<p>Conditional on the existence of a single change-point, we assume the behavior of an individual’s survival time is governed by a piece-wise exponential distribution and test for homogeneity of the change-point across the population. We first describe a graphical diagnostic of this heterogeneity. We then present two distinct modeling approaches to test this homogeneity: first by modeling a smooth, continuous change in the hazard function and then again by examining heterogeneity in the change-point through the density function. We describe likelihood ratio tests of the null hypothesis all subjects experience the same change-point. This problem is motivated by the time to transplant failure times in a group of female renal patients. In a second example, we examine a marked increase from low risk of disease progression after five months following diagnosis in a database of women with breast cancer.</p>