71 Patient-relevant health outcomes: A delphi study 3 Based on the participant characteristics (i.e., a patient, caregiver or healthcare professional), the participant was directed to the specific survey questions. Before the first Delphi round, a long-list of health outcomes was developed based on the long-list used by the CoreHEM and Haemovalue initiatives on health outcomes for patients with hemophilia 12,19. These long-lists were enriched with the findings of a systematic review on patient-reported outcomes in patients with autosomal inherited bleeding disorders 14. A plain language definition was provided for each health outcome. The full Delphi process consisted of three rounds. During the first round, both panels were asked to rate each health outcome on the long-list using a fivepoint Likert scale, ranging from one (not important) to five (very important) in terms of each outcome’s importance for patients with autosomal inherited bleeding disorders. To capture uncertainty, an “I do not know” option was also available. Participants were able to motivate and add explanatory notes on their decisions, to provide recommendations on the definition and to add any additional health outcomes that they deemed important. All participants who filled in the informed consent were invited to complete the second and third round survey, regardless of their participation in previous rounds. In these rounds, both panels’ collective rating of each outcome’s importance was presented alongside the individual’s rating of the previous round. Participants were able to view the collective round rating as the percentage of participants or the absolute number of participants that rated an outcome on each of the options of the 5-point Likert scale. Additionally, any comments made by the participants in the previous rounds were displayed. This enabled participants to consider the opinions of other participants while deciding on the importance of the various health outcomes. Health outcomes were excluded from the second round if there was consensus on importance or consensus on unimportance during the first round, unless meaningful adjustments were suggested to the definition. Consensus on importance was defined as a median score of five and an interquartile range of ≤ 1. In other words, to achieve consensus on importance at least 75 percent of the panel had to score the health outcome with a five. The other 25 percent of the panel had to score the health outcome with a four. Consensus of unimportance was defined as a median score below five and an interquartile range of ≤ 1. To achieve consensus on unimportance no more than 25 percent
RkJQdWJsaXNoZXIy MTk4NDMw