Search

Drug-related problems, particularly those that result from sub- or overtherapeutic doses of high-alert medications, have become a growing concern in clinical medicine. In this paper, we use a model-tree-based regression technique (namely, M5) and support vector machine (SVM) for regression to develop learning-based systems for predicting the adequacy of a vancomycin regimen. We empirically evaluate each system's accuracy in predicting patients' peak and trough concentrations in different clinical scenarios characterized by renal functions and regimen types. Our data consist of 1099 clinical cases that were collected from a major tertiary medical center in southern Taiwan. We also examine the use of Bagging for enhancing the prediction power of the respective systems and include in our evaluation a salient one-compartment model for performance benchmark purposes. Overall, our evaluation results suggest that both M5 and SVM are significantly more accurate than the benchmark one-compartment model in predicting patients' peak and trough concentrations across all investigated clinical scenarios. M5 appears to benefit considerably from Bagging, which has a positive but seemingly smaller effect on SVM. Taken together, our findings indicate supervised learning techniques that are capable of effectively supporting clinicians' use of vancomycin or similar high-alert drugs in their patient care and management. Index Terms--Bagging, decision support in health care, management of clinical use of vancomycin, model-tree-based regression, pharmacokinetic data mining, supervised learning, support vector machine (SVM).

Purpose To analyze and characterize data regarding the prevalence and types of outpatient drug-related problems (DRPs) found by clinical pharmacists after implementation of the Virtual Medicine Record in Cloud System (VMRCS). Methods A cross-sectional study regarding outpatient pharmaceutical care was conducted at a medical center in Taiwan. Patients aged >20 years old with multiple chronic diseases and polypharmacy were enrolled. In Stage I (1 October-31 December 2014), patients received pharmaceutical care according to prescription data accessed online in the VMRCS. In Stage II (1 June-31 August 2015), the VMRCS were pre-download and arranged to the institute's required format, facilitated DRP detection. Clinical pharmacists then reviewed and evaluated the prescription data through pre-downloaded VMRCS. Overall, 1539 and 1600 prescriptions were evaluated in these two stages, respectively. DRPs were recorded using the Pharmaceutical Care Network Europe (PCNE)-DRP. Results DRPs were found for 50.2% of patients in Stage I and 55.2% in Stage II ( p < 0.05) and were most frequently encountered for 'Drugs for the cardiovascular system' and caused by 'Inappropriate duplication of therapeutic group or active ingredient.' In terms of problems, incidence of 'Unnecessary drug treatment' was highest. Duplicate medications were most frequently seen for 'Drugs for acid-related disorders.' The efficiency to identify DRPs was at least 2.4 times higher with pre-downloaded prescription data than with real-time online queries. Conclusions With VMRCS, DRPs were more easily identified whether patients received medical care in the same hospital or not. DRPs could be efficiently prevented through the use of pre-downloaded patient prescription data. Copyright © 2016 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Background: An adaptive-network-based fuzzy inference system (ANFIS) was compared with an artificial neural network (ANN) in terms of accuracy in predicting the combined effects of temperature (10.5 to 24.5°C), pH level (5.5 to 7.5), sodium chloride level (0.25% to 6.25%) and sodium nitrite level (0 to 200 ppm) on the growth rate of Leuconostoc mesenteroides under aerobic and anaerobic conditions. Methods: The ANFIS and ANN models were compared in terms of six statistical indices calculated by comparing their prediction results with actual data: mean absolute percentage error (MAPE), root mean square error (RMSE), standard error of prediction percentage (SEP), bias factor (Bf), accuracy factor (Af), and absolute fraction of variance (R2). Graphical plots were also used for model comparison. Conclusions: The learning-based systems obtained encouraging prediction results. Sensitivity analyses of the four environmental factors showed that temperature and, to a lesser extent, NaCl had the most influence on accuracy in predicting the growth rate of Leuconostoc mesenteroides under aerobic and anaerobic conditions. The observed effectiveness of ANFIS for modeling microbial kinetic parameters confirms its potential use as a supplemental tool in predictive mycology. Comparisons between growth rates predicted by ANFIS and actual experimental data also confirmed the high accuracy of the Gaussian membership function in ANFIS. Comparisons of the six statistical indices under both aerobic and anaerobic conditions also showed that the ANFIS model was better than all ANN models in predicting the four kinetic parameters. Therefore, the ANFIS model is a valuable tool for quickly predicting the growth rate of Leuconostoc mesenteroides under aerobic and anaerobic conditions. [ABSTRACT FROM AUTHOR]

Objective: Computer-prescriber order entry (CPOE) systems that lack clinical decision components actually increase errors and cause harm rather than the opposite. Recent studies have also demonstrated that dosing errors, typing errors, or miscommunication with other systems are the most common CPOE errors. Our objective was to develop an antibiotic dosing calculator and implement it in the CPOE system while integrating the role of the clinical pharmacist in the CPOE in order to minimize dosing errors in the prescription of antibiotics., Methods: A database was prepared using dosage information for 13 renal function-related antibiotics. The dosages in the database ranged from the standard to the maximum dosage based on various creatinine clearance (CL(cr) levels. The antibiotic dosage monitoring system was developed to screen the entire inpatient database for inappropriate antibiotic dosage regimens and record the results as an Excel document., Main Outcome Measure: We tracked the frequency of calculator utilization by physicians, the acceptance rate of recommendations from the calculator and pharmacists, the inappropriate antibiotic dosage regimen prescriptions, and the antibiotic-related renal function deterioration. The relative risk (RR) with 95% confidence intervals (CI) was used to calculate the risk of inappropriate antibiotic dosage prescription, the deterioration in renal function when antibiotics were used., Results: From 2005 to 2008, 38,647 antibiotic prescriptions were recorded in the CPOE system. The instances of inappropriate antibiotic dosage prescriptions were decreased by ~ 80% after the calculator was implemented (RR, 0.18 – 0.23; p < 0.001), and the incidence rates of renal function deterioration were lowered from 12.39% to 9.47%. The frequency of antibiotic calculator utilization by physicians (from 239 times/ year in 2005 to 3,480 times/year in 2008) and the acceptance rate of the calculator’s dosage recommendations (from 68.2% in 2005 to 94.7% in 2008) both increased during the study period. The average acceptance rates of pharmacist recommendations by physicians were 97.65%., Conclusions: Integration of the CPOE decision-supporting system and the clinical pharmacist monitoring practice can help physicians provide appropriate antibiotic dosage regimens and decrease the incidence of dosing errors that could be decreased concerned patients with impaired renal function.