WS IV - 3 Kasus Dari Dr. Tjan Sian - Six Sigma KOnas Case Study FINAL
Uploaded by
Mustakim DuharingWS IV - 3 Kasus Dari Dr. Tjan Sian - Six Sigma KOnas Case Study FINAL
Uploaded by
Mustakim DuharingShould Six Sigma be Calculated
for Internal Quality Control in Clinical Laboratory
CASE STUDY
Tjan Sian Hwa
1
• Choose TEA
• Calculate bias from 3-4 stable EQAS
• Peer group : same method, instruments, reagent
• Calculate CV/SD from 3-6 months IQC data
• Remove outliers
INACCURACY IMPRECISION
2
1
• EQAS
A. BIAS = ……
B. BIAS = …… %
• EQAS
A. BIAS = (126,04-122) = 4,04 .
B. BIAS = (126.04-122)/122*100 = 3,3%
IQC PEERGROUP
Sigma = ( TEA – Bias ) / S
PRESISI (CV, SD)
Presisi Hasil PMI :
• 3-6 bulan
• Menggambarkan berbagai lot reagen, kalibrator dan kondisi
Pada kadar normal dan abnormal
8
Sigma in L1 and L2 significantly differ
SIGMA 2,7 GOOD QC RULES :
SIGMA 5,76 ❖ High error detection
AVERAGE SIGMA : 2.7+5.76 = 4.2 ❖ Low false rejection
SIGMA CALCULATION
• MAKE IQC PLAN : Concentration, run, and
rules
• FOCUS ON PROBLEM TESTS
Alb
Ca
Mg
Case 2
2
GLUCOSE
IQC
Kit Insert
Range: 63.1-85.3 mg/dL
Mean : 74.2 mg/dL
SD : 3.7 mg/dL
13
Kit Insert
Range: 63.1-85.3 mg/dL
Mean : 74.2 mg/dL
SD : 3.7 mg/dL
Peer Group Range
Range: 67.5-80.0 mg/dL
Mean : 73.75 mg/dL
SD : 2.1
CV 2.8 %
Lab Range
Range: 69.6-79.9 mg/dL
Mean : 74.75 mg/dL
SD : 1.7
CV : 2.3 %
CLSI : LABORATORY SHOULD CALCULATE AND USE LAB IQC
RANGE
Peer Group Range Lab Range
Range: 67.5-80.0 mg/dL Range: 69.6-79.9 mg/dL
Mean : 73.75 mg/dL Mean : 74.75 mg/dL
SD : 2.1 SD : 1.7
CV 2.8 % CV : 2.3 %
BIAS : ( 74.75-73.75)/73.75*100% = 1.3%
SIGMA : (10-1.3)/2.3= 3.8
15
QGI: CV > →Imprecision → Random error
=Bias /(1,5*CV) Bias > → Inaccuracy → Systematic error
= 1.3/1.5*2.3 = 0,4
CAUSE OF RANDOM ERROR:
• Unstable Electricity, Grounding, Vibration,
QGI PROBLEM
• Pipetting ,
< 0.8 Imprecision
• Probe contamination
0.8-1.2 Imprecision and
Inaccuracy • Unstable reagent
> 1.2 Inaccuracy • Etc… etc
16
Sigma = ( TEA – Bias ) / S
LOW SIGMA
• Random error
• Systematic error
3
SGOT
LAB XXX
- Use 2 control level
- Calculate Lab own Mean and SD
- Monitor with LJ chart
- Use Westgard multi rules multi
levels
Problem with frequent IQC
rejection
18
EQAS
SGOT
Bias1 = 0,4
CV1 = 1,9 %
Sigma = ( 20-0,4)/1,9 = 10
Bias = 1,1
CV2 = 3,3%
Sigma = ( 20-1,1)/3,3 = 5,8
20
What do you think is the cause :
A. Inaccuracy
B. The EQAS result is just a coincidence
C. Imprecision
D. Lab use inappropriate rules
21
22
TAKE HOME MESSAGE :
• Do the right IQC right • Effective and efficient QC
• Use lab own mean management by having high error
• Choose the right EQA provider detection and low false IQC program
• Calculate sigma for lab
performance • Improve accuracy and patient safety
• Use the right rule • Cost saving
• Trouble shoot and solve the
problem
SERVE WITH QUALITY
1901
