Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
@ All forum members , Biostatistic help needed . - bugguy19
#51

Good morning all Smile

Sensitivity = true +ve/all diseased
Sensitivity = 80/ 20+80 = 0.8
Sensitivity = 80%


Regression toward the mean is an statistical phenomena , usually the extreme values ( highest and lowest ) tend to change , this is why we have to take more than One NBME to make sure it's stable looool

Thank you cardio , waiting for your help Smile
Reply
#52
I got it thank u bugguy so much
Reply
#53
@ bdrusa. http://youtu.be/Qw8b_igoMTM

I hope this is helpful.
Reply
#54
Nice @bug. RT. 80%
@innlake, I try my best to help you on that with limited time we got.

__________________________________________________________________________________
Reply
#55
http://i.imgur.com/gxgVsei.png?1 using info ans the 3Qs.
Reply
#56

Hi cardio , help me with my u matched answer looool


Q1---PPV ---46

Q2---PPV----50

Q3---+LR---1.0

I construct 2 different 2❌2 table for each country

I used the the prevalence to calculate the diseased in each country ; assume the population is 100,000

For country x -------

sick people =25000 , healthy people =75000
TP =24.500
FP=500
TN=71.250
FN=28.750

then calculate the PPV = true +/ all +


For country y ------
Sick people 5000
Healthy 95.000

TP=4.900
FP=100
TN=90.250
FN=4.750

then calculate the PPv

For +LR =sensitivity /1-specificity


Reply
#57
q1 PPV =57
q2 PPV =55
q3Positive likelihood ratio=Sensitivity/(1-Specificity)=19.5
For PPV ,I used,TP=SENSI*PREV and FP=(1-Specificity)*(1-Prevalence)
Thanks for the questions Cardio!
Reply
#58
Q1)
Sensitivity = 98%
Specificity = 95%
Prevalence country X= 25%
Applied to 1000 pat a + b + c + d = 1,000
Because the prevalence of disease 25%. 250 pat with disease have (+) and 750, are disease free
Sensitivity of 98% means that 98% of the 250 pats with disease have a (+) test result (0.98 X 250 = 245) and 2% have (-) result (0.02 X 250 = 5)
Specificity of 95% mean that of the 750 who are disease free have a (-) test result ( 0.95 X 750 = 712.5) and 5% have a (+ ) result ( 0.05 X 750 = 37.5)
-------Disease Present------ Disease absent
POS--------- 245-------------------37.5
NEG----------5--------------------712.5
Prevalence = (a + b) / (a + b + c + d) = 250/1000 = 25%
PPV = 245/245 + 37.5 = 87%
NPV = 712.5 / 5 + 712.5 = 99%
*Likelihood ratio/ LR for (+) “test result”: Sensitivity (1-specificity) 98% / 5% = 19.6* ( ANS TO Q3) but that where you made mistake.
Pretest odds: prevalence/1-prevelancee = 0.25 /1- 0.25) = 0.33
Posttest odds: pretest X LR -> 0.33 X 19.6 = 6 .5
Posttest probability: posttest odds /posttest odds + 1) = 6.5/7.5 = 87%

Q2)
Sensitivity = 98%
Specificity = 95%
Prevalence country Y = 5%
Applied to 1000 pat a + b + c + d = 1,000
Because the prevalence of disease 5%. 50 pat with disease have (+) and 950, are disease free
Sensitivity of 98% means that 98% of the 50 pats with disease have a (+) test result (0.98 X 50 = 49) and 2% have (-) result (0.02 X 50 = 1)

Specificity of 95% mean that of the 950 who are disease free have a (-) test result ( 0.95 X 950 = 902.5) and 5% have a (+ ) result ( 0.05 X 950 = 47.5)
-------Disease Present------ Disease absent
POS--------- 49-------------------47.5
NEG----------1-------------------902.5
Likelihood ratio for (+) test result: Sensitivity (1-specificity) 98% / 5% = 19.6 ( Again ANS to your Q3 )
Pretest odds: prevalence/1-prevalence = 0.05 /1- 0.05) = 0.05
Posttest odds: pretest odds X LR -> 0.05 X 19.6 = 0.98
"Posttest probability": posttest odds /(posttest odds + 1) = 0.98/1-0.98 = 49%

@bug dont worry for Posttest porb but make sure you understant LR +.


Reply
#59
Q3) as I mention *D*
Reply
#60
Hundred pats with lung cancer and hundred pts without lung cancer were enrolled in a study. They were questioned about tobacco use over the past 20 years. 90 of the lung cancer pats reported smoking as did 10 of the pats without lung cancer.

@bug show me your work for OR cal?
Reply
« Next Oldest | Next Newest »


Forum Jump: