Evidence-Based Medicine

10/20/97


Click here to start


Table of Contents

Evidence-Based Medicine

Objectives

Logic + training + experience IS NOT ENOUGH.....

When to use Evidence

Study Design

Study types

Study types

Study types

Study types

Hierarchy of proof

True Experiment

Prospective vs. Retrospective

Harmful Exposure

Rare Disease

Disease Takes Years to Develop

Fishing Expeditions

Rules of evidence - diagnosis

Rules of evidence - treatment

Diagnosis

Diagnosis - the ideal test

Diagnosis - the usual dilemma

Diagnosis

Diagnosis

Above all.... do no harm.

Diagnosis - the box approach

Sensitivity

Diagnosis - Sensitivity (PID)

Specificity

Diagnosis - Specificity (NIH)

Diagnosis - the box approach

Diagnosis - fill in the numbers

Changing the cut-off

Diagnosis - changing cut-off

Changing the cut-off

Changing the cut-off

But the test was positive...

The test was either + or - .

Diagnosis - Pos. Pred. Value

Diagnosis - the box approach

Diagnosis - Neg. Pred. Value

Negative Predictive Value

How to go from Pre-to Post-test Probability

How to go from Pre-to Post-test Probability -

How to go from Pre-to Post-test Probability - STEP 1

How to go from Pre-to Post-test Probability - STEP 2

How to go from Pre-to Post-test Probability - STEP 2

How to go from Pre-to Post-test Probability - STEP 3

How to go from Pre-to Post-test Probability - FINAL

Threshold Decision-Making

Changing Prevalence & Sensitivity (example)

Diagnosis - the box approach

Diagnosis - the box approach

Diagnosis - the box approach

Diagnosis - the box approach

Pre- to Post-Test Probabilities different sensitivities

50% Prevalence

Prevalence < 50%

Changing prevalence

Changing sensit., specificity

Sensitivity (PID)

Pos. Pred. Value

Neg. Pred. Value - FN (Sens.)

Treatment

What was the treatment or exposure effect?

RCT/Cohort vs. Case Control

Odds Ratios & Relative Risks

Odds Ratio

Relative Risk

Odds Ratio ­ Relative Risk

RRR, ARR, NNT

RRR, ARR, NNT

RRR, ARR, NNT

RRR, ARR, NNT

Significance

p Value & Power

Is there a difference?

95% Confidence Interval

95% Confidence Interval

Clinical vs. Statistical Signif.

Summary

APPENDIX

Rules of evidence - diagnosis

Rules of evidence - treatment

Rules of evidence - treatment

Rules of evidence - treatment

Patients - Name that Bias

Example - febrile seizure

Example - febrile seizure

Rules of evidence - treatment

Rules of evidence - treatment

Rules of evidence - treatment

Rules of evidence - treatment

Encainide & Flecainide

Encainide & Flecainide

Author: Family Medicine

Email: chessman@musc.edu

Home Page: http://familymed.musc.edu