Guidelines
Suggest edit

Separate baseline cases from treatment-added cases in risk pictographs

For comparing treatment side effects against preexisting risk at a single time point, use separate visual encoding for baseline and incremental cases on a risk display to improve fidelity and mitigate the mistake of attributing the full treated-group risk to the treatment for patient decision-makers with low numeracy.

  • purpose:refine
  • basis:empirical
  • task:compare
  • time:timepoint
  • quality:fidelity
  • lever:encoding
  • operator:difference
  • audience:decision-maker

advice

Incremental risk encoding

Show baseline cases and treatment-added cases as separate visual groups within the same denominator. For example, in a pictograph, color the cases that would occur anyway in one color and only the extra cases caused by treatment in a second color instead of showing only the total treated-group risk.

reason

Why incremental risk encoding works

A total-risk-only display encourages readers to blame the whole treated-group risk on the treatment. Separating baseline from added risk makes the treatment-caused increment visible and reduces over-attribution.

Mechanism: Readers can see which cases are part of the preexisting baseline and which cases are added by treatment, so they judge the treatment’s contribution more accurately.

Evidence: The paper describes studies showing that separating baseline and incremental risk within pictographs reduced worry about side effects and reduced perceived likelihood of experiencing a side effect; it also notes that the knowledge benefit of the incremental format required pictographs (Fagerlin et al., 2011).

Notes: The paper illustrates this move with one color for baseline cases and another color for the additional treated cases.

context

Use when treatment changes an existing baseline risk

  • User Goal: Judge how much extra risk a treatment adds.
  • Task: Compare baseline risk with treatment-caused incremental risk.
  • Data: A side effect or complication that can occur both without treatment and with treatment.
  • Chart Setting: A static risk graphic that explains side effects for a treatment decision.
  • Audience: Patients weighing risks and benefits of an intervention.
  • Success Criterion: Readers can distinguish what would happen anyway from what is added by treatment.

exceptions

Do not use the incremental format without a pictograph

Break it when: The incremental risk framing is shown without a pictograph. Why: The paper reports that the knowledge advantage of the incremental approach depended on combining it with pictographs.

costs

Tradeoffs of incremental risk encoding

Sacrifice: You give up some visual simplicity by adding another visual group. Risk: If the added cases are not clearly separated from baseline cases, the treatment effect can still be missed. Mitigation: Keep the same denominator visible and add only the extra treatment-caused cases in a distinct visual encoding.

mistakes

Common failure mode in side-effect graphics

Mistake: Show only the total complication rate in the treatment group. Why it fails: Readers may treat the entire displayed risk as if it were caused by the treatment alone.

check

How to check the encoding

Failure Sign: A reviewer cannot tell which cases belong to baseline risk and which cases are added by treatment. Quick Check: Ask whether each highlighted case would still be present without the treatment. Stronger Test: Compare the display with a total-risk-only version and check which one makes the treatment-caused increment directly visible.

fix

How to fix the display

  • Start with the baseline risk in the full denominator.
  • Add a second visual encoding for only the extra cases caused by treatment.
  • Avoid a treated-group total that does not distinguish baseline from added risk.

References

Fagerlin, A., Zikmund-Fisher, B. J., & Ubel, P. A. (2011). Helping Patients Decide: Ten Steps to Better Risk Communication. JNCI Journal of the National Cancer Institute, 103(19), 1436–1443. https://doi.org/10.1093/jnci/djr318