Some health advice won’t fit your body, and some exceptions are just excuses

Key takeaways

• In a 2023 trial of 213 adults, about 73% had lower blood pressure after a week of very low sodium than after a high-sodium week; about a quarter did not, though diet slips and normal swings can blur any one person’s result¹.
• People vary widely in how sensitive they are to caffeine and how fast they clear it, the FDA says, and an expert presentation to Europe’s food safety agency put caffeine’s half-life at up to 16 hours as pregnancy goes on²³.
• About 68% of people worldwide have trouble digesting lactose, yet many can handle about 12 grams, roughly a cup of milk, with mild or no symptoms⁴.
• In a 2017 study, adults who seemed not to respond to exercise all improved once they added two more workouts a week⁵.
• Some exceptions are excuses: smokers tend to rate their own lung cancer risk below the average smoker’s, yet smoking raises that risk 15 to 30 times⁶⁷.

In a trial published in 2023, 213 adults aged 50 to 75 in Chicago and Birmingham, Alabama, ate two very different weeks of food¹. In one week, they added about 2,200 mg of sodium a day to their usual diets. In the other, they ate a diet designed to supply only about 500 mg a day in total; urine tests showed they took in more than planned, though far less than in the high-sodium week¹.

For most people, the result matched the standard advice. About 73% had lower mean arterial pressure, a blend of the top and bottom numbers, in the low-sodium week¹. In the first week, people who started on low sodium had systolic pressure (the top number) about 8 points lower than people who started on high sodium¹.

The effect showed up whether or not people had high blood pressure or took medicine for it¹. But about 26% had higher pressure in the low-sodium week, and about 5% saw it climb by 7 points or more¹.

The authors offered several possible reasons. A few people may truly respond in reverse, some may not have stuck to the diet, and blood pressure naturally drifts from day to day¹.

That is the trouble with every average. It describes a group well and any one person only roughly. This article explains why health advice rests on averages, where people really differ, and how to tell a real exception from an excuse.

Advice is built for groups, and you are one person

Randomized trials, in which people are assigned by chance to one treatment or another, report the average effect across everyone. A methods guide from the federal Agency for Healthcare Research and Quality (AHRQ) points out that this average may not apply to any single patient⁸. Some people benefit more than average, and some benefit less⁸.

Researchers call this spread “heterogeneity of treatment effect.” In plain words, it is the range of individual responses hiding inside one average result⁸. The AHRQ guide says no one knows how wide that range is for most common treatments, though some studies hint that it is large⁸.

Even nutrient targets are built around variation. The Recommended Dietary Allowance (RDA) for a vitamin or mineral is set to meet the needs of about 97% to 98% of healthy people⁹. A lower figure, the Estimated Average Requirement, covers only half⁹.

So the RDA sits above what most people need, on purpose. These values are also meant for healthy people, which is itself a built-in exception⁹.

Many natural-health readers have long insisted that bodies differ. The science agrees. What it adds is a way to measure how much, and that measurement is what separates a real exception from a hunch.

Salt sensitivity is a sliding scale, not a club

The American Heart Association (AHA) issued a scientific statement on salt sensitivity in 2016. An AHA commentary on it, citing earlier research, estimated that about 51% of people with high blood pressure and 26% of people with normal blood pressure are salt sensitive, meaning their pressure rises noticeably when they eat more salt¹⁰.

There is no clean line between sensitive and not. The statement describes salt sensitivity as a continuous trait that varies by degree across the population, so sorting people into “sensitive” and “resistant” depends on cutoffs researchers choose¹¹. It tends to be more common in Black people, older adults, and people with higher blood pressure, chronic kidney disease, diabetes or metabolic syndrome¹⁰.

You can’t easily test yourself. Measuring salt sensitivity takes carefully controlled high-salt and low-salt diet periods that are slow, costly and mostly limited to research¹⁰¹¹. Researchers haven’t settled on one standard definition¹¹.

There’s also a reason not to shrug it off. Salt sensitivity predicts heart and blood vessel events and deaths regardless of a person’s blood pressure level, according to the AHA commentary¹⁰.

Back to the 2023 trial. Using a 5-point cutoff, the team classified 46% of participants as salt sensitive, and they suggested that number may undercount how many people benefit from eating less sodium¹. One week in which your pressure didn’t drop is not proof that salt doesn’t affect you.

Caffeine: same cup, very different bodies

The FDA has cited 400 mg of caffeine a day, about two to three 12-ounce cups of coffee, as an amount not generally linked to negative effects in healthy adults². The same page warns that people vary widely in how sensitive they are and how quickly they clear it from the body². Some health conditions and medicines make people more sensitive².

The numbers behind that warning are large. A presentation at a 2015 European Food Safety Authority meeting put caffeine’s half-life, the time the body takes to clear half of a dose, at about 2 to 8 hours in adults³. A systematic review posted in 2023 put it at about 4 to 6 hours in most adults and noted that one liver enzyme, called CYP1A2, handles more than 95% of the work¹².

Hormones change the picture. According to the 2015 presentation, estrogen and progestin hormones, like those in birth control pills, slow that enzyme, and as pregnancy goes on, caffeine’s half-life can stretch to 16 hours, three to four times longer than in women who aren’t pregnant³. The FDA advises talking with a health care provider about caffeine if you are pregnant, trying to become pregnant or breastfeeding².

What about gene tests that promise to label you a “fast” or “slow” caffeine metabolizer? That review of 22 studies found that variations in two genes, CYP1A2 and ADORA2A, were linked to caffeine’s effects on thinking, anxiety and sleep, but the specific gene versions that matter for each effect are not yet clear¹². The 2015 presentation added that a widely tested CYP1A2 variant affects how readily the enzyme can be revved up, rather than directly making someone a fast or slow metabolizer³.

Your own records are more direct. If an afternoon coffee keeps you awake, a two-week sleep diary with an earlier cutoff time will tell you more than a gene report.

Most of the world digests milk differently

Nearly everyone is born able to digest lactose, the sugar in milk. For most people, the enzyme that breaks it down, lactase, drops after infancy⁴. The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) estimates that about 68% of the world’s population has lactose malabsorption, along with about 36% of people in the United States⁴.

It is most common in Africa and Asia and less common in northern Europe⁴. In the United States, it is more common among Black, Hispanic or Latino, Asian American and American Indian people, and lactose intolerance is least common among people with European roots⁴.

Advice built around a daily glass of milk may not fit you, but that doesn’t mean dairy is off the table. NIDDK uses the term “lactose intolerant” only for people who get symptoms, and it says many can handle about 12 grams of lactose, roughly one cup of milk, with no or mild symptoms⁴. Yogurt and hard cheeses carry less lactose, and lactose-free milk products are widely sold⁴.

If you avoid dairy, plan for calcium from foods such as canned fish with soft bones, leafy greens, tofu and fortified products, and for vitamin D from sources such as salmon, eggs and fortified foods⁴. Digestive symptoms can have many causes, so bring persistent ones to a clinician, who may order a hydrogen breath test⁴.

Exercise “non-responders” may need a different dose

The HERITAGE Family Study has produced more than 180 research papers on how people respond to exercise¹⁴. In a 2022 analysis, 564 adults completed a 20-week endurance training program, and researchers tracked seven heart and metabolic measures, including fitness, body fat, insulin and HDL cholesterol¹³.

Responses varied for every measure, and people who responded strongly on one measure often didn’t on others¹³. Only one person responded at the extreme, all high or all low, across all seven¹³. Only 2.5% landed in the average range on every measure¹³.

That sounds like proof that some people simply don’t respond. Other researchers urge caution. A 2017 critical review in Sports Medicine noted that most studies of individual fitness responses had no untrained comparison group, so no one could tell how much of the spread was real and how much was ordinary random variation¹⁴. In the one older study that had a comparison group, the spread in the untrained group was larger¹⁴.

A 2017 study tested the dose question directly. Researchers assigned 78 healthy adults to one to five hour-long endurance sessions a week for six weeks⁵. Apparent non-response was 69% with one session a week, 29% with three and zero with four or five⁵. When non-responders added two sessions a week for another six weeks, every one of them improved⁵.

The lesson for you: if a fair trial of a plan doesn’t move the numbers, change the amount or the type before you conclude that exercise doesn’t work for your body. The study was small, so treat it as a strong hint rather than a final word.

Medicines are where genes matter most clearly

For some prescription drugs, genetic differences are strong enough to appear in the official label. The FDA keeps a table of drugs whose labeling includes this kind of genetic information¹⁵. For the heart drug clopidogrel and the pain medicine codeine, that information appears in the label’s boxed warning¹⁵.

Don’t stop or change a prescription on your own because of a hunch or a home DNA report. Ask your prescriber or pharmacist whether testing makes sense. Our article on consumer DNA kits explains what those tests can and can’t read.

When “that doesn’t apply to me” is an excuse

Some risks are large and consistent enough that personal exceptions are rare. Smoking is the clearest case. According to CDC, smoking cigarettes makes a person 15 to 30 times more likely to develop lung cancer or to die of it, compared with not smoking⁷.

Yet many smokers believe the odds don’t apply to them. In a national telephone survey of 6,369 Americans, including 1,245 current smokers, researchers writing in the journal Tobacco Control found that smokers judged their own lung cancer risk to be lower than the average smoker’s⁶. Their sense of risk barely rose with the number of cigarettes they smoked⁶. More than half of smokers and former smokers agreed that exercise undoes most of smoking’s effects⁶.

You’ve probably heard someone say their grandfather smoked and lived to 95. Some people beat long odds, but health advice describes odds, not guarantees, and a lucky few don’t change your chances. Our article on absolute risk shows how to read those odds.

Exception or excuse? A sorting guide

Use this guide whenever you suspect a piece of advice doesn’t fit you. The more answers land in the middle column, the more likely you’ve found a real exception.

Question	Points toward a real exception	Points toward an excuse
Does the guidance itself name people like you?	Yes: pregnancy, a named condition, a medicine or an age group	No, and the risk is broad and large, as with smoking
Is this a trait known to vary between people?	Yes: salt, caffeine, lactose, exercise dose	The effect is large and consistent across people
Did you measure, or just notice?	Repeated readings, a diary, a baseline	One reading, one day, a feeling
Did you test fairly?	One change at a time, for a set period, at a dose the evidence supports	Several changes at once, or stopped early
What happens if you’re wrong?	Little: an extra cup of coffee or a glass of milk	A lot: blood pressure, heart disease or cancer risk
Would your records persuade a clinician?	Yes, and you’ve shared them	You haven’t shown anyone

If your answers point toward a real exception, test it properly. Our self-trial guide walks through a baseline, one change at a time, a daily score and a stop rule. For anything involving blood pressure, the heart, pregnancy or prescription drugs, plan the test with your clinician.

What the evidence shows

Claim	Evidence rating	What it rests on
Eating less sodium lowers blood pressure on average, with or without high blood pressure	Strong evidence	Randomized crossover trial; similar effect across blood pressure groups1
Everyone’s blood pressure responds to salt the same way	Evidence says no, or unsafe	About half of people with high blood pressure and a quarter without are salt sensitive10
You can reliably find out at home whether you’re salt sensitive	Not enough evidence	Testing needs controlled research diets; no standard definition1011
People differ widely in caffeine sensitivity and how fast they clear it	Strong evidence	FDA; half-life of about 2 to 8 hours, up to 16 in late pregnancy23
Caffeine-related gene variants shape how caffeine affects your sleep, anxiety or thinking	Promising but early	22-study review found links, but the specific gene versions that matter are not yet clear12
People with lactose malabsorption must avoid all dairy	Evidence says no, or unsafe	Many tolerate about 12 grams of lactose4
Some healthy adults can’t improve their fitness with exercise	Not enough evidence	Responses vary at a fixed dose, but most studies lacked comparison groups, and in one small trial every apparent non-responder improved with more training51314
Genes change how some prescription drugs work	Strong evidence	FDA drug labels, including boxed warnings for specific drugs15
Smoking’s lung cancer risk doesn’t apply to some people	Evidence says no, or unsafe	15 to 30 times the risk of getting or dying from lung cancer7

Questions readers ask

The bottom line

Health advice rests on averages because that’s what studies measure. Averages hide real differences in how people respond to salt, caffeine, lactose, exercise and some medicines, and those differences can matter for you.

The test of a real exception is evidence, not preference. Check whether the guidance names people like you, measure your own response over time, test one change fairly, and bring your records to a clinician. When the stakes are high and the risk is large and consistent, as with smoking, assume the advice applies to you.

Related reading

• Feeling better after trying something isn’t a test, but you can run one
• Your response to food is personal, but the payoff from diet algorithms is modest
• DNA kits read a sliver of your genes, and the data can be sold
• A 22% risk cut can mean helping 2 or 3 people in 1,000
• Wellness advice is written for an average person who doesn’t exist

Sources

1. Gupta DK, Lewis CE, Varady KA, et al. “Effect of Dietary Sodium on Blood Pressure: A Crossover Trial.” JAMA, published online November 11, 2023. pubmed.ncbi.nlm.nih.gov
2. U.S. Food and Drug Administration. “Spilling the Beans: How Much Caffeine Is Too Much?” FDA Consumer Update, current as of August 28, 2024. fda.gov
3. Gundert-Remy U. “Pharmacokinetic and Pharmacodynamic Effects of Caffeine.” Presentation, European Food Safety Authority Stakeholder Meeting on the Safety of Caffeine, Brussels, March 5, 2015. efsa.europa.eu
4. National Institute of Diabetes and Digestive and Kidney Diseases. “Lactose Intolerance.” NIDDK, last reviewed February 2018. niddk.nih.gov
5. Montero D, Lundby C. “Refuting the Myth of Non-Response to Exercise Training: ‘Non-Responders’ Do Respond to Higher Dose of Training.” Journal of Physiology, 2017. pubmed.ncbi.nlm.nih.gov
6. Weinstein ND, Marcus SE, Moser RP. “Smokers’ Unrealistic Optimism About Their Risk.” Tobacco Control, vol. 14, no. 1. pubmed.ncbi.nlm.nih.gov
7. Centers for Disease Control and Prevention. “Lung Cancer Risk Factors.” CDC, May 18, 2026. cdc.gov
8. Kravitz RL, Duan N, eds. “Design and Implementation of N-of-1 Trials: A User’s Guide.” Agency for Healthcare Research and Quality, Publication No. 13(14)-EHC122-EF, February 2014.
9. NIH Office of Dietary Supplements. “Nutrient Recommendations and Databases.” ODS, accessed October 2026. ods.od.nih.gov
10. Carey RM. “Salt Sensitivity of Blood Pressure: An Enigmatic Hypertension Phenotype” (commentary on the AHA scientific statement by Elijovich et al., Hypertension, 2016). American Heart Association Professional Heart Daily, July 21, 2016. professional.heart.org
11. American Heart Association. “Top Things to Know: Salt Sensitivity of Blood Pressure.” AHA Professional Heart Daily, July 21, 2016. professional.heart.org
12. Kapellou A, King A, Graham CAM, Pilic L, Mavrommatis Y. “Genetics of Caffeine and Brain-Related Outcomes: A Systematic Review of Observational Studies and Randomised Trials.” Nutrition Reviews (accepted manuscript, St Mary’s University repository, deposited April 2023). research.stmarys.ac.uk
13. Barber JL, Ruiz-Ramie JJ, Robbins JM, et al. “Regular Exercise and Patterns of Response Across Multiple Cardiometabolic Traits: The HERITAGE Family Study.” British Journal of Sports Medicine, 2022;56(2):95-100. profiles.wustl.edu
14. Williamson PJ, Atkinson G, Batterham AM. “Inter-Individual Responses of Maximal Oxygen Uptake to Exercise Training: A Critical Review.” Sports Medicine, 2017. research.tees.ac.uk
15. U.S. Food and Drug Administration. “Table of Pharmacogenomic Biomarkers in Drug Labeling.” FDA, updated August 12, 2026. fda.gov