Documentation of Child Physical Abuse: How Far Have We Come?

DISCUSSION

Since Kempe's⁵ initial description of the battered child syndrome in 1962, the recognition and evaluation of child abuse have become necessary skills for medical personnel caring for children. Although pediatric and emergency medicine residency programs provide training in these skills through didactic lectures and workshops, an often overlooked but essential skill is the accurate and thorough recording of the medical evaluation. In 1980, Solomons'⁴ review of hospital charts of young children with traumatic injuries found that the majority of records contained inadequate data to determine retrospectively if the diagnosis of child abuse had ever been considered by the evaluator. He states that “it is possible that the appropriate questions were asked and satisfactory answers were given, but they were not recorded.” We sought to determine at our institution if the questions and answers were more completely recorded in 1995 than they were in 1980. Furthermore, are the physical examination, diagnostic procedures, assessment, and disposition better documented in 1995 compared with 1980?

To investigate these questions, we focused specifically on cases of suspected physical abuse evaluated in our emergency department. In circumstances when the diagnosis is child abuse, omissions and inaccuracies in the medical record are critical when legal decisions must be made. From our review, we conclude that most aspects of emergency department documentation in cases of suspected child physical abuse have changed little between 1980 and 1995. Although there have been interval improvements in the documentation of the plan and disposition, there continue to be omissions of important historical and physical examination items. First, a developmental history was missing from every record reviewed. Although not as crucial in older children, a developmental history in an infant or young toddler can provide valuable information. For example, knowledge of the child's gross motor abilities can help assess whether a history of climbing out of the crib is plausible.¹ Other historical items of information that help determine the reliability of a given history are the presence of witnesses to injury, the time elapsed between the injury and the emergency department visit, and a history of past injuries.⁶ All were inconsistently documented in both 1980 and 1995. Documentation of even negative information such as “there were no witnesses to the injury” or “there is no history of past injuries” is relevant and can be helpful in assessing the history provided.

There also seemed to be little improvement in the documentation of the physical examination between 1980 and 1995. Although an evaluation for abuse should include a complete physical examination, including examination of the genitalia,⁷ only half of children in each year had a documented genital examination. Furthermore, we were unable to show a significant increase in the number of medical records containing detailed descriptions and sketches of physical injuries. Photographs were uncommon in either year although four were documented in 1995. Without photographs or a clear diagram, a judge or jury may not be able to fully appreciate the extent of injuries despite a detailed description.^{1 8 9}

Radiologic skeletal surveys were obtained significantly less often in 1995 compared with 1980. Although the mean age of children with a documented skeletal survey did not differ significantly between the 2 years, a greater proportion of children in 1995 were <5 years of age. A probable explanation is the changing policy of obtaining skeletal radiographs in cases of suspected abuse. Obtained more often as part of the evaluation for abuse in 1980, skeletal radiographs are now obtained routinely only when evaluating the young (<2 years of age), preverbal child.¹⁰ Otherwise, radiographs are performed as part of the child abuse evaluation only when there is a specific indication such as a swollen, tender extremity.

Areas of documentation that improved between 1980 and 1995 were the diagnosis and disposition of the child and public agency involvement. Since the late 1980's, housestaff have been advised to replace the equivocal phrase “rule out child abuse” with more definitive statements of “probable/suspected physical abuse” or “history of abuse” followed by a descriptive summary of findings. This teaching is evident in the different discharge diagnoses written in 1980 compared with those in 1995. Although “history of abuse” by itself may be nebulous, it can be strengthened by an assessment of whether the physical findings are consistent or inconsistent with the history, and may be a reasonable statement if the physician is unsure of the diagnostic significance of the physical findings. The increased interactions among law enforcement, Child Protective Services, and the evaluating physician are also reflected in the documentation: every medical record in 1995 noted the involvement of police and social services compared with 77% of records in 1980.

Recent articles have suggested that a standardized clinical form may be helpful in improving documentation. In Bar-on and Zanga's³1996 study, the use of a structured clinical form significantly increased the information collected and documented in evaluations of both physically and sexually abused children. Furthermore, the authors noted an added benefit: residents felt more comfortable in the assessment of abused children and had less anxiety about potentially disregarding or not including important information. Our findings suggest that whereas a structured form greatly improves documentation of selected items such as illustrations of injuries, it does not guarantee complete documentation. These differences in our findings may reflect the differences in the structured forms that were evaluated or the hospital settings in which the forms were used.

One of the limitations of our study is the small number of patients, particularly in 1995. It is possible that we were unable to show a difference in documentation between the 2 years because of our sample size. Among the possible reasons for the decreased number of cases in 1995 is that children with suspected abuse were going to other facilities for evaluation. Primarily a referral center for sexual abuse, the hospital's Child Crisis Center began evaluating children with suspected physical abuse in the late 1980's. In addition, a member of our faculty with expertise in child abuse began evaluating referrals from law enforcement and community agencies in a medical group setting rather than in the emergency department. Another possibility for the decreased diagnoses of child abuse in 1995 is the increasing use of descriptive phrases such as “5-month-old infant with blunt head trauma and retinal hemorrhages” without an explicit statement of child abuse. Such medical records would not have been included in our study although these children would clearly have undergone an evaluation for abuse. We limited our study to children with the discharge diagnosis of abuse. Had we also studied the medical records of young children with blunt head trauma or long bone fractures, we might have observed a different pattern of documentation. However, one of our major objectives was to focus on documentation in those particular cases when the suspicion of child abuse was explicitly stated in the diagnosis and to observe how this index of suspicion was reflected in the medical record. As demonstrated by Solomons'⁴ study, such cases are difficult to determine retrospectively.

Another limitation of our study is our inability to determine at the individual resident level the extent of participation in the child abuse educational activities, leading to possible bias. For example, were the poorly documented records completed by residents who were or were not involved in child abuse training? Furthermore, we did not include level of training or specialty department (pediatrics, emergency medicine, or family medicine) in our analysis. Such variables may influence the completeness of documentation.

Despite these limitations, the deficits in our current emergency department medical records are evident. How can our increasing knowledge of child physical abuse be translated into a clearer and more complete medical record? First, education and training can focus on areas of the history and physical examination that are consistently poorly documented. For example, training in medical documentation can be incorporated into the already existing child abuse workshops, child abuse electives, and emergency department rotations. One project in Ohio using self-instructional programs to improve physician training in child abuse found increases in both physician knowledge and case recognition after use of the program.¹¹ Another educational intervention consisting of a month-long child maltreatment course taught to small groups of residents also found short-term increases in clinical knowledge; however, residents failed to retain this knowledge several months after course completion.¹² These findings suggest that any educational efforts must be ongoing. This can be especially challenging in the emergency department where there are physicians of many different backgrounds and levels of training, staff may be present on a temporary or rotating basis, and schedules often exclude housestaff on certain shifts from educational sessions.¹³ Innovative teaching strategies such as interactive computer programs or videos may be promising options to address these needs.

In addition, our findings suggest changes in emergency department procedures and paperwork that may further improve medical documentation. Small additions in the emergency department can lead to significant improvements in documentation. A simple checklist can remind medical staff about what should be documented in the medical record in cases of suspected abuse. In addition, an easily-completed structured form may also prove valuable. In the future, these forms will likely be computerized and accessible to various public agencies. A camera should be readily available in the emergency department and physicians should have a fundamental understanding of camera operation, picture composition, and medicolegal implications.¹⁴Lastly, teaching rounds in the emergency department might be used to discuss errors, omissions, or ambiguities identified in actual records.¹⁵ Record review by faculty members is an excellent opportunity to educate housestaff, and residents should be encouraged to discuss records with the attending physician.

Optimally, each emergency department that evaluates children with suspected abuse should have a special child abuse and neglect team available to perform examinations or consultations and arrange for follow-up. Although this is not practical in many institutions, each emergency department should be aware of referral centers where comprehensive, multidisciplinary evaluations for child abuse can be performed.