Some text in this section has been reproduced from Wade et al.41 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.
Principal findings and achievements per project outcome
The CARE pilot trial is, to the best of our knowledge, the first RCT to address the top uncertainty for brain cavernoma, with an embedded QRI to understand recruitment barriers and facilitators and develop corresponding actions to optimise informed consent and recruitment.
In the QRI, we found that the main barriers related to equipoise and patient eligibility, with conventions of usual care making a position of equipoise challenging for some, multidisciplinary teams preventing approach of some patients with symptomatic brain cavernoma, and logistical issues surrounding diagnosis and assessment for stereotactic radiosurgery adding further barriers.41 The main facilitators included the high level of commitment of the chief investigators, TMG and the PIs who were prepared to change practice within the research context to offer intervention to people who would have been offered medical management alone outside the trial. Multidisciplinary team equipoise and skilful information provision to facilitate patient understanding (so that the trial could be viewed as a solution to uncertainty about management) were key in optimising recruitment.
The pilot RCT met its targets for participation by neuroscience hospitals, investigators and people with symptomatic brain cavernoma.40 Estimates of recruitment rate, retention, surgical preference, adherence, outcome event rates, healthcare utilisation and intervention effects are informative for determining the feasibility of a definitive RCT. Although the generalisability of the findings outside the UK is unknown, a definitive RCT will require an extensive international multicentre network of sites and prolonged follow-up in view of the rarity of symptomatic brain cavernoma, the willingness of one-fifth of approached eligible patients to take part and the frequency of outcome events. The data collected were adequate to allow the conduct of both a within-trial and a model-based economic evaluation alongside a definitive RCT.
Contribution to existing knowledge
The non-randomised cohort studies comparing outcome after intervention for brain cavernoma with concurrent or historical control groups of people with brain cavernoma who underwent medical management alone have found a wide range of associations between intervention and poor outcome (see Table 1). These associations have ranged from a non-significant association with better outcome (RR 0.6, 95% CI 0.2 to 1.7) over ≈5 years23 to a significant association with worse outcome (RR 7.8, 95% CI 1.1 to 57.4) over ≈4 years (see Table 1).22 The similarity in outcome between medical management and intervention versus medical management alone in the CARE pilot RCT justifies the uncertainty about the benefits of management of symptomatic brain cavernoma with intervention and includes the possibility of harm or benefit.
Considering the challenges of conducting research during the COVID-19 pandemic and during the recovery of the research infrastructure afterwards, we had exceptional engagement from sites, investigators and participants.42 They demonstrated that this RCT could be done and that one-fifth of eligible patients who were approached were recruited, which is close to the finding of a similar RCT.31 The QRI added to the body of work on facilitators and barriers to recruitment to RCTs in general,33 RCTs involving children34 and RCTs comparing surgical with medical management.35,36
Strengths and weakness of the study in relation to other studies
The main strengths are that this project was a methodologically rigorous RCT with an embedded QRI and the first to address brain cavernoma. This was endorsed by the opening comments from two peer reviewers of the main trial report submitted to The Lancet Neurology;40 the first stated, ‘As a general comment, I much appreciated this study and do believe that methodologically sound feasibility studies like this should be performed more than is currently done’, and the second stated, ‘The authors have thoroughly investigated the importance of the research question. The methods are sound, the trial is very well conducted, and the paper is very well written. The authors are to be commended for involving patients and carers involved from the development of the research idea and question to the conduct of the trial. I think this manuscript should serve as an example of how to develop and conduct a trial and report on it’.
There was good engagement from the neurosurgical community, including neurosurgical trainees via the NIHR associate PI scheme. We had exceptional involvement from our Patient, Carer and Public Involvement (PCPI) collaborators, who supported many patients with their individual decisions about participation. The RCT exceeded its recruitment target. We did not find evidence of selection bias based on the similarities between eligible patients who declined and those who participated. Characteristics of participants by intended type of intervention reflected their appropriate selection for each type of intervention. Characteristics of intervention reflected usual practice in the UK. The trial achieved good retention and adherence, and provided estimates of preference for type of intervention (we found that stereotactic radiosurgery was chosen as the type of intervention twice as often as neurosurgical resection, contrary to prior expectations), and outcome event rates to inform the planning of a definitive trial. No participants died, clinical outcome event rates were similar to observational studies and no SAEs were reported. A definitive trial would benefit from repeating the collaboration with patient support organisations and implementing the actions from the QRI.
However, there were some limitations. Participants in qualitative data collection were largely supportive of the RCT; this may be representative, although negative perceptions may be less well documented, and this meant that within the qualitative data collection, we were unable to reach saturation on the influences of some demographic and clinical features on recruitment. The QRI was not a randomised study, and other factors may have contributed to the RCT exceeding its recruitment target (e.g. COVID-19 recovery, chief investigator engagement, trial marketing that signalled a time-bound recruitment target and our highly engaged PCPI partners). It was difficult to evaluate the numerical impact of PCPI support for patients from the patient advocacy organisations, because we had not arranged and sought regulatory approval for the identities of patients on the CAUK call log to be compared with the identities of participants in the RCT. The RCT sample size was small given the frequency of brain cavernoma, recruitment at 28 sites and 2-year duration of recruitment. Only 4 of the 72 participants were aged < 18 years. Participants’ ethnicities (92% White, 1% Asian, 1% mixed and 6% other) were less diverse than the UK. In a recent project assessing readiness for RCTs of drugs for brain cavernoma at five neuroscience centres in the USA, restricted to people aged > 18 years with symptomatic ICH from a brain cavernoma over the preceding year, participants seemed younger, more ethnically diverse, more likely to have a brainstem cavernoma and more likely to have a family history than in the CARE pilot RCT, but other characteristics and the outcome event rate were similar.43 Generalisability of our findings outside the UK is unknown. Finally, correspondence following the publication of the CARE pilot trial illustrated anecdotal misunderstanding or concern among neurosurgeons about ‘lumping’ both types of intervention in one arm of the trial,44,45 even though randomisation was stratified by type of intervention, and outcomes for these two subgroups were reported separately, which we clarified;46 reassuringly, the correspondents agreed that a definitive main phase RCT is needed.
Take-home message(s)
Fundamentally, it was possible to conduct a RCT of intervention for symptomatic brain cavernoma. This pilot phase RCT met its targets for participation by neuroscience hospitals, investigators and people with symptomatic brain cavernoma. Recruitment to a definitive RCT would benefit from training and support to encourage more clinicians to be comfortable approaching patients for whom medical management alone would conventionally be offered in usual care in the UK and broadening the pool of surgeons and multidisciplinary team members prepared to offer surgical interventions, particularly stereotactic radiosurgery, within a trial context. A definitive RCT would need to address the lack of consensus from the outset and use some key supporters of recruitment to the RCT within the neurosurgical community to address the issues head-on before and throughout recruitment, via repeated discussion of issues that challenge equipoise, as well as materials to encourage investigators to shift their usual practice towards recruitment to a RCT. Estimates of recruitment rate, retention, surgical preference, adherence, outcome event rates and intervention effects inform the feasibility of a definitive randomised trial. Although the generalisability of the findings outside the UK is unknown, a definitive trial will require an extensive international multicentre network of sites and prolonged follow-up in view of the rarity of symptomatic brain cavernoma, the willingness of one-fifth of eligible patients to take part and the frequency of outcome events.
Reflections on the project and what could have been done differently
It was worthwhile to begin to address the top uncertainty from brain cavernoma with a pilot phase RCT, which benefited from being conducted with PCPI and qualitative researchers. The QRI worked particularly well by involving qualitative researchers with discussions about equipoise at site initiation visits, contribution of general and disease-specific insights to the creation of participant information materials, engagement with the PAG and communications with investigators. In retrospect, the requirement to have a PI at each site who was a neurosurgeon helped recruitment, but neurologists (who see many patients with brain cavernoma) could have been more explicitly engaged with referring, assessing and approaching potential participants for the RCT; neurologists’ greater involvement with a definitive RCT would be desirable. In retrospect, we created materials successfully to support investigators who were already comfortable with enrolling participants in the RCT, but there were fewer materials orientated to encourage neurosurgeons to rethink their usual practice and introduce the RCT into it.
Challenges faced and limitations
The main logistical challenges we encountered were because of the effects of the COVID-19 pandemic on the research infrastructure and recruitment to research. Investigators navigated some of these challenges by using remote consultations and e-consent, for which we had sought regulatory approval from the outset.
The customs of usual care, even though they have been based on poor-quality observational studies, were hard to shift. We encountered entrenched views about the effects of stereotactic radiosurgery for brain cavernoma. However, there was a 2 : 1 preference for stereotactic radiosurgery over neurosurgical excision as the intended type of intervention, which the QRI found to be due to investigators’ perceptions of lower morbidity after stereotactic radiosurgery and participants’ greater comfort with a non-invasive procedure; this finding provides some reassurance which could help build consensus in support of a definitive RCT.
Engagement with partners and stakeholders
This pilot RCT would not have been successful without the involvement of PCPI partners, the QRI and collaboration with the Society of British Neurological Surgeons research network.
Individual training and capacity-strengthening activities
The NIHR associate PI scheme provided training in RCT conduct and oversight for many neurosurgeons in training, who obtained certificates for their involvement. We have created a web-based archive of training videos that could be reused, or adapted, for a definitive RCT. CAUK, Cavernoma Ireland and the PAG received training and first-hand experience from their involvement with trial oversight, conduct, dissemination and engagement.