The current good should precede the future best: A response to a response

We are pleased that Catania and colleagues (this issue) have noted the importance of behavioral surveillance for HIV/STD (sexually transmitted disease) prevention. As we indicated in the title of our article and reiterated in its final sentences, we consider our efforts to be a work in progress and feel that “input and feedback from the potential users of these questions are critical for the success of these endeavors” (Rietmeijer, Lansky, Anderson, & Fichtner, 2001). From that perspective, the response to our article provided by Catania and colleagues in this issue of AIDS Education and Prevention is not only very much appreciated but is also a necessary step toward reaching consensus in this area. With regard to using the last sexual encounter as a critical event on which to frame a number of sexual behavior questions, we do not disagree with many of the hypothetical concerns Catania and colleagues raise. We agree that further methodological studies are needed to develop the best set of questions for HIV/STD-related behavioral surveillance. It is important to point out, though, that we consider standard questions as a foundation for behavioral surveillance, a key tool for public health practice, whereas Catania and colleagues see the standard questions from a research perspective. In accordance with our perspective, we have used Webster’s definition of standards as “something established for use as a rule or basis of comparison in measuring” (Guralnik, 1972) rather than “standard of scientific quality” as Catania and colleagues suggested. Standard measures are at the core of all public health surveillance systems, as they are critical for information exchange within and between jurisdictions and systems (Morris, Snider, & Katz, 1996). Data collected through surveillance are used as a basis for making population estimates. Thus, in the context of behavioral surveillance for HIV/STD prevention, the extent to which an event is characteristic of any individual’s behavior over time and across partners is less important than the ability to generalize to or characterize a population; events that are uncharacteristic of an individual will even out over a population. Condom use at last sexual encounter as a measure in the Youth Risk Behavior Survey (YRBS) is a good example. An analysis of the 1991-1999 YRBS data (Centers