Appraising convergent validity of patient-reported outcome measures in systematic reviews: Constructing hypotheses and interpreting outcomes

Abstract

Purpose: Convergent validity is one type of validity that is commonly assessed for patient-reported outcome measures (PROMs). It is assessed by means of "hypothesis testing": determining whether the scores of the instrument under study correlate with other instruments in the way that one would expect. Authors of systematic reviews on measurement properties for PROMs may encounter validation articles which do not state hypotheses by which convergent validity can be tested. The information in these articles can therefore not be readily used to determine the adequacy of convergent validity. We suggest that in these cases, reviewers construct their own hypotheses. However, constructing hypotheses and interpreting outcomes is not always straightforward, and we wish to aid reviewers based on our own recent experiences with a systematic review on measurement properties. Recommendations: We have the following recommendations for authors of a systematic review on measurement properties who wish to construct hypotheses for convergent validity: take an active role in judging the suitability of the comparator instruments of validation articles; be transparent about which hypotheses were constructed, the underlying assumptions on which they are based, and whether they were constructed by the authors of the validation article or by the reviewer; discuss unmet hypotheses, especially if convergent validity is judged to be inadequate; and when synthesizing data, add up the results of all hypotheses for one instrument, rather than judging convergent validity per study.