Hematology 425 Quality Assurance

1
Hematology 425 Quality Assurance

• Russ Morrison
• September 15 2006

2
Why Quality Assurance?

• Monitor and weigh the balance between quality and
  costs in health care
• Satisfy stringent regulatory controls
• Participation in federal funding programs
• Evaluation of clinical outcomes
• Reduce risk to the organization
• Risk of suboptimal outcomes
• Risk of litigation

3
Quality Control / Quality Assurance

• Both terms refer to the control of the testing
  process to ensure that test results meet
  requirements
• Diagnostic utility
• Therapeutic validity
• Quality Assurance is overall, coordinated effort
  to satisfy patients and care givers
• Quality Control is activity related to monitoring
  test systems and making corrections, as needed

4
Quality Control Definitions

• Accuracy exactness, measured by periodic
  proficiency testing challenges
• Precision reproducibility, indicator of random
  error, daily (or as required) quality control
  replicates
• Results that exhibit both precision and accuracy
  is the goal
• Avoid results that are precise, but not accurate
  (being precisely wrong)

5
Quality Control Definitions

• Delta Checks compare results from a prior sample
  analysis with the result of the same analyte
  using a current sample
• significant variation is user defined
• Technologist investigates (ie MCV)
• Reference Interval/Normal Range
• often refers to the acceptable range of values
  for the analyte in a healthy person, reference
  range is more appropriate
• Reference population used to establish reference
  range

6
Quality Control Definitions

• Reliability refers to the extent that a method
  can maintain both accuracy and precision over a
  period of time. Lab methods are chosen to have a
  high degree of reliability.
• Primary Standard is reference material of fixed
  and known composition, traceable to national
  performance characteristics
• Secondary Standard is reference material of
  composition traceable to a primary standard

7
Statistics in Quality Measures

• Normal Distribution Curve

8
Statistics in Quality Measures

• Coefficient of Variation (CV)
• is a measure of precision within a test system,
  the smaller the CV value, the more precise the
  analytic test system
• CV is calculated as 100SD/mean
• It is expressed as percent

9
Statistics in Quality Measures

• Diagnostic Sensitivity
• The proportion of patients with a disease who
  will have positive test results
• Defined by the ratio (a /(a b)) 100
• a is the number of true positive results and b is
  the number of false negative results
• THINK ABOUT THIS

10
Statistics in Quality Measures

• Diagnostic Specificity
• The proportion of patients who are correctly
  identified by the test as not having the disease
• Defined by the ratio (c /(c d)) 100
• c is the number of false positive results and d
  is the number of true negative results
• Highly sensitive tests are used to screen for a
  condition, highly specific tests are used to
  confirm the screen.

11
Types of Laboratory Error

• Systematic-errors within the test system or
  method (incorrect calibration, broken equipment,
  failure of the test process to perform accurately
  or precisely, 2 types)
• Constant systematic magnitude of error remains
  constant throughout the range, also called a
  constant bias
• Proportional systematic magnitude of error
  increases with the concentration of the substance
  being tested

12
Types of Laboratory Error

• Random errors that occur without predictability
  or regularity

13
Internal Quality Control

• Levey-Jennings Control Charts

14
Internal Quality Control

• Normal Levey-Jennings Control Chart
• Results randomly distributed about the mean over
  time
• Distribution as expected for normal distribution
  curve
• 1 in 20 points may be between 2 and 3 SD
  (caution), 1 in 100 above 3 SD (act)

15
Internal Quality Control

• Levey-Jennings Control Chart
• Shift is a drift in values, may be sudden or
  gradual
• Trend is continuous movement of values in one
  direction for 6 or more consecutive values
• A trend may be thought of as a gradual shift

16
Internal Quality Control

• Levey-Jennings Control Chart
• Shift is a drift in values, may be sudden or
  gradual
• Trend is continuous movement of values in one
  direction for 6 or more consecutive values
• A trend may be thought of as a gradual shift

17
Westgard Rules

1. 1 2s Rule one control value outside the 2s
   limit - review
2. 1 3s Rule one value is outside the 3s limit,
   random error, investigate
3. 2 2s Rule 2 consecutive values outside the same
   2s limit, investigate
4. R 4s Rule 2 consecutive values are more than 4s
   apart involving both control materials, 1 2s, 1
   -2S, investigate

18
Westgard Rules

• 4 1s Rule 4 consecutive values have been
  plotted on the same side of the 1s range, may
  indicate shift or trend
• 10x Rule 10 consecutive values fall on the same
  side of the mean either with same control or 2 or
  more controls, indicates shift.
• 1 3s and R 4s rule violations suggest random
  error while 2 2s, 4 1s and 10x rules suggest
  systematic errors within the test system.
• Computer programmable with warnings to
  technologists when violations occur.

19
Other Internal Error Monitors

• Moving averages of red blood cell indices (drifts
  of the moving average outside a defined limit
  suggest instrument error)
• Automated Differential Counting has improved the
  precision of differentials because it overcomes
  the non-random distribution of cells on a slide
  and classifies thousands of cells versus 100 to
  200 cells in a manual differential

20
External Quality Control

• Internal Controls monitor precision
• External Controls indicate accuracy
• Subscription program of qualitative and
  quantitative controls
• Analyzed by hundreds of peer labs using same
  equipment, same methods
• Evaluation is based on how close results compare
  with the mean opinion of the peer group

21
Technical Staffing Quality Control

• Active continuing education program maintains
  technical performance at a high level
• Imperative to optimal performance
• Required by regulatory agencies for accredited
  laboratories

22
Overarching Quality Assurance Plan

• Goal is to ensure that the quality of laboratory
  services contributes to the overall delivery of
  excellent medical care
• Focuses on high risk and problem prone areas of
  system performance
• Includes internal and external QC, staff
  proficiency, pre-analytic, analytic and
  post-analytic activity, computer functionality,
  critical values notification and customer
  satisfaction

23
QA/QC Group Question Exercise

• Proper specimen collection and patient
  satisfaction are components of
• 1. Quality control
• 2. Quality assurance
• 3. Reliability
• 4. Precision

24
QA/QC Group Question Exercise

• A patients white blood cells are counted on an
  automated cell counter ten times. The mean white
  count is 8000/µL and the standard deviation is
  300. What is the coefficient of variation (CV)?
• 1. .04
• 2. 2.6
• 3. 3.8
• 4. 26

25
QA/QC Group Question Exercise

• The CV in the above question indicates that the
  white blood counts on this cell counter are
• 1. Accurate
• 2. High quality
• 3. Properly calibrated
• 4. Precise

26
QA/QC Group Question Exercise

• What is the correct interpretation if a 10.0 g/dL
  hemoglobin standard reads 13.0 g/dL?
• 1. This is acceptable agreement
• 2. Imprecise
• 3. Insensitive
• 4. Inaccurate

27
QA/QC Group Question Exercise

• Interpret the following platelet counts from Mr.
  Billy Rueben.
• Day 1 Day 2 Day 3__
• 478 ? 109/L 163 ? 109/L 501 ? 109/L
• 1. Day 2 exceeds delta check limits
• 2. Day 1 and day 3 results are too high
• 3. Day 1 results were probably collected on the
  wrong patient
• 4. All are invalid

28
QA/QC Group Question Exercise

• Is the variation in the above platelet counts
  possible?
• 1. No
• 2. Yes, but unlikely
• 3. Yes, platelet counts typically vary this much
  from day to day
• 4. Yes, the patient probably received a platelet
  transfusion

29
QA/QC Group Question Exercise

• Which of the following is the most likely
  explanation for the platelet counts in Mr.
  Rueben?
• 1. Patient is developing thrombocythemia
• 2. The sample from day 2 contained clumped
  platelets from a traumatic phlebotomy
• 3. Patient received a platelet transfusion on day
  3
• 4. This is normal day-to-day platelet variation

30
QA/QC Group Question Exercise

• Which of the following is true regarding
  reference ranges?
• 1. Should be derived from reference books
• 2. Need be determined only for adults
• 3. Can be established by running the test
  procedure on 10 healthy people
• 4. Are ranges of values for an analyte in normal
  healthy people

31
QA/QC Group Question Exercise

• A test that is positive in all patients who have
  the disease but also in some who do not have the
  disease is
• 1. Sensitive
• 2. Specific
• 3. Accurate
• 4. Reliable

32

• The antinuclear antibody test (ANA) is positive
  in almost all people who have systemic lupus
  erythramatosus (SLE). It is also positive in some
  patients who do not have SLE. The anti-DNA is
  positive only in people with SLE but not in all
  that do. Which of the following is true?
• 1. ANA is a good screening test and anti-DNA a
  good confirmatory test
• 2. Anti-DNA is a good screening test and ANA a
  good confirmatory test
• 3. Both are good screening tests for SLE
• 4. Neither of these tests is valid

33
QA/QC Group Question Exercise

• A purchased hemoglobin standard is used to adjust
  a hemoglobinometer. This standard is being used
  as a
• 1. Control
• 2. Reference
• 3. Delta check
• 4. Calibrator

34
QA/QC Group Question Exercise

• The tubing that brings the lyse reagent to the
  hemoglobin cuvette on an automated cell counter
  is pinched and not delivering any reagent. All
  hemoglobin values are greater than 20 g/dL. This
  represents what type of error?
• 1. Random
• 2. Imprecision
• 3. Constant systematic
• 4. Proportional systematic

35
QA/QC Group Question Exercise

• On day 29 in a month one of the two controls was
  between 2 and 3 SD. What is the correct procedure
  to follow?
• 1. Stop running patients until an investigation
  is completed
• 2. Rerun the control and if acceptable continue
  with patients
• 3. Recalibrate the instrument
• 4. Open new vials of controls and repeat both
  controls

36
QA/QC Group Question Exercise

• The control values for the prothrombin test were
  ranging between the mean and minus one standard
  deviation. Starting on day 20 of the month the
  values consistently were between 1 and 2 SD.
  This is an example of a
• 1. Shift
• 2. Trend
• 3. Random error
• 4. Deviation

37
QA/QC Group Question Exercise

• Which of the following would most likely be
  associated with the above situation?
• 1. Operator error
• 2. Burned out light source
• 3. Instrument miscalibrated
• 4. Starting a new lot number of thromboplastin
  reagent

38
QA/QC Group Question Exercise

• Which of the following patients should not be
  included in indices moving averages?
• 1. Chemotherapy recipients
• 2. Infants
• 3. Obstetrical
• 4. Surgical

39
QA/QC Group Question Exercise

• A laboratory comparing its results to other labs
  is an example of
• 1. Precision monitoring
• 2. Internal QC
• 3. External QC
• 4. Delta checks

40
QA/QC Group Question Exercise

• A laboratory gets numerous complaints regarding
  the length of times it takes hematology results
  to get to the emergency room. What would be an
  appropriate response?
• 1. Make this a quality assurance project
• 2. Ignore the complaints
• 3. Explain why it takes so long
• 4. Tell the employees to work faster

41
QA/QC Group Question Exercise

• The best way to prevent errors in the laboratory
  is to assure
• 1. Quality instruments
• 2. Well trained, conscientious workers
• 3. Quality management
• 4. QC consistently being between 2 SD