Sir.—I must question the conclusion reached by Sawada et al in their article "Mass Screening of Neuroblastoma in Infancy" (Journal 1982;136:710-712). Their data show that they screened more than 78,000 children and found four "unsuspected" cases of neuroblastoma. They failed to detect one case that was later found by physical examination. There were 48 infants with false-positive tests and 2,927 infants with possible false-positive tests that required retesting.
Analysis of their data demonstrates a sensitivity for the urine screen of 4/5, or 80%. Three of these five cases were readily detectable by physical examination. Thus, physical examination has a sensitivity of 60%. Sawada and colleagues' screening test was 99.94% specific. Assuming that their stated incidence of 1/20,000 is correct, I obtain a predictive value for a positive test of only 0.13%, improving to only 7.69% on retesting. It can be shown, therefore, that of every 744 initially positive urine