Longevity Basics, Part 2: The Study Behind the “12–14 More Years of Life” Claim

Part 1 of this series covered a 2009 study linking four basic habits to a 78% lower risk of chronic disease. That's a striking number, but it's also a little abstract. A percentage reduction in risk doesn't tell you what it means for your actual life. A follow-up question naturally arises: if these habits lower your disease risk this much, does that translate into more years lived?

In 2018, a team at Harvard set out to answer exactly that question, and their answer is where the “12 to 14 more years” figure comes from.

This wasn't a study that followed people for 40 years and watched them live out full lifespans; no study could run that long and still be useful today. Instead, researchers used a statistical technique called life-table modeling.

Here's the idea in plain terms: using data on who died and when during the follow-up period, researchers calculated each group's mortality rate at every age. They then used those rates to project forward, essentially asking, “if a 50-year-old with this risk profile continued to experience mortality rates like the people we observed, how many more years would we expect them to live, on average?”

This is the same general method actuaries use to build life insurance tables and the same method national statistics agencies use to calculate standard life expectancy figures. It's a well-established technique, but it's worth understanding that the “14 years” isn't a number anyone actually observed. It's a population-level projection built from observed mortality patterns.

Researchers analyzed data from two of the longest-running health studies in the U.S.: the Nurses’ Health Study (78,865 women) and the Health Professionals Follow-up Study (44,354 men). Both cohorts have tracked health professionals for decades, making them a well-characterized, if not perfectly representative, slice of the U.S. population.

Five factors were scored:

Four of these five overlap closely with the habits covered in Part 1 of this series. The activity and smoking factors are nearly identical; diet quality and BMI are measured somewhat more precisely here than in the 2009 study. The fifth factor, moderate alcohol, is new, and it's worth its own section.

At age 50, women meeting all five low-risk factors were estimated to have 43.1 remaining years of life, compared with 29.0 years among women meeting none of the factors, a difference of 14.0 years. For men, the estimates were 37.6 years versus 25.5 years, a difference of 12.2 years.

Put another way: a 50-year-old woman meeting none of the five factors would be projected to live to about 79. A 50-year-old woman meeting all five would be projected to live to roughly 93. For men, the comparable figures are about 75½ versus 87½.

As with Part 1's findings, benefit didn't require perfection. The researchers found a graded relationship. Each additional low-risk factor was associated with progressively lower mortality and longer projected life expectancy, not just the full set of five.

Scientific study: Li et al., Circulation, 2018 — full text

The 2018 study counted moderate alcohol consumption as one of its five low-risk factors. At the time, this reflected a genuine, if contested, reading of the evidence: earlier observational research had repeatedly associated light-to-moderate drinking with lower cardiovascular risk, and some of that research shaped how “low-risk lifestyle” was defined in studies like this one.

That evidence base has shifted substantially since 2018.

The World Health Organization now states that no safe level of alcohol consumption can be established for cancer risk and classifies alcohol as a Group 1 carcinogen, the same category as tobacco and asbestos in terms of strength of evidence, though not magnitude of risk. Much of the earlier “moderate drinking is protective” research has also faced a specific methodological critique: some comparison groups included former drinkers who had quit due to existing health problems, which could make lifelong abstainers look artificially less healthy by comparison, a bias sometimes called the “sick quitter” effect.

None of this means the 2018 study's overall finding is wrong. The other four factors carry the bulk of the effect, and the researchers were working from the best evidence available at the time. But it does mean the alcohol factor specifically shouldn't be read as a modern recommendation to drink for longevity.

Updated evidence: World Health Organization — Alcohol and Cancer

Unlike Part 1's general-population German cohort, this study specifically followed nurses and other health professionals in the U.S. That has trade-offs in both directions.

On one hand, health professionals tend to have better healthcare access and health literacy than the general population, which could make the “none of the five factors” comparison group healthier than an equivalent group of non-professionals, potentially understating the true gap. On the other hand, studying a relatively homogeneous, well-characterized population made it easier to measure the five factors accurately and follow participants reliably for decades, which is part of what makes the underlying data unusually strong.

The researchers' own interest in this population had a specific angle, too: understanding whether a healthy lifestyle could meaningfully close the life expectancy gap between the U.S. and other high-income countries, where Americans have historically trailed.

The five factors in this study overlap almost entirely with the four from Part 1, which means the same starting point applies.

They're related but different findings from different studies. Part 1's 78% figure describes disease risk reduction over about 8 years in a German cohort. This study projects lifetime years gained, using U.S. health-professional data and a different statistical method (life-table modeling).

That's a personal decision best made with your own healthcare provider, factoring in your individual risk profile. What the current evidence does support is not treating moderate drinking as a health-promoting habit the way the original study's checklist implied.

The Nurses’ Health Study and Health Professionals Follow-up Study are among the longest-running, most rigorously tracked cohorts available, which made it possible to measure the five factors accurately over decades. The trade-off is that health professionals may not perfectly represent the general population.

No. It's a population-level statistical projection based on observed mortality patterns, not a guarantee for any individual. Genetics, healthcare access, and other unmeasured factors all still play a role.

Researchers used an established dietary index that scores overall eating patterns—things like intake of fruits, vegetables, whole grains, and healthy fats versus red/processed meat and refined carbohydrates—rather than any single food or nutrient.

Li, Y., Pan, A., Wang, D. D., Liu, X., Dhana, K., Franco, O. H., Kaptoge, S., Di Angelantonio, E., Stampfer, M., Willett, W. C., & Hu, F. B. (2018). Impact of healthy lifestyle factors on life expectancies in the US population. Circulation, 138(4), 345–355.

World Health Organization Regional Office for Europe. (2025, November 26). Alcohol and cancer.