[PYTHON] I gave a poster presentation at the 36th Joint Conference on Medical Informatics.

I have made a poster presentation on the submitted paper "Efforts to prevent hepatitis B virus reactivation in patients undergoing immunosuppression / cancer chemotherapy using DWH, PostgreSQL, and pandas".

Poster presentation is to first submit a dissertation (abstract) to an academic conference, create a poster for presentation about the adopted abstract, and briefly explain the contents in about 5 minutes. I was quite nervous because I had never made such a presentation at an academic conference, but I managed to get it.

In order to build a check mechanism for a specific disease, the necessary data is extracted from the electronic medical record database (DWH), and the process of creating an alert file is automated using SQL and script languages. I think I'm doing something quite advanced, such as using a jupyter notebook. The purpose of this system is to create alerts that encourage clinical departments to administer nucleic acid analogs or monitor with necessary tests to prevent hepatitis B reactivation. As a procedure

Data extraction from electronic medical record DWH (extraction of patients receiving target drugs) ⇩ Stored in PostgreSQL ⇩ Steroid conversion by case when formula ⇩ Query execution with Python (pandas) ⇩ Collect query results in an alert file (using Python)

It has become. The algorithm is based on the "Guidelines for Countermeasures against Hepatitis B Caused by Immunosuppression and Chemotherapy" formulated by the Japanese Society of Hepatology. http://www.jsh.or.jp/files/uploads/R2-HBV%20Ver2.2.pdf Some of the created SQL statements and Python code have been published on github. https://github.com/riow1983/HBV-R

In particular, it is characterized by being compatible with steroid conversion. Many similar operations have been reported, but we have not been able to accurately grasp the steroid titer of steroids. Any medical institution due to the use of PostgreSQL and jupyter notebook of open source is characterized by the place that it is available. In the future, I would like to introduce a dedicated server in the HIS network with the aim of further saving work and strengthening the backup system.

Recommended Posts

I gave a poster presentation at the 36th Joint Conference on Medical Informatics.
Hannari Python At the LT meeting in December, I made a presentation on "Python and Bayesian statistics".
I did a little research on the class
What I thought after working on the "No comment at all" project for a year
I made a VGG16 model using TensorFlow (on the way)