Prioritising key motivators and challenges influencing informal carers’ decisions for participating in randomised trials: An embedded Study Within A before and after Trial (SWAT 55)

Context of SWAT-55

The SWAT-55 host trial was a planned pilot-feasibility randomised trial (ClinicalTrials.gov identifier NCT03048565, registered 9^th February 2017) based on the following PICO (population, intervention, comparator, outcomes);

Host trial sample size, randomisation and recruitment

The planned sample size for the host trial, based on a recommended sample size for pilot studies of 30 per group⁴ was 80 informal carers, or 40 per group allowing for a 25% attrition rate (10 participants per group) at six-months follow-up, randomised on a ratio of 1:1 using a computer-randomised number generator. Ethical approval to conduct the study was granted by the Research Ethics Committee of the lead researcher’s university. Recruitment to the host trial commenced in March 2017 with intervention delivery planned for April-June 2017. SWAT-55 was planned to commence in June 2017.

An invitation to participate in the host trial was emailed to 538 Dublin based registered Family Carers Ireland members. Intervention delivery was initially planned as face-to-face; however, by the end of August, despite efforts, two carers only were recruited and randomised to the intervention group, one of whom had to withdraw subsequently because he/she was unable to attend the intervention sessions. Following a Trial Steering Group (TSG) meeting, a decision was made to deliver the intervention in an online format and open the study to a wider national base. Ethical approval was granted, and an updated study invitation letter was circulated in September 2017. A social media invitation was also posted to the Family Carers Ireland Facebook page. By October 2017, 11 carers only were recruited to the study; five in the intervention group and six in the control. Following a further TSG meeting, a decision was made to change the trial design to a before and after trial, thus converting the host trial to a non-randomised pilot-feasibility trial⁵, on the basis that a randomised trial was not feasible at recruitment level. Participants recruited to the control arm of the original trial were subsequently offered the intervention, and the revised design was further advertised. Between mid-October and end of December 2017, 17 carers were recruited to the study, 15 returned baseline data, of which seven returned end of intervention data (quality of life, mindfulness and stress outcomes). These challenges further emphasised the importance of SWAT-55 in exploring the reasons why informal carers, as a discrete group within the general population, may or may not decide to take part in a randomised trial. Although the original host trial was redesigned, we proceeded with the SWAT as planned, albeit as a study within a non-randomised trial, accepting that this deviated from our original intention of conducting the SWAT as a study within a randomised trial.

Design

A multi-method two-phase study was conducted. Phase 1 involved a series of face-to-face interviews with a sub-sample of participants from the host trial and a review of systematic reviews that assessed motivators or barriers for taking part in trials^6–12. Resultant data from both the interviews and systematic reviews were used to collate a list of key motivators and challenges. Phase 2 was a national survey of how important these key motivators and challenges were to informal carers in making a decision on trial participation.

Phase 1. The host trial consent form provided participants with an option to agree to future contact for follow-up studies arising from the trial. Of the 17 carers recruited to the host trial, 11 agreed to future contact. These 11 carers were contacted via email (addresses provided as part of the host trial) and invited to take part in an interview designed to ascertain their views of and reasons (motivators and challenges) for participating in research. We aimed to recruit 10 participants for interview; however, only two came forward. The interviews, which were semi-structured using an interview guide (available at: http://www.doi.org/10.5281/zenodo.4029385), were held at a mutually agreed time and venue. The interviews were conducted by the lead author, a female Prof in Midwifery with 10 years’ experience of sensitive interviewing in healthcare. The interviews were recorded and transcribed verbatim. The data was manually coded by two authors (VS and AH), and analyzed, based on a thematic analytical approach using discussion, iteration and consensus, to determine common categories of i) motivators and ii) challenges, for use in phase 2 of the study.

Due to limited participation in the interviews we made a necessary pragmatic decision to additionally draw on data from systematic reviews to inform phase 2. Seven systematic reviews of studies of reasons for participating in trials^6–12 were identified in an evidence mapping exercise conducted as part of the PRioRiTy study¹ based on a search of MEDLINE, EMBASE, Cochrane Database of Systematic Reviews, Social Sciences Citation Index and ERIC using a combination of search terms (e.g."attitudes to trials":kw OR (participat* or recruit* or enrol* or select*) near/8 (trial* or research or study):ti) from the Cochrane systematic review of strategies to improve recruitment to trials¹³ and from the ORRCA project¹⁴. The reviews were identified as eligible for the evidence mapping exercise in the PRioRiTy study¹ by at least two members of the Steering Committee. Data related to motivators/barriers for taking part in a trial were extracted onto an excel sheet for SWAT 55 by one author (VS) and corroborated by a second (AH). These data were combined with the data from the interviews to develop categories, inclusive of aligned motivator/challenge statements associated with these categories, for use in the phase 2 survey. An ‘audit trail’ of developing categories from initial codes was maintained to ensure confirmability of the process. Survey development occurred between January and June 2018.

Phase 2. Phase 2 was a national survey of informal carers in Ireland to prioritise the key motivators and challenges when making decisions to participate in a trial (hypothetical or real). Developed based on the findings from phase 1, the draft survey was reviewed by a carer for user-ability and by the National Adult Literacy Agency (NALA) for plain language and comprehension. Following a number of edits and language ‘tweaks’, the survey received the Plain English Mark and was finalized. Survey distribution, using SurveyMonkey, was planned for July/August 2018; however, as Family Carers Ireland were undertaking a survey of their own at the time, distribution was delayed, initially until September 2018, and subsequently to January 2019. The original agreement was for Family Carers Ireland to distribute the survey link to their database of registered members, accompanied by the study information leaflet and the lead applicant’s contact details to discuss the study further, as needed. The survey was anonymous with no details of participant’s names, locations, email addresses, or any other identifying details requested. At this time, however, uncertainty and concerns related to the new General Data Protection Regulations (GDPR) that had come into effect in 2018 prohibited email distribution, and the survey finally went live at the beginning of February 2019 via the Family Carers Ireland Facebook page, with closure four weeks later.

Outcome measures

The study outcomes were a prioritised list of i) motivators and ii) challenges, based on the level of importance that respondents assigned to each of the survey’s motivator and challenge statements.

Data collection and analysis

Quantitative analytical techniques were used to aggregate individual’s ranking of motivators and challenges. Each participant was asked to rank each motivator and challenge on a 5-point Likert scale of 1=very unimportant, 2=somewhat unimportant, 3=neutral/unsure of importance, 4=somewhat important, and 5=very important. Mean scores and standard deviations (SD) for each item were calculated using SPSS (version 21). The items were subsequently arranged in descending order of importance based on the mean and SD scores attributed to them. The priority list of informal carers’ motivators and challenges for participating in trials was determined.

Ethical statement

Ethical approval for this embedded study was granted by the Research Ethics Committee of the School of Nursing and Midwifery, Trinity College Dublin (Ref: 17/07/2017). Consent for interviews was written and taken prior to commencing the interview. Consent for the online survey was indicated by an ‘I agree’ tab, which indicated the participants gave their consent to take part in the survey.

Interviews, systematic reviews and survey development

The interviews, involving one male and one female, were of 21 and 38 minutes duration, respectively. One participant provided full-time care, and the other part-time care. The care recipients were a child with cerebral palsy and an adult with multiple sclerosis. Although only two carers participated in the interviews, both provided valuable data for survey development. Table 1 presents examples of the categories, and associated codes, which emerged from the analysis of the interview data.

The motivators and challenges data extracted from the systematic reviews, some of which overlapped with the emergent categories from the interview data, are presented in Table 2.

Using the results of the interviews and the systematic reviews, core categories, with related motivator/challenge statements, were developed. The final survey consisted of five core motivator categories with 28 associated statements and two core challenge categories with 17 associated statements. The motivator categories were; carer identity (three associated statements), study design (12 associated statements), altruism/common good (four associated statements) personal interest (six associated statements) and study information (three associated statements). The challenge categories were personal risk (11 associated statements) and study design (six associated statements).

Survey findings

Thirty-six informal carers returned a completed survey. The majority of respondents were ≥36 years of age, female, educated to minimum leaving certificate level and unemployed (Table 3).

When asked to indicate with whom they lived, some respondents’ ticked more than one response option (e.g. lived with partner and children). Two respondents indicated living on their own, and the remainder reported living with a partner (n=22), their children (n=19), their mother (n=7), their father (n=2) and with siblings (n=2). Twenty-nine of the 36 respondents (81%) lived in the same residence as the person they were caring for. The majority of participants were more than 10 years in their caring role (61%; n=22), followed by 2–5 years (19%), 6–10 years (17%) and <2 years (3%). Thirty-three respondents (92%) reported that they provided care on a full-time basis, with the remaining three (8%) providing care part-time. The care recipients (n=38 as some respondents indicated they cared for more than one person) in most cases, were children (n=16), followed by parent(s) (n=10), partner/spouse (n=9) and sibling(s) (n=3). Five care recipients were under the age of 17 years, seven were aged 18 to 29, six were aged 30 to 39 years, and the remaining 18 were aged over 40 years. Of these 18, 14 were 60 years or older. Most respondents (78%; n=22) had not received any training for their role as an informal carer. Of those that did, some indicated previous professional training (e.g. nursing, disability or mental health) or varied short training courses, for example, manual handling, infection control, pain management or courses on autism spectrum disorder. The condition of the care recipients varied widely, with many respondents caring for people with multiple conditions, and comorbidities; for example, Alzheimer’s/Dementia/Parkinson’s (n=8), Motor Neurone Disease/Multiple Sclerosis (n=2), Autism/autistic traits (n=5), Downs Syndrome/other intellectual disability (n=6), brain/spinal injury (n=2), mental ill-health (n=2), emphysema (n=1) and cancer (n=1).

Priority list of motivators. Table 4 provides the list of motivators prioritised by the mean and SD scores attributed to each motivator statement. Helping to create awareness about carers was the top ranked motivator for participating in a trial, followed by four study design categories related to a suitable time at which the study occurs, the study location, format of delivery (i.e. online) and venue. The least important motivators for deciding to participate in a trial, from the carers’ perspectives, related to study information issues; that is how they were informed of or invited to take part in the study, with all three associated statements averaging mean importance scores of 3.5 or less.

Priority list of challenges. Table 5 provides the list of challenges prioritised by the mean and SD scores attributed to each challenge statement. Personal risk, associated with difficulties in planning due to the caring role emerged as the most important challenge for carers when deciding on participating in a trial (mean 4.13, SD 1.25), followed by being unable to leave the care recipient on his/her own. Not knowing the person running the study was deemed to be the least important challenge for carers when deciding to take part in a trial (mean 2.74, SD 1.18).

Implications for a future host trial

The non-feasibility of the host trial in recruiting sufficient participants implies that any plan for a future, similar trial needs to take cognisant of carers’ motivators and challenges, and would require a major rethink as to how the trial would be designed and implemented. Although our study has highlighted important motivators and challenges it is not possible to generalise these findings directly to a host trial as our findings are based on few carers. However, the results provide areas that developers of future trials may consider; for example, how informal carers come to know of a study (e.g. through a family member or friend, or through a leaflet), or their awareness of those conducting the study, were ranked overall, as being neutral or unimportant when deciding to take part in a trial. Thus focusing on aspects of trial design, such as the value the trial has for carers’ identity, and how the intervention is delivered may be of greater value. Furthermore, as carers in this study identified that their caring role leaves it difficult for them to plan ahead, to have time away from their care recipient, or to travel to study locations, this might may prompt consideration for offering the intervention online or, if offering it face-to-face, plan for the interventionist to travel to the participants homes or other location convenient to them. These would likely have resource implications, however, in terms of personnel required and expense; an element that would need to be factored in to the future trial’s processes and budget.

Challenges to conducting the original planned SWAT 55

A number of challenges in conducting the originally planned SWAT were encountered which may limit the results. These included limited recruitment to the host randomised trial and a redesign to a before and after trial resulting in a reduced purposive sampling frame for phase 1 of the planned SWAT. Although rich data were provided by the two interviewees from the before and after trial, a wider pool of participants might have better ensured, or at least increased our confidence in data sufficiency. This was overcome somewhat, however, by the extraction of data from related systematic reviews, combining these data with the interview data in developing the survey.

Despite extensive efforts, 36 participants only responded to the national survey, which is likely to represent a very small proportion of the population of informal carers across Ireland. Although we cannot be sure, the inability of Family Carers Ireland to distribute the survey via their email list, and the move towards distribution via Facebook (with only 12 ‘shares’ and 14 ‘likes’ noted), may have impacted on the survey response rate. Furthermore, other than a request to snowball the survey, we did not consider other networks where carers might have been recruited; this was a probable oversight as had we extended our advertisement to other networks at the time, it may have increased participation. Had a larger number responded, providing greater caregiver representation, the ranked priority list of motivators and challenges might ultimately be different. The difficulties encountered with recruitment, however, could relate to how individuals identify with the concept of caring as evidence suggests that many individuals do not formally identify themselves as a carer or in a caregiving role²³. For example, in a qualitative study of 40 relatives or friends, the researchers concluded that “…self-identification with the role and label of carer is nuanced, shifting and variable”²⁴ and refers to other studies that have shown variation in how relatives identify with the term carer^25,26. In one study, for example, exploring recruitment with carers of people with multiple sclerosis, participants suggested using the term ‘support’ or ‘assistance’ in place of the term ‘caregiver’ and highlighted that many people would not consider themselves to be carers²³. This is an important consideration for researchers when advertising trials to an already hard to reach population, and how carers, including their associated roles, are described may need to be considered within trial participant information.

Underlying data

Access to interview transcripts are restricted under Research Ethics Committee approval as even de-identified transcript data may contain information that could potentially identify a participant based on statements made, phrasing, or personal data contributions which presents a potential breech in GDPR assurances. Transcripts may be made available, in full or in part, on individual request and only with the explicit permission of an interviewee based on the acceptability of the nature and reason for the request. Such requests such be made to the corresponding author (smithv1@tcd.ie) in the first instance.

Zenodo: SWAT 55 survey and dataset. https://doi.org/10.5281/zenodo.4029385²⁷.

This project contains the following underlying data:

- SWAT 55 Survey (participant responses).xlsx

Extended data

Zenodo: SWAT 55 survey and dataset. https://doi.org/10.5281/zenodo.4029385²⁷.

This project contains the following extended data:

- Interview Schedule.docx

SWAT Survey (Final Version).pdf

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).