Proposed Outcomes

What PancreaTrack is designed to measure, and what we believe the data can demonstrate at scale.

Primary Platform Outcomes

PancreaTrack captures longitudinal patient-generated health data (PGHD) across five clinical domains. The following table describes each domain, the data captured, and the measurable outcome it enables.

Domain	What PancreaTrack Captures	Proposed Measurable Outcome
Pain	NRS 0–10 scores, timestamps, free-text notes	Flare frequency, severity trend over time, correlation with dietary events
Nutrition	Meal entries with fat grams (USDA-sourced or manual), enzyme dose per meal	Mean daily fat intake, PERT dose-per-gram-fat ratio, dietary pattern classification
Bowel	Bristol Stool Scale type, oily/floating flag, timestamps	Steatorrhea frequency, dose-response correlation with PERT, treatment response
Labs	Patient-entered fecal elastase, lipase, HbA1c, vitamin panels, CRP, and 8 other markers	Longitudinal trend in nutritional status markers, malabsorption indicators
Glucose	CGM continuous readings via Dexcom API (every 5 minutes)	Time-in-range percentage, post-prandial excursion patterns, hypoglycemia frequency

Clinical Outcome Hypotheses

The following are the primary hypotheses that PancreaTrack data is positioned to test, given adequate sample size and study design:

💊

PERT Optimization

Structured fat-dose-outcome logging will identify a patient-specific lipase unit per gram fat ratio associated with normal stool outcomes.

📉

Pain Reduction

Patients who review longitudinal pain logs with their care team will show greater NRS reduction at 6 months compared to standard care.

🩸

Glucose Control

Post-TPIAT patients using CGM integration will achieve higher time-in-range percentage than those relying on HbA1c alone.

🥗

Nutritional Status

Regular dietary fat logging correlates with improved fat-soluble vitamin levels (A, D, E) at 6-month follow-up labs.

Patient-Reported Outcome Measures (PROMs)

PancreaTrack data is designed to complement validated PROM instruments used in pancreatic disease research:

PANQOLI (Pancreatitis Quality of Life Instrument) — pain and functional status
PROMIS-GI — bowel function and GI symptom burden
CGM-derived Time-in-Range — glucose outcome metric per International Consensus 2019
PERT Adequacy Score — based on fecal elastase, stool type, and fat intake correlation

PancreaTrack does not administer validated PROMs natively but captures the underlying data from which PROM-relevant metrics can be derived or triangulated.

Secondary Outcomes

Appointment preparedness — Patient self-reported confidence entering appointments (pre/post survey)
Clinical communication efficiency — Physician-reported time savings using AI-generated summaries vs. verbal history-taking
Treatment adjustment frequency — Number of between-visit medication adjustments prompted by PancreaTrack data review
EPI diagnostic time — In a referral pathway pilot, reduction in median time from symptom onset to EPI diagnosis

Data Quality Considerations

⚠️

Patient-generated data has known limitations

Logging frequency, accuracy of fat gram estimation, and recall of enzyme doses all introduce noise. Studies using PancreaTrack data should account for adherence rates and include completeness thresholds (e.g., ≥70% logging days during study period) as inclusion criteria.

Long-Term Vision

At sufficient scale, de-identified PancreaTrack data could contribute to:

Natural history datasets for chronic pancreatitis and EPI (conditions with limited published longitudinal data)
Training data for disease-specific AI models that outperform general LLMs on pancreatic disease pattern recognition
Reference ranges for patient-specific PERT dosing based on fat intake and outcomes across the patient population
Registry linkage studies with institutions like the NAPS2 cohort or INSPPIRE consortium