Research

Making Sense of the Next Generation of Tobacco-Related Health Care Costs

June 2018Andrew Schwartz

No one seriously disputes the dangers of tobacco use anymore. Its role in countless early deaths, negative impact on quality of life and dollars-and-cents societal costs are widely documented. That’s why tobacco companies have doled out billions in class action lawsuits, while facing increasingly aggressive tobacco control efforts by federal, state and local entities. Those tobacco control efforts have seen some real successes.

As one example, a paper published in 2010 in Tobacco Control found that for the 14.7 million male California residents alive in 1990, the California Tobacco Control Program prevented the early deaths of hundreds of thousands of them, with total savings of as much as $107 billion in 1990 dollars, including net health care savings and years of life saved.

One of that paper’s lead authors was Wendy Max, director of the UCSF Institute for Health & Aging, part of the UCSF School of Nursing. For decades, Max has led a research team that was one of the first to develop econometric models of smoking-attributable costs – models that continue to evolve and remain enormously influential today. (Econometrics is the application of statistical methods to economic data.) 

With money sometimes serving as the most powerful driver of change, juries have used the Max team’s models to determine the amount of damages tobacco companies must pay. And legislatures across the United States and around the world rely on the models to decide if and how they will invest in tobacco control programs.

Now, however, the team is facing new challenges. Tobacco companies continue to create and market new products – often targeting children and people in communities where tobacco control has more shallow roots. To help policymakers and insurance companies make informed decisions about how to shape more effective policies, Max and her team are now tasked with updating their models with everything from emerging data on the costs of secondhand smoke (SHS) and e-cigarettes through concerns about the possible impact of legalized marijuana.  

A Collaborative, High-Powered Team

Understanding the societal costs of tobacco use has always been an extremely complex challenge, one that requires a team of researchers who create and refine models that rely on massive data sets accumulated over time, growing scientific understanding of the role tobacco plays in various disease states and honing evidence-based assumptions about smoking use and risk behavior. Much as Max collaborated with and was mentored by the late Dorothy Rice – one of the few true pioneers in assessing the costs of health behaviors and related diseases – Max has mentored and collaborated with numerous PhD-trained economists who have become integral members of her research team.

The group now includes longtime Max collaborator Hai-Yen Sung, along with Tingting Yao and Yingning Wang. To help policymakers and payers (Medicare, Medicaid and private insurers) understand the costs associated with tobacco use, Max and her team often divide up the work into categories and then synthesize their findings to answer the larger questions.

Sung, for example, has focused recently on economic evaluations of taxation and tobacco control policies as well as on the association between mental illness and smoking. Part of Yao’s current research estimates the impact of price on the demand for electronic cigarettes in the U.S., as well as on the health care costs of SHS exposure for adults and children. Part of Wang’s work examines nondaily smokers, poly-tobacco use and the health care costs attributable to smokeless tobacco and cigars.

Defending the Data

Their quest to continually refine the econometric models is driven by a commitment to facts and science, as well as by the pressures of having to defend their methods, especially in an era where people are more willing to question academic expertise.

Even during the original tobacco litigation, says Max, “Our statistics were on trial, but we were able to show how we could control for [other factors] and reliably determine the percentage of health costs attributable to smoking behaviors.”

That’s why, says Sung, “When we present our results, we always describe our model.”

They painstakingly demonstrate such things as how they arrive at the health care costs for each type of service, how they arrive at a conceptual framework for how cigarette smoking (versus other potential disease drivers) will influence the likelihood of an individual getting a smoking-related disease, how smoking and smoking-related disease influence health status and how smoking and poor health status impact health care costs. All of those factors and more contribute to the econometric models, which are rooted in one originally created by Leonard Miller, an econometrician who did the bulk of his work at UC Berkeley and was a frequent Max collaborator.

Max says that by consistently holding themselves accountable to the standards of scientific validity Miller set, she and her team have been able to continually improve the models over the last 20 years – and that the evidence over time supports the models’ reliability. That helps Max and her team provide an answer when tobacco companies question the numbers or when tobacco control advocates – who need the data to support their efforts – press for numbers that aren’t yet backed by data.

“I would argue the only way to see the effect of smoking on health is to look at huge data sets, tens of thousands of people,” says Max. And while the cigarette smoking models now have the requisite data sets and enough science behind them to gain wide acceptance, to assess the true costs of tobacco use and smoking in society, the models now must incorporate all of the other emerging pieces of the puzzle.

Today’s Studies

The team is doing just that as part of the Tobacco Centers of Regulatory Science (TCORS) – one of 14 federal centers in the country – with a grant funded by the National Cancer Institute and the Food and Drug Administration (FDA). Specifically, the current work aims to help the FDA, which now has expanded abilities to regulate tobacco products, understand the potential impact on population health of proposed regulatory actions. “Our models have to integrate how the patterns of tobacco use have changed in recent years,” says Max.

Wang, for example, recently found considerable costs associated with smokeless tobacco ($3.4 billion per year in 2014 dollars) and sole and poly cigar smoking ($1.75 billion per year in 2014 dollars). “It’s less than cigarette smoking but still quite significant,” she says.  

An understanding of the costs of SHS is also emerging from various studies the team members have completed together. They estimated the health care costs attributable to SHS exposure at home for children and adults in both the U.S. and rural China. Their studies found that SHS exposure-attributable costs result in a large economic burden in both countries.

“But we haven’t done any large-scale cost models yet for e-cigarettes even though they have a huge market share, because we don’t yet have the needed national and state data,” says Max.

Janine Cataldo “With e-cigarettes, it’s very complicated, because people take one puff or five, so it’s hard to know how to measure use,” says Yao. The team did, however, publish a paper in late 2017, in PLoS One – along with the School’s Janine Cataldo – that used e-cigarette spending as a proxy for e-cigarette use to examine the relationship between e-cigarette use and increases in smoking-related symptoms. 

The FDA and various state legislatures also look to Max and her team to clarify the impact of taxation – and, therefore, prices – on tobacco use. In 2016, for example, as Californians prepared to vote on a proposed $2-per-pack increase, Max and Sung estimated its potential impact and found that it would reduce adult smoking prevalence in California from over 9 percent to about 7 percent by 2020, save more than $4 billion in health care expenditures between 2017 and 2020 because fewer people would get sick – and even create jobs, because people would have more disposable income to pump up the California economy.

“We have learned a lot about how people respond to price changes – how for some goods, there is a huge response,” says Max. In the case of cigarettes, she adds, when the tax rate goes up, the tax revenue money also goes to prevention efforts, and the effect is, in a sense, doubled.

The Policy Impact

Valerie Yerger In a recent paper published in Nicotine & Tobacco Research, the team – along with the School’s Valerie Yerger – found that among African American adults, menthol cigarette smokers were more likely to try to quit than nonmenthol smokers, but these quit attempts did not translate into successful cessation. In addition, Yao spoke at a gathering of the San Francisco Chinese community to discuss the impact of a flavor ban in San Francisco – the measure passed on Tuesday, June 5 – using estimates of smoking-attributable costs for California’s Chinese community that the team had prepared. 

Similarly, Max and Sung and collaborators from other universities recently completed a five-year, Centers for Medicare & Medicaid Services (CMS)-funded study that explored how financial incentives could help Medicaid recipients – who smoke almost twice as much as the rest of the population – quit the habit. “Our economic analysis found that the benefits of a more thorough smoking cessation intervention would outweigh the costs within 10 years,” says Sung.  

“These types of studies are also important for health insurers, who have large risk pools where they don’t ask about smoking history, and they need to understand if there will be a payoff for their wellness programs,” says Max.

The Collaborations

From left: Benjamin Chaffee, Bonnie Halpern-Felsher, James Lightwood Refining the models also involves working with people who work in other disciplines that can shape the assumptions the researchers use.

In addition to the work with Cataldo and Yerger, the team is working with Benjamin Chaffee from the UCSF School of Dentistry to deepen understanding of what smokeless tobacco use implies for health care utilization and how new products and marketing, such as flavors, could affect the results.

“If we ban flavors in smokeless tobacco, what will the impact be on its use?” asks Wang. One key associated question is whether such a ban would delay young people picking up the habit, as some research has shown that preventing young people from beginning the habit can head off tobacco use later in life.

The team is also working with Stanford’s Bonnie Halpern-Felsher and economist James Lightwood of the UCSF School of Pharmacy to understand how people with different risk preferences weigh the costs and benefits of quitting. Such studies deepen the models, even as they offer more insight into how to craft prevention efforts.

Where to Go from Here

And the challenges keep coming. The introduction of legalized marijuana, for example, brings with it a host of unanswered questions about potential new health care costs. New tobacco products such as heat-not-burn products also change the potential landscape of tobacco use.

“Such changes keep the work endlessly fascinating”, says Max. “Every time we’ve felt like we got this thing under control, the world changes.”