The study employed a 2-round Delphi study with a panel of young people who have a parent with a mental illness. The Delphi approach is a structured method to derive consensus on issues for which evidence is scarce. It involves a panel of experts to elicit their direction on the topic. In this study, the panel of experts was young people with lived experience of having a parent with a mental illness [33]. The approach guaranteed participants’ anonymity, thereby preventing possible biases. This is important when dealing with sensitive topics such as mental illness [33].

Young people residing in Australia, aged 16-21 years (inclusive), fluent in English, with a parent or parents with a mental illness (self-reported or diagnosed) were eligible to participate. Participants were from urban or regional settings. “Regional” was used to refer to nonurban participants in the study (ie, inner regional, outer regional, rural, remote, or very remote). Young people aged under 18 years required parental permission to participate. Participants were not required to participate in both rounds.

Recruitment for both rounds of the Web-based questionnaire was conducted via the researchers’ professional networks and through organizations designed for young people (eg, help lines or young people’s carers groups). The response rate is not reported, as the number of people who received the anonymous link is unknown. All Round 1 participants were invited to participate in Round 2. Participants were paid Aus $20 for their involvement in each round.

The study was approved by the Monash University Human Research Ethics committee. Web-based explanatory statements were provided to participants prior to each round. The questionnaire for Round 1 was based on the research questions and consisted of 22 questions in total, mainly of open-ended responses congruent with the Delphi method [35]. Questions related to different features of a Web-based intervention for young people who have a parent with a mental illness. Round 1 comprised of 14 open-ended questions related to intervention facilitation, interaction with peers and how this may function, preferences regarding communicating with a clinician, possible topics of interest, and safety issues and how such issues may be managed. The Round 1 questionnaire posed a closed question regarding whether a Web-based approach was a useful way to support young people who have a parent with a mental illness, along with an opportunity to provide reasons for a yes or no response.

The Round 2 questionnaire was developed from the themes identified in Round 1. It consisted of 39 questions related to the preferred structure of a Web-based intervention, peer interaction, whether the intervention should be time-limited, and preferred frequency and length of clinician-facilitated sessions. There were another 24 questions that required participants to rate the level of importance of the various features identified in Round 1 on an 11-point Likert-type scale from 0 (not important) to 10 (extremely important). These features related to the benefits of accessing a Web-based intervention, topics for Web-based intervention sessions, how Web-based interventions should be facilitated and monitored, internet safety, and issues around ethical issues (eg, informed consent). As discussed in preliminary data analyses, a consensus was reached after 2 rounds of the Delphi method.

Data were collected using the Qualtrics Web-based questionnaire package and analyzed using IBM SPSS Statistics version 24. Delphi studies typically attract similar, smaller numbers of participants in each round (eg, 12-30 participants). However, the Round 2 participant numbers far exceeded those of Round 1, thus providing an opportunity to examine potentially important differences in participant subgroups according to age, gender, and location.

The Braun and Clarke [36] 6-step analytic approach was employed to analyze Round 1 Delphi data. Themes were identified independently by 2 researchers. The researchers then compared and contrasted their interpretations of the main themes by referring back to participants’ responses, to reach a consensus through discussion rather than a numerical level of agreement.

The Round 2 analysis included a rank ordering of preferences, frequencies, and the use of analysis of covariance (ANCOVA) and chi-square test of independence to identify differences across gender and urban and regional locations. In apriori correlations, age was found to be positively correlated with almost all variables, with older participants scoring more highly than those younger. This was subsequently controlled for in the ANCOVAs reported in the results section. Prior to analysis, data were examined to ensure that assumptions pertaining to each of the tests were met, with no violations encountered. Due to the exploratory nature of the analyses, Bonferroni corrections were not adopted (the likelihood of this increasing the type 1 errors is acknowledged) [37], and given the large amount of data generated in Round 2, only the key findings are reported here.

The young people involved in this Delphi study (across age groups and locations) indicated a strong preference for a Web-based approach along with a range of preferred features. Developers should attempt to take these preferences into account when developing Web-based interventions for this target group. An especially important finding was young people’s preferences for a Web-based intervention that was facilitated by professionals. This preference has efficacy support; in a meta-analysis, Andersson and Cuijpers [38] found that Web-based interventions with professional support were much more effective than those without professional support in the treatment of adult depression.

Notwithstanding their preference for professional facilitation, young people preferred to have equal input during weekly sessions, with young people and professional facilitators sharing the role of directing discussions and contributing to session content. For example, a professional facilitator may allow young people to spontaneously direct discussions around their preferred content for up to half of the weekly session based on their needs or interest. Young people also indicated a preference for weekly, scheduled, 1-hour open discussion sessions, according to the particular needs of the group, as well as set topics. Hence, it appears to be important that Web-based interventions are sufficiently flexible to present established topics and be responsive to the immediate needs of young people participating in the intervention. The Kopstoring Web-based intervention has pre-established topics each week but does not appear to allow for open-ended discussion sessions [27]. Additionally, young people indicated a preference for weekly goals to work on between sessions. Others have also highlighted the importance of goal setting with these young people; one study found that through goal setting, young people who have a parent with a mental illness were more likely to achieve their personal goals, which included increasing their understanding of mental illness and improving family connections [39].

Psychoeducation was considered the most important topic to be covered by the intervention. In particular, young people identified a need to learn about their parent’s mental illness, how to support their parent, personal mental health, and how they might deal with situations arising from their parent’s illness. These findings respond to reports of young people’s limited understanding of their parent’s mental illness [40]. Psychoeducation is a key component of a range of interventions for families affected by parental mental illness [3,41] and resonates with previous findings regarding children’s requests to find out more about their parent’s illness [24,42]. Young people also wanted to learn adaptive coping strategies, which is important given the tendency for some young people in these families to employ coping strategies such as avoidance or isolation [43].

As well, young people indicated a preference for a time-unlimited intervention. If it had to be time-limited, young people preferred a 6-week Web-based intervention. Some Web-based interventions for this target group are offered weekly and are generally limited to 8 weeks [26,27]. Survivalkid, in the Netherlands, is unlimited but does not offer structured weekly topics [22]. Further research is required to ascertain the most effective length of Web-based interventions for this target group in terms of efficacy as well as participant satisfaction.

In Round 2, young people identified a need for referrals for additional support. Likewise, others have suggested that multiple interventions may need to be made available to this group of young people [44]. Empowering young people to access their own supports (rather than rely on their parents) potentially addresses the service gaps with identification, referrals, and fears around accessing help due to stigma [45]. Urban females scored higher than all others on referrals for support and wanting open access to chat rooms. This is not surprising given young female adults have reported themselves as their strongest help-seeking influence compared with other social influences, and this was higher for females compared with males [31]. Additionally, young people wanted the opportunity to privately chat with a Web-based facilitator. Survivalkid in the Netherlands has “survival coaches,” not therapists, who provide advice and may refer young people to more formal services if required [25]. Organizations would need to consider the viability of private contact given the time demands, staffing required, and costs involved.

Lastly, young people were asked to identify potential safety issues around providing a Web-based approach and invited to consider how this could then be managed. Young people wanted assurances that participants not adhering to guidelines would be banned or otherwise dealt with appropriately. Likewise, internet safety is an important consideration for young people when engaging in Web-based interventions [20,46]. Young people aged 16 and 17 indicated that they wanted the opportunity to participate in a Web-based intervention without having to obtain their parent’s consent, primarily because they were concerned about their parent’s reaction. Parents may act as “gatekeepers” for their children’s involvement in interventions [47], a concern aligned with the views of young people elicited here. Finally, it is a normal developmental phase for young adults to differentiate themselves from their parents [48], which again is well encapsulated in the preferences of young people in this study.

Generally, there were statistically significant main effects of location type between urban and regional young people. However, the practical significance [37] of this is limited given the small effect sizes and the observation that actual importance ratings for the different variables were still high for both urban and regional young people.

Several limitations apply in interpreting the findings. A smaller number of young people aged 16 and 17 years were recruited compared with the other age groups. Parental mental illnesses were self-reported by those surveyed rather than clinically diagnosed. It is unknown whether participants had their own mental health concerns, which may impact participants’ preferences for a Web-based intervention. The Web-based recruitment strategy might have favored those young people with an existing interest in and preference for Web-based supports. Future studies might differentiate between specific age groups to identify their respective needs for Web-based support, who might also be screened for their own mental health concerns. Consideration is also required to develop strategies for identifying young people at an earlier age. Working in collaboration with parents will be key here. Moreover, ongoing research is needed to investigate and evaluate the efficacy and effectiveness of Web-based interventions. The uptake of Web-based interventions also needs to be gauged, given that others have found that while young people prefer Web-based interventions, engagement is often low [20].

This study constitutes a step in the development of Web-based interventions for young people who have a parent with a mental illness. Young people indicated a preference for a time-unlimited, Web-based intervention that was professionally monitored but with input from both young people and professionals. The results of this study might be used to inform the development of Web-based interventions in regard to topics, intervention length and moderation, and safety requirements. The value in seeking the views of those with lived experience is vital in the design of support. Web-based interventions hold promise to prevent or reduce the risk of adverse outcomes in young people into adulthood and potentially reduce the intergenerational risks of parental mental illness on future generations.