database to shed additional light on the relation between gallbladder (GB) ejection fraction and GB pathology in children who had undergone hepatic iminodiacetic acid (HIDA) scans.In their cohort, GB ejection fraction did not correlate with histologic findings of GB pathology. For these authors, the lack of correlation increased their skepticism that HIDA scans are useful for diagnosing biliary dyskinesia in children and adolescents. Their data also support the idea that physicians depend on HIDA scans to make decisions about cholecystectomy.But it is crucial to consider their findings in light of the HIDA scans and pediatric patients included versus excluded from the analysis. Considering the impact of their inclusion and exclusion criteria raises questions about whether the authors' conclusions are externally valid.The analysis included a small subset of the HIDA scans done in children and adolescents during the study period-only those that led to cholecystectomy. It was remarkable that 70% of the subjects in this study actually had histologic cholecystitis according to their pathology reports. A smaller percentage had stones, but the authors do not note if this was in addition to or overlapping with those that had cholecystitis. If we think that chronic inflammation and/or stones contributes to these patients' symptoms, then this suggests that almost 3 of 4 of the cholecystectomies were done on ''appropriate'' patients with pathologic findings in their excised GBs.The histograms also suggest, as the authors note, that GB ejection fraction data were likely used to help decide about cholecystectomy. The figures show that approximately 50% of children with GB ejection fractions of 0% to 15% underwent cholecystectomy compared with approximately 10% in children with GB ejection fractions >50%. Not surprisingly, providers who ordered HIDA scans to evaluate for biliary issues were more likely to recommend surgery if the imaging showed ''abnormalities.'' It seems reasonable to assume that the gastroenterologists, surgeons, or others were combining information on symptoms and GB ejection fraction to recommend cholecystectomy. If this is true, then children with low GB ejection fraction were probably more likely to undergo cholecystectomy with less ''classic'' or severe symptoms of biliary dyskinesia. The low GB ejection fraction cholecystectomy group may then have had lower-than-expected prevalence of cholecystitis and other histologic findings.On the contrary, children with high GB ejection fraction that went to cholecystectomy probably had more classic or severe symptoms-thus leading them to surgery despite ''normal'' ejection fraction. The high GB ejection fraction group may then have a higher-than-expected prevalence of cholecystitis in their histologic specimens. If these assumptions hold, sampling bias would thus have decreased the observed prevalence of histologic findings in the low GB ejection fraction group and increased the observed prevalence of histologic findings in the high GB ejection fraction group-dim...