Cancer incidence: modest individuals more often develop severe forms, diagnosed later

📜🔗LINK to the source

1. Analytical summary

Social context and cancer profiles

The study analyses the incidence of cancers in France (excluding Mayotte) between 2013 and 2020, by coupling data from the Health Insurance with the EDP-Santé demographic panel, which represents about 4.4% of the population, followed individually. It examines the links between standard of living (deciles D1 to D10), tumour location, prognosis, age at diagnosis, and the detectable or preventable nature of cancers, excluding all individuals already diagnosed with cancer before 2013 to measure a true risk of occurrence (incidence). The authors show that while the overall incidence of all cancers presents a flat "social gradient", certain locations (lung, colorectal) affect the more modest individuals more, whereas others (breast, prostate) are more frequent among the wealthier. Cancers with poor prognosis, preventable cancers (related to lifestyle and environment), and diagnoses at a metastatic stage are overrepresented among individuals with the lowest standard of living. The study employs the Fundamental Cause Theory to interpret these gradients, demonstrating how the wealthier better appropriate prevention and screening.

Operational contributions to prevention

On an operational level, the document quantifies the disparities: a risk multiplied by 1.9 for lung cancer among the most disadvantaged compared to the more affluent, 1.7 for cancers with a poor prognosis, 2.3 for detectable cancers that are already metastasised, and 2.1 for preventable cancers. It provides a clear framework to distinguish between detectable, non-detectable, preventable, and non-preventable cancers, and links these categories to specific social gradients, useful for prioritising actions. The results highlight the importance of early screening in disadvantaged communities and the need to address exposures (tobacco, working conditions, diet, obesity), while simplifying access to organised screening programmes. Professionals can rely on graphs and ratios by decile of living standards to target their territorial interventions and adapt their messages to the social reality of the public.

2. Key points of the document

1. Lung cancers and, to a lesser extent, colorectal cancers are more frequent among the most disadvantaged individuals (D1), whereas breast and prostate cancers are more common among the more affluent (D10). For example, the risk of lung cancer is 1.9 times higher among the most disadvantaged than among the wealthiest, with a ratio of 2.2 for men and 1.7 for women (graph 1, p. 2-3).

2. Conversely, the risk of breast cancer is 1.3 times higher among women in the highest living standard decile than among those in the lowest decile, and prostate cancer is 1.4 times more frequent among the wealthiest men than among the most disadvantaged. These differences are particularly linked to reproductive behaviours and the use of certain hormonal contraceptives for breast cancer, and to PSA testing for prostate cancer (p. 2-3).

3. The most modest individuals have 1.7 times the risk of developing a poor prognosis cancer (five-year net survival of less than 33%), 1.3 times the risk for intermediate prognosis cancers, and 1.4 times less risk for good prognosis cancers, compared to the wealthiest. These gradients persist even after excluding the four most common sites (breast, prostate, lung, colorectal), as shown in graphs 2 and 3 (p. 4).

4. Cancers are diagnosed earlier in life among the most modest: the median age at diagnosis is 6 years lower for lung, colorectal, and breast cancers, and 3 years lower for prostate cancer between the poorest decile and the general population. The median age at diagnosis varies by site and socioeconomic status, linked to screening practices and exposure to risk factors (graph 5, p. 5).

5. The most modest individuals have 2.1 times the risk of developing an avoidable non-screenable cancer, and 2.3 times the risk that their screenable cancer is already metastatic at the time of diagnosis, compared to the wealthiest. The inequalities in severity at diagnosis are mainly observed for screenable sites, while they are almost absent for non-screenable cancers (graphs 6 and 7, p. 6-7)

3. Action points for local stakeholders

1. Strengthen primary prevention of smoking, alcohol, obesity, and occupational exposures in modest environments, relying on figures showing the higher incidence of avoidable cancers in the lowest socioeconomic deciles (D1 to D3). Graphs 1, 2, and 6 (p. 2-4 and 6) provide objective support to convince local elected officials, business leaders, and association partners to prioritise targeting these groups.

2. Adapt the strategies for promoting organised screening (breast, colorectal, cervical) and opportunistic screenings (prostate, skin) to vulnerable populations, taking into account that detectable cancers are more often metastatic among the less affluent. The data from graph 7 (p. 6-7) can be integrated into advocacy efforts to strengthen health mediation, outreach interventions, and support for screening in disadvantaged neighbourhoods.

3. Systematically integrate the dimension of "standard of living" into local health diagnostics, local health contracts (CLS), and territorial projects, using the decile model D1–D10 to analyse disparities in incidence, severity, and age at diagnosis. The methodologies described in box 1 (p. 2) can inspire the construction of local indicators combining socio-fiscal data, healthcare consumption, and hospitalisations.

4. Work on reducing diagnostic delays in low-income settings (access to a general practitioner, use of specialised examinations, understanding of symptoms), as the study shows that cancers are more often invasive, aggressive, or metastatic at diagnosis in these populations. The severity gradient detailed in graph 4 (p. 5) justifies strengthening early detection consultations, rapid pathways, and city-hospital coordination for patients in precarious situations.

5. Co-construct with users and patient associations tools for cancer literacy (incidence, risk factors, screening, severity) tailored to varying levels of understanding. The operational definitions of incidence, preventable cancers, and stages of severity (boxes 1 and 2, p. 2; section on the theory of fundamental causes, p. 5-6) can serve as a basis for simplified educational materials aimed at populations distanced from the healthcare system.

4. Additional References 

🔍➕ For more information, see the articles referenced by "Health Practices" on the topic of screenings ➡️🔗https://pratiquesensante.odoo.com/blog/tag/depistage-17 and inequalities ➡️🔗https://pratiquesensante.odoo.com/2-1-inegalies-sociales-territoriales

1. Public Health France. "Social health inequalities and cancers: surveillance data and prevention issues" (report or thematic file, updated 2024–2025, accessible via Public Health France). Additional resource to cross-reference the Drees study with attributable fraction data and national prevention axes. https://www.santepubliquefrance.fr/docs/article/nombre-et-fractions-de-cancers-attribuables-au-mode-de-vie-et-a-lenvironnement-en-france

2. NACRe Network. "Cancers attributable to nutritional factors" (updated 2025). Methodological and summary document on the share of cancers related to diet, overweight, physical activity, and alcohol, useful for prioritising interventions on modifiable risk factors in the territories. https://www.reseaunacre.eu/prevention-primaire/vous-informer-sur/cancers-attribuables-aux-facteurs-nutritionnels

3. A retrospective on fundamental cause theory: State of the literature, 2021 (update of the literature on the Fundamental Cause Theory, freely accessible on PubMed Central). Allows situating the results of the Drees in contemporary debates on the role of social resources in the production of health inequalities. https://pmc.ncbi.nlm.nih.gov/articles/PMC8691558/

5. Frequently Asked Questions (FAQ)

1. How does this study measure the incidence of cancers?

   Incidence is calculated by relating the new cancer cases observed between 2013 and 2020 to a cancer-free population at the start, identified via ALD, hospitalisations, and drug treatments (notably for prostate) from 2008-2012. Rates are standardised by age and sex to make the living standard deciles comparable (box 1, p. 2).

2. What types of cancers are excluded from the analysis?

   Haematological cancers are excluded because their management modalities do not allow for the definition of severity as with solid tumours, and only the first diagnosed locations during the period are taken into account. The study therefore focuses on invasive solid tumours (ICD-10 codes C) and excludes individuals who had cancer before 2013 (box 1 and 2, p. 2).

3. Why are lung cancers more common among the less affluent?

   The text highlights the central role of smoking, which is more prevalent in less affluent populations and characterised by a later adoption of public health recommendations by these groups. The cited French and international literature (Pouchin et al., 2024; Onwuka et al., 2025; Sidorchuk et al., 2009) confirms socio-economic inequalities in the incidence of lung cancer, more pronounced among men (p. 2-3).

4. How is the severity of cancers approached without TNM staging in medico-administrative databases?

   Severity is inferred from management modalities (procedures, treatments, palliative care) and the distinction between 'in situ' tumours (codes D) and invasive tumours (codes C) in the ICD-10. The authors define several categories: localized, locally advanced, metastatic, particularly aggressive, and untreated non-aggressive (box 2, p. 2; graph 4, p. 5).

5. Why are detectable cancers more often metastatic among the less affluent?

   People with a low standard of living make less use of organised screenings, which delays the discovery of detectable cancers and increases the likelihood of diagnosing them at a metastatic stage. The incidence ratio of detectable metastatic cancers is 2.3 times higher among the less affluent than among the more affluent, while this gradient disappears for non-detectable locations (graph 7, p. 6-7).

6. What does the notion of 'preventable' cancer mean in this study?

   A cancer is classified as preventable when more than 50% of cases are attributable to lifestyle and environmental factors (tobacco, alcohol, diet, overweight, occupational exposures, etc.), according to estimates from Public Health France. This criterion allows for the distinction of cancers on which primary prevention can have a strong impact (p. 5-6).

7. What practical implications are there for public health policies?

   The results align with the Fundamental Cause Theory: the more affluent avoid preventable cancers more and have detectable cancers diagnosed earlier, while the less affluent concentrate the severe and early forms. This argues for strengthening targeted prevention actions on risk factors in lower-income communities and for adapting screening and access to care systems to the constraints of these populations (conclusion, p. 6-7).

6. Rewrite in Easy-to-Read Language 

Easy-to-Read Title

Cancers affect poorer people more severely

Easy-to-Read Summary

• The document discusses cancers in France between 2013 and 2020.

• Researchers look at the link between people's money and their cancers.

• Poorer people more often have severe and advanced cancers.

• Rich people more often have less severe or early detected cancers.

Important points FALC

1. Not all cancers are equal

• Some cancers occur more in modest people, such as lung cancer.

• Others, like breast or prostate cancer, occur more in rich people.

2. Modest people have more severe cancers

• Modest people have more difficult-to-treat cancers.

• They have fewer cancers with good survival rates

• In modest people, cancer is discovered at an advanced stage.

• This makes treatments heavier and less effective.

4. Cancer occurs younger in modest people

• Modest people have cancer several years earlier than others.

• This is especially true for lung, breast, and bowel cancers.

5. Many of these cancers could be prevented

• Some cancers are linked to tobacco, alcohol, weight, or work.

• Modest people are more exposed to these risks

6. Screening mainly helps rich people

• There are tests to look for certain cancers before symptoms.

• Rich people use these tests more and discover cancer earlier.

7. To act

• We need to help modest people stop smoking and eat better.

• We also need to support them in getting screened

7. Cross-sectional analysis — values of health practices

• Literacy : The document remains technical but offers clear definitions of incidence, severity, and preventable cancers, reusable in literacy actions if simplified.

• Empowerment : The beneficiaries are not directly involved in the study, but the results allow for the design of interventions that strengthen the agency of modest audiences over their risk factors.

• Participation : No co-construction approach with users is described; participation is rather at the level of the data-producing institutions.

• Community health : The collective dimension is present through the analysis of social gradients and living conditions, which opens up to community approaches in popular areas.

• Ethics : The text explicitly describes social health inequalities without blaming individuals, emphasising structural constraints (standard of living, access to care, occupational exposures).

• Human rights : The study highlights inequalities in access to screening and early diagnosis, which relates to the principles of equity and equality in prevention and care.

• Intersectorality : The use of fiscal, health, and demographic data suggests the value of collaborations between health, social protection, work, and urban policy, even if these partnerships are not well detailed.

• Partnership : The Drees–Inserm partnership is formalised, but collaborations with field actors are not described; the text mainly serves as a basis for future local partnerships.!

• Fight against discrimination : The authors objectify the effects of living standards on cancer without essentialising groups, which aligns with non-judgement and consideration of social determinants rather than solely individual responsibility.

8. Assessment of the reliability of the resource

• Scientific relevance : The study is based on EDP-Health, a demographic sample coupled with SNDS data covering approximately 4.4% of the population, with a longitudinal follow-up over eight years and a clearly articulated methodology (inclusion criteria, standardisation, categorisation of cancers). The references used are recent (2013–2025), include international work, and are based on solid concepts such as the Fundamental Cause Theory and estimates of cancers attributable to lifestyle and the environment.

• Operational relevance : The results are presented by location, prognosis, stage, avoidability and detectability, with incidence ratios by decile of living standards, making them directly usable to prioritise populations and territories. However, the document does not provide ready-to-use tools (guides, check-lists), which requires work for appropriation and translation by field actors to transform this data into concrete actions.