Prep for USMLEPrep for USMLE
         Forum      |     Resources New Posts   |   Register   |   Login





 statistic quiz  



Post Reply  
  • 0/5
  • 1
  • 2
  • 3
  • 4
  • 5


Author9 Posts
  #1

55 years old man complaint of urinary infrequency ,i cm nodule in prostate gland PSA ordered.By common standard PSA above 4 ng/l considered abnormal. using this standard the test has sensitivity of 80% specificity 90% recently puplished study showed that in cross sectional study 10% of men at this age has prostate cancer. the result of patient PSA is 7 ng/l.what is your best likelihood that this man actually has prostatic cancer ????? 13% 25% 36% 47% 58% 69% 72% 81% i need your help please thanks




  #2

81%


On Jul 05, 2011 - 12:02 AM, deals responded:
please can you explain how did u get this 81% thanks


  #3

i need to know the steps to calculate this pleaseeeeeeeeee cause in kaplan notes it is written 47% and i think it is 81% but please tell me the steps thanks


  #4

gosh anybody knows how to calculate this tuti hami can u tell the steps


  #5

The question is asking for Positive Predictive Value of the test. The answer is 47%. Formula for PPV is as follows

(sensitivity x prevalance)
_________________________________________
(sensitivity x prevalance) + (1-specificity)(1-prevalance)

In the above example it would be

0.8 x 0.1
______________________

0.8 x 0.1 + (1-0.9)(1-0.0.1)

Which is 47%.

Although the question uses the word likelihood it is not asking for likelihood ratio (LR). LR is the odds that a positive test result would be found in a patient with, versus without, a disease. Formula for LR+ is as follows:
Likelihood Ratio Positive (LR+) = Sensitivity / (1 - Specificity).

See this link for a detailed description with an example:


  #6

deals wrote:
gosh anybody knows how to calculate this tuti hami can u tell the steps

Here you go deals. Hope this helps.


  #7

thanks motorolla appreciate it


  #8

 

let's make it simple: assume you have a 1000 patients.

10% with disease that should be on the left column

Now -- the table should be like:

    

80      90       

20      810

 

100    900

 

All you have to do, calculate PV+

 

which is 80/80+90 = 47%


  #9

I hope anyone with a PPV, NPV, sensitivity, specificity, accuracy problem finds this post. motorola is correct but I recommend using monterow technique. In fact, make n=100 (hypothetically) and the math become a little easier.nod

DECOMPOSITION
Prevalence=10% so the number of all those truly diseased is … 10 and those truly healthy are 90.
Of the 10 truly diseased: Sensitivity is 80% so TP=8; FN=2.
Of the 90 truly healthy: Specificity is 90% so TN=81 (90*0.9); FP=9 (90-81).
Now you have the values to calc PPV

REMEMBER
You needed to see that this is a PPV problem. PSA > 4 is considered +; the patient had PSA=7; so, what’s the probability that a person w/ a + test result is truly +? That’s PPV.

This is an EXCELLENT high yield concept biostat question. The question could have been framed to ask about PPV, NPV, sensitivity, specificity, accuracy!!! Hey, so …. what’s accuracy? 10% 47% 58% 69% 72% 81% 89%, 92% or 110%





Bookmark and Share





Login or Register to post messages








show Similar forum topics

a statistic Q
statistic
Statistic problem help
show Related resources









Advertise | Support | Privacy | TOS | Premium | Contact