Importance of consensus methods and involving end users to develop evidence-based materials when evidence is scarce: a Red Cross example
2Belgian Red Cross, Belgium
3Belgian Red Cross; Department of Public Health and Primary Care, Leuven Institute for Healthcare Policy, KU Leuven, Belgium
Up to 20% of children and adolescents worldwide experience mental health problems. Because young people find it easier to talk to their peers than to adult family members, teachers, or professionals, training them in how they can support each other is very important.
Objectives: The project aimed to develop materials to train children and youth in how to support their peers in the context of mental health problems. Because (very) low certainty evidence about the association between peer support and mental health problems was anticipated, the secondary objective was to develop a formal consensus method to collect good practice input from a broad group of stakeholders.
Methods: We systematically collected evidence about mental health risk or protective factors related to how children/adolescents interact with their peers (Q1), and about educational programs aimed at social skills and mental health (problems) (Q2). Next, a modified Delphi consensus method was developed and statements describing the potential content and characteristics of interventions, based on the identified evidence, were sent out to 23 stakeholders, including content experts and mental health peer workers. All statements for which consensus could not be reached were discussed in two face-to-face meetings.
Results: For Q1, we included 160 observational studies with very low certainty evidence about factors, such as peer support, peer influence, peer rejection/acceptance, friendship quality/intensity, and bullying. For Q2, we included 9 systematic reviews resulting in evidence conclusions of moderate to low certainty on the effectiveness of peer support programs. This resulted in 146 Delphi statements; for 77% of the statements, consensus was reached after a first Delphi round, the face-to-face discussions, and a second Delphi round. Forty-one statements were not withheld. The accepted statements were used to formulate the content and characteristics of peer support interventions.
Conclusions: Because the evidence about the content for our peer support program was of very low certainty, the input of various stakeholders, from academic experts to peer workers, was indispensable. To receive input from all profiles (including end users, to assure relevance for practice) and to obtain formal consensus, we successfully used a modified Delphi consensus method.