Clinical Epidemiology

Question 1

Where do questions come from?

Answer 1

Clinical findings (H &P)
Etiology (what causes disease)
Differential diagnosis (and how to rank the likelihood)
Diagnostic test ( how to include or exclude dx)
Prognosis 
Therapy 
Prevention 
Self improvement

Question 2

Options for summarizing clinical research evidence

Answer 2

Reviews

1. SYSTEMIC
2. NARRATIVE - cochran review

Question 3

Narrative Reviews

Answer 3

• Traditional
• expert summary of research evidence to make recommendations
• advantages- can address broad-based topics by authors with experience
• disadvantage- lack of structure, potential for value judgements.

Question 4

Systematic reviews

Answer 4

• Rigorous scientific review of evidence bearing on a specific clinical question set in advance
• Advantages- useful for single, focused questions; finds all quality studies, increased power, improves generalizability.
• disadvantages- expensive and laborious

Question 5

Defining a specific question acronym

Answer 5

"PICO (T)"
PATIENTS
INTERVENTION
COMPARISON
OUTSOMES 
(T) TIME

Question 6

Types of studies for a systemic review

Answer 6

• Randomized trials (strong):
  E.g Antiplatelet drugs for CVD
  -cohort (measure of what people are already doing)
  E.g smoking as a risk factor for breast CA
  -case control
  -diagnostic

Question 7

Potential for publication bias

Answer 7

• non significant findings
• unpublished data
• grey literature (aka fugitive literature; unpublished, sitting in institutional file somewhere)
• disagreement with findings

Question 8

Organizing data for Systemic Review and Meta-analysis

Answer 8

• ODDS RATIO: Proportion/Percent and/or risk in exposed (treated)/risk in unexposed (untreated)
• odds are needed when pooling (combining) samples in the meta analysis section of SR.

Question 9

How to interpret the Odds ratio

Answer 9

• Forest plots and MA use odds ratio to compare studied and combine results.
• an odds ratio = 1 indicated no association
• an odds ratio < 1 indicated the study favors the treatment
• an odds ratio > 1 indicated the study favors the placebo/control

Question 10

FOREST PLOT VS. META ANALYSIS

Answer 10

Forest plots showing individual studies and Meta analysis to combine the results of individual