Objectives: Pediatric gastrointestinal (GI) biopsies represent a disproportionately large fraction of the mucosal biopsies submitted for pathology evaluation relative to the fraction of pediatric endoscopies performed. We sought to explore the rationale for this large sample volume and identify its diagnostic value.
Patients and Methods: Conceptual value-of-information analysis of 100 sequential colonoscopies, including 770 total biopsy specimens. The diagnostic value of each biopsy was evaluated based on its order of appearance in the model and overall contribution to the pathological classification of the disease.
Results: Current practice standards for pediatric GI biopsies provide no guidance on sampling strategy for nonfocal biopsy of suspected inflammatory diseases, resulting in the collection of a nonstandardized number of individual samples and sampling sites that may vary significantly by clinician and institution. We find that this practice adds little to no diagnostic value over more cost-effective protocols, such as a fewer biopsies overall, or fewer specimens with pooled regional biopsies.
Conclusions: In the absence of clear guidance, pediatric GI clinicians and institutions should explore alternative strategies such as the left-right pooled biopsy protocol (common in adult GI practice), or preferably, a more conservative 4-region protocol covering major anatomical landmarks in the colon for nonfocal pediatric colonic biopsies. This will allow individual practitioners and clinical centers to test the hypothesis, supported by our analysis, that such a reduction in resource utilization will have no measurable impact on the quality of care.
*James Homer Wright Pathology Laboratories, Massachusetts General Hospital, Boston
†Kid Risk, Inc, Newton, MA.
Address correspondence and reprint requests to Kamran Badizadegan, MD, Massachusetts General Hospital, 55 Fruit St, WRN219, Boston, MA 02114 (e-mail: email@example.com).
Received 14 February, 2011
Accepted 8 June, 2011
The authors report no conflicts of interest.