Critical Inquiry and Evidence Based Practice

Question 1

Selection bias

Answer 1

Study subjects of diagnostic test in question are not representative of the population on whom the test is typically applied

Ie: control group of ACL diagnostic test has no history of knee pain. Ideally control and experimental group have equal chances of having an ACL injury

Question 2

Verification bias

Answer 2

The reference standard is not applied consistently to all subjects when assessing efficacy of a diagnostic test. Can lead to overestimating diagnostic accuracy of the diagnostic test

Question 3

Incorporation bias

Answer 3

Reference standard includes the diagnostic test being studied. May inflate the accuracy of the diagnostic test

Question 4

Review bias

Answer 4

Bias that occurs if either the reference standard or diagnostic test in question is judged by an individual with knowledge of the other result

Question 5

Sensitivity

Answer 5

Measure of individuals in a population who have a condition. High Sn can help rule out a condition when negative

Sn = (true positive)/(all individual who have condition)

Sn = (true pos)/(true pos + false neg)

Question 6

Specificity

Answer 6

Proportion of people without a condition that tested negative. High specificity can help rule in a condition when test is positive

Sp = (true negatives)/(all individuals who do NOT have condition)

Sp = (true negatives)/(true negatives + false positives)

Question 7

Positive predictive value

Answer 7

Proportion of people who test positive who actually have condition (proportion of patience with positive results who are correctly diagnosed)

Limited clinical value due to susceptibility to selection bias, based on prevalence of a condition within sample group

Question 8

Negative predictive value

Answer 8

Proportion of people who test negative that do not have the condition

Limited clinical value due to susceptibility to selection bias, depending on prevalence of condition within sample group

Question 9

+LR

Answer 9

Very powerful statistic for ruling in a condition if test is positive. Has high potential of improving post-test probability of a condition. Ideally +LR > 10

+LR = Sn / (1 - Sp)

Question 10

-LR

Answer 10

Very powerful statistic for ruling out a condition if the test is negative. Ideally -LR < 0.10

-LR = (1 - Sn)/Sp

Question 11

95% confidence interval

Answer 11

Range of values within which we have 95% certainty. Large confidence interval limits clinical usefulness of a test

Question 12

Chi square statistic

Answer 12

Tests the hypothesis that the diagnostic test result and reference standard have no association

Limited clinical usefulness on its own

Question 13

Kappa coefficient

Answer 13

Measure of reliability. Cannot be used in isolation bc may lead to clinician potentially excluding tests with low reliability when test has good Sn, Sp, or +/- LR. Conversely, may lead to clinician including a very weak diagnostic test that has good reliability

Question 14

External validity

Answer 14

Ability of a measure to predict or relate to outcomes beyond the population of a given trial

Question 15

Construct validity

Answer 15

How well a test measures the concept it was designed to evaluate

Question 16

Face validity

Answer 16

Degree which attest appears effective in measuring the variable or construct that it is supposed to measure

Question 17

Predictive validity

Answer 17

Degree to which test score is accurately predict score or predict a future outcome

Question 18

Research strength for CPG based on strong evidence

Answer 18

Includes at least one level 1 study. Includes preponderance of level 1 and level 2 studies

Clinician “should” perform stated recommendation

Question 19

Research strength for CPG based on moderate evidence

Answer 19

Includes a high-quality RCT or a preponderance of level 2 studies

Clinician “may” perform stated recommendation

Question 20

Research strength for CPG based on weak evidence

Answer 20

Includes a single level 2 study or preponderance of level 3 and 4 studies, including statements of consensus by content experts

Clinician “can” perform stated recommendation

Question 21

Research strength for CPG based on conflicting evidence

Answer 21

Conflicting evidence suggests clinician “should not” perform stated recommendation

Question 22

Research strength for CPG based on theoretical/foundational evidence

Answer 22

Preponderance of evidence from animal or cadaver studies, from conceptual models or principals, or from basic science