Life insurance applicants aren’t exactly forthcoming about tobacco use. Cotinine tests are cumbersome (maybe even beatable), and “solutions” like voice or photo analysis only infer use. Now that Medical Data is here to clear the air, carriers may be surprised to learn just how durable tobacco’s impact really is.
As our hit rates for Medical Data have climbed to an average of 80%, we’re now able to find reality-based evidence for a history of nicotine use in more than 16% of hits. And more than 80% of this data is new information to carriers—it isn’t found in their other risk assessment methodologies and disclosures.
Those climbing hit rates enabled us to do more than just spot present or past use. By running our latest 42M-life mortality study through the Irix® Rules Engine, we identified 2.8 million lives with one or more nicotine-related codes or prescriptions. Then we conducted a painstaking analysis to show that their mortality risks remain elevated for at least seven years—long past the point where carriers have traditionally grouped previous tobacco users with those who never used tobacco.
In this brief Q&A, our study team of actuaries and clinicians shares new insights into tobacco risks that are still of vital interest to insurance carriers.
- Christian Shepley, FSA, MAAA, CERA, Actuary and Product Director—Christian leads research and development on our data interpretation products and proposed this study to test a hypothesis: That the mortality implications of tobacco use last far longer than the one- or two-year lookback periods that most carriers consider.
- Maayan Essak, Senior Actuarial Analyst—Maayan performed the bulk of the mortality analysis.
- Hayley Vrchota, PharmD, CTTS, Clinical Consultant— Hayley is a certified tobacco treatment specialist. She provides a clinical perspective that helps to explain why the tobacco mortality risk is so durable.
- Jenna Fariss, ASA, MAAA, Director and Actuary—Jenna explains that Medical Data brings to light a large number of codes revealing tobacco users or ex-users whose mortality is roughly double that of those with no tobacco related codes.
There’s a lot of excitement at IntelliScript about this study. Do any of you have a personal stake in the subject of smoking or nicotine dependence?
Jenna: I was born in Winston-Salem and grew up in Virginia, with tobacco fields all around. In elementary school we went on tours of cigarette factories. But one of my earliest memories is negotiating with my dad to get him to quit smoking.
Christian: Back when most of our mortality research was focused mainly on prescription histories, we could identify applicants taking smoking cessation drugs, but that only represented people who were trying to quit. We also had much lower incidence, around 3% with a seven-year prescription history, and that rate was even lower for carriers who chose to focus on tobacco treatment medications within the past one or two years. I had researched the long-term effects of smoking as well as the low success rates of quitting and suspected that any history of tobacco use would correlate with significant mortality risk. We didn’t have enough data to back up most of our theories back then, but we do today. This analysis validates our educated speculation and allows us to help carriers better understand those risks.
Hayley: I previously worked as a pharmacist in a tobacco treatment clinic providing care to individuals with varying interest in quitting. So I understand the difficulty not only in quitting, but remaining quit. Like Christian, I suspected risk may extend beyond recent evidence of tobacco-related prescription fills and medical codes.
What is the percentage of smokers in the general population at present?
Hayley: The CDC reported that 19% of the U.S. adult population used tobacco in any form in 2020. Of that group about 12.5% were cigarette smokers. The remainder is made up of people who vape, use various smokeless products, or who smoke pipes or cigars.
When we run Prescription Data and Medical Data on insurance applicants, how often do we discover a history of nicotine use?
Christian: We find that around 16.8% of hits—15.8% in 2022—have medical codes or a prescription history that indicates nicotine use. It’s important to note that’s over the full, seven-year lookback period, so we’re finding both current and past smokers. But what our study shows is that regardless of when those prescriptions were filled or when those codes were billed, the elevated mortality risk is significant.
The Surgeon General’s first report on smoking and health was published almost 60 years ago—and even it was based on an analysis of about 7,000 research papers dating back to the 1920s. How is it that we’re still learning new things about smoking and mortality?
Jenna: Our most recent major mortality study is the first that we’ve been able to correlate with Medical Data. And as we’ve added more data sources, we’re now getting hit rates of over 80% and seeing an average of about 200 codes per person. For the first time we have the reliable data we need to speak to elevated mortality that’s surprisingly durable.
What surprised you most in the data?
Maayan: For years I’ve heard that smoking doubles your relative mortality. So heading into this analysis I was expecting our numbers to be in the ballpark of 200%, but I was surprised just how close to 200% they really were. When looking at the population of people who had evidence of nicotine use, we saw a relative mortality of 212%. However, this includes people on cessation drugs and even those who merely reported a “history” of tobacco use. So as accurate as the rule of thumb is, “double relative mortality” is missing a lot of nuance, and it was fascinating to dig into that nuance.
Jenna: For me it was the double whammy that smoking not only increases the risk for certain diseases but makes so many diseases more serious as well. Some were obvious; you’d expect smoking to complicate asthma because it’s a lung disease. But it also makes a very common and manageable condition like diabetes much more severe and likely to onset.
What should this research mean for underwriting?
Hayley: For a variety of reasons, many clients chose to focus on tobacco use in the past one to two years. As a clinician though, I have wondered how much risk may still be present when evidence of tobacco use is over two years old, given the increased risk of various conditions in this setting. This analysis really highlights the extent to which mortality risk remains elevated over time within the data.
Christian: A history of tobacco use is as important as current tobacco use status.
Does this study make you more curious about other smoking or nicotine-related questions?
Christian: We know that young people are smoking cigarettes less, but research points to an increase in the use of nicotine vaping products. I would love to study the mortality impact of that trend. It’s affecting younger generations, but we just don’t have enough credible data at this point. I also think marijuana deserves focused research, especially as more states legalize it for recreational use.
Maayan: Yes, I’d love to do a parallel analysis, but focusing on cannabis codes instead of nicotine codes. That would help answer a lot of insurers’ questions about cannabis use risk and be more confident in their underwriting decisions.
To learn more about this groundbreaking research, read our article in ON THE RISK and watch our webinar: A New Look at Tobacco Risk: Revelations from Milliman’s 42M-life Mortality Study and Clinical Data Findings
