Reference no: EM132688462
Question 1:
The data below are from a large perinatal mortality study and show the survival of babies born to mothers who did and did not smoke during pregnancy. Low birth weight may be a confounder in the association between smoking and perinatal death. The data have already been stratified by birth weight.
Smoking status
|
Perinatal Deaths
|
Live births
|
Perinatal deaths
|
Live births
|
Smoker
|
378
|
1201
|
118
|
4,856
|
Non-smoker
|
209
|
494
|
118
|
4,414
|
We want to test if low birth weight (LBW) is a potential confounder by applying the three criteria.
Hint: Get organized/ Draw the confounding triangle; then for each part of the question, list the exposure, outcome, and make the 2x2 table.
a) Show that LBW is a risk factor for perinatal deaths in the total population, regardless of smoking status.
This part of the question is asking about the first confounding criteria: is the confounder (LBW) is a risk factor for the outcome (perinatal death)?
b) Show that smoking status is associated with LBW in the total population, regardless of perinatal death
This part of the question is asking about the second confounding criteria: is the confounder (LBW) is associated with the exposure (smoking)?
c) Based on your results above and the third confounding criteria, is birth weight a potential confounder in the relationship between smoking status and perinatal death? Justify your answer.
Question 2:
We conducted a case-control study to investigate a possible positive association between drinking diet soda and kidney disease. We selected our cases from patients hospitalized for kidney disease. We selected our controls from other hospitalized individuals who were diagnosed with obesity-related conditions. Obesity-related conditions have been shown to be positively associated with drinking diet soda.
By selecting these controls, will we likely overestimate or underestimate the true association between drinking diet soda and kidney disease? Or will this control selection have no impact? Justify your answer.
Hint: Draw the confounding triangle and then the 2x2 table for your exposure and outcome. Think about how each cell might be affected by the selection of these controls and what that would do to your calculation of excess risk.
Question 3:
In a previous problem set, we explored exposure to airborne pesticides and the risk of acute respiratory infection (ARI) among workers in a manufacturing plant in Kentucky. We conducted a one-year cohort study, but we were concerned that smoking might be masking the true relationship between pesticide exposure and ARI development.
Following the study, we confirmed the following:
(1) Smoking is a risk factor for ARI development
(2) Because more males work in the pesticide-exposed area and smoking prevalence is higher among males than among females in Kentucky, smoking is associated with pesticide exposure in our study
(3) Smoking is not on the causal pathway between pesticide exposure and ARI development
To deal with potential confounding by smoking, we have stratified the results of our analysis. Use these stratified results to answer the following:
a) Is smoking a confounder in the association between pesticides and the risk of ARI in our study. Justify your answer with the appropriate calculations.
b) Would our results be over- or underestimated if we fail to account for the confounding effect of smoking in our study? Explain in 1-2 sentences.
Non-smokers: Smokers:
|
ARI
|
No ARI
31
|
|
|
MI
|
No MI
|
High pesticide
|
3 7
|
High pesticide
|
19
|
39
|
Low pesticide
|
73
|
Low pesticide
|
7
|
15
|