Personalized drug concentration predictions with machine learning: an exploratory study


  • Danish Shakeel Department of Computer Sciences and Engineering, University Institute of Engineering, Chandigarh University, Mohali, Punjab, India
  • Shakeel Ahmad Mir Department of Clinical Pharmacology, Sher-I-Kashmir Institute of Medical Sciences, Srinagar, Kashmir, India



Drug concentration, Dose, Machine learning, Therapeutic drug monitoring


Background: The dose individualization by therapeutic drug monitoring (TDM) can be improved if population-based reference ranges are available, as there is large inter- and intrapatient variability. If these ranges are not available, dose individualization may not be optimal. Machine learning can help achieve accurate drug dose settings and predict the resultant levels.

Methods: Two random forest models, a multi-class classifier to predict dose and a regression model to predict blood drug level were trained on 320 patients’ data, consisting of their age, sex, dose and blood drug level. The classifier consisted of 1000 estimators (decision trees) and the regression model consisted of 1300 estimators. The model was evaluated on randomly split test set having 10% of the total dataset size. The regression model was compared against k-Nearest neighbor and linear regression models. The classifier was evaluated using accuracy, precision, and F1 Score; the regression model was evaluated using R2, Root mean squared error, and mean absolute error.

Results: The classifier had an out-of-sample accuracy of 68.75%, average precision of 0.7567, and an average F1 score of 0.6907. The regression model had an out-of-sample R2 value of 0.2183, root mean squared value of 3.7359, and a mean absolute error of 2.5156. These values signify an average classification performance, and a below-average regression performance due to small dataset.

Conclusions: It is possible for machine learning algorithms to be used in therapeutic drug monitoring. With a well-structured, rich, and large dataset, a very accurate model can be built.


Walson PD. Therapeutic drug monitoring in special populations. Clin Chem. 1998;44(2):415-9.

Kang JS, Lee MH. Overview of therapeutic drug monitoring. Korean J Int Medi. 2009;24(1):1.

Gross AS. Best practice in therapeutic drug monitoring. Br J Clini Pharmacol. 2001;52(S1):5-9.

Cooney L, Loke YK, Golder S, Kirkham J, Jorgensen A, Sinha I, Hawcutt D. Overview of systematic reviews of therapeutic ranges: methodologies and recommendations for practice. BMC Medi Res Methodol. 2017;17(1):84.

Lucas C, Donovan P. Medications:'Just a repeat': When drug monitoring is indicated. Austr Family Physic. 2013;42(1/2):18.

Hiemke C, Baumann P, Bergemann N, Conca A, Dietmaier O, Egberts K, et al. AGNP consensus guidelines for therapeutic drug monitoring in psychiatry: update 2011. Pharmacopsychiatry. 2011;21(06):195-235.

Umamaheswaran G, Kumar DK, Adithan C. Distribution of genetic polymorphisms of genes encoding drug metabolizing enzymes & drug transporters-a review with Indian perspective. The Ind J Medi Res. 2014;139(1):27.

Varshney E, Saha N, Tandon M, Shrivastava V, Ali S. Prevalence of poor and rapid metabolizers of drugs metabolized by CYP2B6 in North Indian population residing in Indian national capital territory. Springerplus. 2012;1(1):1-7.

Kousar S, Wafai ZA, Wani MA, Jan TR, Andrabi KI. Clinical relevance of genetic polymorphism in CYP2C9 gene to pharmacodynamics and pharmacokinetics of phenytoin in epileptic patients: validatory pharmacogenomic approach to pharmacovigilance. Int J Clini Pharmacol Therap. 2015;53(7):504-16.

Kavakiotis I, Tsave O, Salifoglou A, Maglaveras N, Vlahavas I, Chouvarda I. Machine learning and data mining methods in diabetes research. Computat Structural Biotechnol J. 2017;15:104-16.

Jones G, Barker A. Reference intervals. Clini Biochem Rev. 2008;29(Suppl 1):S93.

W. You. An Algorithmic Approach to Personalized Drug Concentration Predictions. Doctoral dissertation, 2014. Available at: https://pdfs.semanticscholar. org/c003/0746b9f344baa795cce838bc5a238e7cf9b8.pdf. Accessed on 3 January 2020.

Imai S, Takekuma Y, Miyai T, Sugawara M. A New Algorithm Optimized for Initial Dose Settings of Vancomycin Using Machine Learning. Biolog Pharmac Bull. 2020;43(1):188-93.

Hu YH, Tai CT, Tsai CF, Huang MW. Improvement of Adequate Digoxin Dosage: An Application of Machine Learning Approach. J Healthcare Engineer. 2018;2018.

Yao SH, Tsai HT, Lin WL, Chen YC, Chou C, Lin HW. Predicting the serum digoxin concentrations of infants in the neonatal intensive care unit through an artificial neural network. BMC Pediatrics. 2019;19(1):1-1.

Goicoechea M, Vidal A, Capparelli E, Rigby A, Kemper C, Diamond C, et al. A computer-based system to aid in the interpretation of plasma concentrations of antiretrovirals for therapeutic drug monitoring. Antiviral therapy. 2007;12(1):55.

Zhang Z, Ho KM, Hong Y. Machine learning for the prediction of volume responsiveness in patients with oliguric acute kidney injury in critical care. Crit Care. 2019;23(1):112.




How to Cite

Shakeel, D., & Mir, S. A. (2020). Personalized drug concentration predictions with machine learning: an exploratory study. International Journal of Basic & Clinical Pharmacology, 9(6), 980–984.



Original Research Articles