and Widdershoven, 2020) The criteria were subdivided in three categories: diagnostic criteria, treatment criteria and treatment refusal criteria. First, the participants were asked to give their own definition of IPS. Next, they were asked their opinions on 20 criteria using a 5-point Likert-scale (strongly disagree, disagree, neutral, agree, strongly agree). The participants were encouraged to provide arguments for their ratings in an open comment section. Relevant socio-demographic and professional characteristics were also collected (table 1). Finally, participants were asked whether they would perform PAD for PPD themselves. The survey was validated in a pilot phase, during which three senior psychiatry residents from the Netherlands and Belgium filled out the survey in the presence of the corresponding author using the ‘think aloud’ approach. (Priede and Farrall, 2011) After round one, the open definition and accompanying comments for each criterion were coded and categorized using thematic analysis, paying particular attention to indications that the participant had misunderstood any elements of the criteria or desired more details. (Brady, 2015) The Likert-scales were analysed using basic descriptive statistics. Consensus was defined as 70% of the experts agreeing/ disagreeing or strongly agreeing/disagreeing with a statement (i.e. the top or bottom two options on the five-point Likert-scale). (Sumsion, 1998) The round one results were discussed in two project group meetings and summarized in a feedback report (supplement 1). The open definitions served as inspiration for additional criteria for round two. When the comments showed that criteria were misunderstood or valuable suggestions were given for wording changes, the criteria were modified and included in round two accompanied by a summary of the comments. If the comments lacked relevant arguments and no substantial changes were suggested, the consensus or dissensus about the criterion was accepted. The round two survey was piloted again on one of the senior psychiatric residents using the ‘think aloud’ approach. The results of round two were discussed in a project group meeting and summarized in a feedback report (supplement 2). Using the same standards as round one it was concluded that the wording was sufficiently clear for all criteria, arguments for agreeing or disagreeing were similar to round one and became repetitive, and no substantial new viewpoints were introduced by participants, indicating response stability. (von der Gracht, 2012) Therefore, the outcomes of round two criteria were accepted and no third round was performed. A DELPHI STUDY ON IRREMEDIABILITY | 101 6