Avoidable deaths through the primary prevention, early detection, and curative treatment of cancer worldwide: a population-based study

Close to 10 million people died from cancer globally in 2022,¹ and large disparities persist in cancer mortality between countries worldwide. For example, cervical cancer mortality rates range from 3·3 deaths per 100 000 women in countries with a very high human development index (HDI) to 16·3 per 100 000 women in countries with a low HDI.¹ Previous studies have estimated that 44% of cancer cases are preventable through the reduction of exposure to modifiable risk factors² and that 69% of premature cancer deaths could be averted through primary and secondary prevention, in addition to 31% due to improvements in curative treatment.³ The major modifiable risk factor for cancer is tobacco exposure, and eliminating tobacco use could avert up to 2·6 million deaths from cancer globally per year.² Population-based cancer survival—an indicator of the effectiveness of early detection and timely cancer treatment—also varies markedly by country. For example, 5-year net survival of colon cancer ranged from around 12% in South Africa to over 70% in Australia for patients diagnosed between 2010 and 2014.⁴ Such international variations could be due to differences in the availability and access to early detection and screening programmes, as well as the availability of, and access to, curative treatment.⁴

The estimation of avoidable deaths provides insight regarding the burden of deaths that could be averted if such inequalities were eliminated. WHO and countries in the Organisation for Economic Co-operation and Development (OECD) have used avoidable deaths as an indicator to quantify premature mortality that was avoidable through effective primary prevention and other public health measures, and through more effective and timely health-care interventions.^5,6 To calculate avoidable deaths, these two institutions assume that premature deaths occurring at specific ages (eg, <75 years) are avoidable. Others have estimated cancer-related avoidable deaths if societies could eliminate disparities in cancer survival between socioeconomic groups in England and Wales,⁷ and countries in Europe.⁸ In this study, we aimed to expand this approach to estimate avoidable deaths among people diagnosed with cancer globally to assess mortality that is potentially preventable through a reduction of exposure to major risk factors, and deaths that are amenable to improved survival through early detection and curative treatment, by country, cancer site, region, and HDI group.

An estimated 4·5 million (95% UI 4·5–4·5) deaths occurring within 5 years among people diagnosed with cancer in 2022 could have been avoided through primary prevention and improvements in early detection and access to curative cancer treatment (table). This number accounted for 47·6% (95% UI 47·5–47·8) of the expected total number of deaths among people diagnosed with cancer (9·4 million [9·4–9·5]). Of the expected total number of deaths, 3·1 million (3·1–3·1; 33·2% [33·1–33·3]) were preventable and 1·4 million (1·4–1·4; 14·4% [14·4–14·5]) were treatable. The global ASRs were 50·0 preventable deaths (49·8–50·2), 27·0 treatable deaths (26·9–27·1), and 77·0 overall avoidable deaths (76·8–77·2) per 100 000 people (table).

The overall proportion of avoidable deaths, both preventable and treatable, was higher than 40% in 15 of 19 world regions. The highest proportions were found in Africa (ranging from 51·4% [95% UI 50·6–52·2] in southern Africa to 62·0% [60·6–63·5] in east Africa) and Asia (from 47·9% [47·8–48·0] in east Asia to 56·4% [56·2–56·7] in south central Asia; table). Australia and New Zealand had the smallest proportion of avoidable deaths (35·5% [35·1–35·9]), followed by northern and western Europe and North America (all 40% or less). At the national level, the top ten largest proportions of avoidable deaths were all in African countries, including including Sierra Leone (72·8%), The Gambia (70·0%), and Malawi (69·6%; figure 1). By contrast, three Northern European countries (Sweden [28·1%], Norway [29·9%], and Finland [32·0%]|) had the lowest proportions of avoidable deaths. Detailed results by country are presented in the appendix (pp 15–22).

The proportion of deaths that were preventable through primary prevention of the five included risk factors varied across regions from 24·5% (95% UI 24·2–24·9) in central America to 37·8% (37·7–37·9) in east Asia (table). Ten of the 19 regions had proportions over 30%, of which eight were located in Asia and Africa, and the remaining two were North America and eastern Europe. The highest proportions of treatable deaths were found in west Africa (32·2% [31·6–32·8]), middle Africa (29·2% [28·4–30·0]), and east Africa (28·2% [27·5–28·9]), while the lowest were found in North America (7·6% [7·5–7·6]), east Asia (10·1% [10·0–10·1]), Australia and New Zealand (10·3% [10·2–10·4]), and northern Europe (11·0% [11·0–11·1]; table).

The ASRs of avoidable deaths ranged from 54·6 deaths (95% UI 53·9–55·4) per 100 000 people in central America to 102·5 deaths (102·1–103·0) per 100 000 in eastern Europe (table). Rates for preventable deaths differed between regions, with ASRs ranging from 26·2 deaths (25·8–26·6) per 100 000 in central America to 63·6 deaths (63·3–63·8) per 100 000 in eastern Europe. Countries in east Asia and eastern Europe had the highest ASRs of preventable deaths (appendix p 15). The ASRs of treatable deaths also markedly differed between regions, ranging from 16·9 deaths (16·9–16·9) per 100 000 people in North America to 49·9 (45·7–54·7) in Melanesia, Micronesia, and Polynesia.

The five cancer diagnoses that contributed the most avoidable deaths from all causes were lung, liver, stomach, colorectal, and cervical cancers, which collectively accounted for 59·1% of all avoidable deaths globally (figure 2). Among preventable deaths, the five cancer diagnoses with the highest counts (lung, liver, stomach, cervical, and oesophageal) accounted for 74·8% of all preventable deaths. The cancer diagnoses that had the largest proportions of preventable deaths out of their total expected deaths were cervical cancer (82·2% [95% UI 81·6–82·9]; 240 000 preventable deaths [238 100–241 900] of 291 900 total deaths [289 600–294 200]) and Kaposi sarcoma (75·3% [66·8–84·8]; 9000 [7900–10 100] of 11 900 [10 600–13 400]; appendix pp 24–36). For treatable deaths, the five cancer diagnoses with the highest counts (breast, colorectal, prostate, lung, and non-Hodgkin lymphoma) accounted for 54·2% of all treatable deaths. The cancer diagnoses that had the largest proportions of treatable deaths were testicular cancer (70·0% [67·8–72·3]; 5300 treatable deaths [5100–5400] of 7500 total deaths [7300–7800]) and thyroid cancer (39·4% [38·8–40·0]; 23 800 [23 400–24 100] of 60 300 [59 400–61 200]).

The leading cancer diagnosis in terms of the numbers of preventable and treatable deaths differed between countries (figure 3). In terms of number of avoidable deaths, lung cancer had the highest count of avoidable deaths in 97 countries. The other top cancer sites in terms of number of avoidable deaths were cervical (35 countries); stomach (22 countries); liver (14 countries); breast (11 countries); prostate (five countries); and oesophageal cancer (one country). For cervical, breast and stomach cancer, the majority of countries were located in sub-Saharan Africa and south Asia. For preventable deaths, cervical cancer ranked first in many sub-Saharan African and south Asian countries. Lung cancer was responsible for the most preventable deaths in countries in North and South America and Europe, and in Australia and New Zealand. By the number of treatable deaths, prostate cancer had the highest count of deaths in most European countries, Latin America, Australia, and New Zealand, although female breast cancer was associated with the most treatable deaths in south Asian countries and most countries across the African continent.

The burden by proportion of avoidable deaths out of the total expected deaths in each HDI group differed greatly by HDI, ranging from 40·5% in very high HDI countries to 60·8% in low HDI countries. In terms of avoidable deaths by cancer diagnosis in low and medium HDI settings, cervical cancer was the top cancer site by avoidable deaths, accounting for 44 100 avoidable deaths (18·0% of all avoidable deaths) in the low HDI group and 82 500 (10·8%) in the medium HDI group (figure 4). In high and very high HDI countries, cervical cancer was not among the top five cancer sites by number of avoidable deaths; lung cancer accounted for the largest burden, with 623 000 avoidable deaths (31·6% of all avoidable deaths) in the high HDI group and 491 000 (32·5%) in the very high HDI group. Breast cancer also represented a large avoidable burden in low and medium HDI countries, whereas it was absent from the top five cancer diagnoses by avoidable deaths in high and very high HDI countries. Lung cancer had the second largest burden in medium HDI settings, but was absent from the top five cancer diagnoses in low HDI settings. Excluding China (high HDI) and India (medium HDI) changed the cancer profiles in medium and high HDI countries, with breast cancer ranked fourth highest in the high HDI group, and lung and liver cancer ranking first and second, respectively, in the medium HDI group (appendix p 20).

We estimated that 4·5 million (47·6%) of 9·4 million deaths expected to occur within 5 years among people diagnosed with cancer globally in 2022 were avoidable through primary prevention of five major modifiable risk factors and improvements in early detection and access to curative cancer treatments. Among the 4·5 million avoidable deaths, we estimated that 3·1 million (33·2% of total expected deaths) were preventable and the remaining 1·4 million (14·4%) were treatable. Between world regions, the proportion of avoidable deaths ranged from 35·5% in Australia and New Zealand to 62·0% in east Africa, including substantial variation between countries. Lung cancer was the cancer diagnosis with the largest number of preventable deaths globally, followed by liver and stomach cancer. Breast, colorectal, and prostate cancers were the cancer diagnoses with the most treatable deaths. Patterns of avoidable deaths according to cancer diagnosis differed by HDI group.

We found that primary prevention could avoid around a third of all expected deaths among cancer patients, indicating a potentially large impact of prevention programmes aiming to reduce exposure to important cancer risk factors (tobacco, alcohol, excess body weight, ultraviolet radiation, and infections). With lung cancer having such an important contribution to avoidable deaths globally, our study highlights the importance of optimal scale-up of tobacco control. WHO reported that the global average smoking prevalence decreased from 22·8% to 17·0% between 2007 and 2021,²¹ largely thanks to successful tobacco control. Cost-effective measures to reduce tobacco use in the population include implementing price increases through taxation, standard packaging, bans on advertising and promotion, eliminating second-hand tobacco smoke, running mass media educational campaigns, and providing support for cessation.^21,22 However, it is worth noting the importance of other risk factors for lung cancer that we were not able to include in our analysis. A recent study showed a substantial rise in rates of lung adenocarcinoma in many countries since 2005, which is partly linked to air pollution and changes in cigarette consumption,²³ indicating that the preventable burden of lung cancer could be larger than estimated.

In addition to tobacco, the growing number of people with excess body weight poses considerable challenges to global health.²⁴ Overweight and obesity are driven by a wide range of environmental, societal, and biological factors.²⁵ Cost-effective interventions include those which regulate advertising, labelling, and taxes on unhealthy food and beverages.²⁶ For example, Mexico implemented a tax on sugar-sweetened beverages in 2014 which led to a 14% increase in the prevalence of non-consumers of soft drinks compared with before the tax was introduced.²⁷ Similarly, cost-effective strategies to reduce alcohol consumption in the population include increasing the price of alcohol through taxation, reducing the availability of alcohol products, and banning marketing.²² Regarding tackling the preventable burden of cutaneous melanoma due to ultraviolet radiation, educating populations on limiting sun exposure time and encouraging sunscreen use through skin cancer campaigns has been an effective prevention strategy in Australia.²⁸

Infection-related cancers contributed to the large proportion of preventable deaths, particularly in low and medium HDI countries. Cervical cancer was the leading contributor of preventable deaths in most countries in sub-Saharan Africa. WHO's strategy for cervical cancer elimination focuses on three main strategies: vaccination, screening, and providing adequate access to cancer treatment and care.²⁹ Earlier modelling exercises concluded that 97·9% of cervical cancer deaths predicted to occur in 78 low-income and lower-middle-income countries over the period 2020–2120 could be prevented by the realisation of the WHO's cervical cancer elimination strategy.³⁰ However, as of 2020, less than 25% of low-income countries had introduced the HPV vaccine into their national immunisation schedules, in contrast to over 85% of high-income countries.²⁹ In addition, less than 40% of low-income countries had national cervical screening programmes in place for cervical cancer in 2019, compared with over 80% of high-income countries.²⁹ Similar issues apply for other infectious agents, such as viral hepatitis and Helicobacter pylori; ensuring equity in access to strategies to prevent or treat infections are key to reducing their related cancer diagnoses and deaths.³¹

In addition to primary prevention, improved access to early detection and curative treatment is essential, particularly in low and medium HDI countries, which have the highest proportion of treatable deaths. For female breast cancer—which accounted for around 15% of all treatable cancers in this study—the WHO's Global Breast Cancer Initiative lists late diagnosis, inadequate diagnostic services and treatment, and low coverage of treatment by national health services as contributing factors to global inequities in survival.³² The African Breast Cancer—Disparities in Outcomes study found that distances from diagnostic and treatment facilities were associated with delayed breast cancer diagnosis and more advanced stage at diagnosis in sub-Saharan Africa.³³ Suggestions to address these specific barriers include providing transport or travel allowances to attend treatment centres and decentralising treatment by creating more centres.^32,33 Focus should be on implementing early detection strategies to achieve the WHO goals of at least 60% of breast cancers diagnosed in stage I or II, and more than 80% of patients receiving a diagnosis within 60 days of their initial presentation.³² One systematic review found that 40% of countries do not reach this goal.³⁴ While population-based screening can be costly for low-income and middle-income countries, cost-effective alternatives to improve early detection in such settings include promoting breast-health awareness and performing clinical breast examinations.³⁵ Finally, improving access to surgery could improve survival among patients with cancer and reduce the burden of treatable deaths.³⁶ For example, in low-income and lower-middle-income countries, more than 90% of the population do not have access to timely and affordable surgery and anaesthesia care, in contrast to 3·6% in higher-income regions.³⁶ Similarly, for other major cancer sites that contributed the most treatable deaths, such as colorectal cancer and prostate cancer, expanding universal health coverage is essential to address global disparities in access to treatment.²²

The strengths of our study include the comprehensive breakdown of the estimated avoidable deaths among patients diagnosed with cancer and the potential roles of primary prevention, earlier detection, and curative cancer treatment by cancer site globally. To our knowledge, this is the only study to provide estimates of avoidable mortality with details by country, region, and HDI group, while also considering survival, attributable fractions, life tables, and incidence to assess where public health interventions could be the most effective in terms of mortality reduction. The data we used were also as recent as possible, favouring maximum population coverage. Future studies could look at the potential costs and benefits of various public health intervention scenarios to reduce avoidable deaths.

A key limitation of our study is that we provided only the lower limit of avoidable deaths as we accounted for deaths only up to 5 years since diagnosis. However, these estimates might still cover the majority of deaths, as many cancers can be considered to be cured after this time.³⁷ Additionally, 5 years might be an appropriate interval because too long a follow-up would mean that eventually the entire mortality would be accounted for by cumulative all-cause mortality in the population. In terms of the reference scenario, the best survival scenario for each cancer site in this study does not reflect the maximum theoretical net survival, which could be higher or lower than that of the top estimate by country. In addition, the effect on survival of shifting distributions of cancer stage at diagnosis to stage I–II through improved diagnostic tools, and the related effect on avoidable deaths, were not examined. Additionally, our UIs are limited in the use of GLOBOCAN's uncertainties as the only available estimates for uncertainties. This was due to many of the estimators used in calculating avoidable deaths lacking uncertainty estimates. Ideally, we would have estimated the 95% UIs accounting for the uncertainty of every variable; however, this was not possible due to some estimators missing necessary information for this calculation. Furthermore, our current estimates reflect deaths that are avoidable if the prevention, early detection, and treatment measures could be put into place immediately and with full population coverage, but, in reality, the implementation of these measures and their impact would be progressive. Another limitation is that we assumed that preventable and treatable scenarios were independent (ie, that preventing the cancers would not affect the survival probability of the remaining patients). One example is assuming that a patient with a long history of smoking would have the same treatment outcomes as never-smoking patients when treating lung cancer.

For our estimates, we also only accounted for the effects of five cancer risk factors for which we had global estimates. Other risk factors, such as air pollution or dietary risk factors,^2,38 were not considered in the current study. In addition, the cancer and population registry data are likely to be under-represented in low-resource settings. Ideally, our study would be repeated with population-based cancer registry data from all countries; projects such as the Global Initiative for Cancer Registry development aim to increase the global coverage of high-quality population-based cancer registries, especially in low and medium HDI settings.

In conclusion, nearly half of deaths occurring within 5 years among people diagnosed with cancer globally in 2022 could be avoided through primary prevention and improvements in early detection and curative cancer treatment. Large disparities in avoidable deaths persist between cancer sites, countries, regions, and HDI levels, highlighting the importance of scaling up primary prevention and integrating activities into strengthened early detection and cancer treatment interventions as part of operational national cancer control planning.

This project is funded by the Erasmus Mobility programme (paid to OL), and the International Agency for Research on Cancer (IARC). OL was supported by a grant from the French National Cancer Institute (INCa): INCa-ResPP22-002. The work reported by OL was undertaken during a PhD studentship at IARC. Where authors are identified as personnel of IARC and WHO, the authors alone are responsible for the views expressed in this Article and they do not necessarily represent the decisions, policy, or views of IARC or WHO.

Editorial note: The Lancet Group takes a neutral position with respect to territorial claims in published maps.