Why Cochrane says standard check-ups don't work — and what does instead

If you've ever had an annual check-up, you probably assumed it was making you healthier. It's an intuitive assumption: more screening must catch more problems, which must lead to better outcomes. The problem is, the largest body of evidence available says that's not true.

In 2019, the Cochrane Collaboration — the gold standard of evidence-based medicine — published a systematic review of general health checks. The authors included 17 randomized trials; the 15 that reported outcome data covered 251,891 adults. The results were striking.

General health checks did not reduce all-cause mortality. They did not reduce cardiovascular mortality. They did not reduce cancer mortality. What the authors did warn about is that systematic health checks may lead to unnecessary tests and treatments — the classic route to overdiagnosis.

"General health checks are unlikely to be beneficial." — Krogsbøll et al., Cochrane Database of Systematic Reviews, 2019

How can this be?

It seems counterintuitive. Surely finding problems early is better than finding them late. And for specific, targeted screenings, that's absolutely true. But general check-ups — the kind where a doctor orders a broad panel of tests without a clear clinical question — work differently than most people assume.

There are three main reasons.

1. The tests that matter are already targeted

Screening colonoscopy was associated with about 68% lower colorectal cancer mortality in large observational cohorts. Statins reduce major vascular events by about 22% per mmol/L of LDL-C reduction. Blood pressure treatment reduces events by 20% per 10mmHg reduction. These interventions work because they're precisely targeted at specific conditions in specific populations.

A general health check that throws a wide net rarely adds value to these already-proven targeted screenings. If you're 55, your targeted screenings are colonoscopy, lipid panel, and blood pressure. Adding twelve more tests doesn't improve your outcomes — it just increases the chance of a false alarm.

2. Overdiagnosis is a real harm

Overdiagnosis means finding a condition that would never have caused you problems. A common example is thyroid nodules: ultrasound finds them in roughly 20–70% of adults, and around nine in ten are benign (thyroid cancer is found in about 7–15% of nodules evaluated). If you scan everyone's thyroid, you'll find lots of nodules, some of which will trigger biopsies, anxiety, surgeries, and lifelong medication — all for a condition that was never going to hurt them.

3. Asymptomatic ≠ healthy, but the signal-to-noise is low

When you screen 10,000 asymptomatic people for a condition that affects 100 of them, you need an extraordinarily accurate test to avoid drowning in false positives. Most screening tests aren't that accurate. The math doesn't work unless you stratify by risk first.

So what does work?

Preventive medicine is not useless — it's one of the most powerful tools medicine has. The Cochrane review doesn't say prevention is futile. Its findings suggest that untargeted prevention doesn't pay off. The fix is risk stratification: identifying who is actually at elevated risk for specific conditions, and applying proven interventions to those people.

The evidence for targeted preventive medicine is overwhelming:

• Colorectal cancer screening (colonoscopy or FIT; from 45 in the US, from 50 in Germany): about 50% lower colorectal cancer mortality among those actually screened in the randomized NordICC trial; observational studies suggest up to about 68%. The effect of merely being invited was smaller.
• Statin therapy for high cardiovascular risk: about 22% fewer major vascular events per 1 mmol/L LDL-C lowered.
• Blood pressure treatment: 20% fewer cardiovascular events per 10mmHg lowered.
• Diabetes prevention through lifestyle (in prediabetics): 58% reduction in progression.
• Low-dose CT lung screening (high-risk smokers): 20% mortality reduction.
• Mammography (women 50–74): 20–30% breast cancer mortality reduction.

Every one of these works because it's applied to the right people at the right time. That's what evidence-based preventive medicine actually looks like.

What a better check-up should include

The lesson from Cochrane is not to skip the doctor, but to change what the visit is for. A check-up worth having starts with a question — "what is this person actually at risk of?" — rather than with a list of tests.

In practice, that means a focused risk assessment: a structured history including family history, smoking and alcohol, blood pressure, and a short list of laboratory markers such as lipids and blood glucose. The guiding principle is that every test should have a therapeutic consequence if it comes back abnormal. If nobody can say what would be done differently with a result, the test probably shouldn't be ordered.

Further steps should then follow from that assessment. If lipids and family history suggest elevated cardiovascular risk, a closer look — for example coronary calcium scoring — may be reasonable. If they don't, it isn't needed. At 50+ with no relevant history, the colonoscopy that actually saves lives belongs on the list. Women with dense breast tissue may benefit from supplemental ultrasound, which can find cancers mammography misses.

Equally important is what a good check-up leaves out: tests that look impressive but lack evidence. Routine whole-body MRI, NAD+ infusions, telomere panels, or testosterone therapy at normal levels have no place in evidence-based prevention — there is no good evidence of benefit, and some carry real risks or generate incidental findings that lead to further testing.

Questions to ask your doctor

• Which screenings are recommended for someone of my age, sex and risk profile?
• What would we do differently if this test came back abnormal?
• What is the chance of a false alarm, and what would the follow-up involve?
• Are there tests I've been having routinely that I don't actually need?

The Cochrane review was a wake-up call for preventive medicine. It forced the field to ask a harder question than "what can we measure?" The better question — the one worth bringing to your own physician — is: what will actually change your outcomes?

This article is general information and not individual medical advice. Decisions about tests and treatment belong in a conversation with your own physician.