Impact of losses to follow-up on diet/alcohol and lung cancer analyses in the New York State Cohort.

The main objective of the study was to evaluate whether passive surveillance methods can be used in cohort studies without a significant distortion of risk estimates when the active follow-up of every participant is not possible. A nested case-control study including 525 lung cancer cases and 525 controls was conducted among participants of the New York State Cohort Study (n = 57,968 men and women), which allowed the active follow-up of a sample of the cohort and the assessment of the effect of losses to follow-up. Although there were some differences with respect to dietary intake between controls lost to follow-up and those located, the results of the nested case-control study including and excluding losses to follow-up were comparable. Moreover, the results derived from the passive and the active follow-up data were similar. Our findings lent credence to passive follow-up methods and suggested that losses to follow-up did not compromise the validity of the study. Although attempts to trace every participant are preferable in a cohort study, passive surveillance may yield unbiased risk estimates when a rare disease is being investigated.