Premium
Therapeutic drug monitoring: using Bayesian methods to evaluate hospital practice
Author(s) -
Donagher Joni,
Barras Michael A.
Publication year - 2018
Publication title -
journal of pharmacy practice and research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.222
H-Index - 22
eISSN - 2055-2335
pISSN - 1445-937X
DOI - 10.1002/jppr.1432
Subject(s) - medicine , therapeutic drug monitoring , dosing , vancomycin , drug , therapeutic index , emergency medicine , intensive care medicine , pharmacology , biology , bacteria , genetics , staphylococcus aureus
Background Therapeutic drug monitoring (TDM) is used to improve effectiveness and reduce toxicity of various high‐risk medications. It is guided by drug specific protocols and/or TDM management software. Aim The aim of this study was to evaluate the therapeutic drug monitoring ( TDM ) of high‐risk drugs, using Bayesian simulation. Methods Patients prescribed vancomycin, enoxaparin or warfarin and who had plasma concentrations measured for the purposes of TDM were recruited to the study. A Bayesian software package was used to simulate individual patient concentration time curves incorporating patient demographics, dosing data from their course of therapy and all plasma concentrations. The proportion of therapeutic concentrations was determined using the simulated data for each patient and drug. Combined analysis of all patients allowed identification of gaps in TDM at a hospital level. The appropriateness of sampling time and dose adjustments were also investigated. Results In all, 78 patients were recruited who had 279 observed concentrations to allow simulation. Of the 100 observed vancomycin troughs, 66% were sampled incorrectly, as were 34% of 32 observed peak anti‐Xa concentrations (measurement of activity for enoxaparin); all international normalised ratios ( INR s) were taken appropriately. Approximately 50% of simulated concentrations were not therapeutic. Most patients administered vancomycin had subtherapeutic concentrations for the first 24–48 h of therapy and most patients treated with enoxaparin had toxic anti‐Xa concentrations within the first 96 h of therapy. INR s were therapeutic only 68% of the time. Conclusion Bayesian simulation allowed the identification of gaps in the TDM service and future dose recommendations, including the use of loading doses for vancomycin and dose reductions in extended courses of enoxaparin. TDM practices can now be optimised with the aim of incorporating Bayesian software into clinical practice.