[Bias in time delay in ICU admission as a mortality risk factor or "lead time bias"]

To assess if delay in admission to the Intensive Care Unit (ICU), measured according to the prognostic estimation of survival in critical patients (EPEC) system, influences the final outcome of patients admitted to our ICU.

Retrospective and systematic analysis of data collected during six months in 2003.

Nineteen-bed ICU (15 from Standard intensive care and 4 from intermediate care) in a referral teaching hospital.

Four hundred and eighty one patients consecutively admitted to our ICU and followedup to hospital discharge. MAIN INTEREST VARIABLES: Risk of death was estimated with the EPEC, SAPS II and MPM II 0. Variables collected were gender, age, origin of admission, risk of death by means of the 3 methods mentioned, admission time delay (lead time bias) as measured by EPEC and life status on ICU and hospital discharge (alive or dead).

A total of 44 out of 481 patients died during the hospital stay, overall admission delay being 0.7 +/- 1.98 hours (2.96 +/- 3.28, range 0.25-20 hours, for those with delay > 0). No differences were found when comparing delay in admission among those surviving and the deceased, and there was very bad correlation between the prognosis made considering delay time for admission and that established without considering it (SAPS II or MPM II 0).

Our study does not make it possible to relate lead time bias with patient survival. Due to the EPEC design, it is possible to differentiate "physiopathological delay" (inappropriate detection of the critical situation) and "logistic delay" (conditioned by factors such as lack of available beds). Our study as well as the EPEC only considers the latter. It cannot be ruled out that the increase in mortality regarding prognosis is directly related with first type of delay and not with the overall lead time bias.