Heart failure (HF) remains one of the leading causes of hospitalization, healthcare expenditure, and mortality in the United States. In this comprehensive nationwide analysis published in European Heart Journal Open, investigators examined whether outcomes differ between urban teaching hospitals and urban non-teaching hospitals for patients admitted with heart failure. Using data from the Nationwide Readmissions Database (NRD) between 2016 and 2022, the study evaluated more than 7.5 million weighted hospitalizations, making it one of the largest analyses of its kind .
The study focused exclusively on urban hospitals to reduce confounding from rural healthcare disparities. Of the total hospitalizations identified, 76.3% occurred at urban teaching hospitals and 23.7% at urban non-teaching hospitals. Patients were adults aged 18 years and older with a primary discharge diagnosis of heart failure, including both heart failure with reduced ejection fraction (HFrEF) and heart failure with preserved ejection fraction (HFpEF).
To ensure a fair comparison between hospital types, researchers used 1:1 propensity score matching. This statistical approach controlled for demographic characteristics, payer status, income quartile, hospital bed size, comorbidities, heart failure subtype, and markers of severity. By matching over one million hospitalizations in each cohort, the authors minimized baseline imbalances and improved the robustness of outcome comparisons.
The primary outcome was all-cause inpatient mortality. Secondary outcomes included hospital length of stay (LOS), total inflation-adjusted hospitalization cost, inpatient complications (such as cardiogenic shock, cardiac arrest, mechanical ventilation, intra-aortic balloon pump use, extracorporeal membrane oxygenation [ECMO], cardiopulmonary resuscitation [CPR], and blood transfusion), palliative care consultation rates, discharge disposition, and 30- and 90-day readmissions.
The results revealed that heart failure hospitalizations at urban teaching hospitals were associated with a 19% higher odds of inpatient mortality compared with non-teaching hospitals (3.31% vs. 2.76%). Teaching hospitals also demonstrated significantly higher rates of major complications. Cardiogenic shock was more than twice as common, while advanced mechanical circulatory support—including ECMO and intra-aortic balloon pump—was used substantially more frequently. Mechanical ventilation and CPR rates were also elevated.
In addition to higher mortality and complication rates, urban teaching hospitals had longer average hospital stays (4.15 vs. 3.56 days) and higher total hospitalization costs ($11,919 vs. $10,173). Palliative care consultations were more common in teaching hospitals, suggesting either more advanced disease, greater infrastructure for end-of-life care discussions, or both.
Interestingly, despite worse in-hospital metrics, readmission outcomes were similar. Thirty-day and 90-day all-cause readmission rates were statistically higher in teaching hospitals but clinically negligible in difference. Rates of discharge to home were also comparable between hospital types.
Why would teaching hospitals—often considered centers of excellence with advanced technology and subspecialty expertise—show higher inpatient mortality? The authors suggest referral bias and case-mix severity as primary explanations. Teaching hospitals frequently serve as tertiary referral centers and safety-net institutions, treating more critically ill patients and those with higher comorbidity burdens. Indeed, complication rates such as cardiogenic shock and cardiac arrest were markedly higher in teaching institutions, indicating greater illness severity.
Advanced therapies such as ECMO and intra-aortic balloon pump support are predominantly available at teaching hospitals. While these interventions can be lifesaving, they are typically used in the sickest patients and carry substantial risks. Therefore, higher mortality may reflect the complexity of cases rather than inferior quality of care.
The study also acknowledges limitations inherent to administrative database research. The NRD lacks granular clinical data such as admission vital signs, laboratory values, medication use, and heart failure duration. Residual confounding due to unmeasured severity variables may remain despite robust matching. Additionally, socioeconomic and social determinant factors—particularly relevant for safety-net teaching hospitals—may influence outcomes.
Importantly, the authors emphasize that their findings should not be interpreted as evidence of lower quality care in teaching hospitals. Rather, the data are hypothesis-generating and likely reflect differences in patient populations. Teaching hospitals manage more advanced heart failure cases, higher-risk referrals, and patients requiring specialized interventions.
From a healthcare policy perspective, this study highlights the importance of risk adjustment when comparing hospital performance metrics. Mortality rates alone may not capture the complexity of patients treated at tertiary academic centers. Policymakers, clinicians, and healthcare administrators should consider case-mix differences when evaluating hospital quality and reimbursement models.
In summary, this large U.S. nationwide analysis demonstrates that urban teaching hospitals have higher inpatient mortality, greater complication rates, longer hospital stays, and higher costs for heart failure hospitalizations compared with urban non-teaching hospitals. However, readmission rates and discharge outcomes remain similar. These findings likely reflect referral patterns and severity differences rather than intrinsic quality disparities. Further research using more granular clinical datasets is needed to clarify these associations and optimize heart failure care delivery across hospital systems.
