The effect of image quality on the assessment of child abuse photographs.

Objectives: Although child abuse pediatricians are frequently asked to evaluate risk of abuse based on photographs, the effect of photographic quality on this process is presently unknown. Photographs of abused children are often taken by professionals without photographic training, and quality varies widely. This article reports the first study of the effect of image quality on clinical assessment from photographs.

Methods: A total of 120 images depicting 60 cutaneous lesions were selected for the study. Paired images of single lesions varied in quality of focus, exposure, or framing. Seventy medical and nursing professionals were recruited from the Internet listservs focusing on child abuse. Subjects evaluated the images for quality (1-9 scale), opined if the image was "inadequate for interpretation," and answered a clinical question about the type of lesion displayed. Accuracy was defined as concordance between the subject and the live examiner's written documentation. Adequacy was defined as the proportion of subjects that did not indicate that the photograph was inadequate for interpretation.

Results: Mean accuracy among subjects was 64% and ranged from 35% to 84%. Accuracy was not predicted by subject profession, experience, or self-rated computer skill. Image quality and adequacy were independently associated with increased accuracy.

Conclusions: Higher-quality images improved accuracy. An examiner's impression that an image is adequate did not guarantee an accurate interpretation. Reliance on photographs alone is not sufficiently accurate in the assessment of cutaneous trauma.