The role of ultrasonography in the diagnosis of pyloric stenosis: a decision analysis.

BACKGROUND/PURPOSE: The appropriate role for ultrasonography (US) as a replacement for the upper gastrointestinal series (UGI) in vomiting infants remains undefined. The authors have used decision analysis techniques to determine whether the use of ultrasonography as an initial screen in vomiting infants is cost effective when compared with the UGI as the only study.

METHODS

Two diagnostic strategies were compared: 1) UGI alone and 2) ultrasonography followed by an UGI series in 50% of cases when ultrasonography scan was negative for pyloric stenosis. The test sensitivity (US, 0.9; UGI, 1.0) and test specificity (US, 1.0; UGI, 1.0) and the incidence of pyloric stenosis among vomiting infants presenting to the community pediatrician (0.30) or after a negative examination by an experienced examiner (0.02 to 0.18) were obtained from a review of the literature. The relative charges for ultrasonography and UGI were obtained from a national survey from which the cost ratio of US to UGI was estimated to range from 0.67 to 1.81 with a median of 1.06.

RESULTS

Under these baseline assumptions, UGI only was the preferred strategy. The results of the decision analysis were sensitive to, or dependent on, assumptions made regarding the incidence of pyloric stenosis, the US to UGI cost ratio, the sensitivity of the US, and the proportion of patients that proceed to UGI when the US scan was negative for pyloric stenosis. When at least 50% of patients whose US scan was negative for pyloric stenosis proceeded to a UGI, UGI remained the preferred strategy for all cost ratios examined (0.6 to 1.7). Even when no patients proceeded to UGI, the cost ratio of US to UGI had to be less than 0.7 under the typical incidence (0.30) of pyloric stenosis among vomiting infants presenting to the community pediatrician for US to be cost effective. Finally, only UGI was indicated when an olive was not appreciated by an experienced examiner.

CONCLUSION

Under assumptions that fit most clinically relevant circumstances, the UGI as the initial study is the most cost-effective radiological diagnostic test in the evaluation of the vomiting infant.