Background Physicians face uncertainty when predicting death in heart failure (HF) leading to underutilisation of palliative care. To facilitate decision-making, we assessed the Seattle Heart Failure Model (SHFM) as a referral tool by evaluating its performance in predicting 1-year event-free survival from death, heart transplant (HTx), and ventricular assist device (VAD) implantation.
Methods We retrospectively reviewed the charts of consecutive patients with advanced ambulatory HF with New York Heart Association Class III/IV HF and a left ventricular ejection fraction of ≤40% from 2000 to 2016. We evaluated SHFM’s performance by using the Cox proportional hazards model, its discrimination using the c-statistic, its calibration by comparing the observed and predicted survival and its clinical utility by hypothetically assessing the proportion of patients adequately or inadequately referred to palliative care.
Results We included 612 patients in our study. During the 1-year follow-up, there were 83 deaths, 4 HTx and 1 VAD. Although SHFM showed very good discrimination (c-statistic=0.71) and adequate calibration in medium to low-risk patients, it underestimated event-free survival by 12% in high-risk patients. SHFM’s clinical utility was limited: 33% of eligible patients would have missed the opportunity for referral and only 27% of referred patients would have benefited.
Conclusion Use of SHFM could result in a high proportion of referrals while capturing the majority of patients who may benefit from palliative care. Though this may be a more encompassing and safer alternative than current referral practices, it could lead to many early referrals.
- Heart failure
- Clinical decisions
Statistics from Altmetric.com
Correction notice This article has been corrected since it was published Online First. The surname of Yu Tong Linda Lu was misspelled and affiliations were incorrect for Ana Carolina Alba.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval This research has received approval from the University Health Network (ID: 15-9368.3).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.