With unblinded sample size adaptations the usual estimates (e.g., the overall treatment difference) can be biased and the usual confidence intervals may not have correct coverage probabilities. Thus, when providing naïve (unadjusted) point estimates and/or confidence intervals in journals or reports, one must be aware of the poor behavior of these quantities. In adaptive designs, this is even more problematic because, e.g., the coverage probability of a fixed sample confidence interval can decrease dramatically, like the Type I error rate increases as described in Sect. 6.3.4. In this chapter, we discuss the construction of valid confidence intervals that have correct coverage probabilities. We will also discuss point estimates for adaptive designs that account for potential estimation bias. Specifically, we start by showing how to construct overall p-values for adaptive designs. In a sense, this is the generalization of the approaches discussed in Chap. 4 for the adaptive case. The focus of this chapter is on adaptive designs with a single interim analysis. This can be easily extended to the multistage case.