These tools will no longer be maintained as of December 31, 2024. Archived website can be found here. PubMed4Hh GitHub repository can be found here. Contact NLM Customer Service if you have questions.


Pubmed for Handhelds

PUBMED FOR HANDHELDS

Search MEDLINE/PubMed

  • Title: Correlation coefficient versus prediction error in assessing the accuracy of digoxin dosing methods.
    Author: Lalonde RL, Pao D.
    Journal: Clin Pharm; 1984; 3(2):178-83. PubMed ID: 6723224.
    Abstract:
    The accuracy of 18 different methods to predict serum digoxin concentrations ( SDCs ) was assessed in a retrospective study of 50 hospital patients. The patients had documented steady-state SDCs while receiving digoxin tablets. Pertinent data, such as age, total and lean body weight, body surface area, estimated creatinine clearance (normalized and nonnormalized ), and digoxin dose, were collected for each patient and used to make the predictions. Correlation coefficients (r) of predicted versus measured SDCs were determined, and the mean error (ME) and root mean squared error ( RMSE ) were determined for each method to reflect bias and precision, respectively. The correlation coefficients varied from 0.389 to 0.536 and were all significantly different from zero. The different methods varied in their bias and precision; however, the methods of Dobbs and Koup (as modified by Koda - Kimble ) appeared to produce the best balance of minimum bias and greatest precision. The use of normalized or nonnormalized creatinine clearance did not greatly affect the predictive performance of the various methods. Although the methods of Dobbs and Koda - Kimble have good predictive capabilities, they should not replace SDC monitoring. Future studies evaluating predictive performance should not rely solely on r values but should include measurements of ME and RMSE .
    [Abstract] [Full Text] [Related] [New Search]