Four in Ten Cancers Are Linked to Changeable Risks: What That Really Means

Studies in different parts of the world keep arriving at a similar answer: roughly a third to two-fifths of new cancer cases are linked to risk factors that could, in principle, be changed. Tobacco is the largest single factor almost everywhere. Infections come second worldwide, and alcohol, excess body weight, sunlight and workplace exposures follow. These are modelled estimates of what prevention could achieve across whole populations. They are not verdicts on anyone's life.

For someone who already has cancer, or who loves someone who does, headlines about "preventable cancer" can land as an accusation. They shouldn't. A population attributable fraction is a statement about a population: how many fewer cases a country might see if an exposure were removed. It cannot say why any one person developed cancer, and many people diagnosed with cancer never had the exposure in question. What these figures do show is where governments, employers and health systems have the most room to act, because tobacco and alcohol prices, air quality, workplace safety and vaccine programmes are set by policy, not by individuals.

Between a third and four in ten cancers are linked to changeable risk factors, in every major estimate

The newest global estimate, from the World Health Organization's International Agency for Research on Cancer (IARC), covers 30 modifiable risk factors, 36 cancer sites and 185 countries. It attributes 37.8% of new cases in 2022 to those factors: 45.4% in men and 29.7% in women^[3]. Our overview of cancer worldwide gives that headline. What sits beneath it is more useful for anyone planning prevention.

First, the share varies widely by region. In women it ranged from about 24% in North Africa and West Asia to about 38% in sub-Saharan Africa. In men it ranged from about 28% in Latin America and the Caribbean to about 57% in East Asia^[4]. These differences reflect different smoking histories, different infection rates and different cancer mixes. They do not reflect differences in personal virtue.

Second, national studies using their own registry data reach similar totals. The table below sets them side by side. They use different lists of factors, different years and different methods, so treat the comparison as showing consistency, not as a ranking.

Within the UK, the estimated share was highest in Scotland (41.5%) and lowest in England (37.3%), with Wales at 37.8% and Northern Ireland at 38.0%^[1]. All three national estimates exclude the very common non-melanoma (keratinocyte) skin cancers^[1][2][5], and the US abstract does not state how many factors it evaluated. The Australian authors concluded that"at least one in three cancers in Australia is attributable to exposure to known modifiable factors"^[5].

An attributable fraction is a model of prevention, not a cause in any one person

A population attributable fraction (PAF) combines two things: how common an exposure is in a population, and how much it raises the risk of a given cancer according to research. From those it estimates the share of cases that would not occur if the exposure were reduced to a chosen "counterfactual" level, usually none at all. In the UK study, the counterfactual was zero exposure for most factors, a reference level for fibre (the UK recommendation of 30 g a day) and for physical activity, and, for breastfeeding, simply whether a woman had breastfed at all, because duration could not be measured reliably^[1].

Several things follow from how the numbers are built:

• They look backwards in time. Cancer takes years to develop, so the UK estimate for 2015 used exposure data from about 2005^[1]. Today's attributable cases reflect yesterday's smoking rates, and today's policies will show up in tomorrow's cases.
• They cannot simply be added up. Some cancers are linked to more than one factor at once, such as tobacco and alcohol together for oesophageal cancer, and the UK authors note their method does not account for combined exposures^[1].
• They depend on which factors are counted. The global study counted 30 factors and Australia's 13^[3][5]. A longer list usually gives a larger total, which is one reason the figures differ.
• They carry uncertainty. The UK authors describe PAFs as "estimates based on the best available data" and judged that their limitations would, if anything, tend to underestimate the true share^[1].

Most importantly, a PAF of 15% for smoking does not mean that 15% of people with cancer brought it on themselves. It means that, across the population, about that many fewer cases would be expected in a world without tobacco. The other side of the same figure is just as important: in every study here, most cancers are not attributed to any known modifiable factor. Ageing, inherited genetic variants and causes not yet understood all play a part. Our study of cancer myths and causes looks at the claims that do not stand up.

A "Group 1 carcinogen" label shows how sure the evidence is, not how large the risk is

News reports often say that something "is in the same category as tobacco". The category they mean comes from the IARC Monographs, which classify agents by the strength of evidence that they can cause cancer in humans. As of October 2026, IARC lists 135 agents in Group 1 ("carcinogenic to humans"), 98 in Group 2A ("probably carcinogenic"), 329 in Group 2B ("possibly carcinogenic") and 498 in Group 3 ("not classifiable")^[6].

WHO is explicit that these categories "indicate the strength of the evidence" and that the classification "does not measure the likelihood that cancer will occur"^[7]. Tobacco smoking, alcohol and processed meat are all in Group 1, yet they differ greatly in how many cancers they are linked to: in the UK estimate, tobacco smoking accounted for 15.1% of all new cases and processed meat for 1.5%^[1]. Nor is Group 3 a clean bill of health: it means an agent is "not classifiable as to its carcinogenicity to humans", which is not a finding of safety, and for agents it has not evaluated at all IARC says that "no determination of non-carcinogenicity or overall safety should be inferred"^[6]. To understand how big a risk is, look at the attributable fraction or the relative risk, never the hazard group alone.

Tobacco is the largest single factor everywhere, and alcohol is often underestimated

Every estimate in this article puts tobacco first. Smoking accounted for 15.1% of all new cancer cases worldwide in 2022^[3]. In the United States in 2019, cigarette smoking was linked to 19.3% of cases among adults aged 30 and over but 28.5% of cancer deaths^[2]. The gap arises because smoking is so strongly linked to lung cancer, which has low survival: in the same US study, lung cancer had more attributable cases (201,660) and more attributable deaths (122,740) than any other cancer^[2].

Tobacco use is now concentrated in poorer countries. WHO estimates there are 1.2 billion tobacco users in the world, around 80% of them in low- and middle-income countries^[8]. Those countries are likely to see most of the tobacco-linked cancers of the coming decades unless control measures reach them.

Alcohol is a smaller factor than tobacco, but a larger one than many people realise. A global study estimated that 741,300 new cancer cases in 2020, or 4.1% of all new cases, were attributable to alcohol, with oesophageal (189,700), liver (154,700) and breast cancer (98,300) contributing most. Men accounted for 76.7% of these cases. The share was highest in eastern Asia (5.7%) and central and eastern Europe (5.6%) and lowest in northern Africa (0.3%)^[9]. Notably, moderate drinking (under 20 g of alcohol a day), not just heavy drinking, contributed 103,100 of those alcohol-attributable cases (13.9% of them)^[9]. WHO's European office reports that around half of all alcohol-attributable cancers in its region occur in people whose drinking does not meet the definition of heavy drinking, and says the link between alcohol and cancer "is still not widely known to the public in most countries"^[10]. That gap in public knowledge is a policy failure, not a personal one.

Infections account for about one in ten cancers worldwide, mostly in poorer regions

Infections are the second-largest group of modifiable factors globally, accounting for 10.2% of new cases in 2022^[3]. Their impact falls unequally. An earlier IARC analysis estimated 2.2 million infection-attributable cancer cases in 2018, at an age-standardised rate of 25.0 per 100,000 person-years worldwide, ranging from 37.9 in eastern Asia and 33.1 in sub-Saharan Africa to 13.6 in northern Europe. For HPV-linked cancers, the age-standardised rate rose from 6.9 per 100,000 in high-income countries to 16.1 in low-income countries^[11].

Four infections dominate. Helicobacter pylori, a stomach bacterium, was linked to an estimated 810,000 cases in 2018, HPV to 690,000, hepatitis B to 360,000 and hepatitis C to 160,000^[11]. WHO notes that stomach cancer is linked mainly to H. pylori and cervical cancer to HPV, and that lung, stomach and cervical cancers together make up nearly half of all preventable cancers^[4]. WHO names vaccination against HPV and hepatitis B as part of the response^[4]. Where cervical cancer remains common, the question is access to vaccination and screening; our study of the HPV vaccine and cervical cancer sets out the evidence.

Excess weight, sunlight and work: large, uneven and shaped by environment

In the UK, overweight and obesity was the second-largest factor, linked to 6.3% of all new cases in 2015; in the US it was linked to 7.6% of cases and 7.3% of deaths among adults aged 30 and over in 2019^[1][2]. Body weight is shaped heavily by food prices, marketing, working hours, income and the places people live, which is why WHO frames the response as "healthier food and physical activity environments"^[4].

Ultraviolet radiation matters most where fair-skinned populations live under strong sun. An estimated 168,000 new melanoma cases worldwide in 2012, or 75.7% of all new melanomas, were attributable to excess UV exposure, rising to 96% in Oceania^[12]. In Australia, solar radiation was the second-largest factor, at 6.2% of all cancers studied^[5].

Workplace exposures are the clearest case of risk that individuals do not choose. In the UK, occupational exposures were linked to 3.8% of all new cancer cases, as many as UV radiation^[1]. The global study also counted occupational exposures among its 30 factors^[3], and WHO calls for safer workplaces as part of cancer prevention^[4]. Air pollution was linked to 1.0% of UK cases^[1], and WHO notes it alongside smoking as a main driver of lung cancer worldwide^[4]. Because these exposures are not spread evenly, prevention policy is also an equality issue; our study of cancer and deprivation follows that thread.

What the evidence asks of us

The figures in this article point to policy, not to personal blame. The factors with the largest attributable fractions are the ones most shaped by law, price, regulation and public services. WHO and IARC's response to the 2026 global estimate calls for "context-specific prevention strategies" and for coordinated action across health, education, energy, transport and labour^[4]. In practice the evidence asks for:

• Full tobacco control. The WHO Framework Convention on Tobacco Control has 182 countries and the European Union as Parties, and over 6 billion people are covered by at least one of WHO's MPOWER tobacco control measures at best-practice level. WHO describes tobacco taxes as "the most cost-effective way to reduce tobacco use"^[8]. Because 80% of tobacco users live in low- and middle-income countries^[8], support for implementation there matters most.
• Alcohol regulation and honest labelling. WHO calls for alcohol regulation^[4] and, in Europe, for "cancer-related health information messages on labels of alcoholic beverages"^[10]. People are entitled to know about a risk that is still not widely known.
• Vaccination against cancer-causing infections. WHO names HPV and hepatitis B vaccination as core prevention^[4]. With the highest infection-linked cancer rates in eastern Asia and sub-Saharan Africa^[11], funding and supply for those regions is a priority.
• Cleaner air, safer workplaces and healthier environments. These are the responsibilities of governments and employers, and WHO lists them alongside tobacco and alcohol policy^[4].
• Better data. The national estimates in this article describe 2010, 2015 and 2019. Regular, comparable estimates, including in countries without established cancer registries, would let governments track whether prevention is working.

The US authors concluded that their findings show the "potential to substantially reduce the cancer burden through broad and equitable implementation of preventive initiatives"^[2]. "Equitable" is the key word. Prevention that relies on individual willpower tends to reach people with the most time, money and information first. Prevention built into prices, laws, workplaces and vaccine programmes reaches everyone. None of it changes anything for people already diagnosed, who need, and deserve, excellent care without any suggestion of fault. For more on the wider picture, see the cancer research library.

Language: this piece says "people with cancer" and "people living with and after cancer", and describes the illness and its treatment plainly, without combat metaphors. Clinical terms such as "stage", "incidence" and "net survival" are used in the sense the cited source defines them.

Why we're publishing this

Prevention figures are often used in ways that hurt the people they should help. "Four in ten cancers are preventable" can be heard as "four in ten people with cancer could have avoided it", which is false and unkind. We are publishing this so that cancer charities, support groups and hospices have a clear, sourced account of what these estimates do and do not say, and can make the case for the policy changes that would reduce cancer for everyone, without adding to the guilt and stigma that many people living with cancer already describe.

How to read this data

Every figure here is modelled. Attributable fractions combine exposure surveys with relative risks from published studies. They depend on the factors chosen, the counterfactual level, the time lag assumed and the quality of exposure data, and different studies make different choices. The global, US, UK and Australian estimates are not directly comparable and are shown together only to illustrate their broad agreement.

Years differ. The global estimate is for 2022, the US for 2019, the UK for 2015, Australia for 2010, the alcohol estimate for 2020, the infections estimate for 2018 and the melanoma estimate for 2012. Exposure levels have changed since; smoking, for example, has fallen in many high-income countries.

Cases versus deaths. Most figures here are shares of new cases (incidence). The US study also reports shares of deaths, which differ because cancers vary in survival.

Some sources were read only as abstracts. For the 2026 global study, the 2024 US study and several others we used only figures stated in the published abstract. Hazard classifications from IARC describe strength of evidence, not size of risk.

Use this data

Free to cite with attribution to Better Health For Everyone. Charities, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research for general information. It is not medical advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's health, please speak to a qualified health professional.

Sources

1. Brown KF, Rumgay H, Dunlop C, et al. The fraction of cancer attributable to modifiable risk factors in England, Wales, Scotland, Northern Ireland, and the United Kingdom in 2015. British Journal of Cancer, 2018. europepmc.org/article/PMC/PMC5931106
2. Islami F, Marlow EC, Thomson B, et al. Proportion and number of cancer cases and deaths attributable to potentially modifiable risk factors in the United States, 2019. CA: A Cancer Journal for Clinicians, 2024. europepmc.org/article/MED/38990124
3. Fink H, Langselius O, Vignat J, et al. Global and regional cancer burden attributable to modifiable risk factors to inform prevention. Nature Medicine, 2026. europepmc.org/article/MED/41634393
4. World Health Organization and International Agency for Research on Cancer. Four in ten cancer cases could be prevented globally (news release, 3 February 2026). WHO, 2026. who.int/news/item/03-02-2026-four-in-ten-cancer-cases-could-be-prevented-globally
5. Whiteman DC, Webb PM, Green AC, et al. Cancers in Australia in 2010 attributable to modifiable factors: summary and conclusions. Australian and New Zealand Journal of Public Health, 2015. europepmc.org/article/MED/26437735
6. International Agency for Research on Cancer. Agents classified by the IARC Monographs, Volumes 1–142 (last updated 30 October 2026). IARC, 2026. monographs.iarc.who.int/agents-classified-by-the-iarc
7. World Health Organization. Cancer: carcinogenicity of the consumption of red meat and processed meat (questions and answers). WHO. who.int/news-room/questions-and-answers/item/cancer-carcinogenicity-of-the-consumption-of-red-meat-and-processed-meat
8. World Health Organization. Tobacco: fact sheet (26 June 2026). WHO, 2026. who.int/news-room/fact-sheets/detail/tobacco
9. Rumgay H, Shield K, Charvat H, et al. Global burden of cancer in 2020 attributable to alcohol consumption: a population-based study. The Lancet Oncology, 2021. europepmc.org/article/MED/34270924
10. WHO Regional Office for Europe. No level of alcohol consumption is safe for our health (news release, 4 January 2023, amended 17 September 2026). WHO/Europe, 2023. who.int/europe/news/item/04-01-2023-no-level-of-alcohol-consumption-is-safe-for-our-health
11. de Martel C, Georges D, Bray F, Ferlay J, Clifford GM. Global burden of cancer attributable to infections in 2018: a worldwide incidence analysis. The Lancet Global Health, 2020. europepmc.org/article/MED/31862245
12. Arnold M, de Vries E, Whiteman DC, et al. Global burden of cutaneous melanoma attributable to ultraviolet radiation in 2012. International Journal of Cancer, 2018. europepmc.org/article/MED/29659012