A dose-response curve is essential for determining the safe dosage of a drug and is widely used in bioassay and in phase 1 clinical trials. It is generally accepted that the probability of death or the probability of dose-limiting toxicity is a non-decreasing function of the dose. This paper proposes and develops an effective point-wise confidence interval for an isotonic dose-response curve, a problem without a satisfactory solution so far. We show that one subset of the observations is informative for the lower limit while the other subset informative for the upper limit. The property of the partial likelihood ratio using only informative observations is fully investigated. The resulting new confidence interval has the following advantages: it has a valid coverage probability regardless of the sample size, it is accurate as it is substantially narrower than the Morris (1988) specific interval, and it is applicable to different sampling plans of the doses. The new interval is extended to one-way ANOVA of a continuous response variable in a straightforward way. An R function for the proposed interval is provided.