Development and validation of an acute heart failure-specific mortality predictive model based on administrative data

Area of Science:

• Cardiology
• Health Services Research
• Medical Informatics

Background:

• Acute heart failure (AHF) presents a significant global healthcare challenge due to high in-hospital mortality rates.
• Accurate hospital performance comparison is hindered by the difficulty in differentiating patient severity from hospital care quality.
• Routine administrative data offers a potential resource for developing predictive models in AHF management.

Purpose of the Study:

• To develop and validate a risk adjustment model for predicting in-hospital mortality in acute heart failure (AHF) patients.
• To utilize routinely collected administrative data for predicting AHF patient outcomes.
• To create a tool that can assist in evaluating hospital performance for AHF care.

Main Methods:

• Extracted administrative data from 86 Japanese acute care hospitals, including 8620 AHF patients (April 2010-March 2011).
• Employed multivariable logistic regression to identify mortality predictors.
• Developed two models: one without and one with New York Heart Association (NYHA) functional class, validated using bootstrapping.

Main Results:

• Overall in-hospital mortality for AHF was 7.1%.
• Predictors of increased mortality included advanced age, NYHA class, and severe respiratory failure.
• Models demonstrated good predictive accuracy (c-statistics 0.76-0.80), with the model including NYHA class performing better.

Conclusions:

• Administrative data contains factors sufficient for accurate prediction of in-hospital mortality in AHF.
• The developed risk adjustment model can facilitate objective hospital evaluations for AHF care.
• This approach aids in distinguishing patient severity from hospital-specific care effects.