Report Request BETA
Ask a clinical question in plain language and receive an AI-generated structured report drawn directly from the patient's logged data.
Report Request is an experimental clinical tool. All output is generated from patient-reported data only and does not constitute a diagnosis, treatment recommendation, or clinical guideline. Always apply independent clinical judgment.
What It Does
Report Request lets you query a patient's longitudinal data using plain English. Instead of manually reviewing each tab and building a picture yourself, you can ask a specific clinical question and receive a structured report with a narrative, key findings, and auto-generated charts — in seconds.
All data is pulled directly from the patient's PancreaTrack record. The AI never fabricates or infers data that isn't present.
How to Use It
Navigate to any linked patient and click the Report Request tab (last tab, marked BETA).
Write your query in plain English — up to 500 characters. Be specific about the time range and the data you want to analyze. Use the example queries as a starting point.
The system pulls the relevant data, constructs a structured prompt, and sends it to Gemini AI. Generation typically takes 10–30 seconds depending on the amount of data.
The report returns a clinical narrative, a key findings list, and up to two auto-generated charts visualizing the most relevant data.
Use the Print Report button to generate a clean, printable version. Use Ask Another Question to run a new query on the same patient.
Example Queries
The following examples illustrate the types of questions Report Request is designed to answer:
| Query | What It Analyzes |
|---|---|
| "Show me enzyme dose vs bowel outcomes over the past 14 days" | Enzyme logs correlated with bowel log outcomes and urgency |
| "What foods appear to correlate with high pain scores?" | Meal logs cross-referenced with pain score timing |
| "Summarize steatorrhea frequency and fat intake for the past 30 days" | Bowel logs (oily stool) and daily fat totals from meal logs |
| "Show pain trends alongside meal fat content for the past week" | Pain score trend chart overlaid with fat intake |
| "Are there patterns between enzyme dose timing and symptoms?" | Dose timing field in enzyme logs vs. outcome and symptom fields |
| "Give me a longitudinal summary of lab trends over the past 90 days" | All lab values logged within the 90-day window |
Time Range Parsing
Report Request automatically detects time ranges mentioned in your query:
- "past 14 days" → 14-day window
- "last 2 weeks" → 14-day window
- "past 3 months" → 90-day window
- No time range specified → defaults to 30 days
- Maximum range: 365 days
Data Sources
Every report draws from the following logged data sets (where available):
| Data Source | Fields Included |
|---|---|
| Pain Logs | Date, score (0–10), location, notes |
| Meal Logs | Date, meal name, fat (g), carbs (g), protein (g), calories, notes |
| Enzyme Logs | Date, meal description, fat (g), brand, lipase units, dose timing, outcome, symptoms, notes |
| Bowel Logs | Date, BSC type, color, urgency, oily stool indicator, notes |
| Lab Values | Date, lab type, value, unit, notes |
| Patient Profile | Age, sex, diagnosis type, known cause, baseline symptoms |
If the patient has no logged data in the requested time range, the system returns an explicit message rather than generating an empty or fabricated report.
Printing Reports
The Print Report button opens a clean, standalone print window containing only the report — not the full patient dashboard. The printed output includes:
- Patient name and report timestamp
- The original physician query
- The full clinical narrative
- Charts rendered as static images
- Key findings list
- Clinical tool disclaimer
Limitations
- All data is patient-reported — accuracy depends on logging consistency
- The AI cannot access EHR data, imaging reports, or external lab systems
- The AI will not make treatment recommendations or provide diagnoses
- Chart data is limited to what the patient has logged — sparse logging produces sparse charts
- As a Beta feature, output quality may vary; always verify findings against the source tabs
For a full discussion of AI limitations and boundaries, see AI Limitations and Clinical Boundaries.