Table 2.

Considerations in the Choice of a Cohort Model

Scope	Type(s) of new product to be evaluated
	Location for which estimates are relevant
	Age groups and sexes for which results required
	Health endpoints to be considered; mortality, loss of life, and quality of life
	Whether results for specific diseases required
Initial population	Year of baseline
	Smoking groups used at baseline
	Age at baseline
Follow-up	Year of end of follow-up
	Add new young individuals and immigrants during follow-up?
	Remove deaths and emigrants during follow-up?
	Follow individuals or groups?
	Tobacco groups to be used in the alternative scenario
	In the null scenario are initiation, cessation, and re-initiation all to be considered?
	In the alternative scenario which of the possible initiations, cessations, re-initiations, and switches are to be considered?
	Can these transition probabilities vary by period of follow-up?
	Can they vary by previous smoking habits?
Estimating health endpoints from smoking histories	What F-factor is to be assumed for estimating the relative increase in risk for MRTP users and smokers?
	What G-factor is to be assumed for estimating the relative increase in risk for dual users and smokers?
	Allow for other factors such as environmental tobacco smoke?
	How is risk for a given smoking history, relative to never exposed individuals and groups, to be determined?
Sources of data used	These include initial population distributions by sex, age, and smoking, immigration, emigration and birth rates, death rates, transition probabilities, F- and G-factors, and current and former smoking relative risks

MRTP = modified risk tobacco product.