Suppose a trial intends to compare the contraceptive effectiveness of two long-acting hormonal methods. Although both methods have previously been shown to be highly effective through 5 years of use, it’s also known that many women become dissatisfied with these methods and discontinue use within the first year. Why might survival analysis be a better method than logistic regression for comparing the 5-year pregnancy probabilities between these methods?

Question

What will be an ideal response?

Answer 1

Ans: Since many women will stop using their method before 5 years, survival analysis would allow researchers to include their data up until the point at which they discontinue their contraceptive method. Logistic regression would only include the pregnancy status for women still using the method at 5 years or those who became pregnant before that point.