Monday, November 29, 2010

Living an Extra Seven Years - Introduction

Part 1  Part 2  Part 3  Part 4  Part 5
In 1912, Dr. Isaac Alder suggested that there was a link between smoking and lung cancer. Immediately everyone stopped smoking. Actually, that’s not true at all. He was promptly and almost universally ignored. It took another four decades before anyone really listened. By 1957, the evidence had accumulated and the Surgeon General declared that smoking caused cancer. One year after this announcement, 44% of Americans believed smoking caused lung cancer. In 1964, “Smoking and Health” was published, and laws started getting passed restricting smoking, from advertisements to taxes (8). In 1968, 78% of Americans believed smoking caused lung cancer. Unfortunately, knowing something is good for you is not the same thing as doing what’s good for you.

For decades we didn’t know for sure if smoking was bad for you. Movie stars and cool kids pushed the habit, and when we found out it was bad, it was too late. People were already addicted. For those who did quit, severe damage was already done. The slowness of the research, the industry lobbying, and the lack of will on the part of the government and public cost countless lives.

Smoking remains to be one of the most important risk factors for heart disease and cancer. It’s not just hypothetical risk. Life insurance, an industry which bets on your life and death, is keenly aware of this risk and charges smokers much more because of this high risk. A smoker has a 70% higher annual risk of death than a nonsmoker (8). Smoking is the single best observable factor in determining your risk of dying. Or at least it was. What if I told you that there’s another behavior that is even better at tracking with lifespan than smoking is?

People who don’t demonstrate this behavior have an 87% higher annual risk of than those who do (9). A 20 year old who did engage in this behavior throughout his life would live 7 years longer than one who didn’t (for blacks, it’s even better: 14 years longer). You might be thinking you know what it is: exercise! Nope. Exercise is good, but not that good. Those who do what the American Medical Association tells them to do (1400 calories per week of exercise; 30 minutes daily) (2) have a 29% lower annual risk of death(1). What about losing weight? Those who are normal weight (BMI <24) have a 25% lower annual risk of death compared to those who are overweight (BMI >26) (1). What could this behavior be? What could be more important than what our doctors talk about when we visit? What’s more important than smoking? What’s more important than exercise? What’s more important than weight loss?

Church attendance. People who never go to church(3) have an 87% increased chance of dropping dead compared to those that attend more than weekly. Statistically speaking, never going to church is worse for your health than smoking a pack a day for the rest of your life (9). I’m not joking. This is an effect that has been seen in dozens of studies over the last thirty or so years. Scientists really don’t like it, but it’s there, and it’s real.

We have lots of evidence that church attendees have a lower risk of dying. We have so much data in these studies, we can even ‘pool’ it, adding up the results of many studies into one “meta” analytic study. The best meta study that did this found, after looking at a total of over 125,000 people in 42 separate investigations, the odds of a churchgoer surviving were better than a non-churchgoer (10).

But how can this be? Perhaps churchgoers are just goody-two-shoes who never smoke or drink or cuss. To evaluate that theory, I need to tell a story.

[Note: this is the first of a 5-part post. The references for all 5 will be posted here in the first post]


(1) Paffenbarger, RS; Hyde, RT; Wing, AL; Lee, IM; Jung, DL; Kampert, JB. The Association Of Changes In Physical-Activity Level And Other Life-Style Characteristics With Mortality Among Men. NEJM, 328 (8): 538-545 Feb 25 1993.
(2) Pate RR; Pratt M; Blair SN; Wilmore JH; et. al. Physical Activity and Public Health: A Recommendation From the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA. 1995;273(5):402-407.
(3) Most of the studies are representative US populations and so are literally measuring church attendance. Others look at ‘religious service attendance’ and find the same thing. I will use ‘church attendance’ in this article as these results do not have enough data to religious services of other faiths (though it is my belief that this would be true).
(4) Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there' – . Also, I am among those who are not happy about the monopoly that the Randomized Control Trial has on science. It’s a good technique, but not the only good technique.
(5) Sloan, RP. Should physicians prescribe religious activities? NEJM 342(25): 1913-1916 JUN 22 2000
(6) Powell, LH; Shahabi, L; Thoresen, CE. Religion and spirituality - Linkages to physical health. AMERICAN PSYCHOLOGIST, 58 (1): 36-52 JAN 2003
(7) I say supernatural not to give up the search, but to focus it. The effect might be mediated through an undiscovered non-classical mechanism; perhaps it is a hyper-dimensional effect, perhaps it is quantum mechanical. With regard to unsolved mysteries, we must always be ready for a paradigm shift to be hiding underneath it.
(8) “The Reports of the Surgeon General” <>
(9) Hummer RA; Rogers RG; Nam CB; Ellison CG. Religious Involvement and U.S. Adult Mortality. Demography 36(2): 273-285 May 1999
(10) McCullough ME, Hoyt WT, Larson DB, Koenig HG, Thoresen C. Religious involvement and mortality: a meta-analytic review. Health Psychol.19(3):211-22 2000 May
(11) FASTSTATS. Accessed on 10/23/2010. Centers for Disease Control. <>
(12) American FactFinder. Accessed on 10/23/2010. US Census Bureau. <>
(13) US Population: 310,546,653. US Mortality rate: 803.6 per 100,000.

(14) Mokdad AH; Marks JS; Stroup DF; Gerberding JL . Actual Causes of Death in the United States, 2000 JAMA. 2004;291:1238-1245
(15) This trend is supported by the meta-study (reference 10) but with a smaller magnitude. It reports that the increased chance of survival for those who attend is 25% after adjustment. The difference is probably because of a dose-effect in the Hummer paper (reference 9); they didn’t just measure attender/non-attender, but the frequency of attendance. Running 50 minutes every day is better for you than running 5 minutes, and so it seems that going to church more than weekly is better for you than going only on Christmas and Easter. The data in the Hummer paper does appear to agree with the more modest 25% if you ignore the dose-effect and pool all church attenders together.


  1. Hey David, look at the chart you posted. 0 preventable deaths in the increased churchgoers. Wow, that's one hell of a protective effect. Among less frequent and non-churchgoers, it is consistently about 1/6 of total deaths, but then COMPLETELY DISAPPEARS among heavy churchgoers. I bet going to church protects you from making [possibly intentionally misleading??] data entry errors as well.

  2. I probably explained this poorly. Or since it's a footnote, I didn't explain it at all. Thank you for the opportunity. And no, I do not think that there church attendance is protective against data entry errors :)

    The "Prev. Deaths" column is the number of deaths that would be prevented by changing a given group's Relative Risk (RR) to the lowest RR. In this case, that's the '> weekly' group which has an RR=1 (by definition). So, by definition, nothing could be prevented by *not* switching and so "Prev. Deaths" = 0 for the highly churched.