What factors influence successful recruitment of siblings of individuals with first episode psychosis to e‐health interventions? A qualitative study

Abstract Background Recruitment to clinical research studies can prove complex. This is particularly true of mental health research, given factors such as confidentiality, capacity and consent, or when attempting to recruit family members as opposed to service users themselves. Aim This study investigated the challenges experienced and strategies employed in the recruitment of siblings of people with first episode psychosis using Early Intervention in Psychosis Services (EIPS) in England. Methods As part of a randomized controlled trial (RCT) of an e‐health intervention for siblings, we conducted a process evaluation study whereby semistructured interview was undertaken with clinical and research staff involved in recruitment of siblings. Data were analysed thematically. Results Twelve participants from six EIPS were interviewed. Data analysis revealed seven key themes: (i) limited comprehensive family data available; (ii) data governance and consent issues; (iii) organizational factors; (iv) convoluted recruitment methods; (v) concerns about service users' opinions; (vi) fluidity in siblings' needs and expectations; and (vii) strategies to enhance recruitment. Conclusions Recruitment challenges identified in this study concerned administrative, organizational, process and attitudinal issues. These are similar to other studies recruiting mental health service users as well as family members. Failure to recruit to target implies that studies are underpowered to detect potential statistically or clinically meaningful changes. Future studies should establish how best to enhance family inclusiveness in clinical practice and research.


| INTRODUCTION
Siblings of individuals with first episode psychosis (FEP) are vulnerable to developing mental health problems, partly due to the impact of psychosis on individuals within the family, and the wider network. [1][2][3] However, siblings also have an important role in supporting service users' recovery and enhancing prognosis. 4,5 Recent evidence has indicated that siblings can benefit from information about psychosis, as well as strategies to enhance coping, and opportunities for peer support. 3,[5][6][7] Hence, the E Sibling Project was developed to provide a FEP sibling-specific intervention, using an Internet-based medium for flexible access and individualized package. 6 The study described here formed part of the process evaluation for the E Sibling Project Randomised Control Trial (RCT) (See also Sin et al., 6,7 for information regarding development and usability establishment of the intervention). The E Sibling Project RCT recruited siblings (either biologically related, step-or half-siblings or related through adoption) of individuals who were using Early Intervention in Psychosis Services (EIPS) in England, between September 2013 and March 2015. EIPS typically provides comprehensive care using an assertive outreach approach for people aged between 18 and 35, experiencing FEP. 8 EIPS commonly promotes family inclusiveness, such as through providing family intervention, and enhanced psychoeducation for family carers. 8,9 The RCT aimed to recruit 144 siblings, in 9 months (September 2013 to April 2014) originally; the sample size was estimated in order to ensure there was sufficient statistical power to detect treatment effect. At the time of undertaking this process evaluation study (April to July 2014), siblings were recruited from 16 EIPS across England, with a combined caseload of approximately 5000 service users. We were mindful that there could be challenges in recruiting to target, due to uncertainty experienced by family members in engaging with mental health services, 10,11 stigma about being involved, 12 or competing demands. [13][14][15][16] Hence, we had set a conservative recruitment rate based on previous regional surveys, 1,2 suggesting that 90% of EIPS service users would have one sibling or more. We hypothesized that at least 50% of siblings would meet the RCT eligibility criteria (living in England, aged 16 or above, and in weekly or regular contacts with service users), based on consultation with Principal Investigators (PIs) and EIPS leads in five recruitment sites as part of our initial trial setting up preparation. 6 Our sampling frame comprised 2250 siblings. Our recruitment procedures involved (i) potentially eligible siblings being identified by local clinicians or research staff in each participating EIPS, (ii) siblings and/or service users and their named carers being given study information from the clinicians or researchers, (iii) then siblings could self-refer or be referred through the local personnel to join the E Sibling Project. Study information materials (localized for each participating EIPS) were devised for direct communication with siblings if they were already known to the service, as well as for informing service users or named carers who were often parents, in order to ask them to help pass the study information to siblings. These recruitment procedures were devised to ensure that service users were informed about the study as is good practice even though they were not required to participate in the study nor to give consent for their siblings to join the study (of note, ethical approvals were in place for such). 6,7 By April 2014, 9 months after the RCT commenced and at the end of our original recruitment period, 58 siblings (40% of the total required) had been recruited. We instigated some strategies as contingency plan to overcome the recruitment shortfall. These included: extending the recruitment period to March 2015; increasing the number of recruitment sites; and undertaking this process evaluation study with clinical and research staff involved in the recruitment of siblings.
Understanding factors associated with recruitment, including those that facilitate as well as impede recruitment, is important in order that intervention trials are feasible. To explore the recruitment strategies, and factors that potentially mediated successful recruitment, this study investigated "reach" and "recruitment," using the components of process evaluation as outlined by Steckler & Linnan. 17 Reach was defined as the extent to which participants were made aware of RCT recruitment, either directly or indirectly (eg via other family members or service users); 17 investigation of recruitment included examining intended and actual recruitment procedures, as well as practical difficulties experienced in recruiting participants.

| Study design
Qualitative semistructured interviews were undertaken with clinical and research staff involved in recruiting siblings into the RCT. Also, we calculated the number of service users per EIPS caseload, the proportion of service users known to have siblings, and the number of siblings offered information about the study. These figures were compared with the number of individuals who consented to take part, so as to estimate reach of the recruitment activities and service recruitment rates.

| Regulatory approvals
Approvals were granted by the National Health Service (NHS) Research Ethics Committee (Reference: NRES 12/LO1537), and by local Research and Development (R&D) departments at participating NHS Trusts. All participants provided written informed consent.

| Participants
We recruited clinical and research staff, including local PIs who were NHS senior clinicians or clinical academics involved in leading local recruitment; or Research Assistants (CSO/CSA) funded by the Clinical Research Network: Mental Health (CRN: MH) to support recruitment from mental health services, 18,19 and clinicians who worked at EIPS.
As the first step to identify the potential participants, we informed the PIs in all the recruitment sites about the process evaluation study.
Second, we asked for a list of clinical and research staff who were involved in recruiting siblings locally. Lastly, we circulated the study flyer and participant information sheet to all the potentially eligible participants.

| Data collection
A topic guide (available from the first author) was developed and piloted, in consultation with the E Sibling Project Sibling Reference Group (The SRG comprised five siblings to provide oversight for the project design and conduct) 20 and the West Midlands Carer Reference Group (The CRG is a Department of Health-funded carer group which meets monthly to give comments on studies focusing on mental health informal carers in England). 21 Questions were designed to prompt participants about their experiences, successful or otherwise, in identifying, approaching and recruiting siblings. Interviews were conducted in person or by phone, depending on participant's preference. First, initial themes were identified by "indexing" the transcript; these themes guided the formation of a framework within which transcribed material was summarized. Second, key categories were identified to describe the data. Finally, patterns of association were sought. Data analysis was performed in conjunction with data collection, in order that the developing framework could be compared, contrasted and validated until data saturation was reached. 23

| Participants' professional characteristics and recruitment experience
Four PIs, four CSOs/CSAs and four clinicians (total n=12), working in six EIPS, were interviewed (See Table 1). Interviews were audio-recorded, and conducted by phone (n=6) or in person (n=6).
With the exception of three participants, there was consensus that clinical research regularly took place in EIPS, with staff recruiting to approximately 10 to 15 studies during the past three years. Studies predominantly included EIPS service users or staff; family members, including siblings, were seldom recruited. While family-focused research took place occasionally, participants reported that the emphasis was on recruiting service users. Conversely, family members were approached primarily to provide informant ratings of service users' symptoms, and occasionally about their own health outcomes.

| Themes
Seven key themes were identified in the analysis, which concerned the process and challenges encountered with respect to recruitment of siblings of service users with FEP into the E Sibling Project RCT.
These were as follows: (i) limited comprehensive family data available; (ii) data governance and consent issues; (iii) organizational factors; (iv) convoluted recruitment methods; (v) concerns about service users' opinions; (vi) fluidity of siblings' needs and expectations; and (vii) strategies to enhance recruitment. See Table 2.

| Limited comprehensive family data available
There was unanimous consensus between participants that in each service, there were limited data about the family structure and relationships between service users and their immediate family members (ie the degree to which service users were in contact with family members). Furthermore, even when named carers were identified, more often than not, their contact details were not recorded or incomplete.
Additionally it was noted that most carers were parents. Hence, even when CSOs/CSAs were able to access electronic records to screen for potentially suitable individuals, there were no straightforward ways to identify siblings. This difficulty is illustrated by one participant:

I feel like I'm forever chasing up the information (from clinicians) that just doesn't get back to you so I would try to
do what I can through the information on RiO (a common electronic record system used in NHS) but again that's very limited a lot of the time so it's not been easy in that aspect.

| Data governance and consent issues
Several participants, including EIPS clinicians and CSOs/CSAs, described ambiguity in Trust data governance policies regarding T A B L E 1 Summary of participants' professional characteristics

CSOs/CSAs Clinicians
Length of time in current positions 2-10 years 5 months-3 years

| Convoluted recruitment method and unintended selective sampling
We contacted the local research and clinical staff on a monthly basis to check the number of potentially suitable siblings and/or service users (or the named carers) who had been identified and any of them were given information about the study. We also offered to follow up on any potentially interested participants to answer any queries

| Concerns about service users' opinions and responses to study information
Participants raised understandable concerns about how service users might think and feel about their siblings being recruited to a study in which they were not involved. Also, it was noted that some service users were viewed as vulnerable, or floridly symptomatic (eg paranoid or delusional), which meant that clinicians did not wish to mention information that could be misconstrued or potentially exacerbate suspiciousness. Several participants mentioned that service users could seem sensitive to changes in their routines, for example, meeting unfamiliar researchers, including the CSOs/CSAs.
Consequently, it transpired that clinicians could tend to avoid conveying information about the study out of the intention of not adding unnecessary pressure onto the service users. One CSO described his observation:

| Fluidity in siblings' needs and expectations
In general, study participants perceived family members of individuals with FEP to have needs that were distinct from those of individuals who have longer term difficulties. For example, family members typically had had no or minimal contact with mental health services, and as such were less familiar with services available, and unclear about how best to advocate for their needs. It was reported that issues associated with diagnostic ambiguity and prognostic fluidity could affect recruitment to research as well as engagement with the service. Participants identified that many of the EIPS carers (and siblings) might also be preoccupied with dealing with crises which arose with the FEP, therefore would often prioritize their attention on the service users rather than themselves. Some might be adopting a waitand-see or a so-called sealing over approach hoping that it was an one-off episode and that they would not need further input from the mental health service beyond the current episode. The paradox was highlighted by one EIPS clinician: Overall, parents were described as seeming positive about the study when they were asked to pass on information to their other children. Yet, several clinicians stated that a minority of parents appeared suspicious and less amenable to being involved in recruitment. Clinicians believed that this might be because parents might be experiencing a sense of denial about their child's illness, or they might be trying to shield their other children from having any contact with the mental health service.
In comparison, while some siblings were identified by clinicians as being potential study participants, and keen to take part when first given the information, it transpired that a significant proportion did not pursue this further, which was reported to be in part attributed to the demands of everyday life. One participant offered her observation:

| Strategies to enhance recruitment
All participants mentioned strategies to enhance integration of research activities into routine clinical practice. Infrastructural factors such as enhancement of cohesive working relationships between research and clinical staff and, potentially, direct remuneration to teams involved in research activities, were suggested. Having a clinician identified as a research champion within each team was also deemed good practice, whereby they could link with the PI as well as CSOs/ CSAs, and R&D. One clinician elaborated on this:

| DISCUSSION
Given the multilayered recruitment process, and various obstacles in clinical areas. 14,15,[28][29][30][31][32][33] Study participants also outlined some mental health specific challenges. A general impression was conveyed that mental health clinicians can seem protective of their caseload and may hold pessimistic assumptions about service users' capacity and motivation to become involved in research activities, a finding which has been reported elsewhere. 14,15,28,30 For example, one recent study investigated recruitment rates of mental health service users to trials found that only 17% (n=131/752) of potentially eligible service users were approached by clinicians. 16 This potentially implies that clinicians may feel reticent to approach service users, thereby indirectly influencing recruitment to research, before service users are even made aware of the possibilities of involvement in research activities. 16  and also, more pressing demands, for example, needing to manage and cope with social and economic constraints or illness in other family members. While these challenges have mainly been investigated in the context of parent carers of FEP service users, 10,11,13 it seems pertinent to extrapolate these findings to siblings, given that they may be attempting to attain independence from the family network, yet still heavily involved in supporting their sibling and family. 1,3,5 Strengths of this study include that the process evaluation coincided with the active recruitment to the RCT, thus enabling participants to reflect on their practice and the current recruitment processes employed. Also, we actively sought to recruit staff from different disciplines, including PIs, clinicians and CSOs/CSAs, and from broad-ranging recruitment sites, which facilitated an insight and multiple perspectives about this topic. However, we also acknowledge several study limitations, including the relatively small sample size and the lack of representation of staff from all recruitment sites. Despite data saturation being reached, it is possible that staff working at other sites may have had additional opinions and experiences different from results reported in here were not included. Exploring the views of siblings who participated in or declined to join the E Sibling RCT could have augmented the study findings, although this was not the remit of this study.

| Clinical and research implications
Several strategies were identified which can enhance reach and recruitment of siblings to research. Arguably, these strategies are equally applicable in terms of boosting family inclusiveness in routine clinical settings. We suggest that undertaking a comprehensive family assessment, such as via the use of a genogram so as to map out familial networks and relationships, would be useful when the service user is first referred to EIPS. 34,35 Obtaining contact details for relevant family members including siblings seems warranted, consent-permitting, partly so that they can be provided with information about services or means of accessing support. It may be that siblings (and family members in general) would benefit from health promotion activities, even offered via primary or voluntary sector where they are regarded as consumers in their own right. 35 In terms of research, it is likely that specific family training with reference to good practice guidance in involving family members in research could be useful. 21 PPI endeavours could help to optimize that study information is communicated in a manner that takes into account the competing demands siblings have, and to enhance the reach and recruitment of studies. 19,22 Finally, issues pertaining to data protection and governance require consideration, so as to ensure that service users' rights are upheld, but also that family members' needs can be adequately assessed and met. 35,36 Future studies should bear these considerations in mind when planning and setting up the recruitment activities.