Problem
Finding the optimal treatment pathway
The client wanted to automate the prescription process by identifying the optimal treatment pathway for a patient based on his/her symptoms with the help of historical data present for similar kind of disease. This would help the client in improving the success rate of the treatments specific to different type of patients and diseases. In addition to this, an accurate future prediction of the treatment would also help optimize the inventory for the prescription medicine or treatment requirements.
Technology Used
Python, TensorFlow, SQL (for extracting historical patient and treatment data)
Solution
A simple machine learning approach was recommended that made use of various clustering techniques and scoring models on the historical data for different type of diseases, their respective symptoms and success rates of the treatments. Also, to be able to create a treatment pathway, we had to perform computation on the time-series data which required a robust hardware and software combination (Python and TensorFlow in conjunction with GPU-based hardware) to optimize the processing capabilities and improve efficiency. The output of these models quantified a treatment pathway based on the cluster to which the test patient belonged.
Contact Us
Call: 020 3573 6917
Email: enquiries@bsbiconsulting