Synthetic data as a tool to quantify heart failure hospitalization burden
K Haichin, G Zhang, N Giannetti, M Liszkowski, A SharmaAbstract
Background
Heart failure (HF) remains a leading cause of morbidity and mortality in Canada, despite significant advances in HF therapies. HF-related hospitalizations impose a substantial economic burden on the Canadian healthcare system and are associated with worse clinical outcomes for affected patients. Further characterization of trends in HF hospitalizations is needed to inform quality improvement initiatives and optimize healthcare delivery.
Purpose
To assess and quantify the HF hospitalization burden at our University Healthcare Centre with the use of a synthetic dataset (MDClone Cohort).
Methods
A synthetic dataset was generated using MDClone to examine temporal trends and clinical outcomes in HF hospitalizations between 2008 and 2025 at our University Healthcare Centre, a large tertiary care centre in Canada. The study cohort comprised patients with a principal diagnosis of HF (ICD-10 code 150) during index hospitalization. Our analysis explored the HF rehospitalization patterns (within 30 days, 90 days, 365 days of first HF hospitalization). The evaluation of clinical outcomes incorporated length of stay (LOS), mortality rates, and ICU admission rates as primary metrics, with analysis conducted at two levels: individual hospitalization events and longitudinal temporal trends.
Results
A total of 21,459 patients were hospitalized with a primary diagnosis of HF between 2008 to 2025. The mean age was 75 years. Among patients discharged after their index hospitalizations, 9% (n=1,705) were rehospitalized within 30 days of discharge, and 42% (n=7,843) required at least one HF-related rehospitalization. The mean LOS was 18.5 days and remained stable over time. In-hospital mortality exceeded 12%, and 38% of patients required ICU admission. Despite modest year-to-year variability, the annual number of HF hospitalizations remained persistently high across the study period.
Conclusion
Synthetic data derived from MDClone effectively assessed and quantified trends in HF hospitalization burden at a tertiary care centre in Montreal, Canada. HF hospitalizations remain a major driver of healthcare burden despite advances in therapy. Effective and cost-efficient strategies are necessary to reduce hospitalization, morbidity and mortality in this population, and using synthetic datasets may serve as a valuable tool for health system surveillance and quality improvement in HF care.Heart failure hospitalization per yearFor image description, please refer to the figure legend and surrounding text.Hospital length of stay per yearFor image description, please refer to the figure legend and surrounding text.