Key result
The TARGET-HF prediction model, utilizing routine primary care data, achieved a C-statistic of 0.853 for detecting incident heart failure, significantly outperforming an existing hospital-based model.
Why the study?
Early stages of HF may be difficult to identify in general practice, creating a need for an integrated clinical support tool using routine health care data to identify high-risk patients.
Does the TARGET-HF prediction model improve the detection of incident heart failure among symptomatic patients in general practice compared to age alone or existing hospital-based models?
Cohort (n=31,905)
Yes
Does the TARGET-HF prediction model improve the detection of incident heart failure among symptomatic patients in general practice compared to age alone or existing hospital-based models?
Effect estimate: C-statistic 0.853 (95% CI 0.834-0.872)
Absolute Event Rate: 0.853% vs 0.824%
p-value: p=<0.001
A prediction model (TARGET-HF) using routine general practice consultation codes can accurately identify symptomatic patients at high risk for incident heart failure, outperforming age alone and existing hospital-based models.
May aid HF risk stratification in primary care; hypothesis-generating pending prospective validation.
Background Timely diagnosis of heart failure (HF) is essential to optimize treatment opportunities that improve symptoms, quality of life, and survival. While most patients consult their general practitioner (GP) prior to HF, early stages of HF may be difficult to identify. An integrated clinical support tool may aid in identifying patients at high risk of HF. We therefore constructed a prediction model using routine health care data. Methods Our study involved a dynamic cohort of patients (≥35 years) who consulted their GP with either dyspnea and/or peripheral edema within the Amsterdam metropolitan area in 2011-2020. The outcome of interest was incident HF, verified by an expert panel. We developed a regularized multivariable proportional hazards model (TARGET-HF). The model was evaluated with bootstrapping on an isolated validation set and compared to an existing model developed with hospital insurance data as well as patient age as a sole predictor. Results Data from 31,905 patients were included (40% male, median age 60) of whom 1,301 (4.1%) were diagnosed with HF over 124,676 person-years of follow-up. Data were allocated to a development (n=25,524) and validation (n=6,381) set. TARGET-HF attained a C-statistic of 0.853 (95%-CI:0.834-0.872) on the validation set, which proved to provide a better discrimination than C=0.822 for age alone (95% CI:0.801-0.842, p<0.001) and C=0.824 for the hospital-based model (95% CI:0.802-0.843, p<0.001). Conclusion The TARGET-HF model illustrates that routine consultation codes can be used to build a performant model to identify patients at risk for HF at time of GP consultation.
No takes yet. Share an insight, caveat, or question.
Clercq et al. (2022) conducted a cohort in Incident heart failure (n=31,905). TARGET-HF prediction model vs. Age alone and Goyal et al. hospital-based model was evaluated on Model discrimination (C-statistic) for incident heart failure on the validation set (C-statistic 0.853, 95% CI 0.834-0.872, p=<0.001). The TARGET-HF prediction model, utilizing routine primary care data, achieved a C-statistic of 0.853 for detecting incident heart failure, significantly outperforming an existing hospital-based model.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: