CORESTA Congress, Kunming, 2018, Smoke Science/Product Technology Groups, ST 36 (also presented at TSRC 2018)

Multi-dimensional tipping point analyses: Assessing simultaneous shifts in tobacco use patterns from a higher to a lower risk product

BACHAND A.(1); SULSKY S.(1); CURTIN G.(2)
(1) Ramboll U.S. Corp., Amherst, MA, U.S.A.; (2) RAI Services Co., Winston-Salem, NC, U.S.A.

The Dynamic Population Modeler (DPM(+1)) employs a birth cohort framework to estimate effects on population mortality if tobacco use patterns shift from a higher- to lower-risk product. DPM(+1) allows for evaluation of changes in use patterns within the context of 'tipping point' analyses, which estimate the magnitude of a beneficial use pattern needed to offset the population effect of harmful use patterns. In a birth cohort followed from age 13 to age 72, in 5-year intervals, we specified transition probabilities for a counterfactual scenario whereby 3 % of individuals who would have never used tobacco instead initiate modified-risk tobacco product (MRTP) use (up to age 27; additional initiation), and 25 % of those MRTP initiators transition to cigarette use (next age interval; gateway effect). Assuming a 92 % reduction in all-cause mortality risk for the MRTP (relative to cigarettes), population harm was offset if at least 1.6 % of smokers who would have continued to smoke instead switched to MRTP use (each age interval, for ages 18+; switching). In addition, ranges of probabilities for up to three transitions can be assessed simultaneously. Assuming 1-10 %, 0-50 % and 0-10 % for additional initiation, gateway effect and switching, respectively, 2 % switching offset the population harm resulting from, for example, 4 % additional initiation with 20 % gateway effect, 5 % additional initiation with 10 % gateway effect, and 6 % additional initiation with 4 % gateway effect. Tipping point analyses allow regulators to assess the magnitude of simultaneous changes in use patterns likely to result in an overall population benefit or harm. Such analyses may reduce the immediate need for empirical projections of beneficial and/or harmful changes in use patterns during regulatory decision making.