Tools, Technologies and Training for Healthcare Laboratories

QC - the Practice

This lesson provides a summary of the overall process of establishing and maintaining a statistical QC procedure. The objective of this lesson is to outline all the activities that are necessary -- QC Planning, QC Implementation, QC Operation, QC Documentation and Review -- without getting bogged down in the details of each of these activities. More detailed information is provided by links to other materials on this website.

PLEASE NOTE: An updated version of this article is available in Basic QC Practices, 2nd Edition

This lesson provides a summary of the overall process of establishing and maintaining a statistical QC procedure. The objective of this lesson is to outline all the activities that are necessary without getting bogged down in the details of each of these activities. More detailed information is provided by links to other materials on this website.

Purpose of QC

QC is intended to help people do good work by giving them a way to check that their work process is functioning properly. People need tools, such as statistical QC, to help them do this. Statistical QC provides a way of looking at the results of a work process and identifying when they exceed the variation expected under stable routine operation, in which case, it is likely that something has gone wrong and that the process needs to be fixed.

In a healthcare laboratory, known samples from a large number of bottles of stable control materials are analyzed to monitor the variation of a testing process. Results on these known samples are expected to fall within certain statistical limits, e.g., 95% within the mean plus or minus 2 standard deviations, 99.7% within the mean plus or minus 3 standard deviations. The means and standard deviations are calculated from laboratory measurements on these control materials, then control charts are constructed to display the variation of these known samples over time. Control limits are drawn to identify results that are unexpected and need to be investigated. [See QC - The Idea.]

QC Planning

The purpose of QC planning is to select a QC procedure that will assure the required quality at the minimum cost. Optimum cost-effectiveness depends on designing QC procedures on a test-by-test basis on each analytical system, taking into account the particular quality required for an individual test and the particular performance achieved with the analytical method on that system.

  • Define the quality required for the test on the basis of clinical quality needed for proper use and interpretation of test results, or on the basis of the analytical quality needed to satisfy regulatory requirements, such as the CLIA proficiency testing criteria for acceptable performance. Quality requirements in the form of medically important changes (or clinical decision intervals) and total analytical errors (allowable total errors) can readily be used with available QC planning tools. [See QC - The Planning Process].
  • Assess method performance from initial evaluation studies or from on-going QC results and proficiency testing surverys. Initially estimate bias from the comparison of methods evaluation experiment; estimate imprecision from a replication experiment over 10-20 days.
  • Utilize QC planning tools, such as OPSpecs charts, to identify the appropriate control rules and the number of control measurements (N). [See Mapping the Road to Analytical Quality with Charts of Operating Specifications]. In the absence of a quantitative planning process, be aware that the use of 2s control limits (i.e., control limits set as the mean plus or minus 2 standard deviations, or the 12s control rule) will cause a high false alarm or false rejection rate - approximately 5% for N=1, 9% for N=2, 14% for N=3, and 18% for N=4. Also be aware that the use of 3s control limits (or the 13s control rule) may not give sufficient error detection. Use of a 12.5s control rule or the 13s/22s/R4s or 13s/2of32s/R4s multirule procedures would be better to reduce the false rejections compared to a 12s control procedure and provide better error detection than a 13s control procedure. [See QC - The Chances of Rejection.]
  • Identify a Total QC strategy that places appropriate emphasis on statistical QC, non- statistical components (such as preventive maintenance, instrument function checks, performance validation tests, patient data QC, and quality improvement). When medically important error can be readily detected by statistical QC, rely on statistical QC and perform the minimum other QC as required by the manufacturer's instructions, government and accreditation requirements, and good laboratory practice. If errors cannot be readily detected, they must be prevented by frequent maintenance, instrument function checks, thorough operator training, operator experience, etc. [See Total Quality Control Strategies.]

QC Implementation

Because statistical QC is a quantitative technique, there are technical details that must be properly implemented, unlike some other aspects of quality management that may be more philisophical and less technical. Statistical QC won't work right and accomplish its intended purpose unless it is properly implemented.

  • Choose appropriate control materials. This includes the number of materials necessary to monitor the critical medical decision levels and working range of the method, as well as the type of material that will best simulate the true patient sample. It is common practice to use two levels of controls for many chemistry tests and three levels for hematology and coagulation tests. [See QC - The Materials, see also Medical Decision Levels.]
  • Analyze the control materials to obtain a minimum of 20 measurements over at least a 10 day period. It is important to collect measurements that will characterize the actual performance of the method under each laboratory's own operating conditions. Ten days is a minimum period to observe factors that affect method performance; 20 days is better.
  • Calculate the mean and standard deviation of the control measurements. While the number of measurements may be minimal in the beginning, the mean and standard deviation can be updated as more measurements are accumulated. [See QC - The Calculations.]
  • Calculate the control limits for the control rules that are to be applied. Contrary to some manufacturers' instructions and some laboratory practices, it is not advisable to use the bottle values appearing on control materials to set statistical control limits. Control limits should be calculated from the mean and standard deviation that were determined when that control material was analyzed in the laboratory. [See QC - The Levey Jennings Control Chart].
  • Prepare control charts or set the appropriate parameters in a computerized monitoring and charting package. A sheet of graph paper is all that's needed for manual implementation, however, computerized implementation is often necessary for high speed multi-test automated analyzers that produce many control results in a short period of time. This computer support may be provided by software in the analyzer, by a PC data management station, or by a laboratory information system.
  • Prepare written guidelines to document the QC procedure. Written protocols are required by accreditation and regulatory agencies, as well as for good laboratory practice. An important part of this protocol is to describe when control samples are to be analyzed, how many samples are to be analyzed, where they are to be located in an analytical run, how the control results are to be interpreted, and what to do when "out-of-control."
  • Teach the QC procedure to ALL personnel who will be performing the test. In point- of-care applications where some personnel perform the test infrequently, statistical QC is particularly important for documenting operator proficiency.

Routine QC Operation

Routine operation depends on obtaining current control results and using them to determine whether the testing process is performing as expected. It is "expected" that the current control results fall within the established control limits if the testing process is working okay. It is unexpected for the control results to exceed a control limit or violate a control rule unless there is a problem with the testing process.

  • Analyze control materials with each analytical run. According to US CLIA regulations, at least two different control materials should be analyzed with each run. According to US CLIA regulations, at least two different control materials should be analyzed with each run [See QC - The Regulations.] Note that the run is defined by the written QC guidelines that were developed above. A run is typically defined in terms of a length of time, a number of samples analyzed, or a physical grouping (or batch) of samples. The definition of a run is subjective and may change as experience is gained and more information becomes available about a method's stability or frequency of problems.
  • Record the control results and plot on control charts (as necessary). According to the government, if you didn't write it down, you didn't do it. More important, documentation and records are an important source of information about the actual performance of the method under the real operating conditions of a laboratory.
  • Review and interpret the control results to determine control status. Here's where the decision criteria or control rules are important for achieving uniform interpretation regardless of the experience of the person performing the test. [See QC - The Westgard Rules.]
  • When control results are "in-control," i.e., none of the control rules are violated, accept the run and report the patient test results. Our philosophy is that each analyst should have sufficient training and experience to do this on their own, however, there may be situations where supervisory review is necessary before patient test results can be released. [See QC - The Multirule Interpretation.]
  • When control results are "out-of-control," i.e., one of the control rules is violated, reject the run and do not report patient test results. Inspect the process, identify the source of difficulty, correct the problem, then reanalyze the patients and controls. Note that the particular control rule that is violated may be helpful in trouble-shooting; the 13s and R4s rules tend to be more sensitive to random errors, whereas rules such as 22s and 2of32s tend to be more sensitive to systematic errors. Random and systematic errors have different causes, thus the problem can be narrowed down to fewer possible sources. Once a problem is fixed, it may be best to reanalyze the controls first to determine control status, then reassay the patient samples. [See QC - The Out-of-control Problem.]
  • Document any control problems and what was done to solve them. For unusual problems, it may be useful to file a detailed trouble-shooting report for future reference and use in training new personnel.

QC Documentation and Review

Because many factors are involved in maintaining quality, a system of records and documentation is needed for periodic review and evaluation.

  • Maintain QC records for an appropriate period of time to document routine maintenance, unscheduled maintenance, reagent lot numbers in use and expiration dates, calibration records that include calibrator lot numbers and expiration dates, control results and summary statistics (monthly means and standard deviations, as well as cumulative means, standard deviations, and control limits in use), QC problems and corrective actions taken, and trouble-shooting reports. Most records need to be kept for two years according to US regulations, except that maintenance records must be kept for the lifetime of an instrument system and transferred with the instrument. [See QC - The Records.]
  • Review monthly means for trends and small shifts that indicate systematic errors or potential accuracy problems; review monthly standard deviations for changes in random errors or imprecision.
  • Correlate these QC changes with other performance data, such as results from proficiency testing surveys, comparison of methods results with real patient samples, changes in reagents, and changes in calibrations. Investigate problems and recommend corrective actions, such as recalibration, increased maintenance, and additional operator training. Make improvements in the testing process when possible.
  • Periodically assess whether the QC design is still appropriate based on the routine performance being observed.

Who's responsible?

QC planning should be the responsiblity of laboratory management, usually the director, manager, or quality specialist. The medical director has a critical role in defining the quality requirements for the laboratory. The actual QC planning function may then be delegated to a manager or quality specialist. Implementation is usually delegated to supervisors and technologists who are in charge of managing specified analytical systems and testing processes. Routine operation is delegated to everyone who performs a laboratory test.

In large laboratories, there may be several quality specialists who spend a large part of their time dealing with quality systems. In small laboratories, the most senior technologist may inherit the responsibilities for quality. In point-of-care situations, the central laboratory may be responsible for the implementation, training, and maintenance of quality systems, but each healthcare worker who performs a test should be accountable for routine QC operations.