Considerations for Quality Control Monitoring of Machine Learning Models in Clinical Practice (Preprint)

Abstract

Integrating machine learning models into clinical practice presents a challenge of maintaining their efficacy over time. While the existing literature offers valuable strategies for detecting declining model performance, there is a need to document the broader challenges and solutions associated with real-world development and integration of model monitoring solutions.

This work details the development and utilization of a platform for monitoring the performance of a production-level machine learning model operating in Mayo Clinic. The aim of this work is to provide a series of considerations and guidelines necessary for integrating such a platform into a team’s technical infrastructure and workflow. We document our experiences with this integration process and discuss the broader challenges encountered with real-world implementation and maintenance. Source code for the platform is also included.

Our monitoring platform was built as an R shiny application; developed and implemented over the course of 6 months. The platform has been utilized and maintained for 2 years and is still in use as of July 2023.

The considerations necessary for the implementation of the monitoring platform center around four pillars: Feasibility – what resources can be utilized for platform development?; Design – through what statistics or models will the model be monitored, and how will these results be efficiently displayed to the end-user?; Implementation – how will this platform be built and where will it exist within the IT ecosystem?; and Policy – based on monitoring feedback, when and what actions will be taken to fix problems and how will these problems be translated to clinical staff?

While much of the literature surrounding machine learning performance monitoring emphasizes methodological approaches for capturing changes in performance, there remain a battery of other challenges and considerations that must be addressed for successful real-world implementation.

RR2-10.1186/s13063-021-05546-5