Patient Outreach: Abnormal Batches

philipgodfrey

Abnormal Batches.dva

1. Which dataset did you use?

Life Sciences

 

2. How did you analyze or prepare the data?

The data was analyzed by focusing on a single product (OracleVax) and then prepared the data by normalizing fatalities and serious adverse events per 100 patients, to enable fair comparison across vaccine batches with different administration volumes.

Each batch was scored based on:

• Serious adverse event rate (per 100 patients)
• Fatality rate (per 100 patients)

Outlier batches with significantly elevated rates were flagged as “abnormal" and included in the dashboard filter.

Further analysis examined historical trends, correlations by country and time-based patterns, which helped identify whether certain geographies or time periods showed elevated risk patterns.

Patients were then categorized into one of 3 groups:

• Urgent Appointment
• Consider Appointment
• Backlog

The classification was driven by various factors including: 

• Country risk level
• Quarter administered
• Demographic indicators

3. Who is the intended audience for your visualization?

GP Practice / Medical professionals can use this visualisation to inform decision-making for patient outreach prioritization.

 

4. What is your visualization about, and what question or problem does it address?

The visualization focuses on identifying and prioritizing patients who received vaccines from “abnormal” batches. These are batches with the highest rates of serious adverse events and fatalities per 100 patients.

The problems addressed are that healthcare providers need a structured, data-driven way to:

• Identify patients potentially at higher clinical risk
• Prioritize limited appointment capacity
• Reduce adverse outcome risk through proactive review

The visualisation aims to answer questions such as:

• Which vaccine batches show abnormal safety signals (those included in dashboard filter)
• Which patients received those batches?
• How should patients be prioritized for follow-up?
• Are there country or time-based trends that increase risk?

 

5. Did you use any Oracle Analytics AI features when building your visualization (ex. AI Assistant)? If so, please describe how they were used

Exploratory analysis using Auto-Insights which identified increased fatalities by country, which then led us to explore by time period using normalized patient rates.

 

6. Did you upload your visualization image and dva file?

Yes - and video included to show interactivity of the canvas.