How many samples for method comparison

Quick question: for a new non-waived chemistry assay verification, what patient-sample count do you use for method comparison to satisfy CLIA/CAP — 20 or 40? I run 40 across the measuring range per CLSI EP09 and document two QC levels daily; has anyone had a CAP inspector accept 20 with solid linearity and bias analysis?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍​‌‌‍​‍‌⁠‌​‌‍‌‌‌‍​⁠‌‍‍​‌‍⁠‍‌‍‍‌‌‍​⁠‌‍‍‌‌‍​‌‌‍⁠‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠‍​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠‌​‌⁠‍​‌‍​‌​⁠‌‌‌⁠‌​​⁠​‍‌​​‌​⁠‌‌‌‍​⁠‌‌​​‌​⁠⁠‌​‍​​⁠‌‌‌‌​‍​⁠‌​‌⁠‌​​‍​‍‌⁠⁠‌​

I’ve had CAP accept “20 patient pairs spread across the medical decision points” when I added a one-page sample distribution table and EP09 bias plots; I still do 40 if the AMR’s wide or matrix effects are likely — what analyte are you validating?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠​‌‍​‌‌‍​‍‌⁠‌​‌‍‌‌‌‍​⁠‌‍‍​‌‍⁠‍‌‍‍‌‌‍​⁠‌‍‍‌‌‍​‌‌‍⁠‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠​‍​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​⁠​⁠​‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‌⁠‌​‍​​‍⁠‌​⁠​⁠‌​‌⁠‌‍⁠‍‌⁠‍​‌‍‌​‌​‌‍‌​‌‌​⁠‌‌‌​⁠⁠‌‍‌⁠‌‌‍​‌‍‍⁠‌​‌‍​‍​‍‌⁠⁠‌​