Well, I'm not really sure what to write in here, but I just feel the need for writing in here prior to starting my work.
Today, I did laundry, showered, dressed myself, made coffee and a snack, and now I'm sitting in front of my computer, readying myself to work until midnight. I was able to read the paper for my finals last night, it was an interesting read. It was about using hormone therapy to prevent the risk of breast cancer among women with the BRCA1 gene mutation.
So the researchers conducted a matched case-control study of 472 post-menopausal women with BRCA1 mutation, and they matched the women on age, age at menopause, and type of menopause. The cases here are women with the breast cancer and the controls do not have the breast cancer. The exposure is the hormone therapy. So if you plot the 2x2 table and compute for the odds ratio:
cases (w/ breast CA) control (w/o breast CA) total
w/ HT 47 68 115 (total exposed)
w/o HT 189 168 257 (total unexposed)
472 (total sample population)
Crude odds ratio: = (47x168)/(189x68) = 0.61 (an OR of 0.61 means that there is 0.61 times the odds of the breast cancer occurring in the group w/o HT. In short, there is a 39% reduction (1-0.61) of breast cancer occurring among women who are exposed to HT.)
The participants were stratified based on the type of menopause they have, the age at menopause, and the age of diagnosis. The adjusted ORs were computed for this to see if there is any confounding based on these factors. I haven't fully analyzed the paper to see if the researchers have found confounding and what they did to account for this, but I would assume they did some statistical adjustments (and they also did stratification) to account for confounding.
Okay, so I went back to the paper and looking at it, they did a multivariable regression analysis. This is a method that can be used to account/adjust for confounding. Based on the type of menopause, the adjusted and unadjusted ORs do not seem to differ greatly, indicating that the type of menopause is not a confounder. For age of menopause and age at diagnosis, adjusted and unadjusted ORs are also close to each other (basically a difference of 0.01), indicating that these two variables are not confounding the measures of association.
Anyway, these are just thoughts about what I read last night that are going on in my head. I need to get back to work now. At least I was able to think of this and write it down.
Until then,
Wish me luck!!!
