Sample Review

What are we trying to measure? The problems of case ascertainment. Robert M Reece. Am J Prev Med 2008;34:S116-S119. (from Boston, MA)

Review from Winter 2009 issue

In this brief article, Dr. Reece pleads for widespread adoption of specific definitional criteria for abusive head trauma (AHT) in future research studies. In support of his final recommendations, he reviews eight published studies of AHT and applies the definitional criteria offered by three other clinician researchers.

In summary, Dr. Reece opines that “When extra-axial bleeding, cerebral edema, and typical retinal hemorrhages are present in an infant or young child with histories as listed below, the diagnosis of AHT is highly probable when one of the following is present: (1) no history of trauma; (2) a history inconsistent with the injuries; (3) a history that changes over time; (4) witnessed shaking and/or impact; (5) confession of shaking or impact; and (6) additional information supplied by the multidisciplinary child protection team.” Dr. Reece concludes his manuscript with a plea for researchers to apply these definitional criteria to future research studies.

Reviewed by Kent P. Hymel, M.D.

Reviewer’s Note:
I applaud Dr. Reece’s effort and I also sympathize with him. Many of us have been struggling with this specific challenge for several years. I remain absolutely convinced that further progress in differentiating between inflicted and accidental pediatric head trauma will be stymied unless and until researchers achieve some degree of definitional consensus. With all due respect to Dr. Reece, I will opine that his recommended criteria are a huge step in the right direction but, nevertheless, they fall short of the elusive goal. Here are a few of my thoughts.

When formulating or rendering a clinical or a forensic opinion about the origin of a child’s specific cranial injuries, I routinely opine that, compared to young victims of accidental head trauma, young victims of inflicted head trauma are more likely to: (1) demonstrate a constellation of specific injuries, such as characteristic retinal hemorrhages, biconvexity, subdural hematoma, and diffuse brain swelling; (2) have experienced a head injury event unlike a simple indoor fall; and (3) manifest evidence of greater injury severity, shown by a lower Glasgow Coma Scale score. These distinctions have been revealed and verified through years of our published clinical experience. I believe that making such distinctions is appropriate and relevant to the clinical or forensic goals of directing a diagnostic evaluation, protecting children from re-injury, and/or formulating a clinical prognosis. On the other hand, I do not believe that these distinctions are appropriate or relevant to our effort to achieve consensus regarding prospective research definitional criteria for AHT.

Which definitional criteria should be included in future research definitional criteria for AHT?  In recent years, a growing number of researchers have included “history of no trauma,” “a history that changes over time,” “witnessed abuse,” and “confession of abuse” in their definitional research criteria for AHT. I will opine that these definitional criteria — offered by Dr. Reece and others — are therefore acceptable to a growing number of researchers. I will also opine that, for many researchers, a “history clearly incompatible with the child’s developmental capabilities” and “evidence of non-cranial injuries highly suspicious for abuse” are also acceptable research criteria for AHT.

My primary concern with Dr. Reece’s recommendation is his inclusion of the criterion “history inconsistent with the injuries.”  I believe that inclusion of this specific definitional criterion in future research studies is inappropriate. It presumes that we already know which histories are “inconsistent” and which are “consistent.”  In my opinion, we need additional research to determine, with increasing precision, which histories are “consistent” or “inconsistent.” 

To accomplish this future research with objectivity, our future research definitional criteria for AHT must be free of circular logic and inherent biases. For research purposes, we cannot assume that any specific explanation for a child’s head injuries is “consistent” or “inconsistent” (e.g., a short indoor fall cannot explain subdural bleeding).  Also, for research purposes, we cannot assume that any specific traumatic cranial injury or constellation of injuries, such as characteristic retinal hemorrhages, biconvexity SDH, and/or diffuse brain swelling, confirms abuse. Finally, we cannot assume that greater injury severity confirms abuse. I recommend that our future, consensus, research definitional criteria for AHT should exclude any and all references to specific cranial injuries, to injury severity, and to specific histories, other than “history of no trauma.”

If I were the “Great and Powerful Oz” my summary statement would read something like this: “For research purposes, a child’s traumatic cranial injuries will be categorized as “inflicted” if and when: (1) there is a caregiver history of “no trauma” in the face of documented traumatic cranial injuries; (2) the caregiver’s account of the child’s head injury event changes significantly with repetition over time;  (3) the caregiver’s account of the child’s head injury event requires the child to have manifested developmental skills not previously demonstrated; (4) the caregiver’s abusive actions directed toward the head-injured child are witnessed independently; (5) the caregiver confesses to abusive actions that caused the child’s traumatic cranial injuries and clinical presentation; and/or (6) the head-injured child manifests other non-cranial injuries considered highly specific for abuse.