PATIENT PREFERENCES FOR COMMUNICATION OF TAILORED SURVIVAL ESTIMATES

Purpose: To explore patient preferences for, reactions to, and understanding of the provision of tailored risk information from the Seattle Heart Failure Model (SHFM).

Method: Team members not involved in the patients’ direct care conducted semi-structured interviews to explore patient perspectives on their decision to view or not view tailored SHFM data. Fear, anxiety, and preferences for presentation of information were measured with scales. Follow-up interviews were conducted after 3–16 weeks to assess the longer term impact of patients’ prognosis decisions. Themes were identified using a team-based, inductive approach, identifying points of integration with quantitative data.

Result: We interviewed 24 patients (mean age=56.2 [23–94], female=8, inpatient=11) with symptomatic heart failure. 1) Preferences: Those who declined to view their SHFM score (n=7) expressed a belief that the information was not applicable to them. For SHFM score receivers (n=17), getting prognostic information shortly after diagnosis was ideal. Fifteen responded that an annual (at minimum) review of prognosis information would be a useful ‘benchmark.’ Patients indicated that the model should be further personalized by taking into account individual characteristics, medical adherence, comorbidities, and family history. 2) Reactions: SHFM score receivers expressed that prognostic knowledge, whether ‘good’ or ‘bad’ and despite its high degree of uncertainty, restores some control and hope. Follow-up interviews showed that preferences largely remained stable with no significant change in fear/anxiety. 3) Understanding: Uncertainty was a deciding factor for those who declined to view their prognostic information. All 17 receivers understood the uncertainty inherent in a populational estimate—‘it’s sort of like predicting the weather.’ Patients viewed the survival estimates as a ‘guideline,’ yet inpatients were more likely to view the score as a ‘wakeup call.’

Conclusion: Patients with symptomatic heart failure often value having access to tailored prognostic survival estimates and generally understand that prognostic information is uncertain. While it remains unknown whether addition of variables would improve risk model performance, the preference for patient-prioritized covariates illustrates the challenge of creating risk models that are both accurate and patient-friendly.