| CPC G16H 10/60 (2018.01) [G06F 40/40 (2020.01); G16H 15/00 (2018.01); G16H 50/20 (2018.01)] | 19 Claims |
|
1. A computer-implemented method for augmenting medical coding, the method comprising:
receiving a plurality of medical records wherein the medical records comprise at least one treatment;
using a machine learning engine having a combination of a plurality of machine learning methods, including natural language processing (NLP), to build a relationship between a plurality of diagnosis codes and a plurality of procedural codes;
using said machine learning engine to convert a portion of said plurality of medical records received into a machine readable first medical code using NLP;
using said machine learning engine, querying a knowledge graph-comprising historical medical records and an associated procedural coding catalog for an output having records with at least one component that is the same as to that of said first medical code,
generating a relationship between said output received and said first medical code to calculate a medical code of higher order;
using a training engine to search evidence in the medical record for the by comparing at least a portion of clear text the medical record; and
providing a confidence score representing an estimation of a probability of a correctness of the generated second medical code, wherein the searching includes using a machine learning model to learn from a plurality of past coding decisions confirmed by a human user and integrating the evidence in the medical record into an ontology.
|