What makes Burak different
We catch it before the fourth report has to.
Standard labs compare one marker to one range. Burak compares markers to each other and to their own trend — so a developing pattern is visible from report 1 or 2, weeks before every marker has crossed its line on its own.
Cross-marker reading
Flags from report 1
Your own disease cohorts, defined as you go
Simulate a report upload
In production this reads a lab file directly. Here, one click simulates a report arriving for Ahmed K.
sample_report_ahmed_k_wk4.pdf
8 biomarkers · week 4 panel
File received
Parsing values
Matched to patient record
Routed to Category Split
Parsed values
Exactly what Category Split receives next.
Status distribution
Where each of the 20 patients currently sits.
Panel load
Patients with at least one flagged marker, per biological line.
The core of the product
Cross-marker intelligence
Any lab reads one marker against one range. Burak reads markers against each other — so the pattern shows up by report 1 or 2, not report 4.
All patients
How this is calculated
Every marker lands in one panel — Musculoskeletal & Inflammation, Micronutrient & Recovery, or Infection — based on what it measures. These are standard clinical and sports-medicine lab groupings: CRP and Creatine Kinase for muscle damage and inflammation, Vitamin D and Ferritin for recovery and iron status, White Blood Cell Count and Procalcitonin for infection. A marker is "flagged" only when its latest reading is both outside the Burak threshold and has moved the wrong way since its first reading — never from a single number taken out of context.
How this is calculated
The "standard" band is the general-population lab range you'd see on any report. The "Burak cohort" band is built from athletes only — a fit, high-training-load population reads differently from the general public on markers like Vitamin D and Ferritin, so the same number can be unremarkable on one scale and worth a second look on the other. The markers themselves (Vitamin D, Ferritin, and the others used across this pipeline) are standard, peer-reviewed clinical biomarkers; this POC's cohort band is illustrative and would be recalibrated against Burak's real athlete population before use.
These are the confirmed, named patterns — validated when a defined group of markers moves together. "Predictive Intelligence" (next in the sidebar) tracks earlier, related signals that often lead into these same cohorts — they are not separate diagnoses, just an earlier read on the same story. This is a proof of concept: more cohorts can be named and added as we deep-dive into your data.
How this is calculated
A cohort confirms only when every marker in its definition is flagged together — e.g. Overreaching / Injury-Risk needs Creatine Kinase, CRP and Cortisol:Testosterone all past threshold and trending the wrong way at the same time. A single flagged marker on its own is held as a lower-urgency "watch", not a confirmed cohort. The specific markers, thresholds and required combinations are set by Burak and can be recalibrated against your own outcomes data — every marker used here (CRP, Creatine Kinase, Cortisol:Testosterone, Vitamin D, Ferritin, Transferrin Saturation, White Blood Cell Count, Procalcitonin) is a standard biomarker already used in clinical and sports-medicine lab panels. This POC uses simulated thresholds for demonstration, not validated clinical cut-offs.
These are early, cross-panel signals — not new diagnoses. They typically precede or pair with the named patterns in "Disease Cohorts" (see sidebar); the names are deliberately different because the confidence level is different. This is a proof of concept: any signal here can be promoted into its own validated, named cohort as we deep-dive into your data.
1
Report 1 landsTwo markers from different panels are already compared against each other — not just one number against one range.
→
2
Report 2 confirms the directionNow there's a trend. Burak projects it forward and flags what it's heading toward.
→
3
Reports 3 and 4 just confirm itBy then it's paperwork — Burak already called it.
Composite signals
Two panels, one patternPrediction vs. what happened
Called early. Checked against reality.How this is calculated
Two checks run on every patient. Report 1: a marker counts as an early sign if it already sits on the worse side of the squad's own median for that marker — judged against peers, not a fixed cut-off. Report 2: if the direction from report 1 to report 2 is adverse, Burak projects that same rate of change forward to report 4; if the projection would cross the clinical threshold, the patient is flagged right there. Reports 3 and 4 are shown afterward only to check the call against what actually happened — they play no part in making the prediction itself.