Algorithm Design to Generate Automated Treatment Protocols for Rheumatoid Arthritis

Project Description

We are looking to hire a machine language engineer to develop algorithms based on health care data which can generate automated treatment protocols given a definitive diagnosis.

Aims

The aim of this study is to explore whether a computer - by means of a machine learning algorithm - can generate a treatment protocol given a definitive diagnosis for a disease that is complex to treat. This is not an attempt at automating diagnosis, but given a definitive diagnosis, predicting a treatment plan. The study will further compare the level of accuracy that can be attained by the algorithm, with an initial benchmark of matching the capability to be on par with that of a first year resident.  The first disease that will be explored for this study is Rheumatoid Arthritis.

As data on RA patients and the treatment plans that are designed by doctors for each patient is gathered, it will be handed over to machine learning engineers who will be responsible for developing the initial algorithm. The initial data points and those collected subsequently will be fed into the algorithm to teach it how to generate treatment protocols. Once a critical mass of data points is reached, these automated protocols will be validated by doctors. The validation or amendments will also need to be fed into the algorithm to train it to be more accurate. Over the next few months, we plan to have a sufficiently robust algorithm that can not only replicate the efforts of a first year resident but will also be able to output an explanation supporting that protocol. The ability to provide an explanation is a key component as a doctor needs to be able to justify signing off on any automatically generated protocol.

For a more detail descrtiption of project please view attachement.