Where did 47 patients go?

Your Phase 2 trial enrolled 120 patients.

Your biomarker analysis? 73 evaluable samples.

Where did 47 patients go?

→ Site did not collect baseline sample (12 patients)
→ Samples shipped at room temp instead of frozen (8 patients)
→ Patient withdrew before key timepoint (15 patients)
→ Insufficient tissue for analysis (7 patients)
→ Processing delays led to degraded samples (5 patients)

This isn't unusual. It's predictable.

The fix isn't hoping for perfect execution.

Build your biomarker plan assuming 20-30% attrition:

✓ Over-enroll based on your minimum evaluable sample size
✓ Train sites on collection protocols before first patient in
✓ Real-time sample tracking with immediate feedback loops
✓ Backup collection timepoints when feasible
✓ Partner with labs experienced in your sample types

Your statistical plan should account for missing data from day one.

Because discovering you're underpowered after database lock isn't a biomarker strategy—it's expensive hope.