Executive Summary

Why this matters

Communicating statistics on risk is an important part of patient information. 61% of people aged 16-65 are unable to understand health information if numeracy skills are required[1]. As information producers, our job is to help people feel confident in making decisions about their health.

Numbers not words

  • Interpretation of words like rare and common varies enormously. Always use a statistic such as 1 in 100 people alongside words like rare or common.
  • Use expected frequencies rather than percentages, for example, 10 in 100, rather than 10%.
  • When giving people two frequencies to compare, make sure they are both expressed as ‘out of’ the same number (e.g. 1 in 100 compared with 2 in 100, NOT 1 in 100 compared with 1 in 50).

Use absolute risk rather than relative risk

  • The absolute risk of an event increases from 1 in 100 to 2 in 100, but the relative risk of the event doubles. Using relative risk in isolation can be misleading.

Illustrating risk

  • A great way to improve users’ understanding of risk and statistics is to use visual aids.
  • Using a mix of numerical and pictorial formats to communicate risk is helpful.
  • Visual displays may be most helpful for giving people an overall pattern, whereas actual numbers can be better for communicating detail.

Perceptions of risk

  • Consider using both positive and negative framing, i.e. ‘3 out of 100 people experienced this side effect, but 97 out of 100 did not’.
  • Consider adding some context to your statistics. What would make a useful comparison to give people a sense of the likelihood, or to help them make their decision?

Explaining uncertainty

  • Communicate the uncertainty of data.

Pre-empt misunderstandings

  • If something is easily misunderstood, or there is a common myth, it is better to address it head-on.



This guide supports the following PIF TICK criteria:

4.3 Where appropriate, risks and benefits have been communicated in an accurate, balanced, unbiased way that the target audience can understand.

5.0 Involving users

6.2 The health literacy needs of users have been considered.



1. A mismatch between population health literacy and the complexity of health information: an observational study. Gillian Rowlands, Joanne Protheroe, John Winkley, Marty Richardson, Paul T Seed, Rima Rudd. https://pubmed.ncbi.nlm.nih.gov/26009533/

Please login or join us to view the full guide