Quality Assurance and Quality Control

Quality assurance (QA) and quality control (QC) together ensure that a clinical laboratory produces accurate, reproducible, and clinically actionable results. QA is system-level: the policies, procedures, training, audits, and proficiency testing that make quality possible. QC is per-run: the measurements that prove quality was actually achieved on a given day. Both are required by CLIA and CAP.

• QA vs QC distinction: QA prevents errors by designing the system; QC detects errors by measuring outputs. Document control, training, validation, and proficiency testing are QA. Daily controls, calibration verification, and Westgard rule checking are QC.
• Validation vs verification:
  • Validation is the formal performance characterization of a new or modified laboratory-developed test (LDT) before clinical use. Required parameters include accuracy, precision, analytical sensitivity, analytical specificity, reportable range, reference interval, and (for qualitative tests) limit of detection.
  • Verification is the (smaller) confirmation that an FDA-cleared/approved IVD performs as claimed by the manufacturer in the local laboratory. CLIA requires verification of accuracy, precision, reportable range, and reference interval before clinical use.
  • Any modification of an FDA-cleared assay (different specimen type, off-label use) converts it to an LDT requiring full validation.
• Proficiency testing (PT): external, blinded samples sent by an approved provider (CAP, CMS-approved alternatives) for the lab to test as if they were patient samples. CLIA requires enrollment in PT for regulated analytes with at least three challenges per year (commonly described as quarterly). For non-regulated analytes (which includes much of molecular and cytogenetic testing), labs must perform alternative assessment such as inter-laboratory sample exchange or split-sample comparisons at least twice per year.
• PT failures: an unsuccessful PT event triggers documented investigation, root cause analysis, and CAPA. Two consecutive or two of three unsuccessful events for the same analyte can result in CMS suspension of testing for that analyte.
• Internal vs external QC: internal QC = controls run within the lab on each batch (positive, negative, NTC, internal). External QC = proficiency testing and inter-lab comparisons.
• Document control: SOPs are version-controlled, dated, signed by the director, reviewed at least every two years (CAP), and accessible to bench staff. Obsolete versions are archived, not destroyed.
• Training and competency: each tester must be trained on each test before independent performance and competency-assessed at six months and annually thereafter, using all six CLIA-required elements (direct observation of patient testing, monitoring of result recording/reporting, review of records, direct observation of instrument maintenance and function checks, blind specimens, and problem-solving).
• Root cause analysis (RCA): structured investigation of a failure (PT miss, control out, wrong-patient result) using techniques such as 5 Whys or fishbone (Ishikawa) diagrams to identify systemic rather than individual causes.
• Corrective and preventive action (CAPA): corrective action addresses the immediate failure; preventive action changes the system so the failure cannot recur. CAPA records are reviewed at management review meetings and during inspections.
• Internal audits and management review: scheduled internal review of QA/QC indicators (turnaround time, amendment rate, specimen rejection rate, PT performance) feeds back into process improvement.

• A lab that develops its own NGS panel must complete a full validation (often 30-60+ specimens spanning expected variant types) before reporting clinically; switching to a new capture kit or library prep triggers revalidation or, at minimum, documented bridging studies.
• Failure on a CAP PT challenge for, say, CFTR variant detection can suspend the lab's authority to bill for that test; corrective action and successful re-testing are required for reinstatement.
• A wrong-patient report identified post-sign-out triggers an amended report, RCA, CAPA, and disclosure to the ordering provider; the underlying preanalytic process is then reviewed.