EFFECT OF ATYPICAL POPULATIONS IN META-ANALYSIS
Juliana Torres Sánchez
PhD. Juan Carlos Correa Morales
AGENDA
• Contextualization
• Problem
• Design and methods
• Data/results
• Next steps….
Contextualization
Statistical Technique
Sum
Unification
Understanding
Response variable
Problem
• Effect Atypical Study
Results
• Recommendations
• Mean, proportions
Design and methods
Mean 1
Location 0,9
Outlier 0,8
Study 0,7
0,6
0,5
0,4
0,3
0,2
0,1
10
20
30
40
n 50
60
70
80
90
100
1.000 simulations
Studies
Sample Size = n
Meta- Analysis
Meta-analysis size = N = 10 Studies + Outlier study
N = 11
Proportion (%) = # No Heterogeneity
1.000 Simulations
µp + k σp
Design and Methods
+ 0,1𝜎
Atypical population
Design and Methods
+0,2𝜎
Design and methods
+0,3𝜎
Design and methods
+0,4𝜎
Design and methods
+0,5𝜎
Design and methods
+0,6𝜎
Design and methods
+0,7𝜎
Design and methods
+0,8𝜎
Design and methods
+0,9𝜎
Design and methods
+1𝜎
In when Not
Increase the doses
Decrease the doses
Out when Yes
Decrease the doses
Increase the doses
True
True
¿Is that or thoseatypical studies
really partof your population
A good approach is
to use the linear regression
outlier detection
proposals
Next steps….
• Other models
random effects model
mixed effects model