May 1997

Clinician Agreement on Physical Findings in Child Sexual Abuse Cases

Author Affiliations

From the Departments of Pediatrics, Bowman Gray School of Medicine of Wake Forest University, Winston-Salem (Drs Sinai, Lawless, and Rainey); Area Health Education Center, Wake Medical Center, Raleigh (Dr Everett); University of North Carolina at Chapel Hill (Drs Everett and Runyan); and Duke University Medical Center, Durham (Drs Frothingham, Herman-Giddens, and St Claire), NC.

Arch Pediatr Adolesc Med. 1997;151(5):497-501. doi:10.1001/archpedi.1997.02170420067011

Objectives:  To measure agreement among experienced clinicians regarding the interpretation of physical findings in child sexual abuse cases and to determine whether knowledge of clinical history affects the interpretation of the physical findings.

Design:  Experienced clinicians rated colposcopic photographs on a scale of 1 to 5 with 1 being normal and 5 being clear evidence of penetrating injury. To answer an additional study question of whether clinical history affected interpretation, 4 clinicians rated 69 cases in which they were blinded to the patients' histories and 70 cases in which the patients' histories were available. The other 3 clinicians then rated the same cases with the presence or absence of history reversed.

Setting:  All clinicians involved perform child sexual abuse examinations at tertiary care centers.

Patients:  A total of 139 girls with Tanner stage 1 or 2 genitalia who were referred to a general pediatric clinic at an academic medical center for examination of possible sexual abuse.

Results:  Half of the photographs were interpreted as indicating little or no evidence of abuse. Of those photographic sets that both readers could interpret, 39% were in perfect agreement and 77% disagreed by 1 category or less. Perfect agreement across all possible pairs of readers was 34.5%. Agreement was better when the patient's clinical history was unknown (29.3% vs 38.9%, P=.005). The κ a measure of interrater reliability, indicated poor agreement among clinicians. The combined κ for the first group of clinicians was 0.22 without knowledge of clinical history and 0.11 with knowledge of clinical history. For the second group of clinicians, the κ was 0.31 without knowledge of clinical history and 0.15 with knowledge of clinical history. The overall κ across all 7 clinicians disregarding clinical history was 0.20. Agreement was best for categories 1 (normal, κ=0.28) and 5 (clear evidence of a penetrating injury, κ=0.39).

Conclusions:  Clinicians educated and experienced in assessing child sexual abuse do not agree perfectly on the interpretation of photographs of genital findings in girls with Tanner stage 1 or 2 genitalia. Clinicians agree less when a patient's clinical history is available. Efforts should be directed at standardizing physical findings and avoiding overemphasis on physical findings in child sexual abuse cases.Arch Pediatr Adolesc Med. 1997;151:497-501