Return to search

The relative predictive performance of three pharmacokinetic programs for aminoglycosides dosing

Commercial pharmacokinetic programs are used to provide clinicians with the tools to predict pharmacokinetic models, estimate pharmacokinetic parameters, and analyze serum data for efficient and consistent drug therapy and, therefore, help clinicians to optimize drug therapy. Many of the commercial programs use different methods of data entry and analysis. In this study, three commercial dosing programs, Kinetidex®, DataKinetics®, and Simkin®, were utilized to evaluate their performance on predicting gentamicin initial dosage regimens and adjusted dosage regimens. The performance of the three programs were compared by measuring the difference in the percentage of the prediction error (PE) as bias and absolute prediction error (APE) as precision using a modified method developed by Sheiner. A clinically significant difference in outcomes was determined to exist if the difference in the calculation of the dosage regimen obtained from the programs as compared to a reference calculated dosage regimen exceeded 1 0 %. A statistically significant difference in the calculation of the dosage regimens obtained from the programs was determined by ANOVA testing (p < 0.05). The results indicated that the Simkin® program had the tendency to overestimate loading doses, and the difference as compared to the reference data was clinically significant. Also the difference observed in calculating loading doses between the Simkin® program and the other two programs was statistically significant. The Kinetidex® program had bias by underestimating daily maintenance doses, and the difference as compared to the reference data was clinically significant. The difference as compared to the reference data for calculating the daily maintenance doses by the DataKinetics® and Simkin® program did not exhibit bias, but the difference was clinically significant. The difference in the performance of predicting daily maintenance doses by the programs was statistically significant in bias but not in precision. The results of computer predicted serum levels versus the measured serum levels indicated that the Kinetidex® program exhibited bias by overestimating the peak and trough concentrations as compared to the other two programs, but there is no statistically significant difference among the programs in precision. In calculating adjusted doses using the measured serum levels, the results obtained from the three programs showed no difference in bias or precision. A comparison of the results of computer-predicted serum levels showed no difference among the programs in bias and precision. In conclusion, the study showed that there was a difference in the predicting ability of the three programs in calculating initial dosage regimens but showed no statistical difference in calculating the total daily-adjusted doses when using patient serum levels.

Identiferoai:union.ndltd.org:pacific.edu/oai:scholarlycommons.pacific.edu:uop_etds-3309
Date01 January 1997
CreatorsChang, Jen-Chieh Jason
PublisherScholarly Commons
Source SetsUniversity of the Pacific
Detected LanguageEnglish
Typetext
Formatapplication/pdf
SourceUniversity of the Pacific Theses and Dissertations

Page generated in 0.0014 seconds