Lecture 23- CI RR OR Flashcards
Calculate the relative risk of stroke in the aspirin and placebo group on slide 428…
1.528 (working on slides)
How do we determine how good our estimate of RR is?
Calculate a confidence interval
What type of distribution is relative risk described as?
Very skewed
How do we get away with calculating a confidence interval for an RR when the distribution is skewed?
Take the log of RR because this is relatively normal
Calculate the 95% confidence interval for the RR 1.528? (original problem with contingency table is on slide 427)
0.96 to 2.43
What does it mean if the 95% confidence interval for a RR contains 1?
Because the interval includes 1 there is no evidence to suggest an elevated risk of stroke, when using aspirin. (remember relative risk of 1 means risk is equal in both groups).
When is a common case you use odds ratio not measures of relative risk?
When its a case control study because in this case the experimenter has chosen the number of cases and controls so relative risk is meaningless.
Calculate the odds ratio for the contingency table on slide 436 and then the 95% confidence interval + interpret the confidence interval…
-OR= 0.52 (2DP)
0.39 < OR < 0.69
We can be 95% sure that the interval 0.39 - 0.69 contains the true
OR. This means the true OR is likely to between 0.39 and 0.69.
The confidence interval is entirely below 1, indicating that people who
use mobile phones regularly are less likely to develop brain cancer
than those who don’t use mobile phones regularly.
(working on slides)
What does an observational study mean in terms of interpreting causality?
You can’t, the experimenter hasn’t directly manipulated anything and therefore can’t rule out all other possible confounders.
When calculating the CI for both RR and OR what do you have to be careful of at the end?
Take the antilog to convert back to original scale
