Malaria and Typhoid Diagnosis Expert System source code

The problem it solves

Computer Science students need a small, explainable rule-based system to demonstrate how medical expert systems combine evidence. The project focuses on malaria and typhoid fever, using fictional cases, transparent rule logic and offline operation.

Project topic: Development of a Medical Diagnosis Expert System for Malaria and Typhoid Fever

Who uses it

What it does

• Fictional case entry and analysis

  Attendant records up to eight observations, each marked not present, possible or certain. The system runs the rule engine and shows the result immediately.

• Explained result with rule trace

  Shows the entered facts, every matched rule with its weight and contribution, the combined score, the score band meaning and a limitation notice.

• Attendant case history

  Lists the signed-in attendant's saved fictional cases, with pagination and a link to the explained result.

• Rule catalogue with versioning

  Admin lists, adds and edits illustrative rules. Each rule stores its source, version, active flag and explanation text. Existing saved cases keep the version they used.

• All saved cases for admin

  Admin reviews every saved fictional case across all attendants, with pagination and a link to each explained result.

• Login and role-based access

  Standard login with hashed passwords, CSRF protection and role checks. The sample logins are demonstration credentials.

Screenshots

Screenshots of the running system, taken by our checks. The records are sample data and are labelled so on each page.

Database tables

• usersLocal accounts and roles for authentication.id, username, password_hash, full_name, role, created_at
• rulesThe editable illustrative rule base.id, name, condition, outcome, weight, source, version, active
• consultationsOne fictional case with its saved rule version and final score.id, user_id, rule_version, final_score, created_at
• case_factsThe eight entered observations and their certainty for one case.id, case_id, observation, certainty
• case_rule_matchesA snapshot of each matched rule so an old case keeps its explanation.id, case_id, rule_id, rule_name, outcome, weight, source, version, confidence, contribution

Technology

• Python. Readable, widely taught, and supported by the student's environment.
• Flask. Lightweight and easy to explain.
• Jinja. Keeps HTML clean and reusable.
• SQLite. Zero configuration and fully offline.
• Bootstrap. Local files in static/vendor, no CDN.
• Chart.js. Local copy in static/vendor, no CDN.
• Werkzeug. Security best practice with a simple API.
• Flask-WTF. Required for safe form submissions.

How to run it: running a Python project on Windows. Not sure which language to pick? Python or PHP.

How it was tested

17 of 17 tests passed, 11 of them automated. The system was also installed from scratch and every screen was opened as the right kind of user before it was accepted.

What it does not do

• All rules and weights are synthetic and unvalidated teaching examples. The system never diagnoses, excludes illness or recommends treatment.
• The score is an illustrative combination of positive evidence, not a calibrated probability and not a measure of diagnostic accuracy.
• Only positive evidence is modelled. Negative evidence would need a separate, tested signed combination.
• No real patient identifiers are stored, and the system does not connect to a medical device, external AI or hospital record system.
• An admin edits the rule base by hand. There is no clinical review or approval workflow.
• The system does not do the following: image or laboratory analysis, prescription support, patient record storage, user self-registration, email or SMS alerts, cloud sync, multi-language output, or any real clinical decision support.
• The Windows run was not verified from this macOS environment. Windows users should run run.bat and follow guide/RUN.md.