elsewhere. More generally, a circular approach in which science, policy, and practice all inform one another, to help each field decide which questions to ask next, constitutes the CDC’s truly unique contribution, says Nock. “We do research all day, and we may think about, but don’t always act on, the way our work can influence practice,” he says. “The long-term goal is to develop ways of breaking down these barriers, of having science influence practice, but also having practice influence science—getting a sense from practitioners of what they need. With the center, we have the resources to do that.”