SlideShare uma empresa Scribd logo
1 de 43
Baixar para ler offline
Mantel Haenszel Method
Dr. S. A. Rizwan, M.D.,
Public Health Specialist (MOH),
Saudi Board of Preventive Medicine
Riyadh, KSA
With thanks to Dr D. Hannoun (National Institute of Public Health, Algeria)
Introduction
Analytical studies in epidemiology aim to assess the
association between two variables
• Is the association valid? à RD – RR – OR
• Is it causal? à Criterion of causality
In most cases, we have to take in account a third (or more)
variable that may affect the relationship studied
• Confounding à bias
• Effect modification (Interaction) à useful information
Introduction
Exposure Outcome
Vaccine efficacy Measles
Third variable
• No effect: sex (boy/girl)
• Intermediary: Antibodies rate
• Confounder: Mother education
• Effect modifier: Age VE is lower for children < 18mo
VE is the same for boy and girl
AR is a consequence of Vaccine
Effect observed is affected by ME
Introduction
We can avoid these complications at two potential
steps
• Step one in the study design
• Randomisation
• Restriction
• Matching
• Step two in the analytical phase
• Standardization
• Stratification
• Multivariate analysis
Focus of
this class
Stratification: Principle
Broad principle:
• Create strata according to categories of the third variable
• Perfom analysis inside these strata
• Conclude about the relationship inside the strata
• Forming adjusted summary estimate: i.e. weighted average
• Assumption: weak variability in the strata (items within strata
should be as similar as possible)
What is achieved?
• To analyse effect modification
• To eliminate confounding
Stratification: Principle
To perform a stratified analysis, we have 6 steps:
1. Carry out simple analysis to test the association between the
exposure and the disease and to identify potential confounder
2. Categorize the confounder and divide the sample in strata,
according to the number of categories of the confounder
3. Carry out simple analysis to test the association between the
exposure and the disease in each stratum
4. Test the presence or absence of effect modification between the
variables
5. If appropriate, check for confounding and calculate a point estimate
of overall effect (weighted average measure)
6. If appropriate, carry out and interpret an overall test for association
Stratification: Step 1 – Example 1
Investigation of the relationship between Vaccine Efficacy and
Measles (cohort study)
1. Crude analysis: Is there any association between vaccine
efficacy and prevention of Measles?
• RR = 0.55 [0.41-0.74] ; p < 0.001 à VE = 1-RR = 45%
• There is an association between Vaccination and prevention of
Measles
Measles + Measles -
Vaccinated 72 79773
No vaccinated 116 71039
Stratification: Step 1 – Example 1
2. Identify potential confounder:
• Is the association real and valid or could be
modified when we take in account a third factor
like age?
• We are interested in how the effects of a third
variable, age at vaccination, may be influencing
this relationship
Stratification: Step 2 – Example 1
Categorize the confounder and divide the sample in strata,
according to the number of categories of the confounder
1. Number of categories of age : <1 year and 1-4 years
2. Create strata according to the number of categories
<1 year
Measles
+
Measles
-
Vaccinated 38 35587
Not Vaccinated 30 24345
1 - 4 years
Measles
+
Measles
-
Vaccinated 34 44186
No vaccinated 86 46694
Stratification: Step 3 – Example 1
Perform analysis inside these strata
1. In each strata
• Calculate the X2 to test the association
• Estimate the RRi/ORi
<1 year
Measles
+
Measles
-
Vaccinated 38 35587
Not Vaccinated 30 24345
1 - 4 years
Measles
+
Measles
-
Vaccinated 34 44186
No vaccinated 86 46694
RRi = 0.87 [0.54 – 1.40], VE= 13%
p = 0.55
RRi = 0.42 [0.28 – 0.62], VE= 58%
p < 0.001
Stratification: Step 4 – Example 1
Test for interaction by the third variable
• Appropriate tests
• Breslow Day & Mantel-Haenszel test: commonly
used
• Woolf test
• Tarone
å
-
=
i i
2
i2
)var(effect
effect)summary(effect
Χ
Stratification: Step 4 – Example 1
Test the presence or absence of interaction between the
variables
• Breslow-Day: Test of homogeneity in strata:
• H0: RR1 = RR2 or OR1 = OR2
• Χ2 test compared observed and expected counts
• It requires a large sample size within each stratum
Stratification: Step 4 – Example 1
Test the presence or absence of interaction between the
variables
Two possibilities
RR1 = RR2 or OR1 = OR2 RR1 ¹ RR2 or OR1 ¹ OR2
No Interaction: Third variable
is Not an effect modifier
Presence of Interaction: Third
variable could be effect modifier
Next step: Look for confounding
and calculate adjusted measure
Stop here: Results reported only
by strata No pooled measure
Stratification: Step 4 – Example 1
Test the presence or absence of interaction
Homogeneity test: H0: RR<1year= RR1-4years (RR population)
P <0.001 à statistical interaction present
• There is interaction between age at vaccination and VE for
Measles (or)
• Age at vaccination modifies the effect of VE for Measles (or)
• Age at vaccination is an effect modifier for the relationship
between VE and Measles
• Not appropriate to try to summarize these two effects, 0.87 and 0.42,
into one overall number
• We should report the two stratum-specific estimates separately and stop
the analysis
0.87 ≠ 0.42
Stratification: Step 1 – Example 2
Investigation of Effectiveness of AZT in preventing HIV seroconversion after a
needlestick (case control study)
1. Crude analysis: Is there any association between AZT and prevention
of HIV seroconversion after a needlestick injury in health care
workers?
• OR crude = 0.61 [0.26 - 1.44], p = 0.25
• No evidence of a benefit from AZT
• The authors stratified by the severity of the needlestick
HIV + HIV -
AZT + 8 130
AZT - 19 189
Stratification: Steps 2 and 3 – Example 2
Divide the sample into strata, according to the number of categories of the
confounder and perform analysis
1. Categories of severity of needlestick : minor and major severity
2. Create strata according to the number of categories
3. In each strata test the association and Estimate the RDi/RRi/ORi
Minor severity
HIV+ HIV -
AZT + 1 90
AZT - 3 161
Major severity
HIV+ HIV -
AZT + 7 40
AZT - 16 28
ORminor = 0.60 [0.06-5.81], p=1.0
No association
ORmajor = 0.31 [0.11- 0.84], p=0.02
Presence of association
Stratification: Step 4 – Example 2
Test the presence or absence of interaction between the
variables
• Test of homogeneity in strata : H0 : ORminor = ORmajor?
• p=0.59 à Breslow-Day test is not significant
à No statistical interaction
• We assume there is no effect modification between
severity of needlestick and AZT on the risk of HIV
• We could try to summarize these two effects, 0.60 and 0.31,
into one overall number à Construct a weighted average
estimate
• Go to Step 5
Stratification: Step 5 – Example 2
If appropriate, check for confounding
– Two steps
» Calculating adjusted summary estimate
» Comparing adjusted summary estimate to crude
estimate
Stratification: Step 5 – Example 2
If appropriate, check for confounding
1. Forming an adjusted summary estimate
• Weighted average measure of the effect of exposure: RDi or
RRi or ORi according to the size of each stratum
• Weight depends upon a lot of factors:
• Measure of association: RD or RR or OR
• Nature of data: qualitative, quantitative
• Purpose of the analysis: follow-up study, case control study
• Methods:
• Mantel-Haenszel
• Woolf, Miettinen
RR/OR
RRa
=
wi
RRi∑
wi∑
, wi
=
ci
n0i
ni
å
å
å
å
=
i i
1i1i
i
i
i i
0i0i
i
i
a
n
m*n
c
n
m*n
a
RR
Strata i of F
Dis+ Dis -
E + ai bi noi
E - ci di n1i
moi m1i ni
Stratification: Step 5 – Example 2
If appropriate, check for confounding
1. Estimation of RRa: Follow up study
Strata i of F
Dis+ Dis -
E + ai bi noi
E - ci di n1i
moi m1i ni
Stratification: Step 5 – Example 2
If appropriate, check for confounding
1. Estimation of ORa : Case control study
ORMH = S ai di S bi ci
ni ni
ORMH = S wi ORi / S wi
wi = bi ci / ni
Stratification: Step 5 – Example 2
If appropriate, check for confounding
2. Identify confounding
• Compare crude measure with adjusted measure:
• H0: RRMH=RRcrude (or) ORMH=ORcrude
• No statistical test available
• Confounding can be judged present when adjusted
RRMH or ORMH is different from crude effect
• D = (ORMH - ORcrude ) / ORcrude
• Arbitrary cut-off: >10%
• Interpretation
Stratification: Step 5 – Example 2
If appropriate, check for confounding
Two possibilities
D < 10% D > 10%
No confounding Presence of confounding
Use RRcrude or ORcrude Use RRMH or ORMH
Stratification: Step 5 – Example 2
If appropriate, check for confounding
Be careful! We should report the adjusted measure:
• Only if we haven’t detected interaction: RRi or ORi are
homogenous among strata
AND
• If we have detected confounding
Stratification: Step 5 – Example 2
Effectiveness of AZT in preventing HIV seroconversion after a
needlestick in health care workers
1. Estimation of ORa adjusted
ni = 255; OR = 0.60 ni = 92; OR = 0.31
Minor severity
HIV+ HIV -
AZT + 1 90
AZT - 3 161
Major severity
HIV+ HIV -
AZT + 8 40
AZT - 16 28
ORMH = 0.38 [0.14 – 0.87]
Stratification: Step 5 – Example 2
2. Identify confounding
• Compare the ORMH=0.38 with ORcrude=0.61
• D = (ORMH - ORcrude) / ORcrude = 44 %
• D > 10% è We conclude that severity of
needlestick is a confounder
• After adjusting for severity of needlestick, we
obtain a reduction of the magnitude of the
relation between AZT and prevention of the HIV
seroconversion
• Conclusion: The good summary measure to use is the
adjusted ORMH = 0.38
Stratification: Step 6 – Example 2
If appropriate, carry out and interpret an overall test for association
1. Verify the relationship between exposure and outcome after
adjusting for the third variable
• H0: RRMH = 1 (or) ORMH = 1
• Statistical test à Mantel-Haenszel
• It follows a chi-square distribution of 1 df, regardless
of the number of strata
2. Confidence interval of adjusted RRa or ORa
2
MHχ
1,96
1
RR
± 2
MHχ
1,96
1
OR
±
=
Stratification: Step 6 – Example 2
Mantel-Haenszel Chi square test
• It follows a chi-square distribution of 1 df,
regardless of the number of strata
• MH test statistic is defined as
Stratification: Step 6 – Example 2
• Verify the relationship between AZT and HIV seroconversion after
adjusting for severity of needlestick
• H0 : ORMH = 1
• p = 0.036 à Mantel-Haenszel test is significant
• Conclusion:
• After adjustment for severity of needlestick, we
have a significant association between AZT and
HIV
• When we adjust for severity of needlestick the OR
decreased from 0.61 to 0.38 but also became
significant (from p=0.25 to p=0.036)
Confounding: Definition
Be careful
• Factor responsible for confounding is called a
confounder or a confounding variable
• Confounder factor confounds the association of interest:
It confuses an estimate
Examples
• Needlestick severity confounds the effect of AZT in
preventing HIV seroconversion
Confounding: Definition
When we have confounding:
• The observed association between exposure and disease can
be attributed totally or in part to the effect of confounder
• Overestimation (+) of the true association between exposure
and disease occurs:
• Underestimation (-) of the true association between exposure
and disease occurs:
• Qualitative confounding: Direction of observed effect could
change
Crude effect > Adjusted Effect
Crude effect < Adjusted Effect
Confounding: How to identify confounder
Compare:
• Crude effect of association RD - RR - OR with adjusted
measure of effect RDA - RRMH - ORMH
How?
• Take in account only D = (ORMH - ORcrude ) / ORcrude
• If D >10% à Presence of confounding
• If D <10% à No confounding
Statistical test must be avoided to identify confounding
Effect modification
Variation in the magnitude of measure of effect across levels of a
third variable
• Tetracycline discolours teeth in children but not in adults
Tetracyclines
Age: children/adults
• Effect modification is a concept, also called effect measure
modification, interaction or heterogeneity of effect
• Factor responsible for effect modification is called an effect
modifier à it modifies the effect of exposure on the outcome
Teeth discoloration
Effect modification: Additive/multiplicative
• For risk DIFFERENCE:
Absence of interaction is RAAB = RAA + RAB
Interaction is called Additive interaction
OR
• For risk RATIO:
Absence of interaction RRAB = RRA X RRB
Interaction is called Multiplicative interaction
OR
RDAB > RDA + RDB
RRAB > RRA X RRB
RDAB < RDA + RDB
RRAB < RRA X RRB
Effect modification: Properties
Effect modification is not a bias but useful information
• Identification of subgroups with a lower or higher risk
• Targeting public health action
• Better understand of the disease: biological mechanism
Effect modification: How to assess it?
Is there any statistical test to help us to assess effect
modification?
• Yes: many tests to verify the homogeneity of the strata
• But not sufficient
» Clinical/biological decision rather than statistical
» Taking in account the magnitude of the effect
modification
» Statistical tests depend on the size of the study
When to report effect modification?
Effect modification & Confounding
Effect modification
• Belongs to nature
• Rare
• Effects in strata different
• Must report stratum-specific
estimates separately
• Useful information
• Controlled in the study design
phase
• Statistical test for interaction
Confounding
• Belongs to study
• Frequent
• Specific effects ≠ crude
measure
• Should report an adjusted
weighted estimate
• Distortion of effect: bias
• Prevented in the study design
and controlled in the analytical
phase
• No statistical test for
confounding
Effect modification & Confounding
• Both confounding and effect modification
• Must be interpreted and taken in account
according to the knowledge of pathophysiologic
mechanism
• Determination is dependent on choice of effect
measure: RD – RR – OR
• Effect modification and confounding can exist
separately or together
An exercise
Strata
1
Strata
2
Crude
OR
Adjusted
OR
Confounding EM
4.0 4.0 4.0 4.0 ? ?
4.0 0.25 1.0 1.0 ? ?
1.0 1.0 8.4 1.0 ? ?
4.0 0.25 1.0 2.0 ? ?
Crude analysis
Specific estimates not equal across
strata
Yes
= Effect modification
No
= No effect modification
Adjusted estimate not equal to
Crude estimate
Yes = Confounding No = No Confounding
Report stratum-
specific estimates –
No pooled measure
Report adjusted
estimate, 95% CI, p
value of χ2MH
Report crude estimate,
95% CI, p value
Stratification
Specific estimates in each strata
A strategy to
check for
interaction &
confounding
Take home messages
• Stratification is a useful tool to assess the real effect of
exposure on the disease
• But, it has some limits:
– Possibility of insufficient data when we have several strata
– Tool developped only for categorical variable
– Only possible to adjust for a limited number of confounders
simultaneously
• Advanced learning: Simpson’s paradox, non-collapsibility
THANK YOU

Mais conteúdo relacionado

Mais procurados

Stratification and Mantel-Haenszel estimation
Stratification and Mantel-Haenszel estimationStratification and Mantel-Haenszel estimation
Stratification and Mantel-Haenszel estimationVignesh Loganathan
 
Validity & reliability of screening & diagnostic tests
Validity & reliability of screening & diagnostic testsValidity & reliability of screening & diagnostic tests
Validity & reliability of screening & diagnostic testsTanveerRehman4
 
Epidemiology lecture3 incidence
Epidemiology lecture3 incidenceEpidemiology lecture3 incidence
Epidemiology lecture3 incidenceINAAMUL HAQ
 
Analyzing the randomised control trial (rct)
Analyzing the randomised control trial (rct)Analyzing the randomised control trial (rct)
Analyzing the randomised control trial (rct)Vikash Keshri
 
Standardization of rates
Standardization of ratesStandardization of rates
Standardization of ratesHalyna Lugova
 
P value, Power, Type 1 and 2 errors
P value, Power, Type 1 and 2 errorsP value, Power, Type 1 and 2 errors
P value, Power, Type 1 and 2 errorsRizwan S A
 
What does an odds ratio or relative risk mean?
What does an odds ratio or relative risk mean? What does an odds ratio or relative risk mean?
What does an odds ratio or relative risk mean? Terry Shaneyfelt
 
Case control study
Case control studyCase control study
Case control studyswati shikha
 
Odds ratios (Basic concepts)
Odds ratios (Basic concepts)Odds ratios (Basic concepts)
Odds ratios (Basic concepts)Tarekk Alazabee
 
Measures of association 2013
Measures of association 2013Measures of association 2013
Measures of association 2013dinahoefer11
 
Outbreak investigation
Outbreak investigationOutbreak investigation
Outbreak investigationAmandeep Kaur
 
Epidemiological studies
Epidemiological studiesEpidemiological studies
Epidemiological studiesBruno Mmassy
 
Case control study - Part 2
Case control study - Part 2Case control study - Part 2
Case control study - Part 2Rizwan S A
 
3.5.2 selection bias
3.5.2 selection bias3.5.2 selection bias
3.5.2 selection biasA M
 
Measures Of Association
Measures Of AssociationMeasures Of Association
Measures Of Associationganesh kumar
 

Mais procurados (20)

Stratification and Mantel-Haenszel estimation
Stratification and Mantel-Haenszel estimationStratification and Mantel-Haenszel estimation
Stratification and Mantel-Haenszel estimation
 
Validity & reliability of screening & diagnostic tests
Validity & reliability of screening & diagnostic testsValidity & reliability of screening & diagnostic tests
Validity & reliability of screening & diagnostic tests
 
Estimating risk
Estimating riskEstimating risk
Estimating risk
 
Epidemiology lecture3 incidence
Epidemiology lecture3 incidenceEpidemiology lecture3 incidence
Epidemiology lecture3 incidence
 
Analyzing the randomised control trial (rct)
Analyzing the randomised control trial (rct)Analyzing the randomised control trial (rct)
Analyzing the randomised control trial (rct)
 
Chi square test
Chi square testChi square test
Chi square test
 
Standardization of rates
Standardization of ratesStandardization of rates
Standardization of rates
 
P value, Power, Type 1 and 2 errors
P value, Power, Type 1 and 2 errorsP value, Power, Type 1 and 2 errors
P value, Power, Type 1 and 2 errors
 
What does an odds ratio or relative risk mean?
What does an odds ratio or relative risk mean? What does an odds ratio or relative risk mean?
What does an odds ratio or relative risk mean?
 
Case control study
Case control studyCase control study
Case control study
 
Odds ratios (Basic concepts)
Odds ratios (Basic concepts)Odds ratios (Basic concepts)
Odds ratios (Basic concepts)
 
Sample size calculation
Sample size calculationSample size calculation
Sample size calculation
 
Measures of association 2013
Measures of association 2013Measures of association 2013
Measures of association 2013
 
Outbreak investigation
Outbreak investigationOutbreak investigation
Outbreak investigation
 
Epidemiological studies
Epidemiological studiesEpidemiological studies
Epidemiological studies
 
Case control study - Part 2
Case control study - Part 2Case control study - Part 2
Case control study - Part 2
 
Epidemiology Study Design
Epidemiology Study DesignEpidemiology Study Design
Epidemiology Study Design
 
3.5.2 selection bias
3.5.2 selection bias3.5.2 selection bias
3.5.2 selection bias
 
Epidemic Curve
Epidemic CurveEpidemic Curve
Epidemic Curve
 
Measures Of Association
Measures Of AssociationMeasures Of Association
Measures Of Association
 

Semelhante a Mantel Haenszel methods in epidemiology (Stratification)

Sample determinants and size
Sample determinants and sizeSample determinants and size
Sample determinants and sizeTarek Tawfik Amin
 
study design Case Control Study , sample size estimation.pptx
study design Case Control Study , sample size estimation.pptxstudy design Case Control Study , sample size estimation.pptx
study design Case Control Study , sample size estimation.pptxThe university of Punjab
 
observational analytical study
observational analytical studyobservational analytical study
observational analytical studyDr. Partha Sarkar
 
Case-control study un.uob.pptx
Case-control study un.uob.pptxCase-control study un.uob.pptx
Case-control study un.uob.pptxKifluKumera
 
unmatched case control studies
unmatched case control studiesunmatched case control studies
unmatched case control studiesMrinmoy Bharadwaz
 
UAB Pulmonary board review study design and statistical principles
UAB Pulmonary board review study  design and statistical principles UAB Pulmonary board review study  design and statistical principles
UAB Pulmonary board review study design and statistical principles Terry Shaneyfelt
 
Choosing appropriate statistical test RSS6 2104
Choosing appropriate statistical test RSS6 2104Choosing appropriate statistical test RSS6 2104
Choosing appropriate statistical test RSS6 2104RSS6
 
Dr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptx
Dr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptxDr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptx
Dr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptxPriyankaSharma89719
 
scope and need of biostatics
scope and need of  biostaticsscope and need of  biostatics
scope and need of biostaticsdr_sharmajyoti01
 
Imran rizvi statistics in meta analysis
Imran rizvi statistics in meta analysisImran rizvi statistics in meta analysis
Imran rizvi statistics in meta analysisImran Rizvi
 
Parametric tests seminar
Parametric tests seminarParametric tests seminar
Parametric tests seminardrdeepika87
 
2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...
2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...
2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...Cytel USA
 
Common statistical pitfalls & errors in biomedical research (a top-5 list)
Common statistical pitfalls & errors in biomedical research (a top-5 list)Common statistical pitfalls & errors in biomedical research (a top-5 list)
Common statistical pitfalls & errors in biomedical research (a top-5 list)Evangelos Kritsotakis
 

Semelhante a Mantel Haenszel methods in epidemiology (Stratification) (20)

Epidemiologic methods.pptx
Epidemiologic methods.pptxEpidemiologic methods.pptx
Epidemiologic methods.pptx
 
Sample determinants and size
Sample determinants and sizeSample determinants and size
Sample determinants and size
 
study design Case Control Study , sample size estimation.pptx
study design Case Control Study , sample size estimation.pptxstudy design Case Control Study , sample size estimation.pptx
study design Case Control Study , sample size estimation.pptx
 
observational analytical study
observational analytical studyobservational analytical study
observational analytical study
 
Case-control study un.uob.pptx
Case-control study un.uob.pptxCase-control study un.uob.pptx
Case-control study un.uob.pptx
 
Confounding.pptx
Confounding.pptxConfounding.pptx
Confounding.pptx
 
Confounding.pptx
Confounding.pptxConfounding.pptx
Confounding.pptx
 
unmatched case control studies
unmatched case control studiesunmatched case control studies
unmatched case control studies
 
UAB Pulmonary board review study design and statistical principles
UAB Pulmonary board review study  design and statistical principles UAB Pulmonary board review study  design and statistical principles
UAB Pulmonary board review study design and statistical principles
 
Screening of diseases
Screening of diseasesScreening of diseases
Screening of diseases
 
Choosing appropriate statistical test RSS6 2104
Choosing appropriate statistical test RSS6 2104Choosing appropriate statistical test RSS6 2104
Choosing appropriate statistical test RSS6 2104
 
Hypo
HypoHypo
Hypo
 
Dr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptx
Dr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptxDr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptx
Dr. RM Pandey -Importance of Biostatistics in Biomedical Research.pptx
 
Risk Comparison
Risk ComparisonRisk Comparison
Risk Comparison
 
scope and need of biostatics
scope and need of  biostaticsscope and need of  biostatics
scope and need of biostatics
 
Imran rizvi statistics in meta analysis
Imran rizvi statistics in meta analysisImran rizvi statistics in meta analysis
Imran rizvi statistics in meta analysis
 
Biostatistics.pptx
Biostatistics.pptxBiostatistics.pptx
Biostatistics.pptx
 
Parametric tests seminar
Parametric tests seminarParametric tests seminar
Parametric tests seminar
 
2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...
2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...
2014-10-22 EUGM | WEI | Moving Beyond the Comfort Zone in Practicing Translat...
 
Common statistical pitfalls & errors in biomedical research (a top-5 list)
Common statistical pitfalls & errors in biomedical research (a top-5 list)Common statistical pitfalls & errors in biomedical research (a top-5 list)
Common statistical pitfalls & errors in biomedical research (a top-5 list)
 

Mais de Rizwan S A

Introduction to scoping reviews
Introduction to scoping reviewsIntroduction to scoping reviews
Introduction to scoping reviewsRizwan S A
 
Sources of demographic data 2019
Sources of demographic data 2019Sources of demographic data 2019
Sources of demographic data 2019Rizwan S A
 
Effect sizes in meta-analysis
Effect sizes in meta-analysisEffect sizes in meta-analysis
Effect sizes in meta-analysisRizwan S A
 
Presenting the results of meta-analysis
Presenting the results of meta-analysisPresenting the results of meta-analysis
Presenting the results of meta-analysisRizwan S A
 
Heterogeneity in meta-analysis
Heterogeneity in meta-analysisHeterogeneity in meta-analysis
Heterogeneity in meta-analysisRizwan S A
 
Overview of the systematic review process
Overview of the systematic review processOverview of the systematic review process
Overview of the systematic review processRizwan S A
 
Biases in meta-analysis
Biases in meta-analysisBiases in meta-analysis
Biases in meta-analysisRizwan S A
 
Moderator analysis in meta-analysis
Moderator analysis in meta-analysisModerator analysis in meta-analysis
Moderator analysis in meta-analysisRizwan S A
 
Fixed-effect and random-effects models in meta-analysis
Fixed-effect and random-effects models in meta-analysisFixed-effect and random-effects models in meta-analysis
Fixed-effect and random-effects models in meta-analysisRizwan S A
 
Inverse variance method of meta-analysis and Cochran's Q
Inverse variance method of meta-analysis and Cochran's QInverse variance method of meta-analysis and Cochran's Q
Inverse variance method of meta-analysis and Cochran's QRizwan S A
 
Data extraction/coding and database structure in meta-analysis
Data extraction/coding and database structure in meta-analysisData extraction/coding and database structure in meta-analysis
Data extraction/coding and database structure in meta-analysisRizwan S A
 
Introduction & rationale for meta-analysis
Introduction & rationale for meta-analysisIntroduction & rationale for meta-analysis
Introduction & rationale for meta-analysisRizwan S A
 
Types of correlation coefficients
Types of correlation coefficientsTypes of correlation coefficients
Types of correlation coefficientsRizwan S A
 
Checking for normality (Normal distribution)
Checking for normality (Normal distribution)Checking for normality (Normal distribution)
Checking for normality (Normal distribution)Rizwan S A
 
Analysis of small datasets
Analysis of small datasetsAnalysis of small datasets
Analysis of small datasetsRizwan S A
 
A introduction to non-parametric tests
A introduction to non-parametric testsA introduction to non-parametric tests
A introduction to non-parametric testsRizwan S A
 
Kruskal Wallis test, Friedman test, Spearman Correlation
Kruskal Wallis test, Friedman test, Spearman CorrelationKruskal Wallis test, Friedman test, Spearman Correlation
Kruskal Wallis test, Friedman test, Spearman CorrelationRizwan S A
 
Kolmogorov Smirnov good-of-fit test
Kolmogorov Smirnov good-of-fit testKolmogorov Smirnov good-of-fit test
Kolmogorov Smirnov good-of-fit testRizwan S A
 
Use of checklists in critical appraisal of health literature
Use of checklists in critical appraisal of health literatureUse of checklists in critical appraisal of health literature
Use of checklists in critical appraisal of health literatureRizwan S A
 
Critical Appraisal of health literature
Critical Appraisal of health literatureCritical Appraisal of health literature
Critical Appraisal of health literatureRizwan S A
 

Mais de Rizwan S A (20)

Introduction to scoping reviews
Introduction to scoping reviewsIntroduction to scoping reviews
Introduction to scoping reviews
 
Sources of demographic data 2019
Sources of demographic data 2019Sources of demographic data 2019
Sources of demographic data 2019
 
Effect sizes in meta-analysis
Effect sizes in meta-analysisEffect sizes in meta-analysis
Effect sizes in meta-analysis
 
Presenting the results of meta-analysis
Presenting the results of meta-analysisPresenting the results of meta-analysis
Presenting the results of meta-analysis
 
Heterogeneity in meta-analysis
Heterogeneity in meta-analysisHeterogeneity in meta-analysis
Heterogeneity in meta-analysis
 
Overview of the systematic review process
Overview of the systematic review processOverview of the systematic review process
Overview of the systematic review process
 
Biases in meta-analysis
Biases in meta-analysisBiases in meta-analysis
Biases in meta-analysis
 
Moderator analysis in meta-analysis
Moderator analysis in meta-analysisModerator analysis in meta-analysis
Moderator analysis in meta-analysis
 
Fixed-effect and random-effects models in meta-analysis
Fixed-effect and random-effects models in meta-analysisFixed-effect and random-effects models in meta-analysis
Fixed-effect and random-effects models in meta-analysis
 
Inverse variance method of meta-analysis and Cochran's Q
Inverse variance method of meta-analysis and Cochran's QInverse variance method of meta-analysis and Cochran's Q
Inverse variance method of meta-analysis and Cochran's Q
 
Data extraction/coding and database structure in meta-analysis
Data extraction/coding and database structure in meta-analysisData extraction/coding and database structure in meta-analysis
Data extraction/coding and database structure in meta-analysis
 
Introduction & rationale for meta-analysis
Introduction & rationale for meta-analysisIntroduction & rationale for meta-analysis
Introduction & rationale for meta-analysis
 
Types of correlation coefficients
Types of correlation coefficientsTypes of correlation coefficients
Types of correlation coefficients
 
Checking for normality (Normal distribution)
Checking for normality (Normal distribution)Checking for normality (Normal distribution)
Checking for normality (Normal distribution)
 
Analysis of small datasets
Analysis of small datasetsAnalysis of small datasets
Analysis of small datasets
 
A introduction to non-parametric tests
A introduction to non-parametric testsA introduction to non-parametric tests
A introduction to non-parametric tests
 
Kruskal Wallis test, Friedman test, Spearman Correlation
Kruskal Wallis test, Friedman test, Spearman CorrelationKruskal Wallis test, Friedman test, Spearman Correlation
Kruskal Wallis test, Friedman test, Spearman Correlation
 
Kolmogorov Smirnov good-of-fit test
Kolmogorov Smirnov good-of-fit testKolmogorov Smirnov good-of-fit test
Kolmogorov Smirnov good-of-fit test
 
Use of checklists in critical appraisal of health literature
Use of checklists in critical appraisal of health literatureUse of checklists in critical appraisal of health literature
Use of checklists in critical appraisal of health literature
 
Critical Appraisal of health literature
Critical Appraisal of health literatureCritical Appraisal of health literature
Critical Appraisal of health literature
 

Último

❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...
❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...
❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...chandigarhentertainm
 
💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋
💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋
💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋Sheetaleventcompany
 
Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★
Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★
Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★indiancallgirl4rent
 
Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetNanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetCall Girls Service
 
❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR Call G...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR   Call G...❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR   Call G...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR Call G...Gfnyt.com
 
Hot Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In Chandigarh
Hot  Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In ChandigarhHot  Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In Chandigarh
Hot Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In ChandigarhVip call girls In Chandigarh
 
Call Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar Suman
Call Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar SumanCall Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar Suman
Call Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar SumanCall Girls Service Chandigarh Ayushi
 
❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...
❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...
❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...Gfnyt.com
 
Ozhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Ozhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetOzhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Ozhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetCall Girls Service
 
Call Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In Raipur
Call Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In RaipurCall Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In Raipur
Call Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In Raipurgragmanisha42
 
Dehradun Call Girls Service 08854095900 Real Russian Girls Looking Models
Dehradun Call Girls Service 08854095900 Real Russian Girls Looking ModelsDehradun Call Girls Service 08854095900 Real Russian Girls Looking Models
Dehradun Call Girls Service 08854095900 Real Russian Girls Looking Modelsindiancallgirl4rent
 
Call Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service Available
Call Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service AvailableCall Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service Available
Call Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service AvailableDipal Arora
 
pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...
pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...
pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...Call Girls Noida
 
Udaipur Call Girls 📲 9999965857 Call Girl in Udaipur
Udaipur Call Girls 📲 9999965857 Call Girl in UdaipurUdaipur Call Girls 📲 9999965857 Call Girl in Udaipur
Udaipur Call Girls 📲 9999965857 Call Girl in Udaipurseemahedar019
 
Call Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In Faridabad
Call Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In FaridabadCall Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In Faridabad
Call Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In Faridabadgragmanisha42
 
VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591
VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591
VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591adityaroy0215
 
raisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
raisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meetraisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
raisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetCall Girls Service
 
VIP Kolkata Call Girl New Town 👉 8250192130 Available With Room
VIP Kolkata Call Girl New Town 👉 8250192130  Available With RoomVIP Kolkata Call Girl New Town 👉 8250192130  Available With Room
VIP Kolkata Call Girl New Town 👉 8250192130 Available With Roomdivyansh0kumar0
 
❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF ...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF  ...❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF  ...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF ...Gfnyt.com
 
VIP Call Girl Sector 32 Noida Just Book Me 9711199171
VIP Call Girl Sector 32 Noida Just Book Me 9711199171VIP Call Girl Sector 32 Noida Just Book Me 9711199171
VIP Call Girl Sector 32 Noida Just Book Me 9711199171Call Girls Service Gurgaon
 

Último (20)

❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...
❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...
❤️Call girls in Jalandhar ☎️9876848877☎️ Call Girl service in Jalandhar☎️ Jal...
 
💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋
💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋
💚😋Mumbai Escort Service Call Girls, ₹5000 To 25K With AC💚😋
 
Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★
Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★
Enjoyment ★ 8854095900 Indian Call Girls In Dehradun 🍆🍌 By Dehradun Call Girl ★
 
Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetNanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Nanded Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
 
❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR Call G...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR   Call G...❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR   Call G...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Meghna Jaipur Call Girls Number CRTHNR Call G...
 
Hot Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In Chandigarh
Hot  Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In ChandigarhHot  Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In Chandigarh
Hot Call Girl In Chandigarh 👅🥵 9053'900678 Call Girls Service In Chandigarh
 
Call Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar Suman
Call Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar SumanCall Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar Suman
Call Girl Price Amritsar ❤️🍑 9053900678 Call Girls in Amritsar Suman
 
❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...
❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...
❤️♀️@ Jaipur Call Girl Agency ❤️♀️@ Manjeet Russian Call Girls Service in Jai...
 
Ozhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Ozhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real MeetOzhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
Ozhukarai Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
 
Call Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In Raipur
Call Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In RaipurCall Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In Raipur
Call Girl Raipur 📲 9999965857 ヅ10k NiGhT Call Girls In Raipur
 
Dehradun Call Girls Service 08854095900 Real Russian Girls Looking Models
Dehradun Call Girls Service 08854095900 Real Russian Girls Looking ModelsDehradun Call Girls Service 08854095900 Real Russian Girls Looking Models
Dehradun Call Girls Service 08854095900 Real Russian Girls Looking Models
 
Call Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service Available
Call Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service AvailableCall Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service Available
Call Girls Hyderabad Just Call 9907093804 Top Class Call Girl Service Available
 
pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...
pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...
pOOJA sexy Call Girls In Sector 49,9999965857 Young Female Escorts Service In...
 
Udaipur Call Girls 📲 9999965857 Call Girl in Udaipur
Udaipur Call Girls 📲 9999965857 Call Girl in UdaipurUdaipur Call Girls 📲 9999965857 Call Girl in Udaipur
Udaipur Call Girls 📲 9999965857 Call Girl in Udaipur
 
Call Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In Faridabad
Call Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In FaridabadCall Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In Faridabad
Call Girls Service Faridabad 📲 9999965857 ヅ10k NiGhT Call Girls In Faridabad
 
VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591
VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591
VIP Call Girl Sector 25 Gurgaon Just Call Me 9899900591
 
raisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
raisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meetraisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
raisen Call Girls 👙 6297143586 👙 Genuine WhatsApp Number for Real Meet
 
VIP Kolkata Call Girl New Town 👉 8250192130 Available With Room
VIP Kolkata Call Girl New Town 👉 8250192130  Available With RoomVIP Kolkata Call Girl New Town 👉 8250192130  Available With Room
VIP Kolkata Call Girl New Town 👉 8250192130 Available With Room
 
❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF ...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF  ...❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF  ...
❤️♀️@ Jaipur Call Girls ❤️♀️@ Jaispreet Call Girl Services in Jaipur QRYPCF ...
 
VIP Call Girl Sector 32 Noida Just Book Me 9711199171
VIP Call Girl Sector 32 Noida Just Book Me 9711199171VIP Call Girl Sector 32 Noida Just Book Me 9711199171
VIP Call Girl Sector 32 Noida Just Book Me 9711199171
 

Mantel Haenszel methods in epidemiology (Stratification)

  • 1. Mantel Haenszel Method Dr. S. A. Rizwan, M.D., Public Health Specialist (MOH), Saudi Board of Preventive Medicine Riyadh, KSA With thanks to Dr D. Hannoun (National Institute of Public Health, Algeria)
  • 2. Introduction Analytical studies in epidemiology aim to assess the association between two variables • Is the association valid? à RD – RR – OR • Is it causal? à Criterion of causality In most cases, we have to take in account a third (or more) variable that may affect the relationship studied • Confounding à bias • Effect modification (Interaction) à useful information
  • 3. Introduction Exposure Outcome Vaccine efficacy Measles Third variable • No effect: sex (boy/girl) • Intermediary: Antibodies rate • Confounder: Mother education • Effect modifier: Age VE is lower for children < 18mo VE is the same for boy and girl AR is a consequence of Vaccine Effect observed is affected by ME
  • 4. Introduction We can avoid these complications at two potential steps • Step one in the study design • Randomisation • Restriction • Matching • Step two in the analytical phase • Standardization • Stratification • Multivariate analysis Focus of this class
  • 5. Stratification: Principle Broad principle: • Create strata according to categories of the third variable • Perfom analysis inside these strata • Conclude about the relationship inside the strata • Forming adjusted summary estimate: i.e. weighted average • Assumption: weak variability in the strata (items within strata should be as similar as possible) What is achieved? • To analyse effect modification • To eliminate confounding
  • 6. Stratification: Principle To perform a stratified analysis, we have 6 steps: 1. Carry out simple analysis to test the association between the exposure and the disease and to identify potential confounder 2. Categorize the confounder and divide the sample in strata, according to the number of categories of the confounder 3. Carry out simple analysis to test the association between the exposure and the disease in each stratum 4. Test the presence or absence of effect modification between the variables 5. If appropriate, check for confounding and calculate a point estimate of overall effect (weighted average measure) 6. If appropriate, carry out and interpret an overall test for association
  • 7. Stratification: Step 1 – Example 1 Investigation of the relationship between Vaccine Efficacy and Measles (cohort study) 1. Crude analysis: Is there any association between vaccine efficacy and prevention of Measles? • RR = 0.55 [0.41-0.74] ; p < 0.001 à VE = 1-RR = 45% • There is an association between Vaccination and prevention of Measles Measles + Measles - Vaccinated 72 79773 No vaccinated 116 71039
  • 8. Stratification: Step 1 – Example 1 2. Identify potential confounder: • Is the association real and valid or could be modified when we take in account a third factor like age? • We are interested in how the effects of a third variable, age at vaccination, may be influencing this relationship
  • 9. Stratification: Step 2 – Example 1 Categorize the confounder and divide the sample in strata, according to the number of categories of the confounder 1. Number of categories of age : <1 year and 1-4 years 2. Create strata according to the number of categories <1 year Measles + Measles - Vaccinated 38 35587 Not Vaccinated 30 24345 1 - 4 years Measles + Measles - Vaccinated 34 44186 No vaccinated 86 46694
  • 10. Stratification: Step 3 – Example 1 Perform analysis inside these strata 1. In each strata • Calculate the X2 to test the association • Estimate the RRi/ORi <1 year Measles + Measles - Vaccinated 38 35587 Not Vaccinated 30 24345 1 - 4 years Measles + Measles - Vaccinated 34 44186 No vaccinated 86 46694 RRi = 0.87 [0.54 – 1.40], VE= 13% p = 0.55 RRi = 0.42 [0.28 – 0.62], VE= 58% p < 0.001
  • 11. Stratification: Step 4 – Example 1 Test for interaction by the third variable • Appropriate tests • Breslow Day & Mantel-Haenszel test: commonly used • Woolf test • Tarone å - = i i 2 i2 )var(effect effect)summary(effect Χ
  • 12. Stratification: Step 4 – Example 1 Test the presence or absence of interaction between the variables • Breslow-Day: Test of homogeneity in strata: • H0: RR1 = RR2 or OR1 = OR2 • Χ2 test compared observed and expected counts • It requires a large sample size within each stratum
  • 13. Stratification: Step 4 – Example 1 Test the presence or absence of interaction between the variables Two possibilities RR1 = RR2 or OR1 = OR2 RR1 ¹ RR2 or OR1 ¹ OR2 No Interaction: Third variable is Not an effect modifier Presence of Interaction: Third variable could be effect modifier Next step: Look for confounding and calculate adjusted measure Stop here: Results reported only by strata No pooled measure
  • 14. Stratification: Step 4 – Example 1 Test the presence or absence of interaction Homogeneity test: H0: RR<1year= RR1-4years (RR population) P <0.001 à statistical interaction present • There is interaction between age at vaccination and VE for Measles (or) • Age at vaccination modifies the effect of VE for Measles (or) • Age at vaccination is an effect modifier for the relationship between VE and Measles • Not appropriate to try to summarize these two effects, 0.87 and 0.42, into one overall number • We should report the two stratum-specific estimates separately and stop the analysis 0.87 ≠ 0.42
  • 15. Stratification: Step 1 – Example 2 Investigation of Effectiveness of AZT in preventing HIV seroconversion after a needlestick (case control study) 1. Crude analysis: Is there any association between AZT and prevention of HIV seroconversion after a needlestick injury in health care workers? • OR crude = 0.61 [0.26 - 1.44], p = 0.25 • No evidence of a benefit from AZT • The authors stratified by the severity of the needlestick HIV + HIV - AZT + 8 130 AZT - 19 189
  • 16. Stratification: Steps 2 and 3 – Example 2 Divide the sample into strata, according to the number of categories of the confounder and perform analysis 1. Categories of severity of needlestick : minor and major severity 2. Create strata according to the number of categories 3. In each strata test the association and Estimate the RDi/RRi/ORi Minor severity HIV+ HIV - AZT + 1 90 AZT - 3 161 Major severity HIV+ HIV - AZT + 7 40 AZT - 16 28 ORminor = 0.60 [0.06-5.81], p=1.0 No association ORmajor = 0.31 [0.11- 0.84], p=0.02 Presence of association
  • 17. Stratification: Step 4 – Example 2 Test the presence or absence of interaction between the variables • Test of homogeneity in strata : H0 : ORminor = ORmajor? • p=0.59 à Breslow-Day test is not significant à No statistical interaction • We assume there is no effect modification between severity of needlestick and AZT on the risk of HIV • We could try to summarize these two effects, 0.60 and 0.31, into one overall number à Construct a weighted average estimate • Go to Step 5
  • 18. Stratification: Step 5 – Example 2 If appropriate, check for confounding – Two steps » Calculating adjusted summary estimate » Comparing adjusted summary estimate to crude estimate
  • 19. Stratification: Step 5 – Example 2 If appropriate, check for confounding 1. Forming an adjusted summary estimate • Weighted average measure of the effect of exposure: RDi or RRi or ORi according to the size of each stratum • Weight depends upon a lot of factors: • Measure of association: RD or RR or OR • Nature of data: qualitative, quantitative • Purpose of the analysis: follow-up study, case control study • Methods: • Mantel-Haenszel • Woolf, Miettinen RR/OR
  • 20. RRa = wi RRi∑ wi∑ , wi = ci n0i ni å å å å = i i 1i1i i i i i 0i0i i i a n m*n c n m*n a RR Strata i of F Dis+ Dis - E + ai bi noi E - ci di n1i moi m1i ni Stratification: Step 5 – Example 2 If appropriate, check for confounding 1. Estimation of RRa: Follow up study
  • 21. Strata i of F Dis+ Dis - E + ai bi noi E - ci di n1i moi m1i ni Stratification: Step 5 – Example 2 If appropriate, check for confounding 1. Estimation of ORa : Case control study ORMH = S ai di S bi ci ni ni ORMH = S wi ORi / S wi wi = bi ci / ni
  • 22. Stratification: Step 5 – Example 2 If appropriate, check for confounding 2. Identify confounding • Compare crude measure with adjusted measure: • H0: RRMH=RRcrude (or) ORMH=ORcrude • No statistical test available • Confounding can be judged present when adjusted RRMH or ORMH is different from crude effect • D = (ORMH - ORcrude ) / ORcrude • Arbitrary cut-off: >10% • Interpretation
  • 23. Stratification: Step 5 – Example 2 If appropriate, check for confounding Two possibilities D < 10% D > 10% No confounding Presence of confounding Use RRcrude or ORcrude Use RRMH or ORMH
  • 24. Stratification: Step 5 – Example 2 If appropriate, check for confounding Be careful! We should report the adjusted measure: • Only if we haven’t detected interaction: RRi or ORi are homogenous among strata AND • If we have detected confounding
  • 25. Stratification: Step 5 – Example 2 Effectiveness of AZT in preventing HIV seroconversion after a needlestick in health care workers 1. Estimation of ORa adjusted ni = 255; OR = 0.60 ni = 92; OR = 0.31 Minor severity HIV+ HIV - AZT + 1 90 AZT - 3 161 Major severity HIV+ HIV - AZT + 8 40 AZT - 16 28 ORMH = 0.38 [0.14 – 0.87]
  • 26. Stratification: Step 5 – Example 2 2. Identify confounding • Compare the ORMH=0.38 with ORcrude=0.61 • D = (ORMH - ORcrude) / ORcrude = 44 % • D > 10% è We conclude that severity of needlestick is a confounder • After adjusting for severity of needlestick, we obtain a reduction of the magnitude of the relation between AZT and prevention of the HIV seroconversion • Conclusion: The good summary measure to use is the adjusted ORMH = 0.38
  • 27. Stratification: Step 6 – Example 2 If appropriate, carry out and interpret an overall test for association 1. Verify the relationship between exposure and outcome after adjusting for the third variable • H0: RRMH = 1 (or) ORMH = 1 • Statistical test à Mantel-Haenszel • It follows a chi-square distribution of 1 df, regardless of the number of strata 2. Confidence interval of adjusted RRa or ORa 2 MHχ 1,96 1 RR ± 2 MHχ 1,96 1 OR ± =
  • 28. Stratification: Step 6 – Example 2 Mantel-Haenszel Chi square test • It follows a chi-square distribution of 1 df, regardless of the number of strata • MH test statistic is defined as
  • 29. Stratification: Step 6 – Example 2 • Verify the relationship between AZT and HIV seroconversion after adjusting for severity of needlestick • H0 : ORMH = 1 • p = 0.036 à Mantel-Haenszel test is significant • Conclusion: • After adjustment for severity of needlestick, we have a significant association between AZT and HIV • When we adjust for severity of needlestick the OR decreased from 0.61 to 0.38 but also became significant (from p=0.25 to p=0.036)
  • 30. Confounding: Definition Be careful • Factor responsible for confounding is called a confounder or a confounding variable • Confounder factor confounds the association of interest: It confuses an estimate Examples • Needlestick severity confounds the effect of AZT in preventing HIV seroconversion
  • 31. Confounding: Definition When we have confounding: • The observed association between exposure and disease can be attributed totally or in part to the effect of confounder • Overestimation (+) of the true association between exposure and disease occurs: • Underestimation (-) of the true association between exposure and disease occurs: • Qualitative confounding: Direction of observed effect could change Crude effect > Adjusted Effect Crude effect < Adjusted Effect
  • 32. Confounding: How to identify confounder Compare: • Crude effect of association RD - RR - OR with adjusted measure of effect RDA - RRMH - ORMH How? • Take in account only D = (ORMH - ORcrude ) / ORcrude • If D >10% à Presence of confounding • If D <10% à No confounding Statistical test must be avoided to identify confounding
  • 33. Effect modification Variation in the magnitude of measure of effect across levels of a third variable • Tetracycline discolours teeth in children but not in adults Tetracyclines Age: children/adults • Effect modification is a concept, also called effect measure modification, interaction or heterogeneity of effect • Factor responsible for effect modification is called an effect modifier à it modifies the effect of exposure on the outcome Teeth discoloration
  • 34. Effect modification: Additive/multiplicative • For risk DIFFERENCE: Absence of interaction is RAAB = RAA + RAB Interaction is called Additive interaction OR • For risk RATIO: Absence of interaction RRAB = RRA X RRB Interaction is called Multiplicative interaction OR RDAB > RDA + RDB RRAB > RRA X RRB RDAB < RDA + RDB RRAB < RRA X RRB
  • 35. Effect modification: Properties Effect modification is not a bias but useful information • Identification of subgroups with a lower or higher risk • Targeting public health action • Better understand of the disease: biological mechanism
  • 36. Effect modification: How to assess it? Is there any statistical test to help us to assess effect modification? • Yes: many tests to verify the homogeneity of the strata • But not sufficient » Clinical/biological decision rather than statistical » Taking in account the magnitude of the effect modification » Statistical tests depend on the size of the study
  • 37. When to report effect modification?
  • 38. Effect modification & Confounding Effect modification • Belongs to nature • Rare • Effects in strata different • Must report stratum-specific estimates separately • Useful information • Controlled in the study design phase • Statistical test for interaction Confounding • Belongs to study • Frequent • Specific effects ≠ crude measure • Should report an adjusted weighted estimate • Distortion of effect: bias • Prevented in the study design and controlled in the analytical phase • No statistical test for confounding
  • 39. Effect modification & Confounding • Both confounding and effect modification • Must be interpreted and taken in account according to the knowledge of pathophysiologic mechanism • Determination is dependent on choice of effect measure: RD – RR – OR • Effect modification and confounding can exist separately or together
  • 40. An exercise Strata 1 Strata 2 Crude OR Adjusted OR Confounding EM 4.0 4.0 4.0 4.0 ? ? 4.0 0.25 1.0 1.0 ? ? 1.0 1.0 8.4 1.0 ? ? 4.0 0.25 1.0 2.0 ? ?
  • 41. Crude analysis Specific estimates not equal across strata Yes = Effect modification No = No effect modification Adjusted estimate not equal to Crude estimate Yes = Confounding No = No Confounding Report stratum- specific estimates – No pooled measure Report adjusted estimate, 95% CI, p value of χ2MH Report crude estimate, 95% CI, p value Stratification Specific estimates in each strata A strategy to check for interaction & confounding
  • 42. Take home messages • Stratification is a useful tool to assess the real effect of exposure on the disease • But, it has some limits: – Possibility of insufficient data when we have several strata – Tool developped only for categorical variable – Only possible to adjust for a limited number of confounders simultaneously • Advanced learning: Simpson’s paradox, non-collapsibility