Questions

Repeat Patient QC was a well-attended webinar of 2023. But while the speakers answered dozens of questions live, there were dozens more that came in later. Here are the questions and answers that followed the webinar.

Questions from the Repeat Patient QC webinar

January 2023

Susan Daly, BSc (Hons), MSc
Kathleen Freeman, DVM, MS, PhD, DipECVCP, FRCPath, MRCVS
Bente Flatland, DVM, MS, DACVP (Clinical Pathology), Professor of Clinical Pathology, University of Tennessee, Knoxville, TN

We couldn't answer all the questions live, but we did look at all the questions submitted during the Repeat Patient QC webinar. Here are the answers to the most relevant questions.

Q1: How should we run 20 times of controls? Should we do it in 20 days? If we have some days off in between is it ok? Or should we do that 2 times in 10 days?

Frequency is up to you for data collection. Collecting over a period of time that reflects stable performance but also includes changes in calibration, reagens and operators makes sure that those sources of variation are covered. For the initial start up, you can speed things up by running more than one sample per day if you are getting sufficient samples into your lab daily.

Q2: Do you need to select normal and abnornmal patient samples?

I have used "normal" or as close to normal samples, I haven't yet validated using abnormal samples but would like to! This is possible by removing or adding plasma for hematology samples. Abnormal specimens for biochem or endocrinology may be from pooled specimens or spiked patient specimens.

Q3: How can we set Rpt-QC when we only couple ml of serum or plasma. It is not possible to get 20 replicates?

You are not running one patient sample 20 times. You are running 20 individual patient samples. See the chapter on RPT-QC in the Westgard. Basic QC Practices book, 4th edition.

Q4: If you're dealing with short specimens (say, from neonates), could you used pooled patient specimens for RPT QC?

Yes, you can pool specimens to achieve the volume needed.But be sure to run the pooled specimen at the start and after the specified time interval. Use your HCT as the target to achieve this. Need to avoid extremes of SDs/CVs! See answer above. 

Q5: For hematology,we need to do RPT-QC for one parameter or for all the cbc parameters to make sure all parameters are in control?

That is really up to the individual laboratory's needs and QC design. However, as a minimum, we suggest RBC, HCT, WBC, and PLT.

Q6: What is the maximum delay in performing repeat value of a patient testing for RPT-QC?

Based on studies done so far, 7 days for plasma chemistry and routine endocrinology testing and 4 days for hematology (CBC) testing.

Q7: We used pooled plasma for low level vitamin D, stability of result wasn't encouraging, and we stopped it. Do you think thr RPT_qc may work for certain meaurands?

Depends on the stability and time interval that you want to use. You EXPECT some deterioration of t he sample, but it needs to be measurable at the starting time and at the specified time interval for repeat measurement.

 Q8: Is this expected to replace commercial QC or be supplimentatl to it? How doe the requirement to have two different levels of QC affect RPT-QC and how it is used?

The projects we have done with RPT-QC so far have involved only normal or approximately normal samples (meaning, patient results for measurands of interest were within, or barely outside, institutional reference values). To enable human diagnostic laboratories to meet the regulatory requirement of running "normal" and "abnormal" QC, documentation that RPT-QC also works with abnormal samples (i.e., outside institutional reference values) is needed. Our suspicion is that this will work, but we have yet to document use of abnormal samples for RPT-QC in a scientific publication. RPT-QC could be used as your 'normal level' of control until 'abormal' levels have been validated.

Q9: Can you please repeat on how to calculate upper and lower RPT-QC limits?

The upper and lower RPT-QC limits are based on mean difference +/- SD-dup. (SD-dup is standard deviation of the duplicates). So for example, if you wish to apply a 1-3s control rule, the limits would be "mean difference +/- 3*SD-dup". If you wish to apply a 1-2.5s control rule, they would be "mean difference +/- 2.5*SD-dup".

Q10: Hi I am from The Gambia West Africa, I feel using this system would be useful for resource-limited laboratories In Africa. As using commercial haematology controls is a real challenge, mainly with stability.

Definitely would be a good choice for a resource-limited lab in Africa. As long as you are able to standardize storage conditions and have suitable specimen stability.

Q12: We have several analysers in our laboratory for the same measurand - do you have any experience of using RPT-QC over more than one analyser or would you advise RPT-QC for each analyser separately?

RPT-QC control limits must be calculated for each instrument individually (even if the instruments are all of the same make and model). The reason is that individual instruments (even ones of the same make and model) may have varying imprecision, and SD-dup (standard deviation of the duplicates) may vary from instrument to instrument.

Q13: Can we assume RPT-QC an extension of idea behind Moving Averages?

No. It is not related to moving averages. It is an application of SD of duplicate measurements.

Q14: How to decide on the frequency of repeat patient QC? Can it be run alongside the iqc?

Yes, RPT-QC can be conducted alongside iQC and other quality-monitoring procedures. RPT-QC should be done daily (just like QC based on commercial quality control materials) (daily meaning, each day the laboratory is open for business). Length of the repeat measurement interval (1-day repeat vs. 3-day repeat vs. 7-day repeat, for example) is dicated by known patient sample stability and laboratory logistics (e.g., is the laboratory open 7 days per week vs. closed on the weekends).