CLIA Proficiency Testing Rules for Small Labs
Proficiency testing is the CLIA duty surveyors cite most at condition level. This guide covers which of your tests need it, how to enroll, how to run the samples without breaking the referral rule, and what happens after an unsatisfactory or unsuccessful event, with the limits in force since July 2024. To score an event as you read, open the PT score calculator.

CLIA proficiency testing on one card
738 labs
cited at condition level for unsuccessful PT (D2016, § 493.803) in 2024, the most cited condition
548 labs
cited for not checking the accuracy of tests without PT twice a year (D5217, § 493.1236(c)(1))
Proficiency testing is the outside exam for a lab's nonwaived menu: an approved program mails unknown samples, the lab tests them like patient specimens, and the program grades the answers. Here is the whole rule in one tray, then each part in the order a small lab meets it.
The rule at a glance
- Testing events a year, per regulated analyteat least 3
- Samples per testing eventat least 5
- Passing score, most analytes and the whole event80%
- Passing score, ABO group and D (Rho) typing, unexpected antibody detection, compatibility testing100%
- Unsuccessful: unsatisfactory in two consecutive, or two of three consecutive, events2 of 3
- Stay with a program for an analyte before switching1 year
- Keep PT records and the signed attestation2 years
- Accuracy check for tests with no PTtwice a year
- Current acceptance limits in force sinceJuly 11, 2024
Which tests need PT, and which do not
Only the analytes listed in Subpart I need graded PT. CMS calls them “regulated” analytes: “PT is only required for the limited number of tests found in Subpart I”, and you must enroll “and perform PT on each of the regulated analytes your laboratory tests” (CMS PT brochure). The list runs by specialty: routine chemistry from ALT to uric acid in § 493.931, hematology and coagulation in § 493.941, and lists for endocrinology, toxicology, general immunology, the blood bank tests and the microbiology subspecialties in the sections around them.
Everything else falls into one of four groups:
- Waived tests: no PT. CMS's answer to “Is PT required if I only perform waived tests?” is “No”. Check a test's category in the FDA CLIA database. A lab that enrolls its waived tests anyway is bound by the referral rules for them.
- A second analyzer for the same test: no PT on it. “PT is required for only the test system, assay, or examination used as the primary method for patient testing during the PT event” (§ 493.801(b)(7)). The backup is tied to the primary by the twice-a-year comparison in § 493.1281(a).
- Nonwaived tests not on the list, such as urine drug screens (toxicology PT lists serum, plasma or blood drug levels only) or the nine PPM procedures: the twice-a-year accuracy check below.
- A listed analyte no approved program offers samples for on your method: the same twice-a-year check (§ 493.1236(c)(2)).
Takeaway
Tests with no PT program: the twice-a-year check
“At least twice annually, the laboratory must verify the accuracy of” every test it performs that is not in Subpart I, and every listed test with no compatible PT samples (§ 493.1236(c)). “All proficiency testing evaluation and verification activities must be documented” (§ 493.1236(d)).
CMS names two ways to do it:
- Split a patient specimen with another lab that runs the same test, and have your director review both sets of results for acceptability. Split a patient specimen only, never a PT sample.
- Enroll in an approved program for the test even though PT is not required. Once you do, every PT referral rule applies to it, “regardless of certificate type or whether the analyte is listed in Subpart I”.
Write your acceptance criteria before you test, spread the two checks across the year, and file them with your PT records. Surveyors do look: D5217 (§ 493.1236(c)(1)) was cited at 548 labs in 2024, sixth on CMS's standard-level top-ten list. Our inspection checklist carries it as its own line.
Ungraded PT is part of the same duty
When a program cannot grade a challenge, or you score 0 for missing or late results, the score “does not reflect laboratory test performance” and you “must verify the accuracy” of that analyte yourself (§ 493.1236(b)(2)). CMS warns that the program may print an “artificial score of 100%” for a non-consensus sample.
How to enroll
You enroll with a program, not with CMS. The steps in CMS's PT brochure are three: look up the approved programs, choose one (or more) that offers the tests you run, and call or email it. On CMS's 2026 list there are seven approved programs; our list of CMS-approved PT programs shows which subspecialties each covers, and CAP vs API proficiency testing compares four of them (CAP, API, AAB-MLE and WSLH PT) on schedules, result entry and waived menus.
- Cover every regulated analyte on your menu, for each specialty and subspecialty on your certificate. If you use more than one program, designate which one counts for each analyte (§ 493.801(a)(2)).
- Let the program report to CMS. The program notifies CMS of your enrollment, and you authorize it to release your results to HHS (§ 493.801(a)(4)).
- Stay for a year. You participate in a program for an analyte “for one year before designating a different program” and notify CMS before any change (§ 493.801(a)(3)). CMS's brochure reads this as one calendar year.
- Enroll as soon as possible when you start under a new certificate or add a regulated test mid-year, then complete the program for the rest of that calendar year.
PT runs per CLIA certificate. Several sites under one certificate (the multiple-site exception) enroll once and rotate events between the sites; CMS says never to order the same event for several sites or instruments under one certificate. Sites with their own certificates enroll separately, and CMS suggests a different program for each to avoid any appearance of referral.
Accredited labs are not excused. A lab meeting CLIA through accreditation must “Meet the proficiency testing (PT) requirements in subpart H” and authorize its program to send results to its accrediting organization (§ 493.551(b)(3), (4)); the accreditor may add rules of its own (see CAP vs COLA).

Testing the samples like a patient
Subpart H asks for one thing above all: the PT sample goes through the lab exactly as a patient specimen would.
- Tested “with the laboratory's regular patient workload by personnel who routinely perform the testing”, with routine methods. The analyst and the director sign an attestation saying so (§ 493.801(b)(1)).
- Tested the same number of times you test patients: no duplicates if patients get singles ((b)(2)).
- No talk about results with another lab, or between your own sites, until after the program's reporting deadline ((b)(4)).
- Every step documented, from handling to reporting, and kept with the program's result form and signed attestation “for a minimum of two years from the date of the proficiency testing event” ((b)(6)).
CMS's list of what that file holds: preparation instructions, every testing step, instrument printouts and raw data, the submitted result form with the attestation, a print screen when results were keyed online, the program's evaluation, and any corrective action. CMS also suggests rotating PT among your testing staff, and it counts a transcription slip as a wrong answer, so a second person should check the form before it goes.
Takeaway
The no-referral rule
The laboratory must not send proficiency testing samples or portions of proficiency testing samples to another laboratory for any analysis for which it is certified to perform in its own laboratory.
This rule carries the heaviest penalties in Subpart H. A lab CMS finds intentionally referred a PT sample “may have its certification revoked for at least 1 year”. Where a patient specimen would go out for reflex, confirmatory or distributive testing, you test the PT sample “up until the point it would refer a patient specimen” and stop. CMS's brochure tells you to mark those lines “Would refer” or “Test not performed” on the result form.
CMS lists what can follow a referral:
- Revocation, suspension or limitation of the lab's CLIA certificate.
- The director losing the ability to direct a laboratory for two years.
- The owner or operator losing the right to own or operate a laboratory for two years.
- A directed plan of correction and/or a civil money penalty, and the lab's name on CMS's Laboratory Registry.
If another lab sends you its PT sample
Do not test it. Tell CMS, your State Agency or your accrediting organization which lab sent it and what it asked for. The regulation requires the notice “regardless of whether the referral was made for reflex or confirmatory testing, or any other reason” (§ 493.801(b)(5)).
Train every person who opens the mail on this. A courier routine that sends all send-outs to the reference lab will send the PT vial with them unless the kit is pulled at the door.
How an event is graded
For each sample the program compares your answer with the target, and a quantitative answer is acceptable when it falls inside the analyte's acceptance limit around it. A sample can only be graded when enough labs agree: “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 sample is left ungraded (§ 493.931(c)(1)).
The analyte score is the share of its challenges that were acceptable; the event score is the share across all challenges. “Failure to attain a score of at least 80 percent of acceptable responses for each analyte in each testing event is unsatisfactory analyte performance” (§ 493.841(a)), and the overall event needs 80% too. ABO group and D (Rho) typing, unexpected antibody detection and compatibility testing need 100%; antibody identification needs 80%.
With five samples, one miss is 80% and a pass; two is 60% and a fail. The PT score calculator does the sums for a whole event, flags which analytes are one miss away from unsuccessful, and checks a single result against its limit.
Read every evaluation, passes included. CMS's example: “if you receive 80%, you should investigate why one of the five samples was outside the acceptable range” and document what you found.

After an unsatisfactory event
“Unsatisfactory proficiency testing performance means failure to attain the minimum satisfactory score for an analyte, test, subspecialty, or specialty for a testing event” (§ 493.2). Missing an event or returning results late counts: either scores 0.
The regulation asks for two things. Remedial action, “taken and documented”, with the documentation kept for two years; and, unless the failure was a missed event, “appropriate training” and “the technical assistance necessary to correct problems” (§ 493.841(e)). CMS's brochure turns that into a working order:
- Re-check the submitted form for clerical errors. A transcription slip is a wrong result, but it points to a different fix.
- Compare your answers with the program's results for your peer group, with the director and the analyst who ran the samples.
- Look at the test system: QC results, calibration, maintenance, and the manufacturer if needed. A Levey-Jennings chart of the weeks around the event, read with the Westgard rules, shows whether a shift was already building. The QC statistics calculator gives the mean, SD and CV to set beside your peer group, and the rules for how often to run QC set the floor.
- Look back at patient results from the period the PT covered and identify any that may have been affected.
- Write it up: cause, fix, patient review, and the evidence it worked. File it with the event.
Takeaway
After an unsuccessful 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.
Successful participation is a condition of certification (§ 493.803), and failing it brings sanctions under Subpart R. The first time, CMS “may direct the laboratory to undertake training of its personnel or to obtain technical assistance, or both” instead, unless there is immediate jeopardy, the lab cannot show it has corrected the problem, or it has a poor compliance history.
You may choose to stop testing the failed analyte. CMS says to tell your State Agency before a cease-testing letter arrives. To come back, the lab shows it found and corrected the cause, then passes two consecutive events on reinstatement samples, which it can buy from any approved program once the fix is in. Under § 493.807, a lab suspended or limited for PT failure needs “sustained satisfactory performance on two consecutive proficiency testing events”, and the suspension runs “not less than six months”.
There is a bill as well. CMS charges unsuccessful PT desk reviews at “the average State Agency hourly rate multiplied by the number of hours (including travel time)”, $111.81 an hour for FY 2024 on its fee schedule.
The most cited condition
D2016, successful participation under § 493.803, was the most cited condition-level deficiency in CMS's 2024 figures, at 738 labs. It is also one of the four conditions a surveyor must cite whenever non-compliance is found, harm or no harm, along with PT enrollment and PT referral (SOM Chapter 6, § 6130.5). Our guide to the CLIA inspection covers what else they pull.
The 2022 rule: what changed in July 2024
The final rule “CLIA Proficiency Testing Regulations Related to Analytes and Acceptable Performance” (87 FR 41194, July 11, 2022) took effect August 10, 2022, except the changes to § 493.2 and §§ 493.801–493.959, the new analytes and acceptance limits, which took effect July 11, 2024. Any evaluation you get today is graded on them.
It added 29 analytes and deleted five: LDH isoenzymes and four drugs, ethosuximide, quinidine, primidone and procainamide (with its metabolite N-acetyl procainamide). Among the additions a small chemistry lab notices first are hemoglobin A1c, troponin I and T, BNP and proBNP, PSA, ferritin and LDL by direct measurement. It also required programs to try both referee and participant agreement before leaving a sample ungraded, and removed the option of sending PT samples to HHS for on-site testing.
Many limits tightened. A sample, quoted as each edition of the regulation prints it:
| Analyte | Until July 10, 2024 | Since July 11, 2024 |
|---|---|---|
| Glucose | Target value ±6 mg/dl or ±10% (greater) | Target value ±8% or ±6 mg/dL (greater) |
| Potassium | Target value ±0.5 mmol/L | Target value ±0.3 mmol/L |
| Creatinine | Target value ±0.3 mg/dL or ±15% (greater) | Target value ±10% or ±0.2 mg/dL (greater) |
| Cholesterol, HDL | Target value ±30% | Target value ±20% or ±6 mg/dL (greater) |
| Hemoglobin A1c | not listed | Target value ±8% |
| Sodium | Target value ±4 mmol/L | Target value ±4 mmol/L |
| Hemoglobin | Target ±7% | Target ±4% |
| Hematocrit | Target ±6% | Target ±4% |
| Leukocyte count | Target ±15% | Target ±10% |
| Cell identification | 90% or greater consensus on identification | 80% or greater consensus on identification |
The rest are in § 493.931 (routine chemistry) and § 493.941 (hematology), with the other specialties in the sections around them. The PT score calculator checks one result against a handful of them.
A 2026 proposal, not yet law
CMS's proposed rule of October 8, 2026 (CMS-3478-P, 91 FR 64480), would among other things allow digital images in gynecologic cytology PT. Comments are due December 7, 2026. Nothing in it changes the PT rules above until a final rule says so.
Takeaway