And Psychologist Psychologist/Educational Psychologist Paediatrician Psychiatrist Audiologist Specialist teacher Charity representative Total doi:10.1371/journal.pone.0158753.t001 N and Nationality 32 (15 UK, 6 USA, 3 NZ, 3 Ire, 1 Can, 4 Aus) 7 (3 Can, 2 Aus, 2 UK) 8 (3 UK, 1 US, 3 Can, 1 Aus) 3 (3 UK) 2 (1 UK, 1 Can) 1 (1 NZ) 2 (2 UK) 4 (4 UK)and Snowling, who acted as moderators. These had been identified by the editor, Susan Ebbels, as representing a balance of academics and practitioners who could offer different perspectives on the topic of SLI. At this point, three individuals declined to take part–two of these judged they did not have sufficient expertise in children’s language disorders, and one did not think the Bayer 41-4109 site Delphi approach would be helpful. The moderators then scrutinised the preliminary panel and recruited further members to ensure diversity in terms of gender and ethnicity, and to include representatives of the main English-speaking countries, i.e., Australia, Canada, Ireland, New Zealand, United Kingdom and USA. All those invited were deemed to be experts in children’s language impairments. Summary characteristics of panel members are shown in Table 1. We also gathered information to establish the expertise of panel members: 54 members provided information on request and it was obtained from public sources for the remaining five. The mean number of years’ experience that panel members had working with children with language impairments was 24 (SD = 10.7). 51 panel members had provided training to others in their profession; 43 were involved in clinical or research advisory committees and 49 had published peer-reviewed papers or books relating to children’s language impairments. In addition, 10 had a child or close relative with a language impairment.Ethics approvalThis research was approved by The Medical Sciences Interdisciplinary Research Ethics Committee, University of Oxford (approval number: MS-IDREC-C1-2015-061). The Mdivi-1 mechanism of action committee approved for panel members to give written consent for their ratings to be used to derive a consensus statement.Delphi consensus processThe flowchart for the Delphi process is shown in Fig 1. A Delphi process typically starts with an open-ended brainstorming session to identify topics that can form the basis for an initial pool of statements to be rated. This step had in effect already been undertaken with the publication of target articles and commentaries in IJLCD. We also engaged in internet-based activities designed to encourage further debate and to involve more extensive networks of professionals, including a discussion forum set up by the Royal College of Speech and Language Therapists [28], and a Twitter debate moderated by the first author using the @WeSpeechies curated meeting point [29]. Statements for evaluation were taken from these articles, commentaries and internet sources. These were then circulated to a subset of panel members (shown with ?in the Consortium list in Acknowledgements) whoPLOS ONE | DOI:10.1371/journal.pone.0158753 July 8,5 /Identifying Language Impairments in ChildrenFig 1. Flowchart showing stages in the Delphi consensus process. S3, S4, S5, S6 refer to Supporting Information documents. doi:10.1371/journal.pone.0158753.gadvised on wording and added further suggested statements. On the advice of TG, who acted as adjudicator, we aimed for a pool of less than 50 items, and after further discussion between the moderators a potential pool of 76 items was pruned.And Psychologist Psychologist/Educational Psychologist Paediatrician Psychiatrist Audiologist Specialist teacher Charity representative Total doi:10.1371/journal.pone.0158753.t001 N and Nationality 32 (15 UK, 6 USA, 3 NZ, 3 Ire, 1 Can, 4 Aus) 7 (3 Can, 2 Aus, 2 UK) 8 (3 UK, 1 US, 3 Can, 1 Aus) 3 (3 UK) 2 (1 UK, 1 Can) 1 (1 NZ) 2 (2 UK) 4 (4 UK)and Snowling, who acted as moderators. These had been identified by the editor, Susan Ebbels, as representing a balance of academics and practitioners who could offer different perspectives on the topic of SLI. At this point, three individuals declined to take part–two of these judged they did not have sufficient expertise in children’s language disorders, and one did not think the Delphi approach would be helpful. The moderators then scrutinised the preliminary panel and recruited further members to ensure diversity in terms of gender and ethnicity, and to include representatives of the main English-speaking countries, i.e., Australia, Canada, Ireland, New Zealand, United Kingdom and USA. All those invited were deemed to be experts in children’s language impairments. Summary characteristics of panel members are shown in Table 1. We also gathered information to establish the expertise of panel members: 54 members provided information on request and it was obtained from public sources for the remaining five. The mean number of years’ experience that panel members had working with children with language impairments was 24 (SD = 10.7). 51 panel members had provided training to others in their profession; 43 were involved in clinical or research advisory committees and 49 had published peer-reviewed papers or books relating to children’s language impairments. In addition, 10 had a child or close relative with a language impairment.Ethics approvalThis research was approved by The Medical Sciences Interdisciplinary Research Ethics Committee, University of Oxford (approval number: MS-IDREC-C1-2015-061). The committee approved for panel members to give written consent for their ratings to be used to derive a consensus statement.Delphi consensus processThe flowchart for the Delphi process is shown in Fig 1. A Delphi process typically starts with an open-ended brainstorming session to identify topics that can form the basis for an initial pool of statements to be rated. This step had in effect already been undertaken with the publication of target articles and commentaries in IJLCD. We also engaged in internet-based activities designed to encourage further debate and to involve more extensive networks of professionals, including a discussion forum set up by the Royal College of Speech and Language Therapists [28], and a Twitter debate moderated by the first author using the @WeSpeechies curated meeting point [29]. Statements for evaluation were taken from these articles, commentaries and internet sources. These were then circulated to a subset of panel members (shown with ?in the Consortium list in Acknowledgements) whoPLOS ONE | DOI:10.1371/journal.pone.0158753 July 8,5 /Identifying Language Impairments in ChildrenFig 1. Flowchart showing stages in the Delphi consensus process. S3, S4, S5, S6 refer to Supporting Information documents. doi:10.1371/journal.pone.0158753.gadvised on wording and added further suggested statements. On the advice of TG, who acted as adjudicator, we aimed for a pool of less than 50 items, and after further discussion between the moderators a potential pool of 76 items was pruned.
