Simple Letter Tweaks Nudge 85,000 Screening Non-Responders to Finally Take the Bowel Cancer Test

A single sentence tucked into an invitation letter may be enough to pull thousands of hesitant citizens back into bowel cancer screening. In one of the largest trials of its kind ever conducted, researchers in Tuscany randomized 85,193 people who had repeatedly ignored colorectal cancer screening invitations and found that small, cost-neutral messages drawn from behavioural economics significantly increased the return of faecal immunochemical tests. The work, conducted across ten organized screening programmes and published in SSM – Population Health, offers a strikingly cheap lever for health systems struggling to lift participation toward European quality targets.

Colorectal cancer remains a leading cause of cancer death worldwide, and early detection with faecal tests demonstrably reduces mortality. But the benefit of screening hinges on uptake. European guidelines consider a participation rate of at least 45 percent acceptable and 65 percent desirable, yet reported figures across programmes range from roughly 23 to 63 percent, with persistent socioeconomic gaps. In Italy, national participation averaged 34.9 percent in 2023, ranging from 18.9 percent in Lazio to 64.2 percent in Veneto. Tuscany, despite its long-standing programmes, sat at 45.8 percent, comfortably above the minimum but short of the desirable benchmark, leaving a large pool of chronic non-responders whom standard letters had repeatedly failed to reach.

The new trial, part of the BESTcc project evaluating behavioural interventions in Tuscany’s screening programmes, is grounded in dual-process theory. Traditional efforts to boost screening participation have appealed to reflective, deliberative reasoning, the slow System 2 of psychological models. Behavioural economics instead targets the fast, automatic System 1, using small contextual modifications, or nudges, such as feedback on past behaviour or social norm messages, to shape decisions without restricting choice. Two earlier randomized trials in Florence, Rome and Turin had tested such messages among previous non-responders with encouraging but sometimes inconsistent results, and the Tuscan team set out to replicate the approach in a larger, geographically diverse population spanning both urban and rural screening organizations.

The study employed a 2×2 factorial randomized design. Between January 2022 and April 2024, eligible individuals who had not responded to at least two consecutive screening invitations and had never been screened were allocated in a 3:1:1:1 ratio to one of four arms: a control group receiving the standard invitation letter, a feedback group, a minority-norm group, and a combined group receiving both messages. The messages themselves were identical to those used in the previous trials, co-developed and pretested with public and patient representatives through qualitative focus groups. The feedback message gently informed recipients that they had unfortunately not responded to previous invitations while stressing that it is never too late to take care of their health. The minority-norm message noted that in many areas six out of ten people invited accept the invitation and that the recipient belonged to the minority who did not. Exclusions covered personal history of colorectal cancer, adenomas or inflammatory bowel disease, recent colonoscopy, or a recent faecal occult blood test.

The trial was registered with the ISRCTN registry in 2022 and approved by the Ethics Committees of the Local Health Units in Tuscany, with individual informed consent waived because the trial was integrated into routine screening, an Essential Level of Care within the Italian National Health System. Randomization was performed individually with a pseudo-random number generator, and although blinding of researchers and participants was impossible, participants did not know they were part of a study and enrolment staff were blinded to the allocation sequence. Importantly, the team also verified that the added sentences did not compromise clarity, using automated natural language processing readability tools validated by crowd-sourced comprehension testing; the modified letters proved as easy to understand as the standard version.

The primary outcome was the return of a faecal immunochemical test within 90 days of the first invitation, analyzed by intention to treat, which even included people whose letters were returned undelivered. In the control group, 6.5 percent of 42,732 participants returned a test. Uptake was significantly higher in all three intervention arms: 8.4 percent in the feedback condition, 7.4 percent with the minority-norm message, and 7.7 percent in the combined condition. The 2×2 factorial analysis revealed a significant interaction between the two message types, with an odds ratio of 0.80 for the interaction term, prompting the researchers to report each modality separately against the control rather than pooling them.

Because participation rates varied substantially across the ten screening programmes, the statisticians built a mixed-effects logistic regression with random intercepts and random slopes for each programme, a model whose superiority over a single-level alternative was confirmed by likelihood ratio testing. The intraclass correlation attributable to programme-level clustering was 0.0884. In this adjusted model, feedback carried an adjusted odds ratio of 1.31 for participation, the minority-norm message 1.17, and the combination 1.22, all statistically significant relative to control, with no significant differences among the three intervention conditions. Women participated more than men, and people aged 60 to 70 were less likely to respond than those aged 54 to 59.

The age-stratified results carried the most practical weight. Among 60-to-70-year-olds, all three messages worked, and the combined message was the strongest performer there, with an adjusted odds ratio of 1.49. Among the younger 54-to-59 group, only feedback showed a significant effect, with an adjusted odds ratio of 1.21. Translated into absolute terms, feedback produced the best figures: an absolute risk reduction of 1.84 percentage points overall, meaning roughly 54 invitations bearing the feedback sentence were needed to generate one additional screened person, improving to 48 invitations in the older subgroup. The combined message achieved an absolute risk reduction of 2.40 percent and a number needed to invite of 57 among those aged 60 to 70.

Why did stacking the two nudges not outperform either alone? The authors offer several speculative explanations, carefully flagged as such because no process data on message perception or recall were collected. The minority-norm message already contains an implicit feedback element, since telling recipients they are part of the minority who did not participate inherently reminds them of their past behaviour, so the explicit feedback sentence may have added little new information. Longer or more complex combined messages may also have diluted attention, or ceiling effects may have limited further gains. The researchers recommend that future studies incorporate brief process measures, such as recall, perceived relevance and comprehension, and explore factorial designs, message sequencing or qualitative interviews to disentangle these mechanisms, as well as testing more clearly distinct interventions such as deadlines.

The findings arrive with clear caveats and clear strengths. Household spill-over could not be controlled since members of the same family could be randomized to different arms, message placement within the letter may have limited salience, and the analysis was restricted to age and sex without socioeconomic variables such as deprivation, education or health literacy. Nevertheless, the trial’s scale, its objective participation data, its single-region consistency and its focus on the hardest-to-reach group, people who had ignored two consecutive invitations, make the result compelling for screening organizers. The interventions are essentially free, requiring only small edits to existing letters, and the effects were demonstrated in a mix of long-established and newer programmes. The authors caution that organized, population-based faecal-test programmes differ from opportunistic or colonoscopy-based systems, and that cultural norms and trust in health authorities shape how such messages land, so extrapolation elsewhere should be cautious. Within that frame, the message is simple: a reminder of what someone has not done, and an honest account of what their neighbours have, may be enough to bring the reluctant back into the prevention fold.

Subject of Research: Behavioural economics messages to increase participation in population-based colorectal cancer screening among previous non-responders in Tuscany

Article Title: Testing feedback and social norms messages to increase participation: the BestCC randomized controlled trial among non-respondents in population-based colorectal cancer screening programmes in Tuscany

Article References: Gorini, G., Betti, E., Falini, P., Carreras, G., Mallardi, B., Migliori, M., Giorgi, D., Tramalloni, D., Cosimi, A., Venturi, G., Ravelli, A. A., Iossa, A., Battisti, F., Stoffel, S. T., & Mantellini, P. (2026). Testing feedback and social norms messages to increase participation: the BestCC randomized controlled trial among non-respondents in population-based colorectal cancer screening programmes in Tuscany. SSM – Population Health, Article 101974. https://doi.org/10.1016/j.ssmph.2026.101974

Image Credits: AI Generated

DOI: Not provided

Keywords: colorectal cancer screening, behavioural economics, nudge, social norms, feedback messages, FIT test, randomized controlled trial, Tuscany, screening participation, non-responders, public health, health communication