When calculated for specific time intervals, caution is observed with regard to time factor and particularly to the exact population involved under study.

Time-person years will be a better denominator not only in those instances but also most of the times as there is an imminent danger of loss of the cohort subjects at any time.