In Reply.—Our study was conducted, not with a preexisting bias against data management, but with such high hopes that we filed almost 28 000 facts and each year calculated 552 complication rates for 52 classes of procedures. We did not report these details because in 14 years only one difference was significant, and no calculation revealed a problem that we had not already recognized and remedied by a prompt, incidental approach to every complication.
Between statistical and incidental approaches to quality there should be no conflict. The facts in our computer file and in our morbidity discussions are identical. Neither are underreported or subjective. Their differences are in what they achieve. The continuous incidental approach easily and economically optimizes outcomes by minimizing errors. The statistical approach tediously and retrospectively validates claims of acceptable outcomes. The rare need to validate claims justifies continuous filing of data but, except for infection