The PHARMACOM‐EPI Framework for Integrating Pharmacometric Modelling Into Pharmacoepidemiological Research Using Real‐World Data: Application to Assess Death Associated With Valproate

Abstract

In pharmacoepidemiology, it is usually expected that the observed association should be directly or indirectly related to the pharmacological effects of the drug/s under investigation. Pharmacological effects are, in turn, strongly connected to the pharmacokinetic and pharmacodynamic properties of a drug, which can be characterized and investigated using pharmacometric models. Recently, the use of pharmacometrics has been proposed to provide pharmacological substantiation of pharmacoepidemiological findings derived from real-world data. However, validated frameworks suggesting how to combine these two disciplines for the aforementioned purpose are missing. Therefore, we propose PHARMACOM-EPI, a framework that provides a structured approach on how to identify, characterize, and apply pharmacometric models with practical details on how to choose software, format dataset, handle missing covariates/dosing data, how to perform the external evaluation of pharmacometric models in realworld data, and how to provide pharmacological substantiation of pharmacoepidemiological findings. PHARMACOM-EPI was tested in a proof-of-concept study to pharmacologically substantiate death associated with valproate use in the Danish population aged ≥ 65 years. Pharmacological substantiation of death during a follow-up period of 1 year showed that in all individuals who died (n = 169) individual predictions were within the subtherapeutic range compared with 52.8% of those who did not die (n = 1,084). Of individuals who died, 66.3% (n = 112) had a cause of death possibly related to valproate and 33.7% (n = 57) with well-defined cause of death unlikely related to valproate. This proof-of-concept study showed that PHARMACOM-EPI was able to provide pharmacological substantiation for death associated with valproate use in the study population.

Artificial Neural Network vs. Pharmacometric Model for Population Prediction of Plasma Concentration in Real‐World Data: A Case Study on Valproic Acid

“…We showed that in older patients with epilepsy the LSTM neural network had better predictive performance compared with a traditional pharmacometric model of valproate previously determined to have the highest accuracy in our external evaluation study. 1 Such better performance is expected and shown by others in clinical trials. 3,9 We believe that our results are mostly driven by four reasons.…”

“…15 Second, as machine learning can incorporate high-dimensional data and describe nonlinear relationships between variables and response, 16 artificial neural networks can better explain the variability of plasma concentration between patients. Third, populations on which various pharmacometric models have been developed may differ from the real-world population, as also outlined in our recent study, 1 and thus it is expected that when applied to the realworld population such models may not perform well. 17 This was shown in our previous study, where external evaluation results were poor, 1 as well as by other external evaluation studies.…”

“…Hiie Soeorg 1,2, *, Eva Sverrisdóttir 2 , Morten Andersen 1 , Trine Meldgaard Lund 2 and Maurizio Sessa 1 We compared the predictive performance of an artificial neural network to traditional pharmacometric modeling for population prediction of plasma concentrations of valproate in real-world data. We included individuals aged 65 years or older with epilepsy who redeemed their first prescription of valproate after the diagnosis of epilepsy and had at least one valproate plasma concentration measured.…”

“…Real-world data is emerging as an important source of data for pharmacometric modeling, and a new framework PHARMACOM-EPI was recently proposed to guide researchers in all the necessary steps from pharmacometric model selection to its application in real-world data. 1 A recent study using PHARMACOM-EPI for predicting plasma concentrations of valproate in older individuals with epilepsy highlighted good individual prediction accuracy but poor population prediction accuracy. 1 As PHARMACOM-EPI is mostly intended for real-world data and in such data sources it is not always possible to obtain plasma concentrations of drugs in pharmacologically treated individuals (and consequently obtain individual predictions) we found it compelling to investigate new approaches for improving population prediction when using PHARMACOM-EPI.…”

“…PHARMACOM-EPI 1 was used to retrieve available pharmacometric models of valproate from the scientific literature using a schematic approach, which has been detailed elsewhere. 1 Briefly, a total of eight suitable models describing pharmacokinetics of valproate in older people (≥ 65 years) with epilepsy were selected from the literature through a systematic review. All the models were one-compartment models with first-order absorption and first-order elimination.…”