Abstract
INTRODUCTION
Heart failure (HF) is a clinical syndrome which is caused by either structural and/or functional cardiac abnormality.[1] There are 3 types of heart failure classified according to left ventricular ejection fraction (LVEF): HF with reduced ejection fraction (HFrEF); HF with mid-range EF (HFmREF) and HF with preserved EF (HFpEF).[2] HFpEF is widely regarded to be of the greatest challenge due to increasing prevalence and limited disease modifying medication therapy.
With an expanding diversity of electronic health record (EHR) resources that include clinical data, imaging, banked bioresources and plasma samples, application of artificial intelligence-based machine learning algorithms to electronic health records offers opportunities to generate timely real-world evidence to enhance patient care and facilitate research.
PATIENTS AND METHODS
Retrospective data analysis using clinical data set, routinely stored in health informatic centre (HIC) Tayside from 1994 to 2021. This data was linked to the community-dispensed prescriptions, echocardiography data, hospital admissions (Scottish Morbidity Record), and mortality data (General Registry Office). These filtered data was linked with plasma sample stored in the NHS Tayside SHARE register and Digital Imaging and Communications in Medicine (DICOM) images.
RESULTS
Among 3680 patient records with linkage to echocardiographic DICOMs and blood samples, we identified 236 patients with HFpEF, 156 with HFrEF, and 185 controls. Compared to HFrEF, patients with HFpEF were older, more often women and had less coronary artery disease. Compared to controls, patients with HFpEF and HFrEF had greater LV mass, higher mitral E/e’ ratio, higher pulmonary artery systolic pressure, more impaired LV strain and more RV dysfunction. While most patients with HFrEF were on guideline-directed medical therapy, relatively few mineralocorticoid receptor antagonists were described. NT-proBNP levels were elevated in HFpEF, and more in HFrEF, compared to controls. During a median follow up time of 1089 days, 39% of HFpEF and 63% of HFrEF experienced hospitalization or death, compared to only 5% of controls.
CONCLUSIONS
We demonstrate the feasibility of AI-automated detection and classification of HF from routine surveillance of electronic health data, yielding cohorts with clinical and echocardiographic characteristics similar to epidemiologic studies, with raised natriuretic peptides and poor outcomes expected in these patient populations. Such AI-assisted electronic surveillance may be useful for automated monitoring of clinical records for early disease detection, treatment quality improvement initiatives, and case finding for clinical trials.
Heart failure (HF) is a clinical syndrome which is caused by either structural and/or functional cardiac abnormality.[1] There are 3 types of heart failure classified according to left ventricular ejection fraction (LVEF): HF with reduced ejection fraction (HFrEF); HF with mid-range EF (HFmREF) and HF with preserved EF (HFpEF).[2] HFpEF is widely regarded to be of the greatest challenge due to increasing prevalence and limited disease modifying medication therapy.
With an expanding diversity of electronic health record (EHR) resources that include clinical data, imaging, banked bioresources and plasma samples, application of artificial intelligence-based machine learning algorithms to electronic health records offers opportunities to generate timely real-world evidence to enhance patient care and facilitate research.
PATIENTS AND METHODS
Retrospective data analysis using clinical data set, routinely stored in health informatic centre (HIC) Tayside from 1994 to 2021. This data was linked to the community-dispensed prescriptions, echocardiography data, hospital admissions (Scottish Morbidity Record), and mortality data (General Registry Office). These filtered data was linked with plasma sample stored in the NHS Tayside SHARE register and Digital Imaging and Communications in Medicine (DICOM) images.
RESULTS
Among 3680 patient records with linkage to echocardiographic DICOMs and blood samples, we identified 236 patients with HFpEF, 156 with HFrEF, and 185 controls. Compared to HFrEF, patients with HFpEF were older, more often women and had less coronary artery disease. Compared to controls, patients with HFpEF and HFrEF had greater LV mass, higher mitral E/e’ ratio, higher pulmonary artery systolic pressure, more impaired LV strain and more RV dysfunction. While most patients with HFrEF were on guideline-directed medical therapy, relatively few mineralocorticoid receptor antagonists were described. NT-proBNP levels were elevated in HFpEF, and more in HFrEF, compared to controls. During a median follow up time of 1089 days, 39% of HFpEF and 63% of HFrEF experienced hospitalization or death, compared to only 5% of controls.
CONCLUSIONS
We demonstrate the feasibility of AI-automated detection and classification of HF from routine surveillance of electronic health data, yielding cohorts with clinical and echocardiographic characteristics similar to epidemiologic studies, with raised natriuretic peptides and poor outcomes expected in these patient populations. Such AI-assisted electronic surveillance may be useful for automated monitoring of clinical records for early disease detection, treatment quality improvement initiatives, and case finding for clinical trials.
Original language | English |
---|---|
Pages (from-to) | NP5-NP10 |
Number of pages | 6 |
Journal | Scottish Medical Journal |
Volume | 68 |
Issue number | 1 |
DOIs | |
Publication status | Published - 14 Feb 2023 |