Statistics & QC_Clinical Pathology Boards Review

  1. In Gaussian distributions, what % falls within 1SD, 2SD, and 3SD?
    • 1SD = 68%
    • 2SD = 95.5%
    • 3SD = 99.7%
  2. Calculate the SD for the following data set:
    SD = 1.41
  3. What is coefficient of variation? What is an ideal range for the CV of a typical run?
    • Way to describe the precision of a run or the variability around the mean, expressed as a percentage of the mean:
    • CV = (SD/mean) x 100
    • Ideally CV between 1-10%
  4. If a reference interval is set at ±2 SD, what % of patients would fall outside this ref interval?
    4.5% since 2SD = 95.5%
  5. What does partitioning mean in terms of reference intervals?
    It's setting a ref interval for a particular population
  6. Alkaline tide is a physiologic variation that can cause a patient's result to fall outside the reference interval. What is the cause?
    HCl secretion after a meal → transient alkalosis
  7. What lab value is 5x the upper limit of normal in adolescents?
    Alkaline phosphatase
  8. All of the following are physiologic deviations from normal reference intervals that can be seen in elderly patients except:
    A) ↓Cre clearance
    B) ↓Glu tolerance
    C) ↓ releasing hormones (TRH, ACTH)
    D) ↑ lipids
    C - these go UP
  9. What two lab values are higher in males vs. females?
    CK and LDL cholesterol
  10. A Muslim woman comes in to get labs drawn. Her chem panel shows elevated ketones and bilirubin as well as low albumin, K+, and Mg. Should you be concerned?
    No, patient likely had prolonged fasting
  11. All of the following can be seen with exercise except:
    A) ↑LDH
    B) ↑CK
    C) ↑AST
    D) ↓neutrophils
    D - neutrophils go up from demargination
  12. What is the effect of going from supine to sitting or standing on lab values?
    Proteins, substances bound to proteins, and formed elements (RBCs) increase relative to supine b/c hydrostatic forces increase, forcing water and small molecules out of circulation but leaving big molecules behind
  13. How does prolonged tourniquet use affect lab values? Fist clenching?
    K+, NH3, lactate values go up due to hemoconcentration; increased K+ from release due to contraction of forearm muscles
  14. What should be the order of draws for the following tubes:
    1) clx
    2) Blue citrate tube/Coag
    3) Red/no additive for chem, BB, and serology
    4) serum separator tubes/blood serum for therapeutic drug monitoring
    5) Lavender/EDTA tubes for cell counts
    6) Green/heparin for plasma chem
    • 1 → 3 → 2 → 4 → 5 & 6 (additives last)
    • *Coag is before tubes with additives to prevent activation of coagulation
  15. All of the following cause systematic analytical errors except:
    A) Constant bias (shows similar slope, diff intercepts on correl study)
    B) Interferences
    C) Reagent carryover
    D) Proportional bias (shows diff slope & intercept on correl study)
    E) Sample carryover
    B - this causes random errors
  16. Random errors in lab medicine are mainly caused by what two factors?
    interferences (heterophile Abs, solvent exclusion effect) and technologist factors
  17. If an EDTA tube is accidentally used for chem studies, what lab values are affected?
    ↑K+ and ↓Ca, Mg, CK, Alk phos because EDTA tubes are made of K-EDTA salts so K+ goes up and EDTA chelates the rest
  18. What can happen if serum separator tubes are accidentally used for therapeutic drug monitoring?
    artificial ↓drug levels from adsorption by gel - SST tubes are generally used for serum tests like virology or chemistry
  19. All of the following can cause an artificial ↑K+ except:

    • G)
    • Here are the reasons for the rest:
    • B - glucose used up by RBCs, Na/K ATP pump gives up so K+ leaks out of dying cells
    • C - have muscle contraction and hemoconcentration respectively
    • G - EDTA tube has K-EDTA salt so K+ goes up
  20. What is solvent exclusion effect? What types of analyzers are affected?
    Solid phase of blood sample unusually high (ex. hyperlipidemia or paraproteins) → falsely low analyte concentration; indirect ISE analyzers
  21. What type of distribution is required in a T-test?
  22. What is the calculation for sensitivity, specificity, PPV, NPV, and efficiency?
    • Sensitivity = TP/(FN+TP)
    • Specificity = TN/(TN+TP)
    • PPV = TP/TP+FP
    • NPV = TN/TN+FN
    • Efficiency = TP+TN/All results (total all boxes)
  23. If a population dz frequency is 1 in 1000, and a test has a sensitivity of 85% and specificity of 95%, what is the PPV, NPV, and efficiency?
    • PPV = 85/85+5000 = 1.67%
    • NPV = 95,000/95,015 = 99.98%
    • Efficiency = 95,085/100,100 = 94.99%
  24. In blood draws, what do the blue, green, yellow, black, red, lavender, and gray topped tubes represent? What are they used for?
    Image Upload 1
  25. Image Upload 2
    In the curve above, what do points a, b, and c represent?
    • Image Upload 3
    • A: High sensitivity, low specificity (more ppl without dz will test +)
    • C: Low sensitivity, high specificity (more ppl with dz will test -)
    • B: Ideal but depends on what you want to accomplish with test
  26. Define bias. What is an accurate result? What is a positive and negative bias?
    Bias = True result - analytical result; an accurate result means the bias (difference btw true result & analytical result) does not exceed acceptable limits; positive bias means measured result > true result (over-Dx conditions), negative bias means measured result < true result (under-Dx conditions)
  27. What is calibration? What is it used for?
    Scaling the physical measurement with a known concentration; can use multiple calibration measurements to generate a std curve, then compare the signal of an unknown to signals from the knowns
  28. What is the 1/3 rule in CAP accreditation?
    If your lab's SD is <33% of SD allowed, good chance of passing
  29. How often do labs receive unknowns from CAP for proficiency testing? What is a satisfactory result for lab accreditation by CLIA '88? What must the lab do if it fails?
    3x/yr in batches of 5 samples; ≥4/5 (80%); investigate all unacceptable results, all failed proficiencies, and internal report signed off by staff & med director
  30. In CAP proficiency testing, how are participants graded?
    CAP calculates the std deviation index for each participant, then graded as acceptable (<2SD), needs improvement (<2-3SD) or unacceptable (<3SD) from the mean & SD calculated from all results from labs using same method (peer group)
  31. How is std deviation index calculated?
    SDI = (lab's result - mean of peer group)/SD for peer group
  32. What does an accurate result depend on in a lab setting?
    Calibration & Precision (SD or CV)
  33. How is imprecision calculated?
    • Add up all intra- (same run, same day) and inter-assay (diff runs, diff days) variance
    • √(intra-assay + inter-assay variance)
  34. How can you statistically compare the precision of two analyzers using the same analyte? What type of distribution is req'd?
    Do F-test; Gaussian
  35. What is QC and QA? What regulations enforce QC?
    • QC - part of quality assurance, system of rules to detect analytical errors that could lead to incorrect clinical interpretation of results; enforced by CLIA
    • QA - program of regular assessments of lab activity
  36. What four categories of lab tests are stipulated by CLIA?
    • 1) waived tests - follow mfr and/or good lab practices (U/A, FOBT, urine preg test)
    • 2) provider performed microscopy (PPM) - follow good lab practices (wet mounts, KOH, ferning tests)
    • 3) moderate complexity - have SOP, follow mfr directions, calibrate & run controls (most tests)
    • 4) high complexity - same as mod complex but also require validating procedure (tests dev'd in lab)
  37. T/F: The Joint commission and CAP have defined sets of rules for QC in labs.
    False - they don't tell you what rules have to be used, you just have to have rules and use them consistently
  38. When determining if an assay is working properly, what set of rules are followed when deciding if a control is within acceptable parameters? What if a control sample falls "out of control"?
    Westgard Rules; you can't report pt results
  39. Which of the Westgard Rules are due to random error? Systematic errors?
    • Random error:
    • 1) 1:3s (1 value ±3SD from mean)
    • 2) R:4s (2 seq >4SD away from each other)
    • Systematic error:
    • 1) 2:2s (2 seq >2SD on same side of mean)
    • 2) 4:1s (4 seq >1SD on same side of mean)
    • 3) 10:X (10 seq values same side of mean)
  40. Which Westgard rule is merely a warning and does not require rejection?
    1:2s (1 value ±2SD from mean)
Card Set
Statistics & QC_Clinical Pathology Boards Review
Statistics & QC Clinical Pathology Boards Review