METHODS: We first tested ten traditional machine learning algorithms, and then the three-best performing algorithms (three types of SVM) were used in the rest of the study. To improve the performance of these algorithms, a data preprocessing with normalization was carried out. Moreover, a genetic algorithm and particle swarm optimization, coupled with stratified 10-fold cross-validation, were used twice: for optimization of classifier parameters and for parallel selection of features.
RESULTS: The presented approach enhanced the performance of all traditional machine learning algorithms used in this study. We also introduced a new optimization technique called N2Genetic optimizer (a new genetic training). Our experiments demonstrated that N2Genetic-nuSVM provided the accuracy of 93.08% and F1-score of 91.51% when predicting CAD outcomes among the patients included in a well-known Z-Alizadeh Sani dataset. These results are competitive and comparable to the best results in the field.
CONCLUSIONS: We showed that machine-learning techniques optimized by the proposed approach, can lead to highly accurate models intended for both clinical and research use.
Methods: This is a before-and-after study that took place in a tertiary Malaysian hospital. Discharge medications of patients ≥65 years old were reviewed to identify PIMs/PPOs using version 2 of the STOPP/START criteria. The prevalence and pattern of PIM/PPO before and after the intervention were compared. The intervention targeted the physicians and clinical pharmacists and it consisted of academic detailing and a newly developed smartphone application (app).
Results: The study involved 240 patients before (control group) and 240 patients after the intervention. The prevalence of PIM was 22% and 27% before and after the intervention, respectively (P = 0.213). The prevalence of PPO in the intervention group was significantly lower than that in the control group (42% Vs. 53.3%); P = 0.014. This difference remained statistically significant after controlling for other variables (P = 0.015). The intervention was effective in reducing the two most common PPOs; the omission of vitamin D supplements in patients with a history of falls (P = 0.001) and the omission of angiotensin converting enzyme inhibitor in patients with coronary artery disease (P = 0.03).
Conclusions: The smartphone app coupled with academic detailing was effective in reducing the prevalence of PPO at discharge. However, it did not significantly affect the prevalence or pattern of PIM.