EBM

Question 1

Absolute risk (AR) formula?

Answer 1

Events ÷ Total.

Question 2

Control risk (Rc) definition?

Answer 2

Event risk in control group.

Question 3

Treatment risk (Rt) definition?

Answer 3

Event risk in treatment group.

Question 4

Relative risk (RR) formula?

Answer 4

Rt ÷ Rc.

Question 5

Absolute risk reduction (ARR) formula?

Answer 5

Rc – Rt.

Question 6

Relative risk increase (RRI) formula?

Answer 6

(Rt – Rc) ÷ Rc.

Question 7

Relative risk reduction (RRR) formula?

Answer 7

ARR ÷ Rc.

Question 8

Number needed to treat (NNT) formula?

Answer 8

1 ÷ ARR (ARR in decimal).

Question 9

Number needed to harm (NNH) formula?

Answer 9

1 ÷ absolute risk increase.

Question 10

Odds ratio (OR) 2×2 formula?

Answer 10

(a/c) ÷ (b/d).

Question 11

Odds ratio when using risks?

Answer 11

(Rt ÷ (1–Rt)) ÷ (Rc ÷ (1–Rc)).

Question 12

Hazard ratio (HR) meaning?

Answer 12

Time-adjusted relative risk.

Question 13

ARR interpretation?

Answer 13

Absolute drop in event rate due to treatment.

Question 14

RR interpretation?

Answer 14

Risk in treatment vs control (ratio).

Question 15

RRR interpretation?

Answer 15

Proportional reduction from control risk.

Question 16

NNT interpretation?

Answer 16

Patients needed to treat to prevent 1 event.

Question 17

NNH interpretation?

Answer 17

Patients needed to treat to cause 1 harm.

Question 18

Given Rc=30%, Rt=12%: ARR?

Answer 18

0.18

Question 19

Given Rc=30%, Rt=12%: RRR?

Answer 19

0.6

Question 20

Given Rc=30%, Rt=12%: NNT?

Answer 20

≈6.

Question 21

Given Rc=10%, Rt=8%: ARR?

Answer 21

0.02

Question 22

Given Rc=10%, Rt=8%: NNT?

Answer 22

50

Question 23

Given Rc=20%, Rt=30%: RRI?

Answer 23

50% increase.

Question 24

Given ARR=5%: NNT?

Answer 24

20

Question 25

Sensitivity formula?

Answer 25

TP ÷ (TP+FN).

Question 26

Specificity formula?

Answer 26

TN ÷ (TN+FP).

Question 27

Positive predictive value (PPV) formula?

Answer 27

TP ÷ (TP+FP).

Question 28

Negative predictive value (NPV) formula?

Answer 28

TN ÷ (TN+FN).

Question 29

+LR formula?

Answer 29

Sensitivity ÷ (1–Specificity).

Question 30

–LR formula?

Answer 30

(1–Sensitivity) ÷ Specificity.

Question 31

Diagnostic odds ratio (DOR) formula?

Answer 31

+LR ÷ –LR.

Question 32

AUC (ROC) definition?

Answer 32

Overall test accuracy (0.5 random; 1 perfect).

Question 33

Youden Index formula?

Answer 33

Sensitivity + Specificity – 1.

Question 34

High sensitivity use?

Answer 34

Rule out (SNout).

Question 35

High specificity use?

Answer 35

Rule in (SPin).

Question 36

PPV depends on?

Answer 36

Disease prevalence (directly).

Question 37

NPV depends on?

Answer 37

Disease prevalence (inversely).

Question 38

If Sn=0.90, Sp=0.80: +LR?

Answer 38

0.90 ÷ 0.20 = 4.5.

Question 39

If Sn=0.90, Sp=0.80: –LR?

Answer 39

0.10 ÷ 0.80 = 0.125.

Question 40

Prevalence↑ effect on PPV?

Answer 40

PPV increases.

Question 41

Prevalence↑ effect on NPV?

Answer 41

NPV decreases.

Question 42

Prevalence↓ effect on PPV?

Answer 42

PPV decreases.

Question 43

Prevalence↓ effect on NPV?

Answer 43

NPV increases.

Question 44

Best single summary of test accuracy?

Answer 44

AUC (ROC).

Question 45

Best threshold selection metric?

Answer 45

Youden Index.

Question 46

Meaning of AUC=0.5?

Answer 46

No discrimination (coin toss).

Question 47

Meaning of AUC=1.0?

Answer 47

Perfect discrimination.

Question 48

p-value definition (simple)?

Answer 48

Chance of observed result if null is true.

Question 49

Significance for ratios using CI?

Answer 49

CI not crossing 1.

Question 50

Significance for differences using CI?

Answer 50

CI not crossing 0.

Question 51

Narrow CI implies?

Answer 51

Higher precision.

Question 52

Wide CI implies?

Answer 52

Lower precision.

Question 53

Power formula (concept)?

Answer 53

1 – β.

Question 54

Type I error (α) meaning?

Answer 54

False positive.

Question 55

Type II error (β) meaning?

Answer 55

False negative.

Question 56

Power increases with?

Answer 56

Sample size, events, effect size; lower variance.

Question 57

Event-driven trial power key?

Answer 57

Number of events.

Question 58

Two-sided vs one-sided sample size?

Answer 58

Two-sided needs larger n.

Question 59

Design effect formula (clusters)?

Answer 59

1 + (m–1)×ICC.

Question 60

Effect of ICC on sample size?

Answer 60

Higher ICC → larger required n.

Question 61

Why inflate n in cluster trials?

Answer 61

Within-cluster correlation reduces information.

Question 62

Forest plot diamond meaning?

Answer 62

Pooled effect estimate.

Question 63

Diamond crossing 1 means?

Answer 63

Not statistically significant.

Question 64

I² meaning?

Answer 64

% variability from heterogeneity.

Question 65

I² rough cutoffs?

Answer 65

25/50/75 = low/moderate/high.

Question 66

τ² meaning?

Answer 66

Between-study variance.

Question 67

Fixed-effect model assumption?

Answer 67

One true common effect.

Question 68

Random-effects model assumption?

Answer 68

True effects vary across studies.

Question 69

Funnel plot purpose?

Answer 69

Detect publication bias.

Question 70

GRADE assesses?

Answer 70

Certainty and recommendation strength.

Question 71

Interaction p-value meaning?

Answer 71

Tests if subgroup effects differ.

Question 72

Interaction p < 0.05 implies?

Answer 72

Likely true subgroup difference.

Question 73

Cohort study direction?

Answer 73

Exposure → Outcome.

Question 74

Case–control direction?

Answer 74

Outcome → Exposure.

Question 75

Cross-sectional measures?

Answer 75

Prevalence at one time point.

Question 76

RCT key strength?

Answer 76

Balances known/unknown confounders.

Question 77

Crossover RCT requirement?

Answer 77

Stable disease and washout feasible.

Question 78

Cluster RCT unit of randomisation?

Answer 78

Group (e.g., hospital/school).

Question 79

Pragmatic RCT goal?

Answer 79

Real-world effectiveness.

Question 80

Non-inferiority (NI) aim?

Answer 80

Show new is not unacceptably worse.

Question 81

Delta (Δ) meaning in NI?

Answer 81

Maximum acceptable loss of effect.

Question 82

Why ITT + Per-protocol in NI?

Answer 82

Both needed to avoid bias to NI.

Question 83

NI success by CI rule?

Answer 83

Entire CI above –Δ (within margin).

Question 84

Adaptive trial feature?

Answer 84

Pre-planned changes mid-trial.

Question 85

Platform trial feature?

Answer 85

Multiple arms with shared control.

Question 86

Bayesian posterior concept?

Answer 86

Prior × Likelihood → Posterior.

Question 87

Credible interval meaning?

Answer 87

Probability range for true effect (Bayesian).

Question 88

Selection bias definition?

Answer 88

Systematic differences in who enters study.

Question 89

Performance bias definition?

Answer 89

Co-interventions differ by arm.

Question 90

Measurement bias definition?

Answer 90

Systematic error in data collection.

Question 91

Recall bias context?

Answer 91

Case–control self-reported exposure.

Question 92

Confounder definition?

Answer 92

Third factor linked to exposure and outcome.

Question 93

Effect modifier definition?

Answer 93

True effect differs across subgroups.

Question 94

Handling confounding?

Answer 94

Randomise, match, restrict, adjust.

Question 95

Handling effect modification?

Answer 95

Report subgroup results; do not adjust away.

Question 96

Simpson’s paradox cue?

Answer 96

Adjusted effect reverses direction.

Question 97

Missing data preferred fix?

Answer 97

Multiple imputation (MAR).

Question 98

Lead-time bias definition?

Answer 98

Earlier detection falsely inflates survival time.

Question 99

Length-time bias definition?

Answer 99

Slower diseases overrepresented in screening.

Question 100

Overdiagnosis meaning?

Answer 100

Detection of disease that would not cause harm.

Question 101

Attributable risk (ARexp) formula?

Answer 101

Incidence_exposed – Incidence_unexposed.

Question 102

Population attributable fraction (PAF) idea?

Answer 102

Proportion of cases due to exposure in population.

Question 103

OR approximates RR when?

Answer 103

Outcome is rare (<10%).

Question 104

Case–control: can you get RR directly?

Answer 104

No, use OR.

Question 105

Pre-test probability to odds formula?

Answer 105

Odds = p ÷ (1–p).

Question 106

Post-test odds formula?

Answer 106

Pre-test odds × LR.

Question 107

Post-test probability formula?

Answer 107

Odds ÷ (1+Odds).

Question 108

If Sn=0.95, Sp=0.95: +LR?

Answer 108

0.95 ÷ 0.05 = 19.

Question 109

If Sn=0.95, Sp=0.95: –LR?

Answer 109

0.05 ÷ 0.95 ≈ 0.053.

Question 110

If pre-test p=20% and +LR=5: post-test odds?

Answer 110

0.2/0.8 × 5 = 1.25.

Question 111

If post-test odds=1.25: post-test probability?

Answer 111

1.25 ÷ 2.25 ≈ 56%.

Question 112

CI narrow but crosses null: significance?

Answer 112

Not significant despite precision.

Question 113

p=0.04 but wide CI near null: clinical meaning?

Answer 113

Likely trivial effect.

Question 114

Underpowered negative trial means?

Answer 114

Cannot conclude no effect.

Question 115

Subgroup p not adjusted for multiplicity: caution?

Answer 115

High false positive risk.

Question 116

Incidence definition?

Answer 116

New cases per population per time.

Question 117

Prevalence definition?

Answer 117

Existing cases ÷ population at a time point.

Question 118

Risk difference equals?

Answer 118

ARR.

Question 119

Standard deviation (SD) meaning?

Answer 119

Spread of individual values.

Question 120

Standard error (SE) meaning?

Answer 120

Precision of sample mean.

Question 121

Two independent means (normal): test?

Answer 121

Two-sample t-test.

Question 122

Paired means (normal): test?

Answer 122

Paired t-test.

Question 123

Two independent groups (non-normal): test?

Answer 123

Mann–Whitney U.

Question 124

Paired non-normal: test?

Answer 124

Wilcoxon signed-rank.

Question 125

Two proportions: test?

Answer 125

Chi-square (or Fisher’s if small).

Question 126

Time-to-event comparison: test?

Answer 126

Log-rank test.

Question 127

Binary outcome regression?

Answer 127

Logistic regression.

Question 128

Time-to-event regression?

Answer 128

Cox proportional hazards.

Question 129

Continuous outcome regression?

Answer 129

Linear regression.