Automated Coding of Medical Records

Brief Description:

This system is intended to compute the ICD 10AM codes from the Electronic Patient Record (EPR). It is expected to use various technologies already developed by the Laboratory, namely the methods for identifying SNOMED CT codes and other relevant entities in the clinical notes.

Project Aims:

The two aims of this project are:

• 1. to improve the accuracy of recording the correct ICD10AM codes for a patient journey,
• 2. to improve the saturation of recording ICD10AM codes for a patient journey

Project Objectives:

 The objectives of this project is:

1.  to automatically compute all instances of content that maps to an ICD10AM code that can then be confirmed by a medical coder.

2.  to devise methods so that the automatic processor can learn from the medical coder to improve its performance.

Brief Project Methodology:

 The methodology is:

1. Identify the clinical information systems that contain data appropriate for ICD10AM coding.

2. Identify a collection of records that have already been coded.

3. Use those records to create an ICD10AM recogniser and classifier.

4. Design the method for retrieving the desired data from the host clinical information system.

5. Design an application so that the coders can review the electronic clinical record and correct the computed codes.

5. Test the system with trusted staff and develop methods for operationalising the system.

6. Train all staff to use the system.

7. Collect feedback on the use of the system and ways to improve it.

How will this Project improve Patient Care:

This project will not directly improve patient care.

This project will improve the hospital administration by ensuring that more ICD10AM codes are extracted from the patients' records and that they are more accurately represented.