Developing principles for sharing information about potential trial intervention benefits and harms with patients: report of a modified Delphi survey

Abstract

Background: The way information about potential harms of trial intervention is shared within participant information leaflets (PILs) varies widely and can cause subjective ‘nocebo’ harms. This study aimed to develop principles to improve the composition of information about potential trial intervention benefits and harms within PILs so that variability and avoidable harms are reduced. Methods: We conducted a two-round modified online Delphi survey, followed by a consensus meeting. For the first round of the survey, 27 statements were developed based on previous research and relevant guidance from the UK, the USA and the World Health Organization. Participants included members from each of the following stakeholder groups: patient and public representatives, research ethics committee members, industry representatives, medico-legal experts, psychologists and trial managers. Each participant was asked to rate their degree of agreement or disagreement with each statement on a 9-point Likert scale. In the second round, participants were invited to reappraise their ratings after reviewing the results of the first round. Finally, two members from each stakeholder group participated in a meeting to confirm those statements for which there was agreement. Results: Two hundred and fifty participants completed round 1, and 201 participants completed round 2. In round 1, consensus was reached for 16 statements. In round 2, consensus was reached for an additional three statements. The consensus meeting confirmed the survey results and consolidated the statements. This process resulted in seven principles: (1) all potential harms of a given intervention should be listed, (2) all potential harms should be separated into serious and less serious, (3) it must be made explicit that not all potential harms are known, (4) all potential benefits should be listed, (5) all potential benefits and harms need to be compared with what would happen if the participant did not take part in the trial, (6) suitable visual representations should be added where appropriate and (7) information regarding potential benefits and harms should not be presented apart by one or more pages. Conclusions: Our modified Delphi process successfully generated seven principles that can and should be used to guide how information is conveyed to patients in information leaflets regarding potential trial benefits and harms.