Assessment Guide for Algorithmic Impact: User Instructions
In a significant development, a new user guide has been unveiled, aimed at streamlining the process for teams seeking access to imaging data from the National Medical Imaging Platform (NMIP) for research, training new medical products, or testing existing ones. Developed in partnership between the NHS AI Lab and the platform's partners, this comprehensive guide provides clear and concise step-by-step guidance for project teams on how to conduct an Algorithmic Impact Assessment (AIA).
The guide does not provide direct access to the NMIP but outlines the process for teams to access it for their intended purpose. Completion of the AIA is a prerequisite for access to the NMIP dataset by the NHS AI Lab team.
The AIA process, as detailed in the guide, involves several key steps. Firstly, applicants must submit an AIA that evaluates the potential impact of their proposed use of NMIP imaging data. This includes consideration of ethical, legal, and social implications related to data use.
Next, the intended use of the imaging data must be clearly defined—whether for research, product training, or testing—to ensure appropriate review pathways are followed. Following this, risks associated with the algorithm’s intended application, particularly concerning privacy, bias, and fairness, must be identified and assessed.
Mitigation strategies to address these risks and ensure the responsible use of imaging data are then outlined. The process concludes with confirmation that the proposed work complies with relevant governance frameworks, data protection regulations, and NHS standards. The completed AIA is then submitted for review by the NMIP governance board or a designated committee aligned with the NHS AI Lab protocols before access is granted.
This structured AIA approach is designed to ensure that accessing NMIP imaging data is done responsibly, balancing innovation in AI applications with safeguards against potential harms. The guide also includes a template for conducting an AIA, further simplifying the process for teams.
The partnership is focused on exploring AIAs in healthcare, and the user guide is part of research related to AIAs in healthcare on the NMIP. A project page is available to learn about wider work exploring AIAs in healthcare and to access the full report.
It is important to note that the user guide is not for general usage but is specifically for teams seeking access to NMIP imaging data for the mentioned purposes. Since the exact detailed procedural steps and documentation flow are specified in the official NMIP AIA user guide created by the NHS AI Lab and partners, it is recommended to consult that document for the step-by-step instructions, specific form requirements, templates, and contact points for submissions.
This user guide marks a significant step forward in facilitating responsible and ethical use of NMIP imaging data, paving the way for innovative research and product development in the healthcare sector.
Scientists working in health-and-wellness, particularly those interested in AI applications, may find utility in the new user guide, as it details the process for accessing imaging data from the National Medical Imaging Platform for research or product development. Furthermore, this guide focuses on implementing Algorithmic Impact Assessments (AIA) using technology to ensure ethical and responsible use of the platform's imaging data.
