Table 1.

Similarities and differences between the Levy model and our application of PHIM

Characteristic	Similarity
New product	E-cigarettes
Country	USA
Sex	Both sexes considered, results presented by sex, TPs sex-specific
Groups—null scenario	Never, current and former smokers
Groups—alternative scenarios	Never users, current exclusive smokers, current exclusive vapers and former users, but not dual users
Stability of population in follow-up	Only initial population studied, neither model allowing for immigration or emigration
TPs considered	Quitters cannot re-initiate. Switching to cigarettes not allowed, but switching to e-cigarettes is. Initiation and quitting rates derived from national distributions for same birth cohort in successive five-year periods

Characteristic	Difference
Time period	1990–2040 vs. Levy 2016–2100, to allow some hindcasting, and avoid extrapolation to years where rates very uncertain
Diseases considered	Lung cancer (LC), chronic obstructive pulmonary disease (COPD), ischaemic heart disease (IHD) and stroke vs. Levy all causes combined. These form about two thirds of smoking-related mortality (Weitkunat et al. 2015)
Age range	10–79 years (30–79 years for estimating mortality) vs. Levy 15–99 years. As death certification unreliable at older years, individuals not considered after age 79. PHIM includes those aged 10–14 years as product use may start then
Years quit	Considered only by PHIM, to estimate risk more precisely
Age of initiation	To age 35 years vs. Levy to age 25, to reflect US patterns of initiation
Period over which smoking declines to X	First 11 years vs. Levy first 10 years. This reflects periods for which TPs estimated in PHIM (1986–1990, 1991–1995 and 1996–2000)
Former users	Former users of each product combined, rather than separately. Distinction not relevant as re-initiation disallowed, and PHIM retains full product history
Relative risks	Derived from meta-analyses vs. Levy from two CPS studies
Estimating mortality	Full product history used, while Levy uses smoking status at the time