How to Convolute Unknown Factors to Know the Outcomes of the Patients With Heart Failure

Background: We could not identify the elapsed time until hospitalization in certain patients with heart failure (HF). Therefore, we attempted to predict the elapsed time using mathematical method. To mathematically predict, we sought to solve the equation Y = f(x₁ … x_p), where Y represents the clinical outcomes and x₁ … x_p represent clinical factors affecting for HF. Methods: We followed 151 patients (mean age: 68.6 ± 14.6 years) with acute decompensated HF who were consecutively hospitalized and discharged, and collected clinical factors using this population. The mathematical analysis was performed through a probabilistic modeling of the relational data by assuming a Poisson process for re-hospitalization and by linearly approximating the relationship between the clinical factors and the mean elapsed time to re-hospitalization. We also performed data mining to know the factors to discriminate the patients who were re-hospitalized 1) within 6 months and 2) after 2 years or not re-hospitalized. Results: 151 patients were died or readmitted to our hospital at a median time of 296 days after discharge. We collected 402 clinical factors, and excluded 150 factors having small effects by the regularization method. Finally, we identified 252 factors affecting for CHF and estimated the result of attribute coefficients. Conclusions: This study demonstrated that clinical medicine and practice can use a mathematical formula to predict clinical outcomes or events using current data.