Diagnosis

1
Diagnosis

• Computation of Post Test Probability

2
Sensitivity and Specificity
Test results True Status(Golden Standard) True Status(Golden Standard) True Status(Golden Standard)
Test results Diseased Not Diseased Total
Positive a b a b
Negative c d c d
Total a c b d a b c d
3
Using Sen. and Spec. to Revise Probability
We often use maternal serum a-fetoprotein (MSAFP)
test to screen the fetal abnormalities. Pretest
Probability P(D) 0.20 Sensitivity
P(TD)0.34 Specificity P(T-D-)0.86 ?PV? ?
PV-?
4
Methods to Revise Probability

• 2 by 2 Table Methods
• Decision Tree Methods
• Bayesian Methods
• Likelihood Ratio Methods

5
The 2 by 2 Method (1)
D D-
T
T-
Total 200 800
6
The 2 by 2 Method (2)
D D-
T 68 112
T- 132 688
Total 200 800
7
The 2 by 2 Method (3)
D D- total
T 68 112 180
T- 132 688 820
Total 200 800 1000
PVP(D T)68/1800.378 PV-
P(D-T-)688/8200.839
8
Decision Tree Method (1)
9
Decision Tree Method (2)
10
Bayesian Methods
11
Likelihood Ratio Method (1)
LRSen./(1-Spec.) LR0.34/0.142.43 Pretest Odds0.2/0.80.25 Posttest Odds0.252.430.607 Posttest Prob. Posttest Odds/(1 Posttest Odds) PV Posttest Prob. 0.607/(010.607)0.378
12
Likelihood Ratio Method (2)
LR-(1-Sen.)/Spec. LR-0.66/0.860.767 Pretest Odds0.2/0.80.25 Posttest Odds0.250.7670.192 Posttest Prob. Posttest Odds/(1 Posttest Odds) Posttest Prob.0.192/(10.192)0.161 PV-1-0.1610.839