LAUSR

We are grateful for the constructive letter by Skolasky who sees our work as providing “a framework to advance the operational definition of response shift in such a way that allows for the generation of testable hypotheses about response-shift phenomena” [1]. According to Skolasky [1], a major strength of our work is the clarity of the definition where response shift is explicitly defined as an effect that occurs over time and that influences the measurement of change. With respect to the revised model, he acknowledges the useful distinctions between the construct and the measure as well as between the how and the why response shift occurs. This distinction is particularly important as it allows for linkages with theories about coping and adaptation that provide explanations for why response shift occurs. We agree with Skolasky [1] that such linkages with different theories and methodological approaches are critical to advancing the field of response-shift research. Skolasky [1] indicates that the paper about the healthcare implications of response shift for decision making was of particular interest to him. It is understandable that he felt the lack of concrete guidance on which response-shift method to use in which situation at which level of decision making to be disheartening. Although our paper provides empirical examples of implications of response shift at micro-, meso-, and macro-levels of healthcare decision making, Skolasky [1] importantly draws attention to the limited existing evidence and the need for further research to warrant useful and evidence-based guidelines. Skolasky [1] also points to a confusion in the responseshift literature where response shift may be considered on the one hand as a function of changes in appraisal or meaningful change and, on the other hand, as a source of bias or confounding variable. We are pleased that Skolasky [1] highlights our view that this confusion has impeded the advancement of response-shift research. Indeed, we believe that introducing this dichotomy is both unhelpful and misleading. There is no a priori “either or” as meaningful change and measurement bias are not mutually exclusive. We were, therefore, surprised to read that Rapkin and Schwartz [2] believe that we promote a perspective that response shift is defined as a form of bias in measurement, rather than meaningful change. This is a misunderstanding. We consider response shift as inherently meaningful. We, therefore, intentionally refrained from using the term “true change” (which is a commonly used term in the psychometric literature) in our series of papers because of its undesirable connotation in the response-shift field.The presence of response shift does not cause a change to be less “true.” However, our point is that response shift that is unaccounted for leads to measurement bias in that it threatens the validity of comparisons based on the measurement of change over time. Our inferences, decisions, and actions made on patient reported outcomes (PROs) must, therefore, consider the possibility that measurements of change over time that are based on responses to PRO measures (PROMs) may be influenced by response shift. The confusion may stem from the fact that Rapkin and Schwartz [2] seem to misrepresent the meaning of “measurement bias” by implicitly conflating it with distrust in people’s responses. They seem to also imply that “measurement bias” and “appraisal” are incommensurate perspectives. This is not the case. Information of response processes, which * Mirjam A. G. Sprangers [email protected]