The simplification of a clinical data base: application to gallstone patients

In patients with gallstones, recommendations for cholecystectomy depend on evidence of abnormality in the cholecystogram. Symptoms and signs, usually related as a syndrome, are employed to identify patients who should have a cholecystogram. Triplet combinations of clinical data from gallstone patients and controls were examined to find which combination was the best discriminator between diseased and non-diseased, and between those who should be investigated radiologically and those who need not. The cholecystogram was confirmed as the best overall discriminator, but right costal margin pain by itself was as efficient as any triplet. The performance of a cholecystogram on all patients with right costal margin pain is as efficient a system for identifying patients with gallstones as the consideration of any quantity of clinical data.