Population Modeling Helps Predict the Impact of E-cigarettes on Health

Population Modeling Helps Predict the Impact of E-cigarettes on Health

The Oxford Dictionaries 2014 Word of the Year is “vape” – to inhale and exhale the vapor produced by an electronic cigarette (also known as e-cigarette or e-cig) or similar device. This choice reflects the meteoric rise in e-cig popularity.

E-cigs, which deliver nicotine without carcinogenic tar, hold the promise to save the lives of many smokers who switch to them. However, their potential risks include failure to quit cigarettes (dual use), increased initiation to nicotine products among youth, relapse of former smokers to e-cigs, and e-cigs becoming a “gateway to smoking.” Population modeling approaches can be used to capture these uncertainties and weigh the benefits versus risks to predict probabilistic impacts of e-cigs and other tobacco products on public health.

How do e-cigarettes work?

Unlike cigarettes which burn tobacco, e-cigs contain a cartridge filled with a nicotine solution, propylene glycol, glycerin, water, and flavoring. A battery powers a heating element that vaporizes the nicotine solution. The user inhales the vapor through the mouthpiece.

Could this new technology render cigarettes obsolete?

According to the 2014 Surgeon General report, smoking still kills almost half a million Americans annually and reduces life spans by 11-12 years. Thus, a technology that improves upon past smoking cessation therapy would benefit public health. Sales of e-cigs have been roughly doubling each year since their US introduction in 2007, though a slowdown has been noted in the last few months. Could e-cigs satisfy a smoker’s nicotine addiction and render cigarettes obsolete?

How does the FDA view this emerging trend?

Smoking has been linked to diseases of nearly all organs in the body. While smoking is most commonly linked to lung cancer, chronic obstructive pulmonary disease (COPD), and coronary heart disease, it is also implicated in diabetes, rheumatoid arthritis, and colorectal cancer. Since the long-term impact of e-cigs is unknown, the FDA intends to regulate them to protect public health.

Best case vs. worst case scenarios

A range of outcomes might result from the emergence of e-cig technology. A best case scenario would have several components:

  • There is a massive migration of smokers, but not non-smokers, to vaping
  • Dual use is temporary
  • Many e-cig users work their way from high concentration nicotine cartridges to low and no nicotine e-cigarettes
  • Long-term vapor inhalation is found to be safe

A worst case scenario would include the following:

  • Non-smokers start vaping due to the perception of e-cigs as safe
  • Dual use persists
  • E-cigs serve as a “gateway to smoking” by those who would otherwise never have used nicotine
  • Long-term exposure to fine particles in vapor turns out to be harmful

How population modeling can provide insight into net health impacts of e-cigs

Population models can be used to predict smoking prevalence and mortality under various scenarios. Currently, the US adult prevalence of smoking is around 20%, depending on the survey used. Under the status quo without e-cigs, smoking prevalence has been predicted to fall to around 10-15% by the year 2050. However, modeling shows that halving smoking initiation rates would bring prevalence down only modestly from there, whereas increasing (age-dependent) cessation rates by half could reduce prevalence as much as 80% by 2050.

We also know that mortality risk relative to non-smokers drops when smokers quit. The earlier someone quits smoking, the greater the benefit. In fact, quitting smoking by age 40 may return risk to non-smoker levels. Moreover, modeling gives insight into how the level of smoking impacts mortality. Unsurprisingly, the longer someone smokes and the greater the number of cigarettes smoked per day the higher the mortality risk.

When we add e-cigs to the picture, the model becomes more complex. Let’s assume that users change only one product at a time and that there is no relapse, except former smokers may take up e-cigs. Even with these assumptions, there are still over a dozen highly uncertain transition rates that would need to be assessed (rate that non-users initiate e-cigs, rate that e-cig users become dual users, etc.).

Most population health effect models define Markov states and calculate annual proportions of users in each state. Our work uses quasi-Monte Carlo simulation of individual tobacco use histories across a large population. An advantage of this approach is that in simulating individuals, the number of states considered is no longer a limitation. Also, quasi-Monte Carlo numbers provide stable results from a much lower number of simulated individuals than traditional Monte Carlo simulation.

Future research needed

While modeling approaches provide a conceptual framework to evaluate the risks and benefits of e-cigs, much more work is needed. Due to the large uncertainty in model input, there is a wide distribution of possible net health impacts of e-cigs from positive to negative. To fully evaluate their net health effects, we need several things. Most importantly, we need better estimates and ranges of e-cig transition rates. Snus, a Swedish smokeless tobacco alternative, has been used for more than 40 years and provides insight into the potential of e-cigs to reduce smoking-related harm. Next, mortality rates should be adjusted to account for cigarette use history. While smokers who use e-cigs reduce their smoking, they are generally slow to quit completely. Finally, we should build models that predict the impact of e-cigs and cigarettes on morbidity and quality of life as well as mortality, since these are important and relatively responsive health measures.

All information presented derive from public source materials.

Ready to learn more?

I recently spoke on the promise and peril of e-cigs at the INFORMS Annual meeting here are the slides.

I presented joint work in a webinar, “Population modeling of modified risk tobacco products.” I hope that you’ll watch it and let me know what you think!

Bill Poland

About the Author

Dr. Bill Poland is a Vice President and lead scientist at Certara. He has provided pharmaceutical companies guidance in drug development decisions through scientific and decision-analytic modeling and simulation since he joined Certara in 1998. In over 40 projects for top pharmaceutical companies, he has advised on trial and program design for HCV, HIV, and other therapeutic areas, using integrated treatment adherence, pharmacokinetic, pharmacodynamic, and trial models.