Conversion of standard retrospective patient-reported outcomes to momentary versions: cognitive interviewing reveals varying degrees of momentary compatibility

Victor Brun Boesen, Stine Birk Nissen, Mogens Groenvold, Jakob Bue Bjorner, Laszlo Hegedüs, Steen Joop Bonnema, Åse Krogh Rasmussen, Ulla Feldt-Rasmussen, Torquil Watt

The purpose of this study was to adapt different domains of an existing retrospective questionnaire to momentary versions, to use and assess cognitive interviewing for evaluating the new versions, and to compare momentary compatibility (i.e. an item's potential to be validly converted to a momentary version) across different scales. Initial momentary versions of retrospective patient-reported outcomes were produced by converting present perfect tense wording to present tense wording. Cognitive interviews were conducted iteratively with 21 patients to determine which reference period they actually employed, and to identify problems with new, revised versions. A think-aloud interview protocol was supplemented with non-specific concurrent and specific retrospective probing. The momentary compatibility of each item was evaluated by calculating the proportion of interviews wherein momentary reference periods were identified; problems were categorized according to cognitive aspects of survey methodology taxonomy. The efficiency of various cognitive interviewing techniques was determined by evaluating whether applied reference periods were identified by think-aloud alone or by supplementary probes. The momentary compatibility varied from 5 to 100% across items. Cognitive interviews revealed potential problems of various severities in the majority of items. Think-aloud alone was sufficient at determining the applied reference period in one-third of the cases, and the efficiency of additional concurrent and retrospective probing was 50 and 94%, respectively. Cognitive interviewing techniques proved useful for developing and evaluating momentary items. Researchers should be aware of the applied reference period and of emerging problems when evaluating adapted momentary items, since not all concepts are suitable. We recommend the proposed method in future adaptations of existing instruments.