Modeling the population health impact of nicotine misperceptions

doi:10.21203/rs.3.rs-3612054/v1

Download PDF

Research Article

Modeling the population health impact of nicotine misperceptions

https://doi.org/10.21203/rs.3.rs-3612054/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 09 Aug, 2024

Read the published version in Harm Reduction Journal →

You are reading this latest preprint version

Background: Scientific evidence clearly demonstrates that inhaling the smoke from the combustion of cigarettes is responsible for most of the harm caused by smoking, and not the nicotine. However, a majority of U.S. adults who smoke (AS) inaccurately believe that nicotine causes cancer which may be a significant barrier, preventing AS from switching to potentially reduced-harm, smoke-free products (SFP) like electronic nicotine delivery systems (ENDS), smokeless tobacco, and snus. We quantify the population health impact associated with nicotine misperceptions.

Methods: Using a previously validated agent-based model to the U.S. population, we analyzed nationally representative data from the Population Assessment of Tobacco and Health (PATH) study to estimate base case rates of sustained (maintained over four waves) cessation and switching to SFP use, by sex. Nicotine perception scenarios used data from Wave 4 of PATH as the baseline wave to analyze levels of nicotine misperceptions. The overall switch rate from smoking in Wave 4 to SFP use in Wave 5 (3.94%) was stratified based on responses to the nicotine perception question “Do you believe nicotine is the chemical that causes most of the cancer caused by smoking cigarettes?”, (Four-item scale from “Definitely not” to “Definitely yes”). The relative percent change between the overall and stratified rates, corresponding to each item, was used to adjust the base case rates of switching to reflect the impact if all AS exhibited switching behaviors based on responses to the nicotine perceptions question. The public health impact of nicotine perceptions was estimated as the difference in all-cause mortality between the base case and the four nicotine perception scenarios.

Results: Switch rates associated with those who responded, “Definitely not” (8.39%) resulted in a net benefit of preventing nearly 800,000 premature deaths over an 85-year period. Conversely switch rates reflective of those who responded, “Definitely yes” (2.59%) resulted in a net harm of nearly 300,000 additional premature deaths over the same period.

Conclusions: Accurate knowledge regarding the role of nicotine is associated with higher switch rates and prevention of premature deaths. Our findings suggest that promoting public education to correct nicotine misperceptions has potential to benefit public health.

Nicotine misperceptions

population modeling

tobacco harm reduction

Combustible cigarettes are the most harmful of all tobacco products and the leading cause of preventable death in the U.S (1) which have been attributed to more than 400,000 premature deaths every year (2). The morbidity and mortality from cigarettes are primarily due to serious diseases such as lung cancer, chronic obstructive pulmonary disease (COPD), and heart disease (1). The harm from cigarette smoking is caused by inhaling the smoke from combustion, which contains more than 7,000 chemicals. The Food and Drug Administration (FDA) has identified a number of these chemicals in cigarette smoke as harmful and potentially harmful constituents (HPHCs) and classified many of them as carcinogens, respiratory or cardiovascular toxicants, reproductive and developmental toxicants, or addictive (3). Smoke-free tobacco products (SFP) such as electronic nicotine delivery systems (ENDS)¹; oral tobacco products e.g., smokeless tobacco, snus, and nicotine pouches; and heated tobacco products either do not contain or have substantially lower levels of these HPHCs compared to cigarettes and therefore can be potentially less harmful than cigarettes. Public health authorities, including the FDA, acknowledge that not all tobacco products carry the same amount of risk but instead exist on a risk continuum, with combustible cigarettes considered the most risky and SFPs considered much less so. Adults who smoke (AS), particularly those unable or unwilling to quit, completely, may reduce their risks of smoking-related diseases by switching from cigarettes to SFPs. Therefore, making available products that deliver nicotine in a less harmful manner to AS, particularly those that are unable or unwilling to quit, is critical to achieving harm reduction for such individuals. However, despite the knowledge that smoking cigarettes is harmful, there are nearly 30 million adults in the U.S. that continue to smoke (5). One of many potential factors preventing AS from switching to less harmful alternatives is misperceptions of nicotine harm.

The vast majority of AS believe that nicotine is the chemical in cigarettes that causes cancer (6–8). For example, researchers from FDA report that about 75% of people either were unsure of the relationship between nicotine and cancer or incorrectly believed that nicotine causes cancer (9). The researchers posit that the incorrect beliefs regarding nicotine causing cancer, could discourage AS from switching to less harmful nicotine-containing smoke-free alternatives. The observations regarding misperceptions of the harm from nicotine have been consistently reported across many studies. In a U.S. based study, 65% of people who smoke believed nicotine causes lung cancer and 71% believed it caused oral cancer (10). Furthermore, a nationally representative study found that two-thirds of AS responded that “nicotine is a cause of cancer” (11, 12). The International Tobacco Control (ITC) survey conducted from 2002–2008 found that more than half of U.S. AS incorrectly reported that nicotine is the cause of cancer, and this proportion significantly increased over time (7).

Nicotine misperceptions also impact attitudes and beliefs towards nicotine replacement therapy (NRT). Shiffman et al. report that a majority of adults who currently smoke and those who had formerly smoked (defined as having quit smoking within the last year) perceived NRT products to be just as harmful as smoking and were significantly less likely to use NRT in future quit attempts (13). Moreover, Snell et al., report that nicotine misperceptions could impede efforts to encourage people who smoke to transition to less harmful sources of nicotine, either to support cessation efforts, or to replace cigarette smoking (14). However, the net impact of nicotine misperceptions on a population level has not been investigated.

The purpose of this research was to estimate the public health impact associated with varying AS perceptions of nicotine harm using a computational health impact model. To our knowledge, this is the first analysis of the potential public health impact that focuses on outcomes based on real-world evidence regarding AS holding either accurate or inaccurate risk perceptions of nicotine. We applied a validated agent-based model to simulate individuals representative of the U.S. population and associated smoking behaviors (15). We projected changes in smoking prevalence and all-cause mortality outcomes (the number of premature deaths) among AS over an 85-year period and compared the outcomes associated with accurate perceptions of nicotine harm versus misperceptions.

We first start with a base case scenario which represents the status quo. In the base case, the rate of switching to SFP use is representative of a population of AS with their existing perceptions of nicotine harm. We compare the base case to four nicotine perception scenarios in which the rate of switching to SFP use is adjusted based on stratified transition probabilities associated with the responses to the question “Do you believe nicotine is the chemical that causes most of the cancer caused by smoking cigarettes?” from Wave 4 data of the Population Assessment of Tobacco and Health (PATH) Study (R04_AC9120, Four-item scale from “Definitely not” to “Definitely yes”). Each nicotine perception scenario modeled the population impact expected to occur if all AS held the same level of perception of nicotine harm as those in each response group from “Definitely not” to “Definitely yes”. We modeled nicotine perception scenarios corresponding to each response rather than dichotomizing them into correct versus incorrect (9, 11, 12) to highlight the trend in population health impact associated with each response on this perception scale. The public health impact of nicotine perceptions is estimated as the difference in all-cause mortality and cigarette prevalence among AS, between the base case and each nicotine perception scenario.

In this research, we use a validated Agent-Based Population Model (ABM) as previously described (15–17) which projects future cigarette prevalence and all-cause mortality beginning in the year 2000. The ABM begins by initializing a hypothetical population of 2.81 million agents (1/100th of the year 2000 U.S. population) that is representative of the U.S. population for age, sex and tobacco use status. The model is iterated 100 times to simulate the entire U.S. population. The initial population mirrors the age and sex distribution data from the year 2000 U.S. Census. Each agent in the initial population is assigned one of three tobacco use statuses representative of people who have never smoked (NT), people who currently smoke (CS) and people who formerly smoked (FS). Tobacco use status is assigned by sex and ages ≥ 18 using information from the National Health Interview Survey (NHIS) Sample Adult Questionnaire data for the year 2000. In our analysis of NHIS data, CS are defined as those who reported having smoked at least 100 cigarettes in their lifetime and are currently smoking every day or some days now. FS were defined as those who reported having smoked at least 100 cigarettes in their lifetime and currently not smoking at the time of interview. NT are defined as those who have never smoked or have not reached 100 cigarettes in their lifetime. These definitions are commonly used by the Centers for Disease Control and Prevention (CDC) to estimate tobacco use prevalence for the U.S. adult population (18). Since NHIS does not provide tobacco use information for ages < 18 years, tobacco use status assigned to the younger U.S. population, ages 10–17, by sex were estimated from the 2000 National Youth Tobacco Survey (NYTS). The NYTS is a nationally representative survey of middle and high school students focused exclusively on patterns of tobacco use. In analysis of the NYTS, we define youth who have never smoked or have not smoked 100 cigarettes in their lifetime as NT; those who reported having smoked 100 cigarettes in their lifetime and having smoked in past 30 days as CS; and those who reported having smoked 100 cigarettes in their lifetime but not having smoked in past 30 days as FS.

Each agent in the initial population is assigned tobacco use history which is updated over the 100-year simulation timeframe from year 2001 to 2100. Agents who initialize as CS or FS are assigned with their associated years of smoking and/or years stopped smoking and the age(s) at which the agent initiated and/or stopped smoking. Age and sex-specific probabilities from U.S. birth cohort smoking history data developed by Jeon et al. were used to assign when a CS or FS in the model’s starting population initiated or stopped smoking (19). The age and sex specific cigarette smoking initiation and cessation probabilities were generated by researchers who used NHIS surveys administered from 1964 to 2015 to estimate birth cohort smoking histories (20) and details of the methodological approach and the resulting data are available on the Cancer Intervention and Surveillance Modelling Network website (CISNET, https://resources.cisnet.cancer.gov/projects/ - Publication Support and Modeling Resources).

Once the initial population is generated, the following algorithms are executed in 1-year time intervals throughout the simulation time frame:

Mortality Sub-model:

A mortality sub-model is used to estimate the survival probability of each agent based on their age, sex and current or former tobacco use history. The mortality sub-model was developed using data from a Kaiser Permanente (KP) Medical Care Program Cohort study (21), which included number of deaths, person- years, smoking status, age, sex, years smoked, and years since quitting smoking. The KP Study data were adjusted using the Human Mortality Database (HMD) to be representative of the U.S. population in the Year 2000. Mortality rates throughout the simulation time frame are further adjusted to account for expected age-specific changes in mortality over time using the methodology described by Carter et al., (22).

Transition sub-model:

At each time interval within a simulation, agents are provided with an option to change or maintain their current tobacco use status. These decisions are governed by the agent’s defined current status, age and sex specific transition probabilities.

Population Update:

The age of agents who survive at each time interval is increased by 1-year increments. New agents are added to the population each year to account for birth and net immigration based on U.S. Census projections (23). Agents who enter the population via immigration were assigned tobacco use status and history similar to that used for the initial population.

ABM scenarios are projected through the year 2100. We quantify the public health impact of varying perceptions of nicotine harm as the difference in adult prevalence and cumulative all-cause deaths for ages 35–85 between the base case and nicotine perception scenarios. The 35–85 age range is used since most smoking mortalities are expected to occur within this range (1).

Base Case Scenario Inputs

The base case scenario models AS prevalence and all-cause mortality where cigarette smoking continues to be the predominant tobacco use behavior. Transition probabilities for initiation from NT to CS and cessation of CS to FS by sex and age are taken from CISNET, where the smoking history of U.S. birth cohorts is reported using NHIS data (19). CISNET transition probabilities are available by age (0–99), sex and year through 2015. CISNET initiation probabilities by sex and age are updated each year in the model from 2001 through 2015, at which point they are held constant for the remaining simulation timeframe. Cessation probabilities in CISNET are based on at least two years of successful smoking cessation. This is an important consideration since minimal relapse has been reported (24) after a period of two years which minimizes the relapse transition between FS and CS in our model. The model updates the CISNET smoking cessation probabilities from CISNET each year between 2001–2013 with the corresponding CISNET estimates. Beginning in the year 2014, the FS state is split to differentiate individuals that successfully quit smoking and do not use SFP from those who successfully quit smoking but are using SFP now. The SFP use state represents the proportion of FS who completely switched to smoke-free product use. Figure 1 provides a diagram of the base case tobacco use states and transitions.

To differentiate FS from SFP use, beginning in 2014 transitions from CS to FS and CS to SFP use are estimated using data from Wave 1 (W1) 2013/2014 to Wave 5 (W5) 2018/2019 of the PATH study which is funded by the FDA Center for Tobacco Products (CTP) and administered by the National Institute on Drug Abuse (NIDA). The PATH study was designed to generate longitudinal epidemiologic data on tobacco-use behavior and health in the U.S. population (25). As of 2022, five waves of PATH data are publicly available for analyses (26). In our analysis of PATH, we identified AS (n = 6,349) in W1 as those who reported smoking at least 100 cigarettes in their lifetime and currently smoke every day or some days. FS were identified as W1 AS who no longer smoked in subsequent waves and did not use ENDS (defined in PATH to include products such as e-cigarettes, e-hookahs, e-cigars, e-pipes, personal vaporizers, vape pens, and hookah pens) or smokeless tobacco (ST, defined in PATH to include moist loose snus, snuff, dip and spit or chewing tobacco or snus every day or some days). SFP use was identified as W1 AS who no longer smoked but also indicated every day or someday use of ENDS, ST or snus in subsequent waves. We calculated successful quitting and switching as people who smoked in W1 who transitioned to either FS or SFP use in W2 (2015/2016) and maintained their FS or SFP use status through W5 (an approximate 3-year of follow up). W1 AS who indicated they had not used ENDS, ST or snus every day or some days in W2 but had used in any subsequent wave, W3 through W5, were excluded from our analysis. This was done to ensure the transition rates reflected long-term sustained behavior. Estimates from five waves of PATH were used to obtain successful quitting/switching probabilities, which follows a similar methodology as CISNET successful smoking cessation calculation. We observed higher successful cigarette cessation probabilities based on the PATH study data, which aligns with observations from existing research that indicates a recent increasing trend in cigarette cessation (27). We applied PATH cessation probabilities to account for recent changes in smoking prevalence observed between 2014 and 2018 since they are not captured in the CISNET cessation rates.

Based on our analysis of PATH W1 to W5, 3.92% male and 4.35% female AS transitioned to FS in W2 and remained FS through W5. Adults who smoked in W1 transitioned to SFP use in W2 and remained SFP users through W5 at a rate of 1.32% and 0.44%, male and female respectively.

The all-cause mortality probability assigned to CS that switch to SFP use is based on the excess relative risk (ERR) of SFP use compared to cigarette smoking. We applied an ERR of 10%, which reflects a reasonable aggregate estimate for the various SFPs. An ERR of 10% is slightly higher than the estimate of 9% for current ST product use relative to smoking (28). Applying a multi-criteria decision analysis (MCDA) model developed by an international expert panel convened by the Independent Scientific Committee on Drugs (29), the authors assigned the relative importance of different types of harm related to the use of nicotine-containing products compared to cigarette smoking. The panel estimated a relative risk of 5% for snus and 4% for ENDS products relative to smoking. An ERR value of 10% represents a conservative choice for SFP use in the model given that this tobacco use state is representative of exclusive or other combinations of ST, snus and ENDS use. Additionally, we do not allow for transition to former SFP use, therefore agents entering the SFP use state carry their former CS risk in addition to the ERR for current SFP use throughout the remaining simulation timeframe.

Nicotine Perception Scenarios

Nicotine perception scenarios were simulated in which all AS transition to SFP use at rates associated with different level of perception of nicotine harm. PATH W4 (2016–2018) data were used to assess AS perceptions of nicotine harm based on their responses to the question “Do you believe nicotine is the chemical that causes most of the cancer caused by smoking cigarettes?” (R04_AC9120, Four-item scale from “Definitely not” to “Definitely yes”). We choose W4 as the baseline wave to analyze the levels of nicotine perception because, in addition to also being a true cross-sectional wave due to sample replenishment, it is more recent compared to W1 (2013–2014) and may reflect the current state of nicotine perception among AS. Table 1 shows the demographics characteristics of W4 AS overall and by perception of nicotine harm. We note that striking differences were observed among the different age and race/ethnicity subgroups in response to the nicotine perceptions question. While the nicotine misperceptions of harm were high across all age subgroups, they trended to be higher among the 45 + age subgroup. More than 70% of AS in the 45 + age subgroup responded “Definitely yes” or “Probably yes” regarding harm from nicotine, compared to ~ 50–55% for those in the lower age groups. Additionally, a higher proportion (~ 76%) of AS in the Non-Hispanic Black and Hispanic subgroups indicated higher perceptions of nicotine harm compared to the Non-Hispanic Whites (~ 48%). The data in Table 1 are important as the observed differences may result in different rates of switching to SFP use between demographic subgroups. Modeling can theoretically be used to investigate population health impacts by demographic characteristics however, the sample sizes needed to provide robust input estimates have limited most models to include only age and sex variables. We focused our simulations on the overall group of W4 established AS (18+) due to sample size and model limitations (i.e., current model is not capable of utilizing population inputs by race or economic status) however it is important to consider the data in Table 1 when interpreting the results.

Table 1

AS demographics by nicotine perception held
Characteristic	Total	Perception of nicotine harm: “Do you believe nicotine is the chemical that causes most of the cancer caused by smoking cigarettes?”
Characteristic	Total	Definitely yes	Probably yes	Probably not	Definitely not
	7007	1338	2943	2090	515
Sex
Female	46.38%	19.01%	45.99%	28.95%	6.04%
Male	53.62%	18.06%	43.19%	31.64%	7.11%
Age
18–24	8.15%	13.79%	36.36%	36.85%	12.99%
25–44	45.47%	16.84%	39.97%	35.52%	7.67%
45+	46.38%	21.02%	50.36%	24.18%	4.44%
Race/Ethnicity
Non-Hispanic White	68.78%	13.47%	44.42%	35.02%	7.09%
Non-Hispanic Black	13.08%	33.94%	43.08%	17.57%	5.41%
Non-Hispanic Others	5.42%	19.06%	43.02%	31.21%	6.7%
Hispanic	12.71%	29.63%	46.89%	18.26%	5.21%
Education
LT College	57.39%	22.45%	46.92%	25.19%	5.44%
Some College	31.70%	14.08%	41.74%	36.58%	7.59%
College Grad	10.91%	9.64%	40.32%	40.33%	9.7%
Smoking Behavior
Every Day	76.32%	18.51%	45.52%	29.84%	6.13%
Some Day	23.68%	18.58%	41.12%	32.12%	8.18%

Nicotine perception scenario inputs were developed by first estimating the overall rate of switching from CS to SFP use based on longitudinal analysis of PATH W4 (2016–2018) and W5 (2018–2019). The analysis included W4 established AS (18+) that responded to the nicotine perception question (i.e., “Definitely not”, “Probably not”, “Probably yes” or “Definitely yes”) (n = 6886) following the same definitions used to define CS and SFP users as used in the base case. The overall transition rate for W4 AS who completely switched to SFP use by W5 was estimated to be 3.94%. The overall switching rate was then stratified by W4 AS perceptions of nicotine harm. We did not estimate switching rates by age or sex to preserve sample size for each of the nicotine perception response groups. Figure 2 shows the proportions of W4 AS that comprise each response group and their associated transition rates to SFP use in W5. As shown in Fig. 2, 6.61% of W4 AS (n = 515 shown in Table 1) responded “Definitely not”. Of those AS who responded “Definitely not”, 8.39% quit smoking and switched to SFP use by W5, more than twice the overall rate. This is in contrast with 18.52% (n = 1338 in Table 1) of AS who responded “Definitely yes”, of which 2.59% switched to exclusive SFP use at W5. The error bars shown in Fig. 2 represent 95% confidence intervals for each response group switching rate.

A total of four nicotine perception scenarios were simulated corresponding to each of the four responses to the question “Do you believe nicotine is the chemical that causes most of the cancer caused by smoking cigarettes?”. The model inputs corresponding to each nicotine perception scenario were estimated by adjusting the base case sustained switching rates with the relative percent change between the W4 to W5 overall and stratified transition rates corresponding to each different level of nicotine perception. Relative percent changes are calculated as the stratified rate associated with one level of the nicotine harm perception minus the overall rate divided by the overall² rate of CS to SFP use between W4 and W5. Relative percent changes were used to adjust the base case rates by sex to avoid over estimation of switching behavior. Switching rates between W4 to W5 would not be expected to be comparable to the sustained 1-year switching rate used in the base case since W5 is a follow up study conducted approximately two years after W4 data collection and may not reflect sustained switching. Therefore, we only applied relative percentage changes between the overall and stratified switching rates to the base case to evaluate the population impact associated with various levels of nicotine harm perception. Table 2 provides the base case transition rate inputs by sex, relative percent change adjustments and the final rates used in each of the four nicotine perception scenarios. As shown in Table 2, AS who responded “Definitely not” switched to SFP use at a rate 113% greater (more than twice) than the overall observed rate of switching while those that responded “Definitely yes” switched at a rate 34% lower than the overall. Also included in Table 2 are relative percent change adjustments corresponding to 95% confidence limits of the transition rates calculated as the ratio of the upper and lower 95% confidence limit of the transition rate associated with one level of the nicotine harm perception over the overall transition rate of CS to SFP use between W4 and W5. Sensitivity scenarios were conducted based on these 95% confidence limits of the stratified transition rates. Each scenario assumes that all AS will exhibit the switching behavior associated with the specific response groups.

Table 2

Nicotine perception scenario CS to SFP switch rates derived from adjustment of the base case rate using the relative percent change between W4 – W5 overall and each nicotine perception response group.
Nicotine Perception Scenario	Base case CS to SFP Switch Rates		Relative % change Adjustment*	CS to SFP Switch Rates corresponding to each Nicotine Perception Scenario**
Nicotine Perception Scenario	Male	Female	Relative % change Adjustment*	Male	Female
Definitely not	1.32	0.44	113% (47%,204%)	2.81	0.94
Probably not			28% (1%, 60%)	1.69	0.56
Probably yes			-21% (-1%, -37%)	1.04	0.35
Definitely yes			-34% (-5%, -55%)	0.87	0.29
*Relative percent changes corresponding to 95% confidence limits of the stratified switch rates are shown.
*CS to SFP switch rate calculated as Base Rate + Base Rate Relative % change adjustment.

The difference in overall prevalence and all-cause mortality between the base case and nicotine perception scenarios is used to quantify the population health impact corresponding to AS perceptions of nicotine harm.

Table 3 shows AS prevalence projections between 2014 and 2100 for the Base Case and Nicotine Perception scenarios. Under the Base Case scenario, which reflects a mix of nicotine perceptions, the model predicts a relative change of -58.9% (i.e., a 58.9% reduction) in AS prevalence between the years 2014 and 2100. AS prevalence decreases further over the same period when the transition rate to SFP use is adjusted to correspond to nicotine harm perceptions held by all the AS in the simulation similar to the “Definitely not” and “Probably not” response groups with relative changes of -63.7% and − 60.2%, respectively. The model predicted slightly higher AS prevalence in Nicotine Perception scenarios if all the AS in the simulation held nicotine harm perceptions similar to the “Probably yes” and “Definitely yes” response groups resulting in relative changes of -57.8% and − 57.1%, respectively.

Table 3

AS Prevalence 2014–2100 (Ages 18+)
Scenario/year	2014	2025	2050	2075	2100	Relative Change
Base case	17.0%	11.2%	7.5%	7.1%	7.0%	-58.9%
Definitely not	16.8%	10.2%	6.5%	6.2%	6.1%	-63.7%
Probably not	17.0%	11.0%	7.2%	6.9%	6.7%	-60.2%
Probably yes	17.0%	11.4%	7.7%	7.3%	7.2%	-57.8%
Definitely yes	17.0%	11.5%	7.8%	7.4%	7.3%	-57.1%
Relative change is calculated as(Y2100-Y2014)/Y2014

Figure 3 shows the predicted population health impact as cumulative premature deaths prevented through to the year 2100. The nicotine perception scenarios in Fig. 3 resulting in cumulative premature deaths prevented greater than zero correspond to a net population health benefit. On the other hand, differences less than zero indicate an increase in smoking attributable mortality. Additionally, cumulative premature deaths prevented equal to zero indicate no change from the base case would be expected. The error bars in Fig. 3 show the results of additional sensitivity simulations corresponding to adjustments from the relative percent change of the overall W4-W5 switch rate and the 95% confidence limits of the stratified switch rates by nicotine perception response groups. The scenarios corresponding to “Definitely not” and “Probably not” both result in positive net public health gains of approximately 800,000 and 200,000 premature deaths prevented over the simulation time frame, respectively.

Table 4 shows the predicted population health impact as cumulative premature deaths prevented from 2025 to 2100 in 25-year intervals. As shown in Table 4 the population health risks and benefits associated with varying perceptions of nicotine harm are seen as early as 2025.

Table 4

Cumulative Deaths Prevented (Base - Nicotine Perception Scenario)
Scenario/year	2025	2050	2075	2100
Definitely not	19,688	311,133	569,885	796,530
Probably not	5,395	86,384	161,207	228,390
Probably yes	-2,128	-72,207	-135,723	-194,735
Definitely yes	-11,376	-103,339	-200,652	-290,491

Figure 4 show the ranges corresponding to sensitivity scenarios in which the adjustments to the base case rates are based on the 95% confidence intervals of the switching rates (see Table 2) for the “Definitely not” and Definitely yes” response groups. The upper and lower 95% confidence limits associated with the “Definitely not” response group results in a net population health benefit between 373,000 and 1.3 million premature deaths prevented (shaded region in Fig. 4 labeled Definitely not). The upper and lower 95% confidence limit corresponding to the “Definitely yes” response group results in a net population health risk between approximately 40,000 and 509,000 additional premature deaths (shaded region in Fig. 4 labeled Definitely yes). Figure 4 illustrates that if all the AS exhibit the switching behavior corresponding to those who responded “Definitely not” then a net population benefit can be expected. On the other hand, the behavior associated with AS with misperceptions regarding the role of nicotine in the harm caused by smoking result in net population risk. The shaded regions in Fig. 4 do not overlap indicating a clear difference in the health outcomes of these two groups.

The objective of this study was to estimate the public health impact associated with varying perceptions of nicotine harm using computational modelling. While it is imperative for individuals to understand that nicotine is addictive, it is equally important for them to understand that nicotine is not the main reason why cigarette smoking is harmful to health. A clear trend in the rate of switching to SFP use was observed across the nicotine harm perception response scale. We forecasted substantial benefits to public health if AS have correct perceptions of nicotine harm and negative consequences from continuing to maintain misperceptions regarding nicotine. Applying to all AS, switch rates associated with those who responded that nicotine is “Definitely not” the chemical that causes the cancer caused by smoking, resulted in preventing nearly 800,000 premature deaths with an upper estimate of 1.3 million over an 85-year period to 2100. The lower switch rate associated with those who responded “Definitely yes” correspondingly resulted in increasing deaths attributable to smoking by an additional 300,000 by the year 2100 if all AS held this misperception. Our findings indicate that, if left unchecked, increases in the proportion of AS who hold misperceptions of nicotine harm could have negative impacts on harm reduction efforts and public health. This is particularly important for “older” and Non-Hispanic Black and Hispanic AS. Our observations of a higher prevalence of nicotine misperceptions, particularly in those individuals in the older age subgroups (Table 1) are comparable to those reported recently by Rubenstein et al (30). As indicated by the authors, the higher levels of nicotine misperceptions in these individuals, could be a barrier to switching to SFP. Overall, our results illustrate the potential long-term public health gains that could be made through campaigns aimed at correcting perceptions of nicotine harm among all AS rather than the current status quo which has been described as a “quarantine of information” (31). The trend from risk to benefit across the nicotine perception scenarios indicates that even a modest shift in perceptions of nicotine harm could have positive impacts on public health.

Our analysis of PATH W4 data are consistent with previous research indicating that a majority of AS (63% responded “Definitely” or “Probably yes”) have misperceptions regarding the role of nicotine as a cause of cancer from smoking (6, 7, 9, 11, 12, 14). Borland et al reported nicotine harm perceptions are continuing to trend in the wrong direction with a significant decline observed (OR = 0.97, p = 0.002) among people reporting that nicotine is not the chemical that causes most of the cancer (7).

Incorrectly believing that nicotine causes cancer could discourage AS from switching to safer nicotine containing alternatives from cigarettes. Nicotine misperceptions could be erroneously associated with perceived harm of SFPs. Wilson et al. found that correct perceptions of nicotine harm were associated with correctly perceiving e-cigarettes to be less harmful than smoking (32). Studies have shown that AS who perceive ENDs to be less harmful than smoking are more likely to switch. For example, Kim et al. reported that switching to ENDS was nearly three times higher for AS who perceived ENDS to be less harmful than smoking (33). Similarly, Snell et al. report that misperceiving nicotine to be a main cause of smoking-related cancers was associated with lower odds for ENDS use (adjusted odds ratio (AOR) = 0.59; 95% confidence interval (CI)I: 0.49, 0.71; p < 0.01)(14).

Moreover, misperceptions regarding the harm from nicotine also impacts the attitudes and beliefs of AS towards nicotine replacement therapy (NRT). Shiffman et al. report that a majority of AS (66%) perceived NRT products to be just as harmful as smoking or were unsure (13). These individuals were less likely to have used NRT in the past (odds ratio (OR) = 0.45, 95% CI: 0.39–0.53) and less likely to consider using NRT during future quit attempts (OR = 0.60, 95% CI = 0.51–0.71). Additionally, Snell et al. reported lower odds of NRT use (AOR: 0.84; 95% CI: 0.71, 0.99; p = 0.04) among those individuals who made a quit attempt during the study period (14). Therefore, nicotine misperceptions could not only impede efforts to encourage AS to transition from cigarettes to less harmful sources of nicotine, but also to support cessation efforts.

The results of our findings should be considered in the context of the limitations of the model due to the assumptions. In this study we made several assumptions to simplify the modeling framework. For example, we assumed that initiation rates for smoking to be unchanged between both the base case and nicotine perception scenarios. The model outcomes are not expected to be greatly impacted since the assumptions impact both the base case and the nicotine perception scenarios and the population health impact is evaluated by comparing the outcomes between the scenarios. Thus, changes to initiation rates would impact the total prevalence and all-cause deaths to a similar extent in both base case and nicotine perception scenarios. However, it is important to consider that correcting the misperceptions should be targeted to AS and minimize any “spillover” reach to youth. Additionally, as the focus of this research is on cigarette smoking prevalence and smoking attributable all-cause mortality, we did not consider relapse from SFP use back to cigarette smoking. Because the base case rate of switching to SFP use is based on long-term analysis, spanning over roughly four years, (PATH W2-W5), this is a reasonable assumption since minimal relapse back to smoking is expected after two years (34). Moreover, we assumed cessation from SFP use as nonexistent, a conservative assumption, since it excludes any additional potential population health benefit resulting from long term cessation of SFPs. We also made a simplifying assumption that the impact of initiation to SFP use among those not currently using tobacco products was minimal. While potential increases in initiation of SFP use due to corrected perceptions of nicotine harm alone would be expected to decrease the overall net benefit, a concurrent decrease in smoking initiation and potential increases to SFP cessation could act to counterbalance impacts due to initiation by nonusers. Various population health models examining the impact of ENDS on the population have shown that the public health benefit associated with rates of smoking cessation and switching generally far outweighs possible increases to initiation (35–38). In a recent model evaluating the impact of introducing a novel oral nicotine product, we found that initiation would have to increase by more than 3000% before the benefit of AS switching was outweighed (39).

The benefit associated with correcting perceptions of nicotine harm is dependent on the relative risk of the SFP category as compared to smoking. In our model, we assumed that the ERR of SFPs was a single value of 0.1. This estimate is derived from the epidemiological evidence for smokeless tobacco products (40) which is a reasonable and conservative assumption, representative of other SFPs. Current estimates of the relative risk of the various SFPs (e.g., snus and ENDS products) have been reported to be slightly lower than smokeless tobacco (29, 40, 41).

Accurate knowledge regarding the role of nicotine is associated with higher switch rates and prevention of premature deaths. Our findings suggest that promoting public education to correct nicotine misperceptions has potential to benefit public health. Despite the inherent uncertainty of a modeling approach, this research highlights the potential public health benefit of a harm reduction strategy for current AS, especially those unable or unwilling to quit. Providing information to AS about the role of nicotine and the relative risk of SFPs compared to smoking would be crucial to removing barriers to AS transitioning to SFPs, potentially lower harm sources of nicotine. Thus, creating the opportunity to accelerate the decline in smoking prevalence and substantially reduce premature smoking-related mortality. However, dissemination of this information should be considered in the context of interpretation by nonusers, including youth.

Availability of data and materials

The transition and nicotine perception data analyzed are publicly available and can be downloaded from https://www.icpsr.umich.edu/icpsrweb/NAHDAP/studies/36498. The data used to develop the Agent-Based Population Model are publicly available and described in the previous referenced publication and can be accessed from https://doi.org/10.32388/TVND4Q. Additional information may be provided based on reasonable request.

Competing interests

The authors are employees of Altria Client Services LLC.

Authors' contributions

TH and RMK developed the concept, LW generated the model inputs from PATH dataset, EL and MS provided strategic direction, all authors contributed towards the writing and review of the manuscript.

United States. Public Health Service. Office of the Surgeon General. The health consequences of smoking--50 years of progress : a report of the surgeon general. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Office of the Surgeon General; 2014. 2 volumes p.
CDC. Smoking-attributable mortality, years of potential life lost, and productivity losses--United States, 2000-2004. MMWR Morb Mortal Wkly Rep. 2008/11/15 ed2008. p. 1226-8.
FDA. Harmful and Potentially Harmful Constituents in Tobacco Products and Tobacco Smoke; Established List: Food and Drug Administration, HHS (Notice; establishment of a list). Food and Drug Administration Federal Register. 2012a. p. 20034-7.
Koh HK, Fiore MC. The Tobacco Industry and Harm Reduction. JAMA. 2022;328(20):2009-10.
Cornelius M, Loretan C, Wang T, Jamal A, Homa D. Tobacco Product Use Among Adults — United States, 2020. MMWR Morb Mortal Wkly Rep. 2022;71:397-405.
Siahpush M, McNeill A, Hammond D, Fong GT. Socioeconomic and country variations in knowledge of health risks of tobacco smoking and toxic constituents of smoke: results from the 2002 International Tobacco Control (ITC) Four Country Survey. Tob Control. 2006;15 Suppl 3(Suppl 3):iii65-70.
Borland R, Cooper J, McNeil A, O'Connor R, Cummings K. Trends in beliefs about the harmfulness and use of stop-smoking medications and smokeless tobacco products among cigarettes smokers: Findings from the ITC four-country survey. Harm Reduction Journal. 2011;8(21).
Villanti AC, Byron MJ, Mercincavage M, Pacek LR. Misperceptions of Nicotine and Nicotine Reduction: The Importance of Public Education to Maximize the Benefits of a Nicotine Reduction Standard. Nicotine Tob Res. 2019;21(Suppl 1):S88-S90.
O'Brien EK, Nguyen AB, Persoskie A, Hoffman AC. U.S. adults' addiction and harm beliefs about nicotine and low nicotine cigarettes. Prev Med. 2017;96:94-100.
Mooney ME, Leventhal AM, Hatsukami DK. Attitudes and knowledge about nicotine and nicotine replacement therapy. Nicotine Tob Res. 2006;8(3):435-46.
Bansal MA, Cummings KM, Hyland A, Giovino GA. Stop-smoking medications: who uses them, who misuses them, and who is misinformed about them? Nicotine Tob Res. 2004;6 Suppl 3:S303-10.
Cummings KM, Hyland A, Giovino GA, Hastrup JL, Bauer JE, Bansal MA. Are smokers adequately informed about the health risks of smoking and medicinal nicotine? Nicotine Tob Res. 2004;6 Suppl 3:S333-40.
Shiffman S, Ferguson SG, Rohay J, Gitchell JG. Perceived safety and efficacy of nicotine replacement therapies among US smokers and ex-smokers: relationship with use and compliance. Addiction. 2008;103(8):1371-8.
Snell LM, Colby SM, DeAtley T, Cassidy R, Tidey JW. Associations Between Nicotine Knowledge and Smoking Cessation Behaviors Among US Adults Who Smoke. Nicotine Tob Res. 2022;24(6):855-63.
Muhammad-Kah R, Hannel T, Wei L, Pithawalla YB, Gogova M. A computational model assessing population impact of a new tobacco product. Qeios. 2023.
Lee PN, Abrams D, Bachand A, Baker G, Black R, Camacho O, et al. Estimating the Population Health Impact of Recently Introduced Modified Risk Tobacco Products: A Comparison of Different Approaches. Nicotine Tob Res. 2021;23(3):426-37.
Muhammad-kah M, Pithawalla YB, Wei L, Hannel T, Gogova M, Boone EL, editors. Presentation and poster session on advances in nonparametric statistics: An agent based modeling approach for tobacco product risk assessments. Joint Statistical Meetings (JSM 2016); 2016; Chicago.
Centers for Disease Control and Prevention. NCHS Vital Health Statistics Series 10. In: Health NIo, editor. 2017.
Jeon J, Holford TR, Levy DT, Feuer EJ, Cao P, Tam J, et al. Smoking and Lung Cancer Mortality in the United States From 2015 to 2065: A Comparative Modeling Approach. Ann Intern Med. 2018;169(10):684-93.
Tam J, Levy DT, Jeon J, Clarke J, Gilkeson S, Hall T, et al. Projecting the effects of tobacco control policies in the USA through microsimulation: a study protocol. BMJ Open. 2018;8(3):e019169.
Friedman G, Tekawa I, Sadler M, Sidney S. Smoking and mortality: The Kaiser Permanente experience. Smoking and Tobacco Control Monograph No 8: Changes in Cigarette-related Disease Risks and their Implications for Prevention and Control. Bethesda, MD: National Cancer Institute, U.S. National Institutes of Health; 1997. p. 477-500.
Carter LR, Lee RD. Modeling and forecasting U.S. sex differentials in mortality. Int J Forecast. 1992;8(3):393-411.
U.S. Census Bureau, 2008 National Population Projections Tables United States Census Bureau [Available from: https://www.census.gov/data/tables/2008/demo/popproj/2008-summary-tables.html.
Anderson CM, Burns DM, Dodd KW, Feuer EJ. Chapter 2: Birth-cohort-specific estimates of smoking behaviors for the U.S. population. Risk Anal. 2012;32 Suppl 1(0 1):S14-24.
Hyland A, Ambrose BK, Conway KP, Borek N, Lambert E, Carusi C, et al. Design and methods of the Population Assessment of Tobacco and Health (PATH) Study. Tob Control. 2017;26(4):371-8.
Health USDo, Abuse HSNIoHNIoD, Health USDo, Food HS, Products DACfT. Population Assessment of Tobacco and Health (PATH) Study [United States] Public-Use Files. Inter-university Consortium for Political and Social Research [distributor]; 2022.
Mendez D, Le TTT, Warner KE. Monitoring the Increase in the U.S. Smoking Cessation Rate and Its Implication for Future Smoking Prevalence. Nicotine Tob Res. 2022;24(11):1727-31.
Muhammad-Kah RS, Pithawalla YB, Boone EL, Wei L, Jones MA, Black RA, et al. A Computational Model for Assessing the Population Health Impact of Introducing a Modified Risk Claim on an Existing Smokeless Tobacco Product. Int J Environ Res Public Health. 2019;16(7).
Nutt DJ, Phillips LD, Balfour D, Curran HV, Dockrell M, Foulds J, et al. Estimating the harms of nicotine-containing products using the MCDA approach. Eur Addict Res. 2014;20(5):218-25.
Rubenstein D, Denlinger-Apte RL, Cornacchione Ross J, McClernon FJ. Adoption of E-Cigarettes Among Older Adults Who Smoke to Reduce Harm and Narrow Age-Related Disparities: An Application of the Health Belief Model. Nicotine Tob Res. 2023;25(6):1212-4.
Kozlowski LT, Sweanor D. Withholding differential risk information on legal consumer nicotine/tobacco products: The public health ethics of health information quarantines. Int J Drug Policy. 2016;32:17-23.
Wilson S, Partos T, McNeill A, Brose LS. Harm perceptions of e-cigarettes and other nicotine products in a UK sample. Addiction. 2019;114(5):879-88.
Kim S, Shiffman S, Sembower MA. US adult smokers' perceived relative risk on ENDS and its effects on their transitions between cigarettes and ENDS. BMC Public Health. 2022;22(1):1771.
Herd N, Borland R, Hyland A. Predictors of smoking relapse by duration of abstinence: findings from the International Tobacco Control (ITC) Four Country Survey. Addiction. 2009;104(12):2088-99.
Vugrin ED, Rostron BL, Verzi SJ, Brodsky NS, Brown TJ, Choiniere CJ, et al. Modeling the potential effects of new tobacco products and policies: a dynamic population model for multiple product use and harm. PLoS One. 2015;10(3):e0121008.
Cherng ST, Tam J, Christine PJ, Meza R. Modeling the Effects of E-cigarettes on Smoking Behavior: Implications for Future Adult Smoking Prevalence. Epidemiology. 2016;27(6):819-26.
Levy DT, Borland R, Lindblom EN, Goniewicz ML, Meza R, Holford TR, et al. Potential deaths averted in USA by replacing cigarettes with e-cigarettes. Tob Control. 2018;27(1):18-25.
Warner KE, Mendez D. E-cigarettes: Comparing the Possible Risks of Increasing Smoking Initiation with the Potential Benefits of Increasing Smoking Cessation. Nicotine Tob Res. 2019;21(1):41-7.
Muhammad-Kah M, Hannel T, Cheng H, Sarkar M, editors. Assessing the Potential Population Health Impact of a Market Authorization of an Oral Nicotine Pouch Product in the U.S. Society for Research on Nicotine and Tobacco; 2021 February 24-27, 2021; Virtual Conference,.
Fisher MT, Tan-Torres SM, Gaworski CL, Black RA, Sarkar MA. Smokeless tobacco mortality risks: an analysis of two contemporary nationally representative longitudinal mortality studies. Harm Reduct J. 2019;16(1):27.
Murkett R, Rugh M, Ding B. Nicotine products relative risk assessment: a systematic review and meta-analysis. F1000Research. 2020;9.

Electronic nicotine delivery systems, e-cigarettes, e-vapor products.
Relative Change = (Stratified Rate – Overall)/Overall

Competing interest reported. The authors are employees of Altria Client Services LLC.

Download PDF

Journal Publication

published 09 Aug, 2024

Read the published version in Harm Reduction Journal →

Editorial decision: Revision requested
17 Mar, 2024
Reviews received at journal
05 Jan, 2024
Reviewers agreed at journal
28 Dec, 2023
Reviewers invited by journal
28 Dec, 2023
Submission checks completed at journal
15 Nov, 2023
Editor assigned by journal
15 Nov, 2023
First submitted to journal
14 Nov, 2023

You are reading this latest preprint version

Modeling the population health impact of nicotine misperceptions

Status:

Journal Publication

Version 1

Abstract

Figures

Background

Methods

Mortality Sub-model:

Transition sub-model:

Population Update:

Base Case Scenario Inputs

Nicotine Perception Scenarios

Results

Discussion

Conclusion

Declarations

References

Footnotes

Additional Declarations

Status:

Journal Publication

Version 1