Effectiveness of health mediation to promote organized cancer screening among underserved and under-screened populations in Marseille, France: findings from a repeated cross-sectional survey

A repeated cross-sectional survey in Marseille demonstrates that a multicomponent health mediation program significantly increased breast and colorectal cancer screening uptake and awareness among exposed individuals in disadvantaged neighborhoods, although it did not produce a significant effect at the overall population level.

The Gist

Imagine the healthcare system as a massive, high-tech library. It has all the best books (free cancer screenings) available for everyone to borrow, and the librarians (doctors) are ready to help. But for many people living in the "quiet corners" of Marseille, France, the library feels like a maze. The signs are confusing, the doors feel locked, and they don't know where to start.

This paper is about a special team of guides (called "health mediators") who were hired to walk into those quiet corners, find the people who are lost, and personally escort them to the right books.

Here is the story of what they found, told simply:

The Mission: "13 en Santé"

In 2022, a program called "13 en Santé" launched in the most disadvantaged neighborhoods of Marseille. Think of these neighborhoods as the "forgotten aisles" of the city. The goal was simple: get more people to check for three common cancers (breast, colorectal, and cervical) using free tests.

The guides didn't just wait in an office. They went door-to-door, set up booths on the street, and held workshops. They acted as bridges, translating medical jargon into plain language and helping people fill out forms or find their way to a doctor.

The Experiment: A Before-and-After Snapshot

To see if this worked, researchers took two "snapshots" of the city:

1. Before the guides arrived (2022): They asked thousands of people, "Have you had your cancer screening?"
2. After the guides had been working for a while (2024): They asked the same question again.

They compared the "Guided Neighborhoods" (where the team worked) against "Control Neighborhoods" (similar areas where no guides were sent).

The Big Surprise: The "Neighborhood" vs. The "Individual"

Here is the twist in the story:

• The Neighborhood View (The Big Picture): When the researchers looked at the entire neighborhood as a whole, the numbers didn't change much. The overall rate of screenings in the guided areas was roughly the same as in the non-guided areas. It was like looking at a whole forest and not seeing a difference in the number of trees, even though some trees were getting extra water.
• The Individual View (The Personal Connection): But when they looked at specific people who actually met a guide, the story changed completely.
  • People who shook hands with a guide were much more likely to get screened.
  • Breast Cancer: Those who met a guide were 2.3 times more likely to get a mammogram.
  • Colorectal Cancer: Those who met a guide were 2.8 times more likely to take the test.
  • Cervical Cancer: The numbers went up, but not enough to be statistically certain (it was a "maybe").

The Analogy: Imagine a concert. If you look at the whole stadium, the crowd size might look the same. But if you look at the people who were personally invited by a VIP host, those specific people are almost guaranteed to be in the front row. The program worked brilliantly for the individuals it reached, but it wasn't quite loud enough to change the vibe of the entire neighborhood yet.

Why Did It Work for Some and Not Others?

The study found that the guides were most effective at helping people who faced the biggest hurdles:

• The "Lost" People: They helped people who didn't speak French well, didn't have a permanent address, or didn't have a regular doctor.
• The "Confused" People: For colorectal cancer, many people simply didn't know the test existed or what it was. The guides acted as educators, clearing up the confusion.
• The "Stuck" People: For breast cancer, most people knew about the test but were too busy, scared, or didn't know how to book an appointment. The guides acted as navigators, pushing open the door and walking them through the process.

The One Thing That Didn't Work (Yet)

The program struggled a bit with cervical cancer.

• Why? In France, this test is usually done by a gynecologist or a midwife, not a general doctor. The guides were great at connecting people to general doctors, but the "specialist" path is a different road. It's like the guides were excellent at helping people find the "General Library," but the "Specialist Archive" was in a different building they hadn't mapped out yet.

The Takeaway

This study is a hopeful sign. It proves that human connection works. When you send a friendly, trusted person to knock on the door of someone who feels left out of the healthcare system, that person is much more likely to take care of their health.

However, to change the whole city, the program needs to be louder, reach more people, and maybe figure out how to better connect people to the specialists they need. It's not a magic wand that fixes everything overnight, but it is a very powerful tool for the people who need it most.

Technical Summary

1. Problem Statement

Despite the availability of free, organized cancer screening programs in France (breast, colorectal, and cervical), significant social and territorial inequalities persist. In Marseille, screening rates are lower than regional and national averages, with a marked North–South divide where socioeconomically disadvantaged "priority neighbourhoods" exhibit the lowest uptake.

• The Gap: While Community Health Worker (CHW) models are well-studied in the US and low-income countries, evidence regarding their effectiveness in European contexts, specifically for multicomponent health mediation interventions targeting multiple cancer types simultaneously, is scarce.
• The Intervention: The "13 en Santé" program was launched in October 2022 in nine priority neighbourhoods of Marseille. It employs health mediators to provide education, outreach, and navigation support to bridge the gap between underserved populations and the healthcare system.
• Objective: To evaluate the effectiveness of this multicomponent intervention in promoting breast, colorectal, and cervical cancer screening uptake and awareness among residents of disadvantaged neighbourhoods.

2. Methodology

Study Design:

• Type: Controlled before-and-after study using two repeated cross-sectional surveys.
• Timeline:
  • Baseline (T0): August–December 2022 (before program implementation).
  • Follow-up (T1): April–September 2024 (approx. 1.5 years post-implementation).
• Setting: 13 intervention neighbourhoods (where "13 en Santé" operates) and 26 socioeconomically matched control neighbourhoods in Marseille.

Participants & Sampling:

• Population: Residents aged 18–74 years.
• Sampling: Random door-to-door recruitment.
• Sample Size: 4,523 individuals included in the final analysis (4,095 eligible for at least one screening).
• Weighting: Data were weighted to align with census data regarding age, sex, and stratified sampling to ensure representativeness.

Data Collection:

• Instrument: Knowledge-attitude-practice questionnaires administered via tablets.
• Key Variables:
  • Outcomes: Self-reported screening uptake (up-to-date status) and awareness of screening programs.
  • Exposure: Self-reported encounter with a health mediator (individual level) and residence in an intervention neighbourhood (population level).
  • Covariates: Socioeconomic profiles (constructed via clustering: Native/Non-native speaker, renter/homeowner, student), access to primary care, receipt of invitation letters, and demographic factors.

Statistical Analysis:

• Models: Weighted logistic regression models were fitted for each cancer type.
• Levels of Analysis:
  1. Population-level: Compared uptake/awareness between intervention vs. control neighbourhoods (Time $\times$ Group interaction).
  2. Individual-level: Compared participants who reported meeting a mediator vs. those who did not, regardless of neighbourhood.
• Adjustments: Models were adjusted for age, socioeconomic profile, primary care access, invitation receipt, and gender.

3. Key Contributions

• First European Evidence: This is the first study in Europe to evaluate a multicomponent health mediation program targeting breast, colorectal, and cervical cancer screening simultaneously in disadvantaged urban settings.
• Dual-Level Evaluation: The study uniquely distinguishes between population-level effects (community-wide impact) and individual-level effects (impact on those directly reached), providing a nuanced understanding of intervention reach and efficacy.
• Mechanism Elucidation: It investigates specific mechanisms of action (awareness vs. navigation) and identifies differential effectiveness across cancer types.

4. Key Results

Demographics & Reach:

• The program successfully reached individuals with cumulative socioeconomic barriers (non-native speakers, renters, low education).
• Exposure: In the T1 survey, 12.78% of participants in intervention areas reported encountering a health mediator. Non-native speaker renters were significantly more likely to report an encounter.

Population-Level Effects:

• Null Result: No statistically significant difference was observed in screening uptake or awareness between intervention and control neighbourhoods over time. The intervention did not shift the aggregate population rates.

Individual-Level Effects (Exposure to Mediators):

• Breast Cancer: Significant increase in uptake among those exposed to mediators.
  • Uptake: 74% (exposed) vs. 54% (unexposed).
  • Odds Ratio (OR): 2.26 [1.12–4.54].
• Colorectal Cancer: Significant increase in uptake and awareness.
  • Uptake: 50% (exposed) vs. 30% (unexposed).
  • OR: 2.75 [1.31–5.77].
  • Awareness: 93% (exposed) vs. 83% (unexposed); OR: 8.07 [2.10–30.96].
• Cervical Cancer: No statistically significant increase in uptake (OR 1.79, p>0.05) or awareness, though an upward trend was noted.

Other Determinants:

• Primary Care Access: Strongly associated with breast and colorectal uptake, but not cervical cancer uptake (suggesting a different pathway involving specialists like gynaecologists).
• Invitation Letters: Significantly increased uptake for all three cancers.
• Socioeconomic Disparities: Non-native speaker renters consistently had lower access to primary care, lower receipt of invitation letters, and lower screening awareness compared to native speakers.

5. Significance and Interpretation

• Effectiveness: The study provides evidence that health mediation is effective at the individual level, increasing breast and colorectal screening uptake by approximately 20 percentage points for those directly reached. This aligns with, and slightly exceeds, median effects reported in US systematic reviews.
• Mechanisms of Action:
  • Colorectal Cancer: The intervention likely works through a dual mechanism: raising awareness (which was low) and providing navigation to overcome barriers.
  • Breast Cancer: Since awareness was already high, the intervention likely functioned primarily through navigation support (solving practical barriers to access).
  • Cervical Cancer: The lack of effect suggests that health mediation alone is insufficient for this screening type, which relies heavily on specialist care (gynaecologists/midwives) rather than primary care or general navigation.
• Limitations & Future Directions:
  • The lack of population-level impact may be due to "dilution effects" (mediators reaching people outside target zones), low coverage (only ~6% of the target population surveyed), or the short duration of the intervention.
  • Self-reported data introduces potential recall and social desirability biases.
  • Conclusion: Health mediation is a promising strategy to reduce inequalities in cancer prevention in France, but future implementations must refine targeting (especially for cervical cancer) and utilize medico-administrative data to confirm long-term population impacts.