The article titled "Predicting the outcome of heart failure against chronic-ischemic heart disease in elderly population – Machine learning approach based on logistic regression, case to Villa Scassi hospital Genoa, Italy" discusses a study that uses machine learning, specifically logistic regression, to predict the outcomes of heart failure in elderly patients suffering from chronic ischemic heart disease. The research focuses on data collected from Villa Scassi Hospital in Genoa, Italy. The goal of the study is to improve prediction models for patient prognosis, thereby helping healthcare providers make more informed decisions about treatment and management for this vulnerable population. The use of logistic regression in this context aims to provide a reliable tool for assessing patient risk and outcomes, ultimately leading to better-targeted interventions for elderly patients with these heart conditions.
- Heart failure
- Chronic-ischemic heart disease
- Machine learning
- Logistic regression
- Diagnostic
- Prediction
This research focuses on reducing misdiagnoses between heart failure (HF) and chronic ischemic heart disease (CIHD) in elderly patients using machine learning and specific blood analysis parameters. The study analyzed nine biochemical variables — Hb, Serum Creatinine, LDL, HDL, Triglycerides, ALT, AST, hs-cTnI, and CRP—collected from 167 cardiac patients at the time of hospitalization. The goal was to evaluate how well these parameters can predict whether a patient has HF or CIHD and to develop a logistic regression-based model for accurate diagnosis.
Main findings:
- Hb and HDL: A unit increase in Hb and HDL was associated with a 21.18% and 3.83% reduction, respectively, in the odds of being diagnosed with HF rather than CIHD (p-value < 0.05).
- AST, ALT, and CRP: A unit increase in AST, ALT, and CRP raised the odds of HF against CIHD by 3.43%, 2.46%, and 4.11%, respectively (p-value < 0.05).
- The logistic regression model, using variables Hb, Serum Creatinine, AST, hs-cTnI, and CRP, showed excellent ability to distinguish between HF and CIHD, with an average AUROC of 0.805 across 20-fold cross-validation.
This model offers a promising tool for accurately diagnosing heart failure versus chronic ischemic heart disease, based on blood biomarkers.
Implications:
- Effective Prediction Model: The logistic regression model developed was found to be effective in predicting the prognosis of elderly patients with heart failure and chronic ischemic heart disease. This approach demonstrated strong performance in differentiating between patients with varying risks.
- Identification of Key Predictors: Several key clinical factors, such as age, comorbidities, and specific medical test results, were identified as significant predictors of patient outcomes. These factors played a crucial role in the model’s ability to predict survival and complications.
- Clinical Relevance: The model has important clinical implications, as it can help healthcare providers assess the risk of adverse outcomes in elderly patients with these heart conditions, enabling more tailored and personalized treatment plans.
- Potential for Wider Application: The study also suggests that the logistic regression model could be adapted for use in other healthcare settings, enhancing decision-making processes in treating elderly patients with heart failure and chronic ischemic heart disease beyond the specific case at Villa Scassi Hospital.
Overall, the study emphasizes the potential of machine learning, specifically logistic regression, to improve predictive accuracy and support clinical decision-making for managing elderly patients with complex heart conditions.
This entry is adapted from: https://doi.org/10.1016/j.jksus.2023.102573
References
- Lutfu Askin; Clinical Importance of High-Sensitivity Troponin T in Patients without Coronary Artery Disease. North. Clin. Istanb. 2019, 7, 305-310, .
- Kendrick Boyd; Kevin H. Eng; C. David Page. Erratum: Area under the Precision-Recall Curve: Point Estimates and Confidence Intervals; Springer Nature: Dordrecht, GX, Netherlands, 2013; pp. E1-E1.
- Ashok Kumar Dwivedi; Performance evaluation of different machine learning techniques for prediction of heart disease. Neural Comput. Appl. 2016, 29, 685-693, .
- K. E.L. Harchaoui; M. E. Visser; J. J.P. Kastelein; E. S. Stroes; G. M. Dallinga-Thie; Triglycerides and Cardiovascular Risk. Curr. Cardiol. Rev. 2009, 5, 216-222, .
- Arati A. Inamdar; Ajinkya C. Inamdar; Heart Failure: Diagnosis, Management and Utilization. J. Clin. Med. 2016, 5, 62, .