CAP vs API Proficiency Testing

API's 2026 catalog notes that its second Chemistry-Core event moved two weeks, to May 18–20, with results due June 10. CAP lists each program's shipments on its own page in its online store. Either way, put the ship dates and due dates on the lab calendar the day the order is confirmed: a missed due date scores zero.

The waived lines show something easy to miss. CAP's waived A1c program ships three challenges twice a year, not five three times, because CLIA does not require PT for waived methods. If the same analyte runs on a nonwaived analyzer anywhere in the lab, that method needs the full five-sample, three-event program.

The pass mark is written into each specialty's section, in the same words: "Failure to attain a score of at least 80 percent of acceptable responses for each analyte in each testing event is unsatisfactory analyte performance" (routine chemistry § 493.841, hematology § 493.851, and endocrinology, toxicology and general immunology alike), and the overall event score must also reach 80%. ABO group and D (Rho) typing, unexpected antibody detection and compatibility testing need 100%; antibody identification needs 80%. Check an event score with the PT score calculator.

Two words in every evaluation report are defined in § 493.2. "Unsatisfactory proficiency testing performance means failure to attain the minimum satisfactory score for an analyte, test, subspecialty, or specialty for a testing event." "Unsuccessful proficiency testing performance means a failure to attain the minimum satisfactory score for an analyte, test, subspecialty, or specialty for two consecutive or two of three consecutive testing events." Not taking part, or returning results late, scores zero. More in CLIA proficiency testing, explained.

Unsuccessful is the one that brings sanctions. For a first unsuccessful performance, CMS may direct training or technical help instead, unless there is immediate jeopardy, no evidence of correction, or a poor compliance history (§ 493.803(c)). Nor can a lab hop between programs: it stays in one for an analyte for a year and tells CMS before any change (§ 493.801(a)(3)).

A program can only grade a sample when participants agree on the answer: CLIA asks for agreement of "80 percent or more of 10 or more referee laboratories or 80 percent or more of all participating laboratories", and both must be tried before a program may leave a sample ungraded (§ 493.931(c)(1)). Small method groups miss that bar too, which is why a lab on an uncommon analyzer sees more ungraded results.

An ungraded result is not a free pass. When a score "does not reflect laboratory test performance", because the program could not reach agreement or the lab scored zero for missing or late results, the lab "must verify the accuracy" of that analyte itself and document it (§ 493.1236(b) and (d)).

Two programs spell out how they mark these. AAB-MLE's FAQ uses "#" for a result not graded because too few labs used your method and "?" for a lack of participant consensus; both carry a default score of 100, and AAB-MLE says a documented QA/QC review is the minimum follow-up. WSLH PT's post-event guide lists "Not scored, insufficient peer group" and "Non-consensus: self-assessment needed", each with the review it expects. CAP's PT FAQ explains two codes: 26 marks an educational challenge that is not graded by design, to be self-evaluated against the participant summary report, and 27 marks a challenge left ungraded because fewer than 80% of participants or referees agreed. We did not find the same explanation on the API pages we read; the legend on your own evaluation report is the place to look.

The twice-a-year check comes from § 493.1236(c), which covers any test "not included in subpart I" and any listed test for which no approved program sends compatible samples. None of the nine PPM procedures in § 493.19 is a Subpart I analyte, and a PT program's PPM package is one way to document that check.

One trap for mixed labs: if a moderate-complexity lab sends its waived tests through PT anyway, § 493.20(c) still applies the handling rules of § 493.801(b)(1)–(6) to them. Test the samples like patient specimens, by the people who run them, with no talk between labs before the due date and no referral.

The other trap is a method change. AAB-MLE's catalog warns that if a lab moves a regulated analyte from a nonwaived to a waived method mid-year, it must keep the five-challenge program for any nonwaived tests left in that subspecialty.