Electronic Health Record (EHR)-Embedded Decision Support Platform for Individualized Precision Drug Treatment in Neonates

Aim 1: To successfully implement a morphine PK and individualized precision dosing decision aid into workflows of neonatal prescribing clinicians.

This study will evaluate the steps required to successfully implement a population pharmacokinetic (PK) model-informed real-time decision support interface for the prescribing of neonatal pain and sedation medications. Currently, most opiate and benzodiazepine treatment is done by empirical dosing without any knowledge of drug concentrations, which can easily lead to under or over-dosing of medications. Background data show that over 65% of infants who are placed on morphine experience blood concentrations well above the currently suggested target ranges. This project will incorporate PK model-informed dosing and eventually include real-time drug concentration information feedback directly into prescribers' workflows so that they can make a data-informed decision, thus allowing maximal therapeutic efficacy while minimizing the likelihood of adverse events. The expected outcomes are better clinical efficacy and safety with fewer side effects in the neonatal population.

Study design Open label, prospective, quasi-experimental study that will use data collected during patients' Neonatal Intensive Care Unit (NICU) hospitalization at Cincinnati Children's Medical Center (CCHMC).

The study will use user-centered participatory design methods to establish the optimal means of incorporating this decision support into already existing workflows. These findings will allow the investigators to implement the clinical decision support directly into clinical practice during the study period. During and following implementation, the investigators will use a mixed methods approach through continually evaluating the effectiveness of the precision dosing application through technical analytics, real-time user feedback, and clinical data feedback. Technical process outcomes will consist of usage analysis statistics such as application access logs, time spent viewing the application, and technical error rates (e.g., failure of the application to launch or load data). Users will have the option to provide direct in-app feedback. Such feedback is necessary for successful agile software development and will be incorporated into regular revisions of the tool. In addition, the investigators will use clinical data including subjective bedside pain and sedation assessments as well as objective morphine concentration data to help further refine the PK/PD models.

Prospective evaluation of the decision support tool The investigators will start using the tool to provide dosing guidance based on patient information and the morphine PK model predictions but without using morphine concentrations as feedback for 3-6 months. This is anticipated to be the time required to prospectively test the tool while putting in place the logistics for timed sampling and morphine concentration measurements in real-time. Only members of the study team will have access to the tool at this time. When the clinical team is discussing the morphine usage over the last 24 hours and before the pain/sedation plan is developed, the member of the study team that is reviewing the tool will introduce this information on rounds. Currently the NICU clinical pharmacist on the study team reviews NeoRelief during rounds to assess morphine doses being used and the estimated concentrations based on the model. The tool will be made available to the other members of the study team as part of the roll out. The model-based dosing suggestions are not meant to take the place of clinical judgement. It is impossible to take into account all of the unique variables that go in to developing a pain management plan for a patient. If the clinical team choses to deviate from the recommendations, the clinical pharmacists will make a note as to why the decision was made to deviate from the recommendation.