Validation of Betalactam ML Prediction Models - TDMAide

University Hospital, Ghent600 enrolled

Overview

The goal of this study is to learn about the real wold behavior of developed machine learning models that predict the plasma concentration of piperacillin-tazobactam and meropenem in critically ill patients admitted to the intensive care unit (ICU). The main aim of the study is to validate the performance of these machine learning models. To this end, daily measured plasma concentrations of the investigated antimicrobials will be compared with the predicted concentration by the machine learning algorithms. Additional goals of the study include: * To describe the total plasma concentration over time of piperacillin-tazobactam and meropenem in patients admitted to the ICU. * To quantify the correlation between plasma concentrations of piperacillin-tazobactam and meropenem and the development of side effects. * To evaluate the perceived necessity of therapeutic drug monitoring (TDM) of consultants and physicians in training working in the ICU. * To evaluate the perceived added value of daily TDM. Samples (where possible taken routinely) from participating patients will be analyzed for meropenem and piperacillin-tazobactam plasma concentration. Participating physicians will be asked to fill in a short daily questionnaire during the time a patient under their care is treated with the antimicrobial under investigation.

Study Type

INTERVENTIONAL

Allocation

NON_RANDOMIZED

Purpose

TREATMENT

Masking

NONE

Enrollment

600

Conditions

Infections

Prediction of plasma concentration of piperacillin-tazobactam or meropenemDEVICE

For included patients, a prediction will be made by developed machine learning models about the expected plasma concentration of piperacillin-tazobactam or meropenem by using routinely collected health care data.

Determination of plasma concentration of piperacillin-tazobactam or meropenemDIAGNOSTIC_TEST

For included patients, the total plasma concentration of piperacillin-tazobactam or meropenem will be determined. Were possible, this will be done using a blood sample that was collected during routine daily bloodwork which is performed in the morning. If no routine sample is available, a study specific sample will be drawn at approximately the same time as a routine sample would be drawn.

Daily short questionnaireOTHER

Physicians who care for patients included in the study will be asked to fill in a short daily questionnaire that evaluates the perceived necessity and added value of daily therapeutic drug monitoring.

Eligibility

Sex: ALLMin age: 18 YearsMax age: 100 YearsHealthy volunteers:

Medical Language ↔ Plain English

Patients Inclusion Criteria: * Admission to the ICU. * Age above 18 years old. * Treatment with piperacillin-tazobactam or meropenem for less than 48 hours. Exclusion Criteria: * Pregnant or lactating patients. * Limitation of therapy beyond "Do not resuscitate". * Expected demise within 48 hours after inclusion. * Haemoglobin \< 7 g/dL. * Previous inclusion in this study for a treatment course with the same antimicrobial. Consultants and physicians in training Inclusion Criteria: * Consultant or physician in training working in the ICU. Exclusion Criteria: * None

Outcomes

Primary Outcomes

Difference between predicted (TDMAIde) and measured (via HPLC-MS/MS method) plasma concentrations

The difference between the predicted concentration (from the TDMAide software) and the concentration range based on the measured concentration (measured from the blood sample from the patient and analyzed using a HPLC-MS/MS method, with and without taking into account intra- and inter measurement variabilities of the HPLC-MS/MS method.