Routine general health checks have been evaluated in trials with results that surprise most people.
The trial evidence
General health checks showing little effect on mortality.
Which systematic reviews have found repeatedly.
Why the intuition fails
Finding things not being the same as improving outcomes.
Which is the same logic as overdiagnosis in screening.
Targeted screening
Specific programmes with demonstrated benefit.
Which is a different question.
Private health screening
Extensive testing marketed to individuals.
Which produces incidental findings requiring investigation.
Why the trial results are counterintuitive
General health checks find abnormalities, and finding abnormalities feels like it must help, and randomised trials have not shown a mortality benefit.
Which is because many findings would never have caused harm, investigation of them carries its own risk, and the people most likely to attend are the least likely to benefit.
Targeted screening for specific conditions in specific age groups is a different question with different and sometimes positive evidence.
Incidental findings
Extensive imaging producing findings requiring follow-up.
Which is the main harm from private whole-body screening.
What is worth doing
Participating in evidence-based national programmes.
Which are selected on trial evidence.
Blood pressure
Opportunistic measurement with clear benefit.
A general note
Discuss screening decisions with a clinician who knows your history.
Why reading health evidence properly is worth the effort
Health is the area where people most often act on what they read and where the reporting is least reliable. That is not primarily because anyone is dishonest; it is because a chain running from researcher to press office to journalist to headline writer amplifies at every step, and the incentives at each stage push in the same direction.
The result is a public information environment where genuinely useful findings sit alongside overstated ones, presented identically. The difference between them is visible to anyone willing to ask four questions, and invisible to anyone who is not.
The four questions
What was actually studied, in whom, compared against what, and how large was the effect in absolute terms. Those four dispose of the majority of misleading health coverage, and none of them requires any scientific training to ask.
A study in mice is not a finding about people. An observational association is not evidence of cause. A thirty percent increase in a one in ten thousand risk is a small absolute change. A comparison against doing nothing tells you almost nothing about whether an intervention beats the alternatives.
What actually has good evidence behind it
The interventions with the strongest and most consistent support across health research are unglamorous and mostly free: not smoking, regular physical activity, adequate sleep, a varied diet without excessive processed food, moderate or no alcohol, and social connection.
None of these can be packaged and sold at a margin, which is a substantial part of why they receive less attention than products that can. The evidence for them is stronger than for anything in the wellness aisle, and it has been stable for decades.
A general note
Nothing here is medical advice. It describes how evidence in these areas is produced and reported. Anything concerning your own health belongs with a qualified professional who knows your circumstances, and guidance differs between countries.
Why the boring answer keeps being right
The interventions with good evidence have been tested repeatedly, in different populations, using different methods, over long periods. That is a high bar and very little clears it.
Anything new arriving with dramatic claims and thin evidence should be read against that background rather than in isolation. Most of it will not replicate, which is not cynicism but the base rate in health research. The findings that survive tend to be modest, unexciting and durable.
A note on uncertainty
A great deal of health science is genuinely uncertain, and that uncertainty is real rather than a failure of the researchers. Studying what people eat, how they sleep and how they live over decades is extremely difficult, and the methods available are imperfect.
Coverage that presents findings as settled when they are not does more damage than the uncertainty itself, because it produces the impression that the science keeps reversing when what is actually happening is that preliminary results were reported as conclusions.
Where to look instead
National health services and systematic review organisations publish plain-language evidence summaries, free and without anything to sell. Clinical guidelines publish their reasoning and their assessment of how certain the evidence is.
All of it is less interesting than headlines and considerably more likely to still be accurate next year.
The practical position
Do the things with good evidence, be sceptical of anything requiring a purchase, take symptoms to a clinician rather than to a search engine, and treat any single study as a data point rather than as news.
That is a short list, it is unexciting, and it is what the accumulated evidence actually supports.