Multivariate analysis using a linear discriminant function for predicting the prognosis of congestive heart failure

Area of Science:

• Cardiology
• Medical Prognostics

Background:

• Congestive heart failure (CHF) can be intractable, responding poorly to standard treatments.
• Identifying factors influencing CHF prognosis and treatment response is crucial for patient management.

Purpose of the Study:

• To identify clinical and laboratory factors that differentiate between patients with good treatment response (curative group) and those with refractory CHF (refractory group).
• To develop a predictive model for assessing prognosis and treatment response in hospitalized CHF patients.

Main Methods:

• A cohort of 114 hospitalized patients with CHF was divided into a curative group (77 patients) and a refractory group (37 patients).
• Eight variables were analyzed: heart rate, hemoglobin, serum potassium, serum total protein, albumin/globulin ratio, blood urea nitrogen, hepatomegaly grade, and previous CHF episodes.
• A linear discriminant function was derived using these eight variables to classify patients.

Main Results:

• Eight specific variables significantly contributed to differentiating between the curative and refractory groups.
• The derived linear discriminant function (Y = -9.64 - 0.0686X1 + 0.345X2 + 1.351X3 + 1.513X4 + 1.988X5 - 0.0876X6 - 0.792X7 - 0.737X8) showed excellent discriminatory power.
• A Y value > 0 indicated the curative group, while a Y value < 0 indicated the refractory group.

Conclusions:

• The developed discriminant equation effectively predicts prognosis and treatment response in congestive heart failure patients.
• This model offers a valuable tool for clinicians managing CHF, aiding in treatment planning and patient stratification.