The following are hypothetical data from a case‑control study of oral contraceptives (OC) and cervical cancer. You are interested in determining the potentially confounding effect of education, dichotomized in your data as “low” (< 12 years) and “high” (12+ years):
Table 1: Strata 1: low education level (< 12 years)
|
OC use |
Cervical cancer |
Control |
Total |
|
Yes |
45 |
9 |
54 |
|
No |
47 |
16 |
63 |
|
Total |
92 |
25 |
117 |
Table 2: Strata 2: high education level (12 + years)
|
OC use |
Cervical cancer |
Control |
Total |
|
Yes |
26 |
42 |
68 |
|
No |
12 |
33 |
45 |
|
Total |
38 |
75 |
113 |
A. to evaluate “education” as a confounder based on the criteria of confounder (i.e. strategy 1 in lecture) (4 points)
- Step 1: evaluate the association between confounder (education) and exposure (OC use). Please complete the following table using the above Table 1&2 and calculate the OR between education and OC use (1 pts)
|
Education |
OC use |
|
|
Yes |
No |
|
|
high |
||
|
low |
||
- Step 2: evaluate the association between confounder (education) and disease (Cervical cancer). Please complete the following table using the above Table 1&2 and calculate the OR between education and Cervical cancer (1 pts)
|
Education |
Cervical cancer |
|
|
Yes |
No |
|
|
high |
||
|
low |
||
- If we assume that “education” is not an intermediate factor in the causal pathway of the relationship between OC use and cervical cancer, could you determine if education is a confounder based on the results from part i and ii of A? If yes, what is the direction of confounding of education (negative, positive or qualitative) based on the associations between education and OC use and between education and cervical cancer from part i and ii of A? (2 pts)
B. to evaluate “education” as a confounder based on the method of stratification
- Step 1: evaluate the crude association between OC use and cervical cancer (e. without stratified with education). Please complete the following table using the above Table 1&2 and calculate the OR between OC use and cervical cancer
|
OC use |
Cervical cancer |
|
|
Yes |
No |
|
|
Yes |
||
|
No |
||
- Step 2: evaluate the association between OC use and cervical cancer in each stratum of education; i.e. please calculate the OR between OC use and cervical cancer in each stratum of education respectively.
OR strata-low=
OR strata-high=
- Based on the results of Part i and ii of B, do you think that “education” is a confounder in study of the relationship between OC use and cervical cancer? Why or why not?
