Non-Concordance between Patient and Clinician Estimates of Prognosis in Advanced Heart Failure: Prognostic Estimates in Advanced HF

Laura P. Gelfman, Harriet Mather, Karen McKendrick, Angela Y. Wong, Mathew D. Hutchinson, Rachel J. Lampert, Hannah I. Lipman, Daniel D. Matlock, Keith M. Swetz, Sean P. Pinney, R. Sean Morrison, Nathan E. Goldstein

Research output: Contribution to journalArticlepeer-review

6 Scopus citations


Importance: Despite efforts to enhance serious illness communication, patients with advanced heart failure (HF) lack prognostic understanding. Objectives: To determine rate of concordance between HF patients’ estimation of their prognosis and their physician's estimate of the patient's prognosis, and to compare patient characteristics associated with concordance. Design: Cross-sectional analysis of a cluster randomized controlled trial with 24-month follow-up and analysis completed on 09/01/2020. Patients were enrolled in inpatient and outpatient settings between September 2011 to February 2016 and data collection continued until the last quarter of 2017. Setting: Six teaching hospitals in the U.S. Participants: Patients with advanced HF and implantable cardioverter defibrillators (ICDs) at high risk of death. Of 537 patients in the parent study, 407 had complete data for this analysis. Intervention: A multi-component communication intervention on conversations between HF clinicians and their patients regarding ICD deactivation and advance care planning. Main Outcome(s) and Measure(s): Patient self-report of prognosis and physician response to the “surprise question” of 12-month prognosis. Patient-physician prognostic concordance (PPPC) measured in percentage agreement and kappa. Bivariate analyses of characteristics of patients with and without PPPC. Results: Among 407 patients (mean age 62.1 years, 29.5% female, 42.4% non-white), 300 (73.7%) dyads had non-PPPC; of which 252 (84.0%) reported a prognosis >1 year when their physician estimated <1 year. Only 107 (26.3%) had PPPC with prognosis of ≤ 1 year (n=20 patients) or > 1 year (n=87 patients); (Κ = -0.20, p = 1.0). Of those with physician estimated prognosis of < 1 year, non-PPPC was more likely among patients with lower symptom burden- number and severity (both p ≤.001), without completed advance directive (p=.001). Among those with physician prognosis estimate > 1 year, no patient characteristic was associated with PPPC or non-PPPC. Conclusions and Relevance: Non-PPPC between HF patients and their physicians is high. HF patients are more optimistic than clinicians in estimating life expectancy. These data demonstrate there are opportunities to improve the quality of prognosis disclosure between patients with advanced HF and their physicians. Interventions to improve PPPC might include serious illness communication training.

Original languageEnglish
Pages (from-to)700-705
Number of pages6
JournalJournal of Cardiac Failure
Issue number6
StatePublished - Jun 2021


Dive into the research topics of 'Non-Concordance between Patient and Clinician Estimates of Prognosis in Advanced Heart Failure: Prognostic Estimates in Advanced HF'. Together they form a unique fingerprint.

Cite this