C6
Communicating the results of randomised controlled trials: do patients understand survival data?
Christopher R Davis1, Angus G K McNair2, Michael G Clarke3, Jane M Blazeby1
1University Hospitals Bristol NHS Foundation Trust, Bristol, UK, 2Department of Social Medicine, Bristol, UK, 3Department of Upper Gastrointestinal Surgery, Plymouth, UK
Background
Shared decision-making frequently involves the communication of survival data, most often derived from randomised trials. However, evidence suggests that patients may not understand such statistics. This study assessed methods for improving understanding of survival data.
Method
Patients with colorectal cancer underwent semi-structured interviews based upon hypothetical clinical scenarios describing surgical treatment for cancer leading to various proportions of long-term survival. Scenarios A, B and C presented graphical survival data (bar chart, pictograph, Kaplan-Meier curve respectively). Patients were asked standard questions examining their interpretation of the information. Correct understanding of the scenario was achieved when patients correctly responded to two questions regarding each graph. Scenario D presented narrative survival data only, followed by similar questions assessing interpretation. Logistic regression examined associations with clinical and sociodemographic details and understanding of each scenario.
Results
Some 70 patients (42 male, mean age 66) participated, of whom 43 (61%) had colonic and 27 (39%) rectal cancer. There was a very good understanding of the visual presentation methods: 66 (94%, 95% CI 0.89 1.00) correctly understood Scenario A (bar chart), 65 (93%, 95% CI 0.87 0.99) Scenario B (pictograph) and 67 (96%, 95% CI 0.91 1.00) Scenario C (Kaplan-Meier curve). Patients had a lower understanding of narrative alone, with only 53 participants (75%, 95% CI 0.65 0.86) correctly understanding Scenario D (narrative). Univariable analyses demonstrated that education, age, social class, ASA grade, cancer stage and location were not associated with understanding. Male sex was associated with a higher understanding of Scenario D (narrative) (OR 3.89, p=0.02), retaining significance in the multivariable model (p=0.04).
Conclusion
Most patients with colorectal cancer understand survival data presented using visual methods. It is recommended that visual methods be used to compliment verbal information during shared decision-making in cancer treatment.
