Optimising methods for communicating survival data to patients undergoing cancer surgery

BACKGROUND: Patients undergoing cancer surgery require outcome data to inform decisions, but communication of numerical risk is difficult. This study assessed patient understanding of survival data presented in different formats. METHODS: Semi-structured interviews in which patients interpreted four presentation formats of survival data (three graphical and one narrative) were audio-recorded. The interviewer and a blinded observer (listening to the audio-recordings) scored patients' understanding of each format. Logistic regression examined associations between understanding and clinical and socio-demographic details. RESULTS: Seventy participants with colorectal cancer were interviewed and 67 [95.7%, 95% confidence intervals (CIs) 90.9-100%] correctly interpreted a simplified Kaplan-Meier survival curve. A high proportion accurately understood data presented as a bar chart or pictograph (94.3%, 95% CIs 88.7-99.9% and 92.9%, 95% CIs 86.7-99.0% respectively). Standard narrative alone was least well understood (n=53, 75.7%, 95% CIs 65.4-86.0%). Multivariable analyses demonstrated that older and female patients had poorer overall understanding (OR 0.93 per year, 95% CIs 0.87-0.98, p=0.01 and OR 0.24, 95% CIs 0.07-0.86, p=0.03). CONCLUSION: Patient understanding of survival data was higher when presented with graphs compared to narrative alone. Further work examining understanding in the clinical context and before surgery is recommended before this can be used routinely.