Abstract

Psychometric validation can appear very specific and context sensitive. For the case of patient-reported outcome measures (PROMs), I argue that it falls short of securing the rigorous interpretation of scores, in part due to the different values of respondents and administrators in a new clinical context. I argue that validation practice is overly rigid. The lack of robust theories of constructs and response behaviour makes the standardized interpretation of scores unwarranted. Instead, attention to user values requires that validation practice remain dynamic, where the interpretation of scores must be collaborative. While some theories of psychometric validation attend to the ethics of measurement and the role of values, these are often addressed as bias. I discuss how values are an essential ingredient of validity evidence to secure the clinical interpretation of scores in the case of mental health. I substantiate my claims through my ongoing collaboration with a youth mental health service in British Columbia. There, the requirements of measurement-based care (MBC) and values like patient-centredness are in tension with the need to build and evaluate a learning health system and values like standardization. If the interpretation of scores can only be fixed in a clinical situation, it is not clear how such data can be collected and aggregated to support system level inferences. Different values must be traded off in a measurement practice, which in turn affects the quality of the inferences that different measurements may support for different purposes.