Systematic Approach to QC Problems Clment Ars Sr

Systematic Approach to QC Problems Clément Arès Sr Product Manager Bio-Rad Laboratories (Canada) Quality Systems Division Clement_ares@bio-rad. com Manitoba Congress of Medical Laboratory Sciences October 3 rd, 2012 MCMLS - October 2012

Agenda • It’s out of Control!!! • Collect Information • Systematic Approach – – – Determine type of error Relate error to cause Common factors Relate to recent changes Verify the solution and document Assess effectiveness regularly • Resources MCMLS - October 2012

It’s Out of Control!!! • Assess the magnitude of the problem – Large or small variation are not treated the same • Make sure patient results since the last “in control” event can be retrieved if required • Determine if other means of evaluating data can provide information as to the acceptability of the patient results – Statistical Vs Clinical significance – Fixed mean Vs Peer group mean MCMLS - October 2012

Collect information • Out of control situations can be caused by : – Instrument … – Reagent … – Calibration … – Control (except if Bio-Rad!!!) … …unacceptable performance MCMLS - October 2012

Collect Information • When was the last time the QC was acceptable? • Did you introduce a new lot # of reagent? – Keep track by adding a comment or an action in URT • Did you introduce a new lot # of calibrator? – Keep track using the group function in URT • Was there a service call, maintenance or software update done? – Use the comment by instrument in URT to notify users • Are all levels affected? MCMLS - October 2012

Collect Information • How do you compare to peer group? – Check Instant. QC database • Did you run another control to confirm? – – Old Proficiency sample Patient pool or patient from previous run Old lot number of same control Different lot number of same control • Does the control perform the same on multiple analyzers or reading cells? MCMLS - October 2012

Collect Information • Contact control manufacturer technical support group to investigate if other labs have reported an issue for this analyte • Ask if there is a technical bulletin about this issue • Provide information collected (fax or email) to document issue and obtain help for resolution MCMLS - October 2012

Systematic Approach – – – Determine type or error Relate the type of error to possible causes Consider factors common on multiple test systems Relate the problem to recent changes Verify the solution and document the remedy Assess effectiveness of the quality system on a regular basis MCMLS - October 2012

Determine the types of errors – Random (sudden, unpredictable) • 1 -3 S, R 4 -S – Systematic (over time) • 2 -2 S, 4 -1 S and 10 -x • Shift • Trend MCMLS - October 2012

Causes of Random Errors – Random error are caused by: • • • Bubbles in reagent and reagent lines Inadequately mixed reagent Unstable temperature and incubation Unstable electrical supply Individual operator error in pipetting, timing etc… • Defective unit-test devices – « flyers » MCMLS - October 2012

Random error – Much more difficult to find the cause – Look at charts to see if trending or one time event – 3 suggestions • Inspect the instrument during operation • Perform a precision run using 10 replicates of the same patient sample and compare to IFU data • Run sample in duplicate MCMLS - October 2012

Causes of Systematic Errors – Systematic errors are caused by: • • • Change in reagent or calibrator lot Wrong calibrator value Improperly prepared reagents Deterioration of reagents or calibrators Change in sample or reagent volume due to pipettor misadjustment or alignment • Change in temperature of incubators and reaction block • Deterioration of light source • Change in procedure from one operator to the other MCMLS - October 2012

Systematic error - Shift – If sudden shift • Inspect reagent, calibration, maintenance records • Note recent actions – if the shift occurred immediately following a reagent replacement, verify that the lot number is correct and has been checked out or calibrated, and that the reagent has been prepared properly, that the reagent is indeed the correct reagent. MCMLS - October 2012

Systematic error - Trend – Systematic trend • Trends are caused by: – – Slowly deteriorating reagent Calibration shift Change in instrument temperature Deterioration of a filter or lamp • Review QC records before taking an action MCMLS - October 2012

Common factors on multitest systems – Only one test involved • Determine type of error and relate it to cause – Several tests involved • • Do all tests have small or large sample sizes? Do all tests use the same filter? Do all tests use the same lamp? Do all tests use the same mode of detection? (EIA, FPIA, Rate, End-point) • Do all tests have something in common: mechanical or optical MCMLS - October 2012

Relate the problem to recent event – When a problem arises, ask « What changed? » • Determine if the error is random or systematic • Determine if the is a trend or shift • Use systematic approach to isolate the cause – Change only one thing at a time – Document each action taken MCMLS - October 2012

Verify and Document – Retest the control to verify that the problem has been resolved – If QC is good, repeat patient samples from out-of-control run – Document the out-of-control event along with the corrective action MCMLS - October 2012

Assess effectiveness of your quality system on a regular basis – Use advanced software functions such a « Total Error- Biological Variation » to do a retrospective evaluation of your SPC rules and determine: • Frequency of statistical out-of-control events • Frequency of QC values outside the TEa limits during a period or across periods • The amount of biais present, if any MCMLS - October 2012

Assess effectiveness of your quality system on a regular basis – Are SPC rules too restrictive? – Can you improve error detection with more stringent single rule or complex multi-rule? – Should the mean be adjusted? – How much imprecision is there? Can it/Should it be improved? – How much bias is there? Can it/should it be improved? – Is the consensus group appropriate to determine bias? – Is the performance goal achievable or too lenient? – How frequently is the test being recalibrated? Does it exceed manufacturers recommendation? MCMLS - October 2012

Resources • www. youtube. com – BIORADQC Channel • • www. qcnet. com/ca-en/QC_Education www. westgard. com URT 2. 0 User Guide Articles by Dr C. Parvin et al. in Advance/Laboratory – Jan. 2011 MCMLS - October 2012

THANK YOU! MCMLS - October 2012