Interval censoring in longitudinal data of respiratory symptoms in aluminium potroom workers: a comparison of methods.

In a longitudinal study of workers in seven Norwegian aluminium plants, the time to development of asthmatic symptoms could only be determined to lie in the interval between two consecutive health examinations. In a previous paper we analysed the data by survival techniques for interval censored data. In the present paper the data are reanalysed in two ways and compared to the previous analyses. First, it is assumed that occurrence of symptoms coincided with reporting, in which case the data can be analysed as right censored. Secondly, the follow-up times are completely disregarded and the effects of covariates are analysed on the binary outcomes of symptoms. Comparing the estimated effects of the covariates between the three approaches, only minor differences were found. However, the estimates on incidence were strongly influenced by whether the data were analysed as right or interval censored.