Challenges in the Assessment of Maltreatment

One such problem involves the diagnosis of maltreatment, noted Leventhal. Diagnoses of maltreatment are based on a combination of clinical features rather than a single diagnostic test. Furthermore, these diagnoses have major implications for children and families related to safety, placement, and possible termination of parental rights. Pediatricians continue to struggle with this diagnosis, said Leventhal. “Many of us know physicians who have made the wrong diagnosis and have sent abused children home, and sometimes those children come back with more serious injuries due to abuse or even die from an abusive injury. We take these problems very seriously.”

An additional problem cited by Leventhal is that some so-called experts in court continue to deny that abuse has occurred and propose specious theories of causation, such as vitamin deficiencies or reactions to vaccines.

In contrast to the 750 children described in the 1962 article, Leventhal and his colleagues have estimated that in the United States about 4,500 children annually enter the hospital with serious injuries due to abuse (Leventhal et al., 2012). The majority of these children are younger than age 3, and most of those are less than a year old. The mortality rate for these children is very high, at around 6 percent in the hospital.

Changes in the Assessment of Maltreatment

Partly in response to the problems he identified, the assessment of maltreatment has undergone major changes over the past two decades, Leventhal observed.

First, evaluations have become more collaborative and multidisciplinary. Once done largely in a hospital setting or by child protective services (CPS) or the police, assessments are increasingly done in child advocacy centers (CACs) or by multidisciplinary teams (MDTs). In Connecticut, for example, MDTs can include prosecutors, CPS workers, physicians, forensic interviewers, social workers, mental health treatment staff, and school social workers. In Florida, pediatricians are closely involved in CPS work and provide advice to agencies about the kinds of investigations, medical workups, or other assessments needed.

Second, hospitals have emphasized child protection teams, with support from the National Association for Children’s Hospitals and Related Institutions, which recently published standards of excellence for such teams (NACHRI, 2011).

A new focus on sentinel injuries has directed attention to less serious injuries that often occur before serious injuries. In Connecticut, for example, every child reported to CPS who is less than a year old and has a physical injury triggers a consultation with a pediatrician to decide whether that child needs a more substantial investigation.

Finally, in 2009, 191 pediatricians were certified in the new specialty of child abuse pediatrics, which has changed perspectives in departments of pediatrics, according to Leventhal.

Research Advances Relevant to the Assessment of Child Maltreatment

Leventhal described several research advances that have furthered the assessment of child maltreatment. For example, much research has focused on defining the disease (and nondisease) and generating strong evidence about the range and specificity of clinical findings due to physical abuse. Some research has examined the process of evaluation, such as the use of skeletal surveys as a diagnostic test or the evaluation of the siblings of abused children. However, little research has examined decision making by clinicians, such as their biases and reporting patterns.

As an example, Leventhal described research on bruises as an indicator of abuse in young children. Children can be bruised when they start “cruising,” or pulling themselves upright and walking from object to object, at about the age of 9 months. “These studies have suggested that children who cruise can bruise, but children who are not cruising are less likely to have bruises as part of normal activities,” said Leventhal. Children with unexplained bruises at less than 9 months of age need careful evaluations to determine whether abuse has occurred to the child. Other studies have examined children less than age 48 months and have found different distributions of bruising in the abuse group versus the accident group. This research has led to a mnemonic called TEN-4, where “TEN” stands for the location of bruises that are worrisome for abuse—torso, ears, and neck, and “4” stands for children who are less than 4 years of age or any bruise in a child less than 4 months old (Pierce et al., 2010). “TEN-4 is a great way to teach about this problem,” Leventhal said.

He also described a study of 434 primary care clinicians who collected data on more than 15,000 child injury visits in two national practicebased research networks (Jones et al., 2008). More than 1,600 of these children had a “suspicious” injury, but only 95, or 6 percent, were reported to CPS, and 27 percent of “likely” or “very likely” abuse cases were not reported to CPS. Reasons given for not reporting the children included familiarity with the family (“if we like them we don’t report them”), aspects of the case history (“I kind of believe what the mother said”), the use of available resources (“I’ll handle that on my own”), and negative views of CPS, which is an attitude that Leventhal has heard from primary care colleagues.

Leventhal briefly described research on the evaluation of suspected child sexual abuse. For example, the interview protocol developed by Lamb (the National Institute of Child Health and Human Development Interview Protocol) provides helpful ways of interviewing children and has been studied extensively, as have the various influences on children’s memories (Lamb et al., 2007). Again, decision making by physicians, CPS workers, and police has received less attention, and little research has been done on the value of the multidisciplinary approaches despite the proliferation of CACs, MDTs, and other collaborative efforts.

Finally, individual family variables have received a moderate amount of research. These variables include domestic violence, substance abuse, and the mental health of parents. One variable that has not received enough research, said Leventhal, is the abuse inflicted by males. “A lot of the serious abuse that we see in hospitalized patients comes from men—either fathers, stepfathers, or boyfriends. How to reach that part of the society to prevent some of these serious injuries is an important challenge.” In addition, less is known about combinations of factors, how individual parents respond to a child’s behaviors such as crying, and how to ameliorate the risk of maltreatment if it is elevated.

Future Research on Assessment and Diagnosis

Leventhal concluded by listing several suggestions for future research. Systems of evaluation and care need additional study, including the linkages between child abuse pediatricians and CPS. The role of CACs and MDTs also needs more study, especially because these evaluation systems are so tightly linked to treatment.

Leventhal mentioned the need to fund fellowships in child abuse pediatrics. “We need to figure out ways to train these physicians and enhance their research expertise.”

Finally, research should examine how to improve the decision making of primary care clinicians, emergency room physicians, and child abuse pediatricians, Leventhal said. How can physicians be trained not to overreport or underreport injuries that are reasonably suspicious? The process of evaluation also should be a subject of research. Which children need which diagnostic tests? This question should be studied in multiple sites to yield widely applicable findings, Leventhal said. Later in the workshop, Charles Sabel, the Maurice T. Moore Professor of Law and Social Science at Columbia Law School, discussed the importance of decision making by other frontline workers such as case workers, teachers, and police officers. He said that more research is needed to study innovative systems-level changes that may address challenges associated with decision making by frontline workers in the current system. His presentation is summarized in Chapter 8.

Discussion

During the discussion session, Frank Putnam from Cincinnati Children’s Hospital Medical Center and the University of North Carolina School of Medicine observed that hundreds of thousands or possibly millions of videotaped interviews done for maltreatment assessments exist, but there are no guidelines on how those interviews can be used. They could be a valuable research resource, but they are sensitive tapes that need to remain confidential. “How long do we maintain them? Who has access to them?”

Also, Joy Osofsky of the Louisiana State University Health Sciences Center recommended doing research in cooperation with CACs, which Leventhal labeled an excellent idea. For one thing, he noted, such research could make IRBs more accepting of CAC procedures.

The Nature of Assessments

One of the first questions that needs to be answered, said Saunders, involves the type of assessment. Should it be shorter, easier, and free, which would be easier for the existing child welfare workforce to handle? Or should it be more comprehensive to obtain all of the information that might be needed? Such assessments would require professional skills “that are probably far beyond the child welfare system workforce and therefore would require substantial coordination with community resources,” said Saunders. This is a question that applies not just in assessments, but throughout human services, he added.

A fair number of evaluation tools have been developed and tested. For example, Johnson et al. (2008) reviewed 85 instruments covering patterns of social interaction, parenting practices, parent or caregiver histories, and problems accessing basic necessities, and found 21 to be sound. “Automated” assessment, interpretation, and service planning frameworks have been developed, along with multidisciplinary approaches. “We have a lot of psychometrically sound, useful measures of the common problems exhibited by children in the child welfare system,” Saunders said. “Doing mental health types of assessments for the purposes of service planning is a well-accepted operation within the child welfare system.”

Assessment frameworks point to multiple sources of information, including the child, siblings, parents, other caregivers, teachers, family members, and peers. The “gatherers” of information include departments of social services, guardians ad litem, forensic interviewers, medical providers, mental health providers, victim advocates, law enforcement, and school personnel. Methods for gathering information include open interviews, structure interviews, standardized assessment instruments, observations, and other forms of interaction. Targets for information gathering include family and social history, abuse history, other trauma history, anxiety, depression, behavior problems, delinquency, substance abuse, academic performance, social functioning and support, and family functioning and support.

Outcomes of Assessments

Despite the attention devoted to assessments, not much research has examined whether they improve outcomes for children, particularly mental health outcomes. For example, few studies have assessed whether purported best practices are being followed and, if they are, whether they lead to better outcomes. “We can ask administrators and mental health systems how many people their centers saw in the past year, and they can tell you with extraordinary precision how many new patients they had,” said Saunders. “They can tell you with precision how many units of service they delivered. They can tell you how much they can bill. [But] very few people can tell you how many people actually got better.”

The research that has been done points to the difficulty of implementation, Saunders said. “We know a lot,” he said. “We have some tools to do some very good work. However, actually translating that into the daily behavior of the typical child welfare worker turns out to be extraordinarily difficult.”

Standardized systems that are highly dependent on worker compliance have been criticized as taking away from worker judgment. “Of course, that is exactly what they are intended to do,” said Saunders, but this outcome is a two-edged sword. “The computer does not always make great decisions.” On the contrary, research on the input to standardized systems suggests that they still require judgment and information input by the worker, though this input can vary from person to person based on experience.

Research also demonstrates that many assessment findings are not followed. People may do an assessment, but pay little or no attention to the results in a service plan. The child welfare workforce needs certain levels of knowledge and skill to use assessments effectively, but the question is whether the child welfare system can establish such prerequisites for the workforce. “Our history has not been all that great in this,” Saunders said.

Many stakeholders are involved in making a decision in the child welfare system. Some are doing assessments; others are doing service planning and delivery. Many times a judge is at the center of conflicting inputs. “When it is not coordinated, the judge hears about different treatment plans, some of which are completely contradictory, which then extends the life of the case and makes it more difficult for people to actually get treatment,” Saunders said (Figure 2). A community-based approach may be one way to provide greater coordination for these inputs.

FIGURE 2

Competing treatment plans in child maltreatment cases. NOTE: DJJ = Department of Juvenile Justice; DSS = Department of Social Services; GAL = guardian ad litem; IEP = individualized education program. SOURCE: Saunders, 2012.

Future Research on Assessment

Saunders concluded by listing a number of critical research questions, many of which involve implementation:

• Within the context of frontline child welfare practice, how well do current (and proposed) assessment tools and procedures identify children with particular problems who likely need mental health services?
• What are the major sources of error in child welfare assessment approaches?
• How should assessment approaches be adjusted due to factors such as culture, ethnicity, race, and gender to reduce disparities?
• What is the influence of worker background and experience on the implementation of assessment systems?
  •  What are the most cost-effective and efficient approaches (in terms of financial cost, worker and family time, training, supervision, and compliance effort) to effective assessment?
• What levels of assessment can be reasonably performed by typical child welfare workers, and what levels require additional community professional resources?
• What are the minimal knowledge and skills needed in the child welfare workforce to do the levels of assessment necessary for good practice?
• What sorts of initial and ongoing training, supervision, and monitoring of practice are needed to achieve and maintain effective assessment activity?
• To what degree can technology be used to make the assessment process (and application of assessment results) more efficient and more effective without negating appropriate child welfare worker judgment?
• Does greater coordination of assessment tasks with community resources and the family result in better assessment?

Discussion

During the discussion session, Clare Anderson from the ACYF pointed to research showing that increased mental health assessment can lead to the increased use of psychotropic medications among children in foster care. An important research question, she said, is whether the scalingup of evidence-based practices would affect the use of psychotropic medications in this population.

In responding to a comment about procedures for doing mental health assessments, Saunders argued for a combination of standardized tools and professional judgment. Such a balance would accommodate a realistic view of how well the workforce ever will be trained, he said.