of which the cohort is a part. The expected number of deaths is usually calcu-
lated using published national life tables and the method is similar to that of
indirect standardization of death rates.
Each member of the cohort contributes to the calculation of the expected
deaths for those years in which he or she was at risk of dying during the study.
There are three types of subjects:
- Subjects still alive on the analysis date.
- Subjects who died on a known date within the study period.
- Subjects who are lost to follow-up after a certain date. These cases are a
potential source of bias; e¤ort should be expended in reducing the num-
ber of subjects in this category.
Figure 10.1 shows the sitution illustrated by one subject of each type, from
enrollment to the study termination. Each subject is represented by a diagonal
line that starts at the age and year at which the subject entered the study and
continues as long as the subject is at risk of dying in the study. In the figure,
each subject is represented by a line starting from the year and age at entry and
continuing until the study date, the date of death, or the date the subject was
lost to follow-up. Period and age are divided into five-year intervals corre-
sponding to the usual availability of referenced death rates. Then a quantity,r,
Figure 10.1 Representation of basis of subject-years method.POISSON DISTRIBUTION 353