- Design
- Systematic review and network meta-analysis of randomised trials, with GRADE certainty assessment
- Population
- 76 randomised clinical trials across eight colorectal cancer screening uptake strategies
- Primary outcome
- Colorectal cancer screening uptake, compared as risk ratios with P-score rankings
- Effect
- Patient navigation RR 1.58 (95% CI 1.23-2.02) and mailed faecal test outreach 1.36 (1.07-1.74) beat usual care; in settings below 30% uptake, mailed testing RR 3.12 (1.70-5.71)
Colorectal cancer screening reduces mortality and uptake is poor everywhere, so the question is which intervention to spend money on. This network meta-analysis pooled 76 randomised trials across eight strategies - patient navigation, mailed faecal immunochemical test outreach, educational multimedia, reminders alone, choice-based outreach, colonoscopy outreach, multistep interventions, and usual care - and ranked them.
Two won clearly. Patient navigation raised uptake by 58% relative to usual care (RR 1.58, 95% CI 1.23-2.02) and mailed faecal test outreach by 36% (1.36, 1.07-1.74). Educational multimedia (1.27, 0.91-1.78) and reminders alone (1.24, 0.98-1.57) had intervals crossing 1.0. Choice-based outreach and colonoscopy outreach were no better than usual care, and mailed faecal testing beat colonoscopy outreach directly (1.35, 1.11-1.63).
The subgroup finding is the operationally useful one. Where baseline uptake was under 30%, mailed faecal test outreach was dramatically the most effective (RR 3.12, 1.70-5.71). Where uptake was already 30% or higher, educational multimedia performed best. So the right intervention depends on where the service is starting from, and a strategy that works in a well-screened population is the wrong one for an unscreened one.
That maps directly onto Indian practice, where organised colorectal screening barely exists and baseline uptake is close to zero. Inviting people to a colonoscopy - the instinctive approach, and the one that gets funded - was among the least effective strategies tested. Posting a faecal test with clear instructions and a return envelope, backed by someone whose job is to chase non-responders, is what the evidence supports. Certainty of evidence was mostly moderate to low, so treat the ranking as a guide to where to start rather than a settled hierarchy.
- In a low-uptake population, post the faecal test out; do not lead with a colonoscopy invitation
- Fund a navigator role explicitly - it was the top-ranked intervention across all settings
- Reminders alone and choice-based outreach did not reliably beat usual care - do not rely on them
- Switch to education and multimedia only once baseline uptake is above about 30%
- Measure your own baseline uptake before choosing a strategy; it determines which one works
The statistics, in plain English
A risk ratio of 1.58 means 58% more people screened than with usual care, not 58% of people screened - the absolute gain depends entirely on the starting rate. Educational multimedia's interval of 0.91 to 1.78 crosses 1.0, so despite a promising point estimate it could represent no benefit. Network meta-analysis compares interventions that were often never tested head to head, by linking them through common comparators, which makes the rankings less certain than the individual comparisons - and the authors grade most of the evidence as moderate to low certainty. The RR of 3.12 in low-uptake settings has a wide interval (1.70-5.71) and comes from a subgroup.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for gastroenterology & hepatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free