Author(s): Guyatt GH, Kirshner B, Jaeschke R
Abstract Share this page
Abstract The different measurement properties necessary for instruments whose goal is to detect differences between subjects at a single point in time (discriminative instruments) and those whose goal is to detect longitudinal change within subjects (evaluative instruments) is becoming increasingly recognized. Up to now, requirements for evaluative instruments have been presented as reproducibility, validity, and responsiveness (i.e. the ability to detect change over time). An alternative conceptualization would characterize any instrument as requiring two crucial measurement properties. One is validity, the other a high ratio of signal to noise. For discriminative instruments, the signal to noise ratio can be summarized in a reliability coefficient; for evaluative instruments, in a responsiveness index or coefficient. This formulation can simplify and clarify the understanding and teaching of issues in health status measurement.
This article was published in J Clin Epidemiol
and referenced in Pediatrics & Therapeutics