A method of assessing the sensitivity of the Cochran–Mantel–Haenszel test to an unobserved confounder
2008

Assessing Sensitivity of the Cochran–Mantel–Haenszel Test

Sample size: 500 publication 10 minutes Evidence: moderate

Author Information

Author(s): Yu Binbing, Gastwirth Joseph L.

Primary Institution: National Institute on Aging

Hypothesis

How sensitive is the Cochran–Mantel–Haenszel test to unobserved confounders?

Conclusion

The study presents a method to assess the sensitivity of the CMH test to omitted confounders, indicating that maternal hypertension may not be causally related to low birth weight without considering race as a confounder.

Supporting Evidence

  • The CMH test is significant with p=0.015 and the odds ratio is 2.74.
  • The study indicates that the significant association between maternal hypertension and LBW might be due to hidden biases.
  • Race is identified as a potential unobserved confounder affecting the results.

Takeaway

This study looks at how hidden factors can affect the results of medical tests, showing that we need to consider things like race when studying health issues.

Methodology

The study uses sensitivity analysis to evaluate the robustness of the CMH test results against unobserved confounders, applying it to two datasets.

Potential Biases

Potential bias exists if unobserved confounders are not adequately accounted for in the analysis.

Limitations

The analysis relies on assumptions about the prevalence of unobserved confounders, which may not be accurate.

Participant Demographics

The study includes data from 500 singleton births, with a focus on maternal hypertension and low birth weight infants.

Statistical Information

P-Value

0.015

Confidence Interval

(1.23, 6.14)

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1098/rsta.2008.0030

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication