Risk Calculation Model

This AI model evaluates a patient’s age, heart rate, oxygen level, temperature, and diabetic status to answer one critical question:

“How likely is this patient to deteriorate soon (require urgent care, transfer to higher monitoring, or risk readmission)?”

What the model does

• Does not read clinical notes or make diagnoses.
• Only uses numeric and yes/no inputs provided.
• Returns:
  • A probability (between 0 and 1) quantifying the risk.
    • Example: 0.92 → “92% chance of deterioration according to the model.”
  • A label (low, medium, or high risk) derived from that probability, so humans or other systems can act.

Example

Input (patient data): is_diabetic = 1 spo2 = 96% temp = 98°F age = 11 heart_rate = 112

Output (model response): Probability = 0.91 → "This person is very likely to deteriorate soon." Label = high → "Treat this as a high-priority case."

How to Use the Output

• If probability ≥ 0.75 → Automatic escalation or clinician alert.
• If probability is 0.5 – 0.75 → Flag for closer monitoring or human review.
• If probability < 0.5 → Routine care, no immediate escalation.

Test Framework flow

Read input values and expected output from excel sheet -> Normalize the values -> Pass the values to the Model -> Model reads the input values and returns Probability and Label -> Validates the model output with expected output -> Generates report with the results.

Setup Instructions

1. Download the repository to a local folder.

2. Install Python

3. Create & activate virtual environment
   Windows python -m venv venv venv\Scripts\activate

   macOS/Linux python3 -m venv venv source venv/bin/activate

4. Install dependencies
   pip install -r requirements.txt

How to Run Tests

Run all tests: pytest -v -s

Generate an HTML report (saved under reports/): pytest -v -s --html=reports/report.html --self-contained-html

Logs

• Console output shows test progress.
• A file testlogs.log is created/updated automatically with detailed logs:

Example: 2025-09-28 14:10:05 [INFO] Running test case | input='105620' | expected='10:56:20 AM' | actual='10:56:20 AM' 2025-09-28 14:10:05 [INFO] Running test case | input='1530' | expected='3:30 PM' | actual='3:30 PM'

Viewing Test Reports

After running tests with the HTML option, open the generated report in a browser:

1. Navigate to the reports/ folder in your project.
2. Open report.html in any modern web browser.