Communicating risk and evidence
- Concept
Good risk communication gives absolute figures with the same denominator and time frame, and says how uncertain they are.
What it is
Risk communication turns research results into information that patients, clinicians and decision-makers can use. The same figure looks huge or tiny depending on whether it is stated in relative or absolute terms, as a percentage or a natural frequency, or as a gain or a loss. Good communication shows benefits and harms on the same scale, states the uncertainty, and keeps the evidence apart from the values of whoever decides.
Worked example
- Data (made up): over 10 years, 40 of 10 000 untreated people and 30 of 10 000 treated people have a stroke.
- Risks: 40 / 10 000 = 0.4 % and 30 / 10 000 = 0.3 %; RR = 0.3 / 0.4 = 0.75, a relative reduction of 25 %.
- Absolute reduction: 0.4 − 0.3 = 0.1 percentage points, that is, 1 fewer stroke per 1 000 people.
- NNT = 1 / 0.001 = 1 000 people treated for 10 years to prevent 1 stroke.
- Harm (made up): 5 in 1 000 treated people have a serious adverse effect; for each stroke prevented, 5 people have that effect.
- Uncertainty (made up): the 95 % CI of the absolute reduction runs from 0 to 2 per 1 000; the benefit could be nil.
- Balanced message: of 1 000 people, 996 untreated and 997 treated will have no stroke in 10 years; 5 treated will have the adverse effect.
- Reading: ‘cuts risk by 25 %’ sounds big; natural frequencies with harms and uncertainty allow an informed decision.
The same evidence said four ways
| Form | Example | Effect on the listener |
|---|---|---|
| Relative reduction | Cuts stroke by 25 % | Sounds like a big benefit |
| Absolute reduction | 1 fewer stroke per 1 000 people over 10 years | Shows the real size |
| Number needed to treat | 1 000 people treated for 10 years to prevent 1 stroke | Helps weigh the effort |
| Natural frequencies with harms | Of 1 000: 4 versus 3 strokes; 5 with an adverse effect | Lets benefit and harm be compared |
Markers
- RRR
- ARR
- NNT
- Natural frequencies
- Framing
- 95 % CI
In practice
- In the clinic, decision aids with icon arrays of 100 or 1 000 figures show benefits and harms on the same scale.
- In public health, clear, fast and honest messages about what is known and not known sustain trust during an emergency.
- Decision-makers also need the number of people affected, the cost and the effect on equity.
Common pitfalls
- Reporting only the relative reduction of a benefit and the absolute risk of a harm (or the reverse): it misleads.
- Switching denominators between figures (per 100 in one, per 10 000 in another).
- Leaving out the time frame: a 1-year and a 10-year risk are not comparable.
- Hiding uncertainty for fear of confusing people: trust is lost later when the evidence changes.
Test yourself
Think of the answer, then tap the question to see it.
Why can relative risk reduction mislead?
Because it hides the baseline risk: a 25 % reduction can mean preventing 1 case per 1 000 people or 1 per 10, depending on the starting risk.
What are natural frequencies?
Figures stated as numbers of people in a concrete group (‘4 in every 1 000’); they are easier to understand than percentages or conditional probabilities.
What is the framing effect?
The change in perception and decision depending on whether the same fact is presented as a gain (‘90 % survive’) or a loss (‘10 % die’).
Besides the number, which three elements should a well-communicated risk figure carry?
The denominator (out of how many people), the time frame (over how long) and the uncertainty (interval or certainty of evidence), with harms on the same scale.
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Chapter 15: Interpreting results and drawing conclusions (Cochrane, 2024); Communicating Risks and Benefits: An Evidence-Based User's Guide (U.S. Food and Drug Administration, 2018); Chapter 14: Completing ‘Summary of findings’ tables and grading the certainty of the evidence (Cochrane, 2026).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.