Table 1 Index card for major stakeholders and their tasks
From: An operational guide to translational clinical machine learning in academic medical centers
Role | Main responsibility | Tasks |
|---|---|---|
Principal Investigator | Maintain an overview of the project and coordinate tasks. | Identify and convene relevant stakeholders for the project Ensure smooth information flow among stakeholders across different project phases and modules Review that complete set of information is transferred during stage transitions Vet the infrastructure constraints Approve interface mock-ups Weekly oversight of project timeline and budget |
Machine-learning engineer | Program a tool that pulls data, generates a prediction and delivers it to the end user according to the needs | They are the primary drivers for all tasks in the paper |
Data scientist | Ensure the MLE is able to faithfully transcribe the model into a working platform and modify the model as required | Document the modeling process: data sources, pre-processing steps, model inference, outputs and interpretation Make changes to the model based on evolving needs |
Clinician or user representative | Ensure the model output is valuable to the end user | Confirm availability of model inputs at time of prediction generation Ensure interpretability of model outputs and their relevance given clinical needs Provide feedback on application interface |
Information Technology | Serve as technical expert on data sources, infrastructure hosting and best practices | Help connect to and pull data from the data warehouse Vet hardware constraints for the platform Help host the platform |
