The first aid guidelines were produced using: (a) a systematic search of the relevant evidence and claims made by authors of consumer and carer guides and websites; (b) development of a questionnaire on possible first aid actions which was based on the search; (c) and the consensus of panels of clinicians from each of the countries on which first aid actions should be included in the guidelines.
Systematic search for possible suicide first aid actions in the literature
As part of the project to develop suicide first aid guidelines for developed English-speaking countries, a systematic search for possible first aid actions was carried out in the formal professional literature listed in PubMed and PsycLit and other sources such as existing general mental health first aid manuals [5], other relevant manuals and guides on suicide prevention (e.g. Suicide Prevention Skills Training, [21]; Mental Health for Emergency Departments [22]) and relevant web sites (e.g. Samaritans). This method has been previously published for suicide first aid guidelines in developed English-speaking countries [8].
Construction of the questionnaire
A questionnaire was constructed from a content analysis of the actions indicated in the literature. Only statements that suggested a potential first aid action (i.e. what the first aider should do) or relevant awareness statements (what the first aider should know) were included in the questionnaire. These statements were grouped into common themes and used by a working group to generate questionnaire items specifying what actions a first aider should take. No judgments were made by the working group about the potential usefulness of the statements. Anything was included that fitted the definition of first aid, even if contradictory to other statements.
The questionnaire developed for English-speaking countries had 114 items, each describing a potential action that a first aider could do, which could be put to the panel for rating. These items covered the following broad areas: identification of suicide risk, assessing seriousness of suicide risk, initial assistance, talking with a suicidal person, no-suicide contracts, ensuring safety, confidentiality, and passing time during the crisis. The items are shown in Additional File 1. For the Asian guidelines, we added a few other items based on previous work on suicide prevention in Asian countries [17, 23]. The initial questionnaire contained 140 first aid action items, plus 13 questions on participants' socio-demographics, experience/training, and opinions on suicide first aid. Open-ended questions to generate additional culturally-specific items were also included. Given that this was an exploratory project, we used English-language questionnaires, because the cost of doing it in the experts' native languages would have been prohibitive.
Forming the expert panel
Mental health professionals, including psychiatrists, psychologists, social workers, counselors, and psychiatric nurses currently working in the Philippines, were recruited by DN, HM and EC (see Figure 1) to form the panel. The recruits were a mix of active practitioners from government, private, and non-government organizations providing psychiatric and other psychosocial services, although there was initial preference for active practitioners who were also connected with academia. In order to increase the cultural appropriateness of the guidelines, when forming the expert panels in each country we were careful to include as wide a representation (cultural and geographical) of professionals as possible.
When invitation letters, together with the Plain Language Statement, were sent (by email) to professionals asking them to be involved, they were also invited to nominate any colleagues who they felt would be appropriate panel members. During the recruitment process, potential participants were informed that one of the selection criteria was fluency in written English.
No attempt was made to make panels representative. The Delphi method does not require representative sampling; it requires panel members who are information and experience-rich [15].
The number of panel members in previous Delphi studies has varied considerably from 15 to 60 [24]. We aimed to have a minimum of 25 members in the panel.
Delphi process
In Round 1 of the Delphi process, panel members were asked to complete an on-line questionnaire. This was administered using SurveyMonkey [25], with the option to complete it by email or paper mail if this was not possible (although no participant opted for this alternative). The questionnaire consisted of a list of first aid actions to rate. Only actions that are do-able by mental health first aiders were included in the list of items to be rated. Members of the expert panel were given the following instructions to guide their judgments:
"The following questionnaire asks about the best way a member of the public can help someone who is thinking about, or planning to suicide. Mental health first aid is defined as help given to someone who is experiencing a mental health problem, or is in a mental health-related crisis, until professional help is received or the crisis resolves. It does not include counseling or therapy. In the case of suicide, mental health first aid is given until the person decides to accept professional help, or decides against suicide. People who offer mental health first aid may be friends, family members, colleagues or acquaintances. They may or may not be involved in the person's life before or after offering first aid. For brevity, we will refer to the person offering assistance as "the first aider". When completing this questionnaire, you will read statements describing possible actions that the first aider can take to assist a suicidal individual. You will be asked to rate how important each item is as a guideline for a first aider. Please rate as "essential" or "important" those items which you feel should guide most people, most of the time, when assisting a suicidal person. The statements in this questionnaire were derived from a search of both professional and lay literature in English-speaking western countries. Therefore, there will be actions which would be appropriate for members of the public in your country which are not included and there will be actions that may be appropriate in Western countries but not in your cultural context. At the bottom of each page, there is room for you to add suggestions. Please consider the cultural, social and religious environment where you live, and try to add some relevant suggestions on each page. The more panel members add to this questionnaire, the more relevant and useful the guidelines will be for each individual country. Thank you for taking the time to assist us in this important suicide prevention project!".
The definition of mental health first aid given to the panel in the above instructions distinguishes a first aider's role from that of a clinician. In the case of a suicidal person, the first aider responds by getting professional help for the person, and supporting the person and ensuring their safety until the crisis has passed or until the suicidal person is receiving professional treatment and care. The guidelines needed to focus on the immediate prevention of suicide and not on solving the problems that led to the crisis.
Panel members were asked to rate each statement according to how important they believed it was as a potential first aid action for helping a suicidal person. The response scale was: 1. Essential; 2. Important; 3. Don't know/Depends; 4. Unimportant; 5. Should not be included. The scale was purposefully asymmetric because only items with positive ratings were of interest for the guidelines. This scale has worked well in previous guideline development work [8].
At the end of each block of items the participants were asked to give any comments. In particular, they were invited to comment on items that were in the initial questionnaire that they considered to be culturally irrelevant or unacceptable, or that would not be feasible because of the local health system and other resources. Panel members were also invited to add any additional actions that were not included in the questionnaire. This was the place where culturally specific material could be introduced. The suggestions made by the panel members in response to the open-ended questions were reviewed by the research team and used to construct new items. Suggestions were accepted and added to Round 2 if they represented a new idea, could be interpreted unambiguously and were actions. Suggestions were rejected if they were near-duplicates of items in the questionnaire, if they were too specific, too general or were more appropriate to therapy than first aid.
Responses were analysed to give the percentage of the panel who rated an item as either "essential" or "important". Items for which there was at least 80% consensus were included in the guidelines. Items were re-rated if 70-79% of the panel rated them as "Essential" or "Important".
In Round 2, a second questionnaire was prepared. This consisted of items that were positively rated by at least 70% of respondents but did not reach the 80% criterion for inclusion and new items that were generated from the comments in Round 1. A small number of items that received more than 50% of "Don't know/Depends" or "Not sure" answers were reworded, clarified or specified, and re-rated. For example, "contact the person's spiritual or religious leader" became "contact the person's spiritual or religious leader, if they have one". Participants received an email with an individualized link to the online survey and a Word file that, together with the latter items, fed back a statistical summary of the items that were to be re-rated (i.e. their own original response to the item together with total percentages of endorsement of the item). They were told that they were able to change their responses when re-rating an item if they wished to do so. At the end of this round, any item that reached the 80% consensus criterion was selected for inclusion in the guidelines, those reaching between 70-79% of consensus were re-rated and the rest were rejected. This process was repeated for Round 3.
Ethics
Ethics approval for the study was obtained from the University of Melbourne Human Research Ethics Committee (Project No. HREC 0605537).