There is effect modification of the OR by one or more of the stratification

There is effect modification of the or by one or more

This preview shows page 14 - 18 out of 30 pages.

 There is effect modification of the OR by one or more of the stratification variables  OR is heterogeneous across the strata  ORi≠ORj for some i, j [Test for Homogeneity] H 2 = i = 1 l [ ln ( ^ ¿ i ) ln ( ^ ¿ MH ) ] 2 ^ Var i ln ( ^ ¿ i ) x l 1 2 Pr [ x l 3 2 > ] = ¿ 0.05 H is distributed x 2 with I-1 degrees of freedom. There is not sufficient evidence in these data to reject the null hypothesis of homogeneity of the OR across strata.
Image of page 14
[Matching] Matching can be done in the design phase of a study: 1. Cohort studies: the exposed are matched to the unexposed group on one or more matching factors 2. Case-control studies: the cases are matched to the controls on one or more matching factors Matching is most useful in small studies with several confounding variables of nominal nature Individual matching = Frequency matching [Matched Cohort studies] Matching can be used: 1. To reduce/prevents confounding (even in crude/marginal analysis) 2. To remove bias 3. To prevent an association between the exposure and matching factors We can match on known confounders that are widely distributed in the population (sex) We can also match on factors that are relatively unique to a given study participant: sibship, genotype, multiple factors simultaneously (age, sex, ethnicity, neighborhood) We can evaluate the association between matching factors and the outcome of interest We can evaluate the modification by matching factors between exposure and outcome. [Matched Case-control studies] Improve precision Does not remove bias (matching should be followed by stratification) Confounders that are widely distributed in the population (sex) Factors that are unique to case and control: sinship, genotype, multiple factors (age, sex, ethnicity etc)
Image of page 15
Matching abolishes any association between the matching factors and the outcome of interest Matching => imbalance in # of cases and controls lead to a decrease in precision If matching factors are associated w/ the exposure => Distribution among control = cases (null value) The association between matching factor and the outcome cannot be evaluated (×) [Appropriate matching] Analysis Design Stratified Not stratidy Match Valid Max precision Bias Not match Valid, ↓precise Bias [Unnecessary Matching] Analysis Design Stratified Not stratify Match Valid Sl ↓ precision Valid Max Precision Not match Valid, Sl↓precise Valid Max Precision [Overmatching]
Image of page 16
Analysis Design Stratified Not stratify Match Valid ↓ precision Bias Not match Valid, ↓precise Valid Max Precision [Matching on an Intermediate] Design Stratified Not stratify Match Bias Bias Not match Bias Valid Max Precision [The Mantel-Haenszel x 2 test statistic (McNemar’s test)] H 0 : There is no association between the exposure and the outcome after accounting for the matching factors H A : There is an association between the exposure and the outcome after accounting for the matching factors Case/Contr ol E No E E f 11 f 10 No E f 01 f 00 Z 2 = [ f 10 f 01 ] 2 f 10 + f 01 x 1 2 , Pr [ x 1 2 > ] = ¿ f case ( 1: expose 0: unexpose ) ,control ( 1: expose , 0: unexpose ) We reject the null hypothesis at the 0.05 level of significance.
Image of page 17
Image of page 18

You've reached the end of your free preview.

Want to read all 30 pages?

  • Summer '14
  • FrancisCook

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern

Stuck? We have tutors online 24/7 who can help you get unstuck.
A+ icon
Ask Expert Tutors You can ask You can ask You can ask (will expire )
Answers in as fast as 15 minutes