When Caution Kills: Why Cancer Screening Policy Needs a Radical Rethink

Cancer screening sits in an uncomfortable place in modern medicine: it promises to catch deadly disease early in people who feel perfectly well, yet it carries real risks of harm, cost and anxiety. A provocative new analysis published in eClinicalMedicine argues that the way wealthy countries decide whether to introduce or change screening programmes is itself a hazard — one that may have cost tens of thousands of lives through delays, indecision and a quiet, unexamined bias toward doing nothing. Writing from the United Kingdom, cancer screening experts Peter Sasieni and Stephen John contend that excessive caution in health policy is not neutrality but a choice with a body count, and they propose a new framework for deciding when evidence is good enough to act.

The authors build their case from a series of revealing examples drawn from British screening history. The UK National Screening Committee recommended against population-based prostate cancer screening, concluding that while screening prevented prostate cancer deaths, it also caused harm through overdiagnosis and overtreatment, leaving some men with urinary incontinence or erectile dysfunction. In reaching that conclusion, the committee made a value judgement about how to weigh extended life against morbidity — and, the authors note, it assumed without evidence that individual men would share its values. Meanwhile, when Wales extended the cervical screening interval for women aged 25 to 49 from three years to five in 2022, a scientifically defensible decision triggered a public outcry and a petition signed by more than a million people, prompting England to delay the same change until July 2025.

Bowel screening illustrates how resources quietly shape supposedly clinical thresholds. The faecal immunochemical test, or FIT, applies a binary cut-off to the amount of blood in a stool sample, and that cut-off ranges from 8.5 to 150 micrograms of haemoglobin per gram of faeces across programmes, largely to manage colonoscopy capacity. Historically, Wales used a threshold of 150, England 120 and Scotland 80 micrograms per gram — meaning a result deemed normal in England could have triggered a colonoscopy just across the Scottish border. The stakes of delay are also stark: randomised trials showed guaiac-based faecal occult blood testing reduced colorectal cancer mortality as early as 1996, yet England’s screening programme only completed roll-out in 2010. Screening now prevents roughly 2,000 colorectal cancer deaths annually in England; had it rolled out in 2000, the authors estimate some 20,000 premature deaths could have been averted.

To make sense of such cases, Sasieni and John adapt five principles for decision-making under uncertainty — proportionality, justice, autonomy, feasibility and adaptability — into a framework specifically for cancer screening. Their central claim is that policymakers suffer from an asymmetry of fear: they worry intensely about false-positive decisions, such as introducing a screening programme that turns out to do more harm than good, while largely ignoring false-negative decisions, in which beneficial programmes are delayed or withheld. Because committees periodically revisit recommendations, a cautious ‘not yet’ feels reversible and safe. But delay has consequences that can be quantified even without certainty. If prostate screening in men aged 50 to 65 reduces prostate cancer mortality by 25 percent over 15 years, a two-decade implementation delay could translate into around 20,000 avoidable deaths in the UK. Conversely, had multimodal ovarian cancer screening been introduced for women aged 50 to 74 before the definitive UKCTOCS trial results, roughly 7,000 women each year might have undergone unnecessary surgery. Both sides of the ledger can be estimated, the authors argue, and should be before decisions are made.

The principle of justice exposes another hidden cost of slowness: inequality. Wealthy and well-educated people frequently obtain screening privately long before public programmes roll out. In the United States, colonoscopy use is 50 percent higher in the wealthiest socioeconomic quintile than in the poorest. In England, prostate-specific antigen testing is 25 percent lower in the most deprived quintile — and metastatic prostate cancer rates there are 11 percent higher, while overall prostate cancer incidence is 19 percent lower, plausibly reflecting inequitable access to asymptomatic PSA testing. Justice also complicates the details: some researchers argue FIT referral thresholds should differ for women and men, but the right answer depends on which measure of equity one chooses to prioritise, a value judgement that should be made transparently. Even overdiagnosis falls unevenly — introducing prostate screening for men in their seventies would generate more overdiagnosis among the most deprived, whose shorter life expectancy means more of them die of other causes before screening could ever help.

Autonomy poses a deeper philosophical problem. Public health typically overrides individual preferences for the collective good, while clinical medicine demands informed consent. Screening sits awkwardly between: society seeks consent before screening anyone, yet takes a paternalistic stance on what screening is offered at all, rarely considering the ethics of denying screening to those who want it. The authors suggest a more democratic division of labour, inspired by the Dutch model: expert committees could summarise evidence and uncertainties without issuing recommendations, forcing elected politicians to make the value-laden choices explicitly. Where experts disagree, politicians should decide. They even sketch a personalised option for prostate screening — inviting men aged 50 to 69 for triennial PSA testing without encouraging participation, letting personal risk tolerance guide the decision, much as oncology patients weigh their own treatment choices.

Feasibility, the authors stress, already governs screening policy whether it is admitted or not. FIT thresholds and age ranges are set by colonoscopy capacity; MRI-based prostate screening would demand vast new infrastructure of scanners and staff; and in low- and middle-income countries, cervical screening is constrained by the lack of facilities to triage screen-positive women and treat precancerous lesions. Public opinion constrains policy too — hence the continued offering of cervical screening at age 25 to women vaccinated against HPV as adolescents, despite their extremely low cervical cancer risk and the likelihood that screening does them more harm than good. The authors argue the public is mature enough to understand these trade-offs, provided they are communicated honestly.

Adaptability may be the most urgent principle in an era of artificial intelligence and liquid biopsies. Traditional screening trials are ruinously expensive — the National Lung Cancer Screening Trial cost 256 million dollars in 2002, roughly 400 million today — and a binary framework of full national implementation or complete rejection leaves no pathway for pragmatic pilots that generate evidence while delivering benefit. Multicancer detection tests sharpen the dilemma: by the time one test’s clinical utility is fully evaluated, the technology will be outdated, and trial designs that test each cancer type separately would require randomising millions of people. The authors call for publicly funded pilots designed to fill evidence gaps, with implementation beginning in health-deprived regions to address injustice rather than widen it. They are candid about the risks of permissiveness — a pilot ovarian screening programme before 2021 would have harmed some healthy women with false positives — but insist those harms and benefits could and should have been estimated in advance, and that any new decision-making system should itself be monitored and reformed if it fails.

The paper’s conclusion is disarmingly simple: presenting screening policy as a straightforward application of evidence-based medicine obscures profound ethical and political choices, and delaying a decision is itself a decision against change — one measured in preventable deaths, avoidable morbidity and widening inequality. As AI begins to exceed human radiologists and pathologists, quantifying future disease risk and optimising screening intervals, governance structures built around static population interventions and decade-long trials will be unable to evaluate products superseded every five years. Who decides, on what basis, and how quickly — the authors argue — are questions that can no longer be left unasked.

Subject of Research: Ethical and policy frameworks for cancer screening decision-making under uncertainty

Article Title: The clinical consequences of excessive caution: rethinking how cancer screening policy is made

Article References: Sasieni, P., & John, S. (2026). The clinical consequences of excessive caution: rethinking how cancer screening policy is made. eClinicalMedicine, Article 104188. https://doi.org/10.1016/j.eclinm.2026.104188

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104188

Keywords: cancer screening, screening policy, decision-making under uncertainty, health inequalities, prostate cancer screening, bowel cancer screening, multicancer detection tests, artificial intelligence, overdiagnosis, public trust, clinical, consequences