Longevity and Health: The Big Effects Are the Ones Nobody Sells

There is an unusual gap in health information between what is well established and what gets discussed.

The things with the largest effects on how long people live in good health have been known for decades, are unglamorous, and are broadly agreed on. The things that dominate attention — specific supplements, protocols, devices, dietary systems — mostly have small or uncertain effects.

This is not a conspiracy. It has a simple cause: the large-effect interventions cannot be sold. Nobody profits from telling you to sleep enough and walk more, so nobody spends money telling you. The result is an information environment weighted almost inversely to the size of the effects.

Worth being clear about scope: this is a general orientation, not medical advice, and anything specific to your situation belongs with a doctor who knows it.

What has the largest effects

A small number of factors account for most of the modifiable risk in most populations.

Not smoking. The single largest, by a wide margin, for anyone who does.

Regular physical activity. Consistently associated with lower risk across essentially every major cause of death. Two measures come up repeatedly in the research: cardiorespiratory fitness, and muscular strength. Both are trainable, which is what makes them interesting — unlike age or genetics, they are things you can change.

Sleep, in adequate quantity. Chronic short sleep is associated with a long list of poor outcomes. The mechanism is not fully mapped and the association is robust.

Not being severely deconditioned or severely overweight. The relationship here is more complicated than headline coverage suggests, and fitness appears to carry independent weight — but the extremes are clearly associated with worse outcomes.

Alcohol in moderation or not at all. The older idea that moderate drinking was protective has weakened considerably under better-controlled analysis, largely because early studies compared drinkers to abstainers who included people who had stopped due to illness.

That last point illustrates the general problem with this field, and it is worth a section of its own.

Why health findings are so often reversed

Most nutrition and lifestyle research is observational: it follows groups of people and looks for associations. This is often the only ethical or practical option — you cannot randomly assign people to smoke for thirty years.

Observational studies have a structural weakness. People who do a healthy thing tend to differ in many other ways: income, education, stress, access to care, and dozens of behaviours nobody measured. When a study finds that people who take a supplement live longer, the supplement is one of many candidate explanations, and often not the best one.

Researchers adjust statistically for the confounders they can measure. They cannot adjust for the ones they did not think of.

There is also reverse causation. People who are already becoming ill often change their behaviour — reducing exercise, losing appetite, stopping drinking. A study can then find that the changed behaviour predicts illness, when the illness caused the change.

Add publication pressure, which favours striking findings, and the base rate problem that most tested hypotheses are false, and you get a field where reversals are common and confident headlines are usually ahead of the evidence.

The practical filter: be most confident where many types of evidence agree — observational data, trials, plausible mechanism, and consistency across populations. Smoking clears that bar comprehensively. Most individual nutrients do not.

Sorting advice sensibly

Two questions handle almost everything.

How strong is the evidence? A randomised trial with a hard outcome beats an observational study, which beats a mechanism argument, which beats a testimonial. Most claims marketed directly to consumers rest on the weakest tiers.

How large is the effect if true? Many real findings are real and tiny. A statistically significant difference in a marker is not the same as a meaningful difference in how long or well you live.

Things scoring well on both are worth doing. Things scoring well on neither are worth ignoring, however enthusiastic the presentation.

Three more points worth holding.

Dose matters, and more is not better. Many beneficial stressors follow the pattern described in hormesis: an upside-down U where a moderate dose helps and a large dose harms. Exercise, heat, cold, and fasting all appear to work this way, and the recovery period is part of the intervention rather than an interruption to it.

Consistency beats intensity. Because effects accumulate over years, a moderate routine sustained for a decade outperforms an intense one abandoned in six weeks. This is compound interest applied to physiology.

Beware advice derived from outliers. Studies of unusually long-lived people are prone to survivorship bias — the traits shared by centenarians tell you little unless you know how common those traits are among people who died earlier.

The uncomfortable summary is that the high-value actions are known, dull, and mostly free, and the interesting-sounding ones are mostly marginal. That is a poor basis for content and a reasonable basis for decisions.