The same result, described in two ways
Imagine 100 people with a particular cancer. Without an additional treatment, 10 would be expected to have a recurrence within five years. With treatment, that falls to 5.
The absolute reduction is 5 percentage points: 5 fewer recurrences for every 100 people over those five years.
The relative reduction is 50%, because 5 is half of 10. Both descriptions are mathematically correct. They answer different questions.
These are invented numbers to explain the arithmetic, not results for a particular cancer or treatment.
Why the starting risk matters
Now imagine the starting risk is 2 in 100 rather than 10 in 100. If a treatment halves that risk, it becomes 1 in 100. The relative reduction is still 50%, but the absolute reduction is now 1 percentage point.
That is why a striking percentage, without the starting risk, can give an incomplete impression. It also helps explain why the same treatment may have different expected benefits for people with different clinical circumstances.
Ask what is being measured
A figure might describe recurrence, death from cancer, overall survival, tumour shrinkage or the chance of a side effect. These outcomes are not interchangeable. The comparison also needs a time period: a five-year risk is not the same measure as a one-year risk.
Ask what treatment is being compared with. “Without this additional treatment” may still include surgery or other care; it does not necessarily mean no cancer treatment at all.
Numbers support a decision; they do not make it
Expected benefit needs to be considered alongside possible harms, treatment burden and what matters to you. Estimates come from groups of people and may not predict an individual's outcome precisely. Your team can explain how well the evidence fits your diagnosis and health.
Useful questions include:
- Out of 100 people like me, how many would have this outcome with and without treatment?
- Over what period, and how uncertain is that estimate?
- What are the important side effects and alternatives?
This discussion is part of informed consent. Read more about what a prognosis can and cannot tell you.
General information only. These examples cannot estimate your own recurrence risk or recommend treatment.