Digitalising the past decades: automated ICD-10 coding of unstructured free text dermatological diagnoses.

ABSTRACT:

Background

Current digital medical databases record systematically coded diagnoses, but many legacy databases are full of hand-written, free text diagnoses, which can only be meaningfully analysed after mapping them to a coding system. While diagnoses can be extracted from full medical notes with good accuracy, no algorithm using only an unstructured free text diagnosis with no additional data has been published to date.

Objectives/methods

Therefore, we sought to create an algorithm which maps hand-written German diagnoses from our clinical photography database to ICD-10 diagnosis codes, validate its output manually by dermatologists and analyse diagnosis counts over time as a proof-of-concept of its application.

Results

Our rule-based algorithm mapped 50,884 unproces

SUBMITTER: Sitaru S 

PROVIDER: S-EPMC11520645 | biostudies-literature | 2024 Oct

REPOSITORIES: biostudies-literature