ArcMAP: Machine Learning Assisted Medical Concept Mapping for NHS Data Standardization

A two-month longitudinal real-world evaluation of the ArcMAP system using medication and laboratory test data from five UK-based NHS hospitals. The dataset is a PDF document describing an end-to-end application that integrates a biomedical representation model into a human-in-the-loop workflow to accelerate medical concept mapping. It was authored by Joseph Cronin and last updated on 2026-04 14.

Use Cases

• Benchmarking machine learning models for medical concept mapping based on the described evaluation of top-1 accuracy.
• Designing human-in-the-loop workflows for clinical data validation based on the described graphical user interface.
• Studying the impact of continuous learning on model adaptation based on the described feedback pipeline.
• Analyzing variability in data coding practices across healthcare providers based on the described results from multiple NHS hospitals.

Strengths

• Includes a real-world evaluation conducted over a two-month period.
• Reports specific performance metrics, such as top-1 accuracy increasing from 37.0% to 91.6% for laboratory test names.
• Evaluates the system across multiple deployment scenarios, including the onboarding of a new hospital.

Limitations

• The dataset is a 99.5 KB PDF document, which is a small-scale report rather than a primary data resource.
• Row count and column-level documentation are absent; the content is a research description.
• The data described may reflect biases inherent to the five specific UK NHS hospitals used in the evaluation.

Provenance

Source

Joseph Cronin via figshare.

Collection Method

Likely derived from a research project evaluating the ArcMAP system.

Time Range

The evaluation period covered a two-month longitudinal study.

Freshness

Last updated 2026-04-14 05:36:35.

Geography

Data from five UK-based NHS hospitals.