Automated Classification of Diseases from Medical Reports

74% of main diagnoses automatically detected

Challenge

Medical patient reports need to be classified with disease codes. At the University Hospital in Berne, this process was manually performed by a coding team consisting of highly trained doctors.
The hospital wanted to develop and evaluate machine learning algorithms to automatically classify patient reports in order to support the team in their daily work.

Value

Zühlke's machine learning based solution enables the automated classification of medical patient reports and can be employed in an expert in-the-loop setting to increase the efficiency of the coding team.
In just three weeks, Zühlke's team of data scientists and consultants develioped a first prototypical machine learnng solution and demonstrated its feasibility.

Results

Key decision makers decided to invest in further development Zühlke developed a state-of-the-art solution based on deep learning in another four months.

Substainability

Zühlke enabled the UNiversity Hospital Bene to establish a data science team and corresponding infrastructure.