Diagnosis

Question 1

Diagnosis

Answer 1

exam- history, systems review, test and measures–

evaluation- clinical judement, synthesis, decision and/ or action–

diagnosis

cant make decision if 70% sure

Question 2

Diagnosis goals-

Answer 2

1. assess the quality of a diagnostic study
2. introduce quantitative interpretation of diagnostic accuracy (sens/spec, +-LR)
3. decide- can the test be applied to CONFIRM or EXCLUDE a suspected diagnosis (rule in/out)

Question 3

Diagnostic study- PDRA

Answer 3

population of patient of interest

diagnostic test being studied

reference standard

accuracy of diagnostic test

Question 4

Population of patients

Answer 4

demographic decriptors (age, sex)

include those with suspected target disorder (type of disorder, severity of disorder, diagnostic dilemma)

Question 5

Diagnostic test

Answer 5

well described-needs to be repeatable. Adequate rationale for test

study should be good enough to replicate

Question 6

Reference standard

Answer 6

gold stand, criterion stand, diagnostic stand

result is widely accepted as “the diagnosis”

if unacceptable, study is invalid

Question 7

Accurace

Answer 7

of correct matches (reference and diagnostic tests agree-both pos and neg)

of incorrect matches (ref and diag tests disagree (one pos and one neg)

use 2x2 table to compare test to gold standard

**review chart in notes

Question 8

Diagnostic test valid? Question 1

Answer 8

independent, blind comparison with a reference standard

-clinicians performing tests have no knowledge of the other test results (masked or blinded)

“independent assessments”

ex. hs basketball players

all have ant drawer and arthroscopy

physician arthroscopy and PT ant drawer tests

each unaware of other test results when performing their own test

Question 9

Diagnostic test valid? question #2

Answer 9

Was the diagnostic test evaluated in an appropriate spectrum of patients?

all common presentations of target disorder

complete range of illness of target disorder

commonly confused target disorders

ex. acute vs chronic suspected ACL injuries. Include injuries you want to distinguish from meniscus and PCL injuries. Exclude those with known ACL injury

Question 10

Diagnostic test valid? question #3

Answer 10

reference standard applied regardless of the diagnostic test result

-may be tempting to avoid reference standard when negative diagnostic test result

eiminated info on negative (fase pos rate, true pos rate)

Everybody needs ref standard and gold standard for test

Question 11

Distinguishing accuracy-3 ways

Answer 11

sensitivity

specificity

likelihood raios

Question 12

Sensitivity

Answer 12

proportion of patients WITH the target disorder and POSITIVE diagnostic test result

One portion of table-A and C

Indicated how often a diag test detects a disease or condition when it is present

TP/FN

Can rule out disease if test is neg and good sens

clinically useful= high send, neg result, rules out

essentially tells clinician how good the test is at correctly identifying patients with the condition of interest

Question 13

Specificity

Answer 13

proportion of patients WITHOUT target disorder and NEGATIVE diagnostic test result

absence of disease in those without it

positive result=rule in

B and D

high spec= very high proportion of neg tests do not have the target disorder

clinically useful= high spec, pos result, rules in

essentially tells the clinician how good the test is at correctly identifying the absence of disease

Question 14

Limitations of SENS and SPEC

Answer 14

epidemiological- dependent on overall prevalence of disease

statistically- only focus on target disorder + or -

clinical- can’t be used to change post test odds

Question 15

Likelihood ratio (LR)

Answer 15

quantifies relationship

-pretest odds or probability,
-test results (+ or -)
post test odds or probability

“Good” LR result in large shifts from pre to post test probability

Question 16

+ LR

Answer 16

the strength of a positive test finding

ratio of true positives to false positives

used to RULE IN or confirm

50% pretest-don’t know anything yet

+LR= SN/ (1-SP)

ex. +LR=1.25

means that a person with the disease is about 1.3x more likely to have a positive test than a person who has not got the disease

Question 17

• LR

Answer 17

the strength of a negative test finding

ratio of false negatives to true negatives

used to RULE OUT or exclude

50% pre test
Want lower post test to rule out.

-LR= (1-SN) / SP

ex. -LR=.63
means that the probability of having a neg test for individuals with the disease is about .63x of that of those without the disease

Question 18

Interpretation of + LR

Answer 18

> 10- lrage or conclusive changes in post test odds

from 5-10 moderate shift in post tests odds

from 2-5 small (sometimes important) changes in post test odds

from 1-2 small (rarely important) in post test odds

Question 19

Interpretation of -LR

Answer 19

less than 0.10 large or conclusive changes in post test odds

from 0.1-0.2 moderate shift in post test odds

from 0.5-0.2 small (sometimes important) changes in post test odds

from 0.5-1 small (rarly important changes in post test odds

Question 20

Fagan’s nomogram

Answer 20

pre test 50%/post test 83%

means that the probability of having the disease in this person incrases from 50-83% when he or she had a positive test result

ratios close to 1 do little to alter diagnostic probability

small LR values substantially reduce the odds of the diagnosis

large positive LR values substantially increase the odds of the diagnosis