A rule-based clinical decision support demo built with Python, SQLAlchemy, and Jupyter Notebook. Loads a JSON symptom database into SQLite, applies a scoring formula (Frequency × Evoking Strength), and returns the top probable diagnoses in descending order.
A simplified clinical decision support system using Python, SQLAlchemy, and Jupyter Notebook.
Loads a JSON symptom database into SQLite, applies the scoring formula:
Score = Frequency × EvokingStrength
and returns the top-N probable diagnoses in descending order.
- JSON → SQLAlchemy ORM → SQLite pipeline
- Many-to-many mapping (symptoms ↔ conditions)
- Transparent rule-based scoring (no ML “black box”)
- Outputs top-N probable diagnoses for patient symptoms
Click the badge to run in the cloud (no install):
Input symptoms:
["Fever", "Cough", "Headache", "Shortness of Breath"]
This project demonstrates how a simple rule-based clinical decision support system can be built by combining structured data, databases, and algorithmic scoring. The goal was to replicate the workflow of mapping symptoms to conditions and producing ranked outputs — similar to how decision support modules in EHR systems work.
- How to design and query a many-to-many schema using SQLAlchemy and SQLite
- How to transform JSON clinical data into a normalized relational structure
- How to implement and test a rule-based scoring function (Frequency × EvokingStrength)
- How to present results in Pandas DataFrames and visualize outputs for interpretation
- Shows transferable skills in data modeling, database management, and algorithm design
- Reflects how informatics professionals build clinical decision support (CDS) prototypes
- Demonstrates practical use of Python + SQL for healthcare data workflows