Diagnostic Validity

Question 1

What are the steps to assessing diagnostic validity?

Answer 1

1. Ask a clinical question:
   - patient (knee pain)
   - test (Lachmans)
   - condition (ACL injury)
2. Can I believe this study?
   - bias
   - internal + external validity
3. What is the diagnosis telling me?
   - sensitivity
   - specificity
   - prevalence
   - +ve LR
   - -ve LR
4. Apply results to practice
   - Bayes nomogram to estimate post test probability

Question 2

What factors affect internal validity?

Answer 2

1. Was the study conducted in a population in whom there was doubt of the diagnosis?
2. Was the index test blinded?
3. Valid reference standard used?

IDEAL = prospective blind comparison with consecutive series of patients

Question 3

What factors affect bias of DTA studies?

Answer 3

1. Patient selection (consecutive/random sample/case control avoided/inappropriate exclusions avoided)
2. Index + reference test blinded (assessor shouldn’t know the result of the other test)
3. Timing of index test and reference test

Question 4

What factors affect bias of DTA systematic reviews?

Answer 4

1. Inclusion/exlcusion criteria for studies?
2. Relevant studies included? (language, date, databases)
3. Methodological quality accounted for? (STARD guidelines)

Question 5

Interpret the LR scores

Answer 5

>10 = conclusive increase in likelihood of disease
5-10 = moderate
2-5 = small
1-2 = minimal

0.5-0.1 = minimal decrease
0.2-0.5 = small decrease
0.1-0.2 = moderate decrease
<0.1 = conclusive decrease in likelihood of disease

Question 6

Briefly describe SNS and SP for SMR testing

Answer 6

SP is high but SNS is low - so good to rule IN if positive but not good to rule out

Question 7

Are Laslett tests for SIJ good to rule in or out SIJ problems?

Answer 7

Good to rule out - SNS = 91; SP = 78