Discussion

Principal findings

We found that MPH, at a dose individually tailored to each person according to perceived benefits and side effects, was not superior to placebo at relieving fatigue after 6 (± 2) weeks’ treatment. Although we detected some improvement in fatigue (over that offered by placebo) at most times during the study, none of these improvements reached a level that was considered to be clinically important (i.e. enough to be noticeable by patients and to make a significant impact on their QoL), and given the neutral primary outcome, the findings on secondary outcomes must be considered nominal/exploratory. This conclusion is supported by the finding that after 6 weeks’ treatment, participants receiving MPH did not report statistically significant improvements in QoL or health status and were as likely to state that they considered MPH to be effective (MPH 27/72, 37.5%) as placebo (23/67, 34%).

In addition to evaluating the effectiveness of MPH, our study also paid close attention to monitoring potential AEs. We did this because we were aware that, even if MPH had been found to be effective as a treatment for fatigue, its usefulness may have been limited by toxicity. In fact, we found that MPH was neither very good at relieving fatigue nor did it cause any great increase in adverse effects. This suggests that while MPH cannot be recommended as a treatment for cancer-related fatigue, it would be safe to continue to explore this medication as a therapy for other symptoms in patients with advanced cancer (such as low mood or opioid-related drowsiness).

Contribution to existing knowledge

Prior to our study, there was contradictory evidence about the benefits or otherwise of MPH for fatigue in patients with advanced cancer. Table 3 summarises the main features of, and results from, previous studies. There have been major variations in key study design features in previous studies. In particular, with respect to: choice of patient population, formulation and dose of MPH, use of as required or regular dosing, fixed-dose or individually titrated dosing schedules, length of study and choice and timing of primary outcome measures.

Our findings support the conclusions of previous systematic reviews and meta-analyses14,32,33 that psychostimulant medications appear to be safe when used in patients with cancer for the attempted relief of fatigue. Gong and colleagues’32 systematic review reported that adverse effects were described in 4.9% of participants (from five included studies) receiving MPH versus 1.6% of participants receiving placebo. The difference in frequency of adverse effects was not statistically significant. The most commonly reported adverse effects (compared to those reported by patients on placebo) were vertigo, anxiety, anorexia and nausea. In Minton and colleagues’ systematic review,33 it was reported that there was no increased risk of SAEs in trial participants receiving MPH versus placebo. In our own study, we similarly found no increase in SAEs or in mortality among participants receiving MPH. We collected frequent and wide-ranging data regarding potential adverse effects of MPH throughout the 10 weeks of the study generally found that the treatment was well-tolerated. The following mild, moderate or severe AEs were reported with a frequency ≥ 10 percentage points higher among participants receiving MPH than placebo at any time over the 10-week study period: nausea (70% vs. 57%), vomiting (38% vs. 28%) and depression (51% vs. 38%). In contrast, mild, moderate or severe weight loss was reported more commonly (≥10 percentage points higher) by participants receiving placebo than by those receiving placebo (57% vs. 46%). The only severe AE to differ substantially between treatment groups was the frequency of ‘feeling drowsy’ (11% in MPH group vs. 22% in placebo-treated participants).

Strengths and weaknesses in relation to other studies

It is noteworthy that although previous systematic reviews and meta-analyses14,15 have suggested that MPH may be effective, the majority of individual studies included in such reviews have not reported benefit.19–25 Indeed, the positive findings from previous meta-analyses have largely been driven by the results of just one relatively large study. Lower and colleagues16 investigated d-MPH, the pharmacologically active portion of the drug, which is approximately twice as potent as MPH. In Lower’s study, participants started at 10 mg/day of d-MPH (equivalent to 20 mg/day of MPH), and the dose and frequency were modified weekly over an 8-week period so that the mean final dose received was 25.5 mg/day. The study involved patients (n = 154) with different cancer diagnoses, the overwhelming majority of whom had either breast or ovarian cancer (90–91% of participants in each arm). Participants had previously had at least four cycles of cytotoxic chemotherapy (a mean of over 2 years previously) and were currently off treatment. The stage of disease was not reported, but 93% of participants had ECOG performance status 0 or 1. A key feature of Lower’s study was that it involved a single-blind (placebo only) 7-day run-in phase. Participants who responded (n = 8) or had adverse effects (n = 2) or other problems (n = 4) with placebo were excluded from randomisation. After 8-week treatment, Lower and colleagues reported that FACIT-F scores had improved by 10.5 points in those receiving d-MPH and 6.8 points in those receiving placebo (a mean difference in change scores of 3.7 points). The discrepancy between Lower and colleagues’ results and the majority of the other randomised controlled trials (RCTs), including our own, may be explained in various ways. It is likely that the population in Lower’s trial was clinically different from the populations studied in other studies of cancer-related fatigue. All participants had a good performance status, and many of them may have even been disease-free at the time of study enrolment (on the basis that they had been off treatment for 2 years and many may have had potentially curative disease). The causes (and potential treatments) for fatigue in disease-free cancer survivors may be substantially different from those in patients with active disease, on treatment or in palliative care. It is also noteworthy that the doses of medication used in the Lower study were higher than those attained in most other clinical trials (although more similar to the doses achieved in our own study). But perhaps the most important difference between Lower’s study and other previous research was the decision to include a 1-week placebo-only run-in period, with placebo responders being subsequently excluded from randomisation. This is likely to have resulted in a smaller response in both treatment groups than would otherwise have been expected. This is particularly important because the reason that most previous studies have failed to demonstrate superiority of MPH over placebo has been, at least in part, due to significant responses34 in the placebo arm. Finally, it should be noted that even though the study reported statistically significant superiority for d-MPH, the magnitude of the difference in change scores between treatment arms was only 3.7 points, which is less than the MCID our study was powered to detect.

Several other studies have tried to replicate Lower and coworkers’ findings in patients with advanced cancer. Bruera and colleagues20 undertook a four-armed (2 × 2 factorial) study comparing MPH versus placebo delivered in combination with either a nursing telephone or control intervention. The participants all had advanced cancer. They used an ‘as required’ dosing schedule of MPH 5 mg tablets, that permitted participants to receive up to a maximum of 20 mg/day for 14 days. The mean daily dose of MPH actually received was 6.4 mg/day. Although fatigue significantly improved from baseline in both groups (5.5-point improvement in MPH and 6.0-point improvement with placebo), there was no significant difference in the change scores between groups. Centeno and colleagues24 also studied patients with advanced cancer, but in this trial, participants received an individually titrated regular dose of MPH over 6 days (starting at 15 mg/day and subsequently adjusted to between a minimum of 10 mg/day and a maximum of 25 mg/day). There was a significant improvement in fatigue in both groups (4.9-point improvement in MPH and 6.4-point improvement with placebo), but no significant difference in the change scores between groups.

Our study contributes to the existing literature by confirming the lack of effectiveness of MPH at reducing fatigue in patients with advanced cancer. One potential explanation for the inability of previous studies to demonstrate effectiveness of MPH may plausibly have been that many previous studies did not use regular dosing at a sufficiently high strength. Of the three previous studies16–18 which have observed MPH to be effective, one involved ‘as required’ dosing of 10 mg at a time18 and one involved regular dosing of 10 mg/day.17 Only Lower and colleagues’ study (previously discussed) used a higher dose (final dose equivalent to MPH 51 mg). In our study, the mean daily dose of MPH was 30 mg/day, with many participants achieving the maximum permitted daily dose of 60 mg/day. Nonetheless, despite these relatively high doses, after 6 weeks of treatment, we did not detect any significant differences in fatigue in the intervention group.

The choice of a 5-point difference on the FACIT-F score as representing the MCID may be considered a potential limitation of our study. There is, after all, no consensus about the MCID for FACIT-F and so we had to make an informed decision about the appropriate cut-off on the basis of previous studies in this area. Two different approaches to determining the MCID have been used previously. Reddy and colleagues31 calculated the MCID by looking at changes in FACIT-F score and comparing it with changes in subjective GBSs, using pooled data from three clinical trials in palliative care patients with advanced cancer. They hypothesised that if participants in these trials had reported a change in their condition that was at least ‘moderately important, consistently beneficial’, then they should be considered to have had a clinically important improvement in fatigue. They found that a change in FACIT-F score of approximately ≥ 10 points corresponded to such an improvement. Rather than using a subjective improvement in perceived global benefit, Patrick and colleagues35 calculated the MCID by using an ‘anchor-based’ approach. That is, they analysed data from a RCT36 in which cancer patients were given either epoetin alfa or placebo and determined what change in FACIT-F score corresponded to at least a 10 g/l improvement in haemoglobin levels. They calculated the relationship between change in haemoglobin and change in FACIT-F score in two ways, either by comparing those with stable versus those with improved (> 10 g/l) haemoglobin levels (MCID 4.75 points) or by using regression (MCID 1.63 points). Similarly, Cella and coworkers used a combination of ‘anchor-based’ (in this case, a combination of haemoglobin levels, performance status and treatment response) and ‘distribution-based’ (1/2 SD or 1 standard error of measurement) and estimated the MCID to be 3 points.

We believe that our choice of a MCID of 5 points is justified on several grounds. Firstly, in an evaluation of a symptomatic treatment to improve subjective fatigue in palliative care patients, it makes more sense to use a MCID based on a subjective measure of improvement (global benefit) as the benchmark, rather than a MCID derived from fatigue scores corresponding to a > 10 g/dl change in haemoglobin levels. Secondly, if we had used the most conservative estimates of a MCID suggested by Cella37 and Patrick35 (i.e. 3 and 2 points, respectively), this would have equated to a GBS corresponding to somewhere between ‘Not beneficial’ and ‘Slightly beneficial’. Whereas a mean change of 5 points on FACIT-F score (the MCID used in our study and the higher estimate provided by Patrick and coworkers) corresponds to a GBS closer to ‘Slightly beneficial’ or ‘Somewhat important, consistently beneficial’. Finally, post hoc support for our choice of at least a 5-point change as representing the MCID was the finding that, despite patients who received MPH having a nominally statistically significant mean improvement of 2 points on FACIT-F score at week 6 compared to placebo, more patients in the MPH than the placebo group considered that their overall condition had ‘stayed the same’ or ‘got worse’ [45/72 (63%) vs. 38/67 (57%)].

Limitations

Although we maintain that the choice of a MCID of 5 points on FACIT-F score was justifiable for the reasons described above, given the key importance of this figure for the design of the study, it may have been helpful to undertake further preparatory work to inform and support our chosen cut-off. This might have included, for instance, undertaking a wider consultation with patient and public involvement (PPI) groups and/or clinicians about the minimum levels of fatigue that are considered consequential, or even further empirical work to explore the relationship between fatigue levels and QoL or self-reported overall benefit.

The participants in our study were highly selected. This was driven by safety requirements and reflected the long list of cautions and contraindications that accompany the use of MPH in clinical practice. As a result, most fatigued patients at participating sites were not eligible for inclusion in the trial because they were gravely ill or else had clear contraindications to participation. Of the patients who were formally screened, only 162/297 (55%) were randomised. This suggests that, even if our study had demonstrated a statistical and clinical benefit for MPH, it would have been unlikely to have become a widely used therapy in this population.

Our study failed to reach the originally planned sample size. Although, following modification of our protocol, permitting an adjustment in sample size and power (see section on protocol modifications), we did still manage to successfully recruit and retain sufficient participants to answer the primary objective.

Reflections on the project

Randomised controlled trials of IMPs [Clinical Trial of an Investigational Medicinal Product (CTIMP)] in palliative care are relatively uncommon, and much palliative care research focuses instead on the evaluation of services, policy-related issues and observational studies.38–40 Most CTIMPs in palliative care populations on the NIHR portfolio are trials of disease-modifying therapies (systemic anticancer therapy or radiotherapy trials) rather than evaluations of symptom-directed interventions. Palliative care studies are hard to plan, fund and deliver. There is consequently a paucity of robust data from clinical trials in palliative care to inform clinical practice about symptom management. The difficulties of undertaking trials in palliative care have been previously described. For example, Chen and colleagues39 undertook interviews with 61 leading palliative care researchers and identified five major obstacles to the successful completion of research studies in this population: research funding; institutional capacity; researcher workforce issues; challenges related to the topic and/or the study population; and public and professional misunderstandings and misperceptions about undertaking research in this area. Despite such difficulties, and the added complication of an intervening global pandemic, methylphenidate versus placebo for fatigue in advanced cancer (MePFAC) became one of the few clinical trials of an IMP to have been successfully completed in this patient population in the UK. The investigators, therefore, consider it a great achievement to have completed this study and to have reached the required (albeit reduced) sample size, thus providing a definitive answer to the research question.

One of the key factors contributing to our success was the involvement of the Priment Clinical Trials Unit (CTU). Prior to leading the MePFAC study, the chief investigator (although previously a PI) had had no previous experience of leading a CTIMP and was reliant on the expertise and experience of the CTU to lead him through the processes required to set up and run a multicentre study. Correspondingly, however, the CTU had also had limited prior experience with undertaking RCTs in palliative care populations and needed to be mindful of ways in which usual ways of operating may need to be modified to reduce the burdens on study participants (e.g. minimising number of data collected and limiting number of face-to-face visits). The chief investigator wishes to acknowledge the support of Professor Michael King as a co-applicant and mentor during the funding application process and the first few years of the study. Professor King was an experienced triallist and, although a psychiatrist by background, had also undertaken clinical research in palliative care populations. He was, therefore, able to advise on how the study design could be optimised for patients with advanced disease. Unfortunately, following a short illness, Professor King died in 2021, before the trial was completed. However, his contribution to the success of the project cannot be overstated.

Reflecting on the issue of the paucity of clinical trials in palliative care, there are relatively few research-active palliative care services in England. This meant that many of the MePFAC PIs and research sites were relatively inexperienced. For a significant proportion of the sites involved in the MePFAC study, this was their first involvement in a RCT of an IMP or even in a multicentre research project. Many palliative care studies are observational and/or use qualitative or survey methodologies. As a result, PIs and research delivery staff at each site required extensive and ongoing training and support, and this placed a large administrative burden on the central research project management team (chief investigator and trial manager).

Identifying, training and supporting PIs only partially addressed the workforce issues. Collaborating sites also needed to access research nurse time. Although, in theory, Clinical Research Network (CRN) nurses were available to support recruitment (at least at collaborating sites based in cancer centres), there were sometimes practical difficulties with accessing research nurse time. Cancer clinical trials are often disease-specific, and recruitment is usually supported by site-specific research nurses (e.g. lung cancer research nurses recruiting to lung cancer studies). However, MePFAC was open to all patients with advanced incurable cancer, rather than to patients with a specific tumour type. This ought to have made recruitment easier (more potentially eligible patients to be recruited), but in practice, it sometimes meant that recruitment to the MePFAC trial was considered a lower priority than recruitment to tumour-specific studies. At some centres, PIs benefited from dedicated palliative care research nurse time specifically to recruit to the MePFAC study; at other centres, responsibility for recruitment was ‘distributed’ across a number of different site-specific nurses. Our highest recruiting site (University Hospitals Sussex) had a PI who was also a co-applicant, a dedicated NIHR-funded research nurse, and recruitment across tumour types was supported by all of the oncology consultants (with much of the recruitment occurring among patients with advanced cancer on the chemotherapy day unit). Some CRN nurses were also able to outreach to recruit participants at geographically remote hospice sites, but because the nurses were not based on site, this meant that (due to lack of availability) potential participants to the study may have been missed. The most effective hospice recruiting sites (John Eastwood Hospice, Sue Ryder Hospice and Pilgrims Hospices) had one or more research nurses on site (supported by a combination of charitable funding, with CRN and some industry-funded studies).

Recruitment sites based in cancer centres were generally supported by experienced research pharmacies and pharmacists who were very familiar with undertaking CTIMPs. In comparison to other members of the local site research team, these services seldom needed additional central support. However, some hospice sites initially expressed an interest in participating in the study but did not have adequate pharmacy support to undertake CTIMPs and were therefore deemed to be unsuitable for study participation. Other sites, who participated in the trial, needed to use research pharmacy services from their local hospital trusts.

After participants were recruited to the study, they needed to be reviewed on a weekly basis to be asked about perceived benefits and adverse effects, so that local PIs could make decisions about whether (and how) to adjust the dose of the study medication. Initially, the reviews at weeks 3, 6 and 10 were scheduled to be undertaken on site by PIs and locally employed research nurses, with the intervening assessments being undertaken remotely by centrally employed research nurses/clinical trial practitioners at University College London. However, following protocol amendments in response to the COVID pandemic, nearly all of these weekly reviews were undertaken remotely by the central team (some sites continued with face-to-face visits depending on local policies). The funding for these two centrally employed members of staff came from the NIHR project grant which funded the MePFAC trial. This additional, centrally funded workforce was essential to the success of the MePFAC trial. It meant that many more sites considered that they could participate in the study despite relative lack of access to research infrastructure support locally. The role of the two centrally employed research staff assumed even greater importance after the trial reopened following a pause for the COVID pandemic. Following a protocol amendment, it became possible for the centrally employed research nurses to undertake all the weekly assessments (including assessments at weeks 3, 6 and 10, which had previously been undertaken face to face at local recruiting sites). This innovation was particularly important to the success of the trial at a time when other services were being stretched very thinly (because of, e.g. re-deployment of research staff to clinical front-line duties).

The COVID pandemic resulted in a significant increase in workload for MePFAC staff and collaborators [e.g. conducting risk assessments; closing and reopening sites; developing new standard operating procedures; submitting Research Ethics Committee (REC), Health Research Authority, and Medicines and Healthcare products Regulatory Agency amendments; retraining staff in new procedures; securing additional funding]. MePFAC was supported throughout by the funder. At the start of the pandemic, while the impact of the disruption was still uncertain, NIHR offered an immediate 6-month-no-cost extension. Later in the year, once the situation was somewhat clearer, NIHR agreed to a longer, funded extension, that allowed the study to reach its recruitment target.

Although the COVID pandemic entailed significant increase in workload, there were some positives to emerge. The changes that were made to the research protocol in response to the pandemic made the study much easier to conduct and less burdensome for participants (e.g. simplification of screening and consent process, fewer face-to-face visits, greater use of remote monitoring and home delivery of study medications). Future studies are much more likely to make greater use of remote monitoring and online data collection.

Major changes over the course of the trial

The MePFAC trial was originally planned to last for 3.5 years (42 months). In fact, the study took 6.5 years (78 months) to complete. This included two funded variations to contract (VTCs) and an approximate 1-year pause in recruitment as a result of the COVID pandemic. As previously described, one of our co-applicants, Professor Michael King, died on 10 September 2021. In order to contain costs, Professor King was not replaced, and the rest of the trial was completed by the original co-applicants. A list of the major changes that occurred to the protocol is provided in Appendix 1, Table 4.

Impact of the COVID pandemic

The COVID pandemic caused major disruption to the study and increased costs. A summary of the actions that the study team took in response to the COVID pandemic is provided in Supplementary Information (see Report Supplementary Material 5). Recruitment was significantly curtailed in early March 2020 as the impact of the pandemic began to affect NHS services. On 19 March 2020, the chief investigator took the decision to cease recruitment completely. At that point, the Trial Management Group (TMG) conducted a risk assessment and approached the funder for support to pause the trial with the aim of reopening when and if the pandemic allowed. NIHR allowed the study to continue with an initial 6-month-no-cost extension to account for the estimated length of time for which the COVID pandemic was (at that time) considered to be likely to affect recruitment. By chance, one of the centrally employed research nurses had resigned prior to the pandemic and had not yet been replaced. We, therefore, took the opportunity to save costs by delaying reappointment until the autumn. New recruitment was suspended between March and October; nonetheless, the remaining NIHR-funded staff were fully employed with the safe close-down of sites, data-checking, remote monitoring, preparation of protocol amendments and reapplications for research ethics and governance approvals for the revised protocol. Although we managed to reopen a few sites to recruitment by November 2020, recruitment did not immediately return to normal. That is because, by the autumn of 2020, COVID once again started to place the NHS under enormous stress, and most research sites did not have the capacity or capability to reopen. Even those sites that were able to remain open to recruitment were badly affected by social-distancing guidelines and then by another national lockdown. Between 19 March 2020 and 31 March 2021, only seven participants were recruited to the trial (from four sites). Recruitment only began to return to something approximating to pre-COVID levels after June 2021.

Protocol changes

The original protocol [v1.0 (5 May 2017] was submitted to the London City and East REC for approval in June 2017 [Integrated Research Application System (IRAS) project ID 215297; REC reference number 17/LO/0871; protocol number 15/0592]. The final Study Protocol is available in Supplementary Information (see Report Supplementary Material 1), and the version history at the start of the document details all of the changes that occurred with each subsequent protocol amendment. A summary of the major changes to the protocol is provided in Appendix 1, Table 4.

Funded variations to contract

Over the course of the study, there were two major variations to the original contract with NIHR. The first variation to contract (VTC1) was a request to extend recruitment by 16 months (from 31 October 2020 to 28 February 2022). Principal changes included a reduction in minimum permissible eGFR from 60 ml/minute to 45 ml/minute and exclusion of patients only if they had uncontrolled (rather than pre-existing) heart failure, angina or drug dependency and if they had had MI or stroke within the previous year (rather than at any previous time). The amendment also included a reappraisal of the original assumptions underlying the randomisation target. Prior to starting the study, we had estimated that there would be 25% attrition. On that basis, we had estimated that we would need to recruit 230 participants to achieve 172 evaluable participants. However, study attrition (at the time of the VTC application) was running at 13%. We, therefore, modified our assumption to estimate that attrition over the course of the study would be no worse than 20%, and therefore the estimated sample size needed to accrue 172 evaluable patients could be reduced from 230 to 215 randomised participants without any change in the study power. VTC1 was submitted on 19 December 2019 and was awarded on 9 July 2020. However, it should be noted that between submission and the award of VTC1, the MePFAC trial (like most other UK research) was severely adversely affected by the COVID pandemic. Indeed, all recruitment to the trial was suspended in March 2020. At the time, it was estimated that it may be necessary to suspend recruitment for 6 months (in fact, the trial was effectively closed to recruitment for 12 months).

The second variation to contract (VTC2) was a request to extend the study by 20 months (from 28 February 2022 to 31 October 2023). Based on recruitment performance up until that point, it was no longer deemed feasible to recruit to the originally planned sample size (providing 90% power). However, recruitment targets were adjusted to ensure that a sufficient number of participants would be randomised (n = 162) to accrue a minimum of 130 evaluable participants (this would provide a minimum of 80%+ power, or higher if attrition was lower than estimated). In efforts to facilitate recruitment and retention to the study in the light of ongoing pandemic conditions, further modifications were made to the protocol, including greater use of remote (rather than face-to-face) follow-up assessments and merger of separate screening and enrolment visits.

Engagement with partners and stakeholders

The TMG consisting of the co-applicants, our PPI representative, the central research team (trial manager, research nurses and/or clinical trial practitioners), representatives from the CTU and other interested members of the extended research team, met on a monthly basis throughout the trial (hybrid meeting to allow PIs to facilitate remote attendance).

Because many of the participating centres in the MePFAC study were relatively research-naive, the central research team made great efforts to engage with and support local PIs and site research teams. So, in addition to the TMG meetings, we held monthly virtual meetings for all participating sites, initially by telephone and later using an online virtual meeting platform [Microsoft Teams (Microsoft Corporation, Redmond, WA, USA)]. At these meetings, the chief investigator briefed sites about developments in the study, planned or approved protocol changes and AE reporting. In turn, sites shared information about the barriers and facilitators to conducting the study at their own sites and shared best practice to improve recruitment across the study. The chief investigator and the trial manager were also directly accessible to members of the wider research team by mobile phone and e-mail. The trial manager produced a newsletter for all participating sites, which was distributed three times per year and which shone a spotlight on areas of good practice and attempted to foster friendly rivalry between sites to boost recruitment. We promoted wider engagement of stakeholders by periodic updates about recruitment on Twitter [X Corp. (formerly Twitter) San Francisco, CA, USA] (tagging in NIHR Twitter account and the relevant local recruitment centres).

David Miodrag (PI at John Eastwood Hospice) reflected thus on his involvement in the study:

I had never been a PI, nor been directly involved in a CTIMP, before my involvement in MePFAC. When the trial started, I felt enthusiastic though daunted due to my lack of previous experience, although Prof. Stone, and Elli Enayat (Trial Manager), were very supportive and accessible throughout, and this put me at ease and gave me assurance and confidence. The regular MePFAC teleconferences and newsletters for updates and information sharing between sites were a great help and kept me motivated and well informed.

John Eastwood Hospice was one of the highest recruiting sites in the trial. There were several factors that contributed to our local success. We had two less than full time Research Nurses working on the study who undertook screening and attended multi-disciplinary team meetings at the hospice to identify potential participants (the nurses are partially funded by the NHS and partly by the John Eastwood Hospice charity). We gave several presentations about MePFAC to different clinical teams within the hospice at the outset to persuade clinicians to identify potential participants. Our Day Hospice was a particularly fertile source of participants, and it reached a point where some Day Hospice patients requested to be screened for the trial having heard about it from their peers who were taking part.

Being a PI in MePFAC was invigorating and rewarding, and I am raring to be involved in future Palliative Care research, especially CTIMPS.

Training and capacity strengthening activities

Prior to opening the study at any collaborating site, we held a site initiation visit to provide training to research staff on the protocol. Sites were also provided with access to an audio-recording and PowerPoint^® (Microsoft Corporation, Redmond, WA, USA) slide set of the training materials for members of staff who could not be present at the site initiation visit. Training logs were kept to ensure that all staff had had the necessary training. Additional remote training was provided on the use of the study database and on the randomisation and pharmacy dispensing processes. Sites were given access to training videos about how to identify potentially eligible patients for the study and how to obtain informed consent. After the temporary pause in recruitment occasioned by the COVID pandemic, it was necessary to effectively relaunch the study at each site and to provide updated training on the new protocol. Ad hoc training was provided to new members of staff and PIs.

The MePFAC also provided an opportunity for more junior staff to gain experience with working on clinical trials. For example, as part of her NIHR Development and Skills Enhancement Award for Clinical Trials, Joanne Bayly (King’s College London) observed some of the Trial Steering Committee (TSC) meetings (under the mentorship of Toby Provost). MePFAC was part of the NIHR Associate PI Scheme, which provides junior staff, who are interested in research, with 6 months’ in-work training and an opportunity to gain practical experience with being involved in a clinical trial. Over the course of the study, we had two Associate PIs, one hosted at Royal Sussex County Hospital and another at University Hospital Southampton.

In June 2023, the chief investigator was invited to give a talk at a conference on research-active hospices which was attended by over 100 participants (academics, clinicians, patient and public representatives and research administrators). It is envisaged that this may lead to a community of practice to encourage and support future multicentre research studies. A report of the meeting has been produced and shared with NIHR and other stakeholders and is available in Supplementary Information (see Report Supplementary Material 6).