A single FIT test detects around 88.7% of colorectal cancers in people with symptoms. Which raises an obvious question: would two tests do better?
The answer is more interesting than a simple yes or no, and it depends on which of two quite different things you mean.
Two approaches that sound the same but aren’t
Dual FIT means collecting two samples from two different stools over a short period, and treating a raised result from either one as positive.
Repeat FIT means doing one test, getting a normal result, and then testing again later because symptoms have not gone away.
These have been evaluated separately and the conclusions differ.
Dual testing: not currently recommended
NICE reviewed dual sampling and concluded against implementing it for now.
The reasoning is a familiar trade-off. Taking two samples does improve sensitivity, because you have two chances to catch intermittent bleeding. But it reduces specificity: more people are flagged who have nothing wrong, each of whom then faces an investigation with its own small risks.
There is also a practical problem that is easy to overlook. Asking people to collect two samples reduces the proportion who complete and return the test at all. A more sensitive test that fewer people finish is not necessarily a better test in the real world.
NICE concluded that a single sample at a single threshold remains the most effective and cost-effective approach on current evidence, and recommended further research before dual testing is adopted.
Repeat testing: genuinely useful in the right situation
This is where a second test earns its place.
If your first result was normal but symptoms have persisted, a repeat test is a reasonable part of continuing to look. A study of patients whose first FIT was negative found that two negative FIT results together bring the risk of colorectal cancer below 0.04%.
NHS England has advised that a second test should be offered where clinical concern remains after a negative first result. The authors of the underlying study were careful to note that fuller economic evaluation is still needed, so this is an area still developing.
The important framing: a repeat test is part of safety-netting, not a way of double-checking a result you did not like.
What this means for you
If you feel well and are testing for peace of mind, one test is the right answer. Doing two will not meaningfully improve what you learn.
If your result was normal and your symptoms have settled, one test is again sufficient.
If your result was normal and symptoms have persisted or worsened, do not simply repeat the test and stop there. A repeat FIT can be part of the picture, but persistent unexplained symptoms with a normal result should prompt reassessment by a doctor, and possibly other tests entirely. Blood tests, a calprotectin test, or referral through a different pathway may all be more informative than a second stool sample.
If you are testing periodically over time, knowing your actual figure rather than a positive or negative answer becomes far more useful. A result of 5 last year and 45 this year is meaningful information. Two “normal” results would have concealed the change completely.
The principle underneath all of this
A test result is a piece of evidence, not a verdict. If symptoms persist without explanation, the right response is to keep evaluating until either the symptoms resolve or a cause is found.
Repeating a test can be part of that. It is not a substitute for it.
References
- National Institute for Health and Care Excellence. Quantitative faecal immunochemical testing to guide colorectal cancer pathway referral in primary care. Diagnostics guidance DG56, 2023.
- Two negative FIT results and subsequent colorectal cancer risk. BMJ Open 2022;12:e059940.
- Bailey JA, et al. Double faecal immunochemical testing in patients with symptoms suspicious of colorectal cancer. British Journal of Surgery 2023.
- NHS England letter to general practices on FIT safety-netting, October 2022.
Figures are drawn from the sources above. Framing and commentary are the author’s own. Last reviewed August 2026.