Paper Image

Improving similarity of patients for precision medicine using ICD codes

Published on:

14 August 2023

Primary Category:

Machine Learning

Paper Authors:

Jan Janosch Schneider,

Marius Adler,

Christoph Ammer-Herrmenau,

Alexander Otto König,

Ulrich Sax,

Jonas Hügel

Bullets

Key Details

Proposes method to improve ICD-based patient similarity using a scale term

Tested 80 algorithms on sets of ICD codes from pancreatic cancer patients

Best algorithm reached 0.75 correlation with expert similarity ratings

Scale term accounts for degree of comorbidity in ICD sets

Shows importance of accounting for comorbidity variance in similarity

AI generated summary

Improving similarity of patients for precision medicine using ICD codes

This paper proposes a method to improve calculating similarity between patients using their ICD diagnosis codes. It accounts for varying degrees of comorbidity by introducing a scale term. The method was tested on pancreatic cancer patients, comparing 80 algorithms. The best performing combination reached a 0.75 correlation with expert ratings. This demonstrates the importance of accounting for comorbidity variance when calculating patient similarity from ICD codes.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up