Volume 28, Issue 6 pp. 1028-1039
Tutorial in Biostatistics

Events per person-time (incidence rate): A misleading statistic?

Helena Chmura Kraemer

Corresponding Author

Helena Chmura Kraemer

Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA 943305, U.S.A.

Professor of Biostatistics in Psychiatry (Emerita).

1116 Forest Avenue, Palo Alto, CA 94301, U.S.A.Search for more papers by this author
First published: 16 January 2009
Citations: 26

Abstract

Results in risk studies based on events per person-time (EPT, technically ‘incidence rate’) often prove non-confirmable. The circumstances in which the EPT-ratio is unquestionably both valid and optimal to compare a high- and low-risk group, a constant hazards situation, are discussed. However, the constant hazards situation seldom applies in medical research. When the constant hazards situation does not apply, even under optimal circumstances, with fixed entry time and follow-up time for all those not experiencing the event and absence of censoring, EPT-ratio yields at best ambiguous, at worst misleading, results. More careful design and survival analyses are recommended in place of use of EPTs. Copyright © 2009 John Wiley & Sons, Ltd.

The full text of this article hosted at iucr.org is unavailable due to technical difficulties.