The trial question and adherence to the Royal College of Obstetricians and Gynaecologists guidelines

For the trial to succeed as intended, the rationale for the research needed to be understood and accepted by all those involved in delivering the trial. As the trial progressed, however, two emerging issues were identified by some site PIs: first, that the trial question had lost its relevance and second, that NVP had a lower priority than other pregnancy complications:

. . . if I’m honest with you I think the general feeling was that practice has moved on, and that the question that this trial was asking was now probably a historical question . . . You know we’d almost like we’d missed the boat.

TS19

. . . when we’re coming to nausea and vomiting the perceived, the sort of the perceived issues or problems coming out of it will not sort of rank alongside stillbirth or neonatal death.

TS18

However, other PIs, patients and midwives/nurses were of the opinion that research in this area was much needed and that the trial question remained important. This was in spite of the issue of randomisation to double dummy being controversial at the initial stages among patient and public involvement advisors:

I thought it was a well-designed study and it was worth definitely worth, a question worth asking . . . I don’t think the study design was a problem which was what we initially thought was going to be a problem.

TS18

There was evidence from both PIs and research midwives/nurses that there had been a change in the practice of prescribing antiemetic drugs since the decision was made to conduct the trial (February 2016), which impacted negatively on recruitment during the course of the trial:

I think because it took us a long while to get it going and I think the whole landscape changed almost just before we started recruiting [. . .] and my, my belief is that was largely to do with the guideline the RCOG guideline . . . the trial was an independent endeavour but meanwhile there were lots of us working in the background to improve care for women with hyperemesis, which meant enabling GPs and hospital doctors to prescribe appropriately for it, and ironically our success in doing that has led to the failure of the trial.

TS16

This shift in practice was evident in both primary and secondary care in spite of RCOG step-up guidelines¹² recommending the use of the trial drugs as a second-line medication and confirmation from all trial sites that local guidelines on antiemetic use were consistent with RCOG guidelines.¹² However, participants indicated that not all clinicians at the study sites and in general practices in their localities adhered to these guidelines when they were treating women with severe NVP because they had become used to prescribing the trial drugs:

It was not easy a trial to start with because of all the bad practice which is all very well set where people would prescribe whatever antiemetic they would feel like at that time rather than follow RCOG standardised step-up protocol.

TS17

Some of the research staff at sites were unaware of the antiemetic-prescribing culture within which the trial was attempting to operate:

. . . we just hadn’t realised how many women were already on those study medications from GPs and things so it’s been a big eye opener to us about what kind of prescribing practices are out in primary care.

TS20

As a result of this culture of prescribing in primary and secondary care, women were frequently presenting with severe NVP having already been treated with one of the trial drugs, rendering them ineligible for the trial. In some regions, this practice was widespread among GPs, whereas in others the problem was in secondary care:

I did think initially when the trial started that a lot will be ineligible because the GPs prescribing. But I was surprised actually we could catch a lot who were eligible and it was, we were messing up their eligibility at secondary care level which was sorted later part. So I don’t think GPs were a major problem here.

TS17

Organising the delivery of the EMPOWER trial

The successful delivery of the trial required organisational co-operation from clinical staff across departments, that is liaison between maternity/obstetrics, gynaecology, emergency and pharmacy departments. Staff at one trial site were more familiar with trials that were based solely in maternity services, which meant additional effort working with gynaecology colleagues:

This trial was a little bit different in that it was more gynae[cology] than obstetrics which gave I think the trial team in every unit a little bit of challenge whereas we’re used to approaching women in pregnancy further on in pregnancy.

TS16

Nevertheless, there was evidence of good working relationships between clinical staff in maternity and clinical staff in gynaecology day units, and research staff, to facilitate recruitment:

We work with quite closely with the early pregnancy unit and also gynae[cology] emergency, and what we decided to do was get the clerical staff that work on reception to contact us, bleep us when the women were coming in. Cos we’ve got quite a good relationship with that area.

TS21

Where there were staff with both a clinical and a research role, this facilitated the recruitment process for the EMPOWER trial:

One of our research nurses works there in her substantive role as well. So she does some research and some early pregnancy and, so that’s brilliant because she knows all the systems she knows all the staff and consequently they’ve got to know us as well through her. We do, they are very good at referring people to us.

TS02

In some sites, the responsibility for the care of the patient was restricted to unit or ward staff, with research staff providing a minimal amount of help. This may be because the number of research staff was limited. In other sites, however, research staff contributed to the care of eligible women while they were considering the trial:

. . . what I’ve tended to do is not just give the information sheet and come away but I’ve usually kind of helped with starting i.v. fluids and you know making sure that if the patient has any questions that I’m there to answer, you know.

TS04

Good working relationships with the NCTU staff at Newcastle University were critical to the operational delivery of the trial. In the initial stages, there were several glitches, mainly to do with the unavailability of trial staff when they were urgently required:

. . . we’ve tried to call the EMPOWER office and there’s never been anybody who actually works on the trial available to talk to us.

TS05

. . . you’re usually be able to ring by phone if not but there was no back-up there was no way of getting a hold of anybody so we just had to scrap it which is not, not great.

TS20

Some staff commented that the trial database was not user-friendly and that it was unhelpful having separate log-ins to recruit and randomise women:

. . . the whole recruitment process and the databases, having separate databases for randomisation and things like that it can be very confusing.

TS08

Nevertheless, as the trial progressed, and as lines of communication were more firmly established, sites reported how helpful the NCTU staff were in supporting the management of the trial:

Yeah we’ve never had a problem we’re in contact regularly we get a lot of e-mails from them if there’s anything on the database that we haven’t done right, or if there’s any discrepancies they’re quite good at liaising with us and managing like that, so yeah all our contact with the team has been really positive, I think it’s a lovely trials team to work with.

TS11

Demands of the EMPOWER trial protocol

The trial protocol stipulated that women had to have taken first-line antiemetic treatment for a minimum of 24 hours before they could be considered eligible for the trial. Most site guidelines for the management of NVP did not recommend the admission of women, especially with moderate symptoms, who were started on first-line antiemetic therapy. Thus, trial staff reported that they were unable to easily determine if potentially eligible women responded to first-line treatment:

It’s, like we’ve got to get her in and out we can’t wait 3 hours and then see how she feels and then put her in EMPOWER because they don’t tend to like admitting them either.

TS03

So if they’ve never had anything before, usually in that time they’ll just have one dose of cyclizine and then go home. So it’s then waiting to see if they come back and then we can consider them.

TS05

Women could return at any time, including out of hours or at the weekend, when site research staff were not available. Nevertheless, sites were informed that women could be approached about future consideration of the trial when they were on first-line treatment. This advice was followed at some but not all sites:

We’ve also seen a few people who have come in and maybe just having their first-line treatment and just mentioned the study to them and just said we have a study running in the unit you wouldn’t be eligible for it now but bear it in mind if this treatment doesn’t work for you we may come and see you again. So just to raise a bit of awareness with these women who might be coming in.

TS13

One patient suggested that information should be provided about the trial at her first visit to the MAU:

I didn’t know whether you could like send something out after I’d been there in the first time, to say if you come back there is a trial. So I would have had prior knowledge.

EMP1003

The trial was perceived to be particularly challenging because it was a CTIMP with a complex design consisting of four groups, as well as requiring a trial population of pregnant women in a fragile physical and emotional state at the point of recruitment. For this reason, research staff were required to adopt a sensitive approach and give eligible women enough time (1 hour) to consider the trial information.

Women’s perspectives provided insight to the recruitment processes. Although one patient felt that she needed more time to think over her decision to participate, another patient declined because she thought that her wait for treatment would be lengthened because of the trial:

I think how it was explained was that obviously she would come back with, like in an hour, and then there was paperwork to be filled out so obviously that took, would have taken time and then that had to be uploaded in to the computer system to decide which treatment I would then go ahead with . . . I mean I think we’d had a chat about it so we’d pretty much made our mind up probably what 15 minutes after she’d left us. So had she maybe been around a little bit sooner than the hour that we were expecting, that might have helped a little bit.

EMP1003

Research staff reported that they felt that the recruitment process could be quite quick once women had agreed, whereas approaching the women and making sure that the information was understood took longer:

. . . obviously going to the patient making sure that they’re eligible, giving them the information sheet and obviously you want them to have the information sheet for the length of time so that they’re happy, that they understand and that I understand or know that they understand what we’re asking of them.

TS04

Some participants reported difficulties in accessing medical staff who were GCP trained. One group highlighted were GP trainees who were perceived to have less interest either in the specialty area or in contributing to the research:

I found I was quite successful with the doctors who were career trainees in obs[tetrics] and gynae[cology] but I didn’t have as much success with the GP trainees because they were here for 4 months or 6 months and I didn’t find that, they didn’t feel that for the time that it took that it was useful for their CVs.

TS19

On the other hand, other PIs reported that they had ensured that there were enough GCP-trained doctors available:

So most of our senior trainees have completed the GCP training and they are on the delegation log for a number of research trials.

TS22

If women agreed to participate in the trial, the protocol stated that an assessment of treatment efficacy had to be made at 12 hours to see if they were responding. However, some women felt that this time period was too long a time to endure:

I mean I do reassure them by saying after 12 hours if the medication wasn’t working you’d be withdrawn from the study and [ML – yeah] but I guess, you know, for the women who are feeling really bad, 12 hours probably seems like a very long time.

TS01

An additional issue was the chance that participants might receive the double dummy for 12 hours, an issue highlighted by both site staff and prospective participants:

And if they come in when they’re really, really ill when they come in, I could sort of see people not wanting to do that really just risking getting the placebo.

TS15

. . . I was really, really worried that I had two placebos.

EMP1702

The need for double blinding meant that participants were not aware of the trial drugs that they were taking. Concerns about unblinding in the event of treatment failure or trial withdrawal were dealt with through the initiatives of the research teams (e.g. by prescription of both trial drugs). However, this did not take away the desire of some participants to know what medication they had received after the trial was completed:

I just think sort of ethically I would although you sign up to do a trial you do have a right to know what you took because now I’m guessing I was on metoclopramide so in a way, or you know guessing that’s the best one for me to take erm yeah I would, I would like to know what I took.

EMP1702

Publicity and education about the trial

For the trial to be successful, relevant health-care professionals had to be made aware of the requirements of the trial so that appropriate referrals were made. Medical staff also had to be trained to approach patients, confirm eligibility, take written consent from the patients and prescribe the trial drugs. In addition, women with severe NVP needed to be made aware that there was an option for them to join a trial that might benefit them and other women with the same condition. If health-care professionals were not offering information about the trial and/or were not adhering to RCOG guidelines¹² on antiemetic use, this could preclude potentially eligible women from participating in the EMPOWER trial.

The impact of the unpredictability of presentation to secondary care

From staff accounts, there did not seem to be predictable patterns of when patients presented to sites at which recruitment would take place. Women could present with symptoms out of hours, at weekends or on bank holidays, perhaps because they could not find the time during the week when they were at work or they had other family commitments. Women might also have to wait until evening for transport to the hospital by their partners who are at work:

. . . if they’re at home and then someone else brings them cos they’ve been at work all day and then they’ll bring them in the evening.

TS13

Women who attended at the weekend as an emergency were often given a weekday follow-up appointment at a specialist MAU or GAU (TS16). Some staff interviewees made reference to the maternity assessment centre (MAC) or the antenatal day unit, whereas others referred to the early pregnancy unit or the early pregnancy assessment unit.

In addition, care pathways differed from hospital to hospital depending on the gestational age at presentation. This influenced when women were referred to specialist assessment units (AUs) at which they could be approached by trial staff. At some sites, to ease the workload at their MAU, women at lower gestations were seen within the gynaecology department (either in an early pregnancy unit or on the gynaecological ward). This meant that research staff had to try to adapt their recruitment strategies accordingly:

I think it depends on the way that your trust is set up so we, we have the obs[tetrics] med team and they go and see the ward every single day to see if anyone’s been admitted with hyperemesis.

TS08

In some units, it worked well that the local A&E department policy was to transfer patients with NVP directly to the specialist MAU or GAU. Some sites provided out-of-hours or 24-hour access to specialist assessment, but several staff made reference to the issue of women presenting out of hours leading to their inability to recruit:

But it just, it’s out of hours you know they tend to come in, I think they battle on through the day and come in on an evening or a weekend or maybe they’ve battled on at work and they they’ve come in out of hours then, you know, so. But it’s like all trials you’re going to miss you cannot be there 24 hours a day, 7 days a week, so it’s hard.

TS09

However, there were also pathways in which women were encouraged to attend in the daytime if they were able:

But we have a system where, doesn’t always happen but if somebody’s got hyperemesis and they’re not too unwell I think certainly from evenings onwards and overnight the appointment is made for them to come in the morning and they can have their treatment in daylight hours and they can speak to the research midwives.

TS15

On the other hand, some women presented with such severe symptoms that they were admitted as inpatients rather than being seen in an AU, which meant that they could potentially be missed by research staff if they focused efforts on screening women referred to the AU. When the protocol was changed to allow trial drugs to be taken orally, inpatients treated intravenously were likely to be missed:

I think once they’ve got to the point where they’ve being admitted to hospital and they’re in and they’re vomiting and they’re being rehydrated they’ve got i.v. access so they tend to get given i.v. drugs.

TS20

The impact of the infrequency of eligible women

Clinicians reported that the number of women who required specialist assessment and treatment for severe NVP was small. Familiarity with trial processes took some time to achieve, and research staff commented that the small number of eligible cases meant that they gained limited experience with the recruitment processes:

And I think the relatively infrequent presentation of women requiring treatment for nausea and vomiting in early pregnancy again makes it quite challenging . . . And I think that made it difficult for people to, for the clinical staff to always be on the ground, and always be on the ball if you like.

TS18

. . . it’s not like you can find your stride very quickly it takes, and then by the time you get another recruit you think what am I supposed to do again.

TS05

However, to some extent, features of help-seeking in the affected population offset this barrier. For example, some women attended the site several times, and staff reported that this gave opportunities for research staff to introduce the trial to them:

Yes to like attend you know the antenatal day unit because for 2 weeks I was constantly going in every other day because my, my, I was being dehydrated so nothing was working, sickness was at very peak. Then I kept going in to have some fluids because I was, my ketones in the urine was really high.

EMP1402

Women who were ineligible

Women who were eligible

Confidence and morale

The amount of paperwork, together with the slower than expected recruitment into the trial, had an impact on the confidence of research staff in carrying out trial processes to the standards required in a CTIMP. Accounts from staff on the delegation log indicated that they could be reluctant to take on the task of recruiting a patient because the process involved felt onerous:

. . . it sounds awful but you kind of dread actually recruiting someone because you know it’s so complicated, you know, it is really hard all the information you need to remember.

TS08

Failure to recruit was because of a number of administrative hurdles and in the early stages was a source of frustration:

And we had given our first recruit the medication orally because she asked for it that way and hadn’t realised that the first dose had to be i.v. so it was that, with that lady . . . and then there’s been some issues with sending things over different e-mail addresses . . . We have had to call to clarify at the point of recruitment on every single occasion.

TS02

Faced with the challenges of recruitment, enthusiasm for the EMPOWER trial waned. However, successfully recruiting new participants encouraged staff to press on:

. . . we’ve found, you know women, we were kind of getting a little bit, that we keep on looking and we can’t find women that are fitting the criteria but having the recent two recruits has been a real boost.

TS06

Morale was also boosted when comparisons were made with other trial sites through the regular updates provided by NCTU, and PIs reassured research staff that they were not alone:

. . . they feel better that it’s not just our site that isn’t doing as well. I mean they know that, you know, it’s a national thing it’s not just one particular site doing, not doing as well.

TS14

When staff encountered an AE, their morale was affected; however, when they saw participants benefiting from the care provided, they were motivated to carry on:

. . . the midwives have been on a shift where women have taken part it feels like they’ve got an enthusiasm for the study they can see, you know, these women kind of felt better in the short term. We know it might just be the fluids, we don’t know what really drug that we give them but it’s kind of a positive feeling which kind of helped, you know, remind the staff that it’s a good study.

TS06

Recruitment and trial processes

Recruiting vulnerable women suffering from severe NVP was frequently reported as challenging:

I would even say that like we’ve done studies on labour ward where you’re speaking to ladies in labour and sometimes that’s been easier than trying to speak to a lady who’s puking her guts up on the, on the assessment unit.

TS10

Trial staff were dependent on their clinical colleagues to assist, but they reported difficulties because of the amount of paperwork involved during recruitment that required them to be present. This made recruitment out of hours virtually impossible:

. . . it’s a little bit more time consuming that clinical staff aren’t able to fit that in. If it’s just a short period of time it’s not quite so bad and they will accommodate but this is a little bit more involved really I think.

TS04

. . . it’s always been really difficult to recruit patients out of hours when you’re relying on your medical teams unless it’s really straightforward and obviously there’s a lot of paperwork that goes along with EMPOWER recruitment isn’t there.

TS10

Staff also reported that data collection during the course of the trial could be demanding:

All the paper work you need to give, all the forms you need to fill out, you have to make sure the patients know, you know, when to fill it in and you know, you’ve got to call the patient, well if they’re discharged at 12 hours, they’re clearly fine at 48 hours, at you know 4 days, 10 days etc.

TS08

However, staff accounts indicated that good teamwork mitigated against the demands of the trial processes. Less confident staff reported that they felt supported by their colleagues:

But before I’d done those two I was a bit like if we get one can someone come with me so we can kind of go through it together.

TS03

Principal investigators in the trial valued their research midwives or nurses, especially if they did most of the liaising with fellow clinicians and co-ordinated the work of the trial:

I think the main staff on MAU they were aware of the study and I think I must give credit to [name 1] who was like the team leader and the co-ordinator there. So she used to liaise with [name 2] who was the research midwife and [name 3]. So [name 2] and [name 3] also co-ordinated closely with [name 1] and all the consultants too were aware of the study.

TS22

Patient counselling and staff empathy

There were reports describing how staff counselling about the risks involved in participating in the trial, particularly reassurance about patient care during and after the trial, was critical in the recruitment process:

. . . sometimes they worry about the placebo–placebo but once you explain to them that you are actually having the i.v. fluids they’re not leaving you with, you know, absolutely nothing and if it doesn’t work if you’re feeling unwell we can take you off straight away, they do feel more reassured.

TS08

The following observations from one of the PIs summarises the importance of staff counselling and care, not just from research staff but from clinical staff managing women at the point of admission:

. . . if, when a women came in with nausea and vomiting, the admitting midwife had a positive word to say for the study in ways, not just of course through the study, but other ways in which the woman would feel assured she was going to be taken care of, that makes it much easier for the recruiting team to come on board and make that, make that case effectively for taking part.

TS18

However, while providing information about the trial, convincing patients that rehydration with i.v. fluids was a form of treatment in itself could prove difficult.

At the same time, staff also expressed empathy for the women whom they were trying to recruit onto the trial, such as for the long wait that women had to endure for treatment:

I do sometimes feel that the women have a rough deal if you like, because they go to the GP and then they come to the hospital and it’s sometimes a while before they have, you know, the treatment.

TS04

If patients were given information about where they were in the queue to see a doctor, some participants felt that this could make women more open to considering the trial:

So we want to make sure that they know that their care isn’t being held up by us as soon as the doctor comes they’ll see them so, we just say we’re just seeing them while they’re waiting for the doctor anyway and I think that makes them feel a bit better cos they don’t feel like we’re like sort of delaying their treatment if you know what mean.

TS11

When approaching women, research staff had to be sensitive to how women were feeling at that point in time, and to approach them appropriately. One patient felt that the research was more important than the care that she was desperate to have:

But it was kind of, felt as though I’m waiting for some treatment which I’m not getting but the research information was more important, it felt to me, cos I had got a visit about three times.

EMP1401

On the other hand, many patients gave positive and appreciative accounts of how they were approached, whether they accepted or declined participation:

She explained everything in detail she gave me the time I needed to process. She, basically sat there and explained it all in all ways, in a way that I understood.

EMP1201

. . . when I spoke to them you know they said it was about my health and quality of life so they were very good at making sure I knew what my options were to withdraw from the trial if I needed to.

EMP1703

Principal investigators mostly had complete confidence in the abilities of their research midwives/nurses to counsel patients about their options:

They don’t have to do obviously you know sort of, in terms of consenting and everything but they’re happy to do all the talking, and you know very objective and unbiased and stuff so.

TS15

From other feedback from interviews that described empathy for women with NVP, it appeared that staff commitment to the trial could be compromised if they felt that participation had a negative effect on the well-being of their patients. To illustrate this, when a SUSAR, a fetal abnormality that could have had a link to one of the trial drugs, occurred research staff reported feeling distress and a sense of responsibility for the patient’s treatment with the trial drugs and the subsequent termination of the pregnancy.

Staffing levels and support

Appropriate levels of staffing were also critical to the trial; recruitment processes could have been affected by staff being on leave:

I only work 15 hours and because we’ve been short of staff in the, in the team, really everybody, although we’ve got about, we had about three staff until somebody just came back off long-term sick. Because we’re so short when we’re at work we’re just mainly focusing on our own studies.

TS10

Seen from the perspective of one PI, as recruitment numbers were small, the workload for midwives or nurses was relatively lighter than for busier trials with a large number of participants:

. . . there were few patients, even if we’d recruited to target it wasn’t many patients per centre and, therefore, the workload for the midwives [they were] working with the women, supporting women and the questionnaire . . .

TS16

From the perspective of other staff, however, much of their work involved keeping the EMPOWER trial on the agendas of their colleagues. Maintaining their familiarity with trial processes that were complex was crucial, and this was helped by the co-operation and support from fellow staff:

I’ve got lots of support though we’ve got a really good research team here and everybody kind of helps out each other. So always kind of felt like we’ve had plenty of support during the whole thing.

TS05

At some sites, because the AUs were so busy, research staff were involved in administering i.v. fluids and obtaining the medication. Good relations were fostered when research staff understood the pressures that clinical staff were under and the added burden placed on them by participation in trials:

. . . they’re working flat out and we’re constantly asking them to remember something new and something extra and doing extra jobs. So the more you can kind of give back and take off them in terms of their workload then yeah like you say the better the relations between the two are.

TS20

There were accounts that when trial staff did manage to recruit a woman to the trial, there was a real sense of achievement, despite the fact that it required a lot of effort and teamwork:

The other day we did get the woman in the study there were three members of the research team that stayed 2 hours late and one of the consultants who stayed an hour over to finally get somebody in. So we certainly feel like we’ve gone above and beyond . . .

TS20

Trial-specific factors

The issue of adherence to the RCOG guidelines¹² was raised as a hurdle to recruitment, but this was not necessarily the case if women were approached by research staff before second-line treatment was prescribed. The problem arose when women attended AUs providing out-of-hours or 24-hour access at times when research staff were not available. One solution was to raise awareness with clinical teams along the care pathway so that women could be signposted to the trial. However, publicising the trial among primary care clinicians was generally not felt to be the way forward because of the scale of the task.

Operational difficulties at the start of the trial were to be expected, but the smooth working of the trial was generally facilitated by good staff and professional relationships between maternity and gynaecology departments, and with the Clinical Trials Unit. At most sites, there was a dogged determination by lead staff to work hard at liaising with colleagues across departments and educating them about the trial and the need to follow RCOG step-up guidelines.¹² Nevertheless, the key issues for all research staff were the unpredictability of women presenting to services and the infrequency of their presentation. These barriers affected the efforts of research staff to recruit women in the best way possible.

The demands of the complex trial design added to the recruitment challenges faced by staff. Staff spent most of their time finding ways and means of maintaining the profile of the trial; reminding their colleagues about trial process requirements; and tracking, identifying and screening women. Recruiting women was made more challenging because women with severe NVP were physically and emotionally vulnerable; therefore, time and patience were required. Recruitment to the EMPOWER trial involved more than just paperwork; it required high levels of empathy and patient care. Time was needed for the patient to consider the information carefully before making an informed decision about consenting to participate or not. Continuity of care for trial participants was provided and, where appropriate, trial drugs were prescribed as patients requested. In this way, the issue of unblinding did not seem to be as problematic an ethics issue as when first considered.

Women who were eligible to participate were discouraged at the thought of their condition not improving because of the risk of receiving a double dummy. Their experience of severe NVP had an impact on their decision; all that they could think of was a quick, effective solution. The effort and time required for participation, to read and understand the information, and to wait to be entered into the trial were significant for several women. However, for altruistic reasons, there were women who willingly participated to help others.

Site-specific factors

The care pathways that women took in seeking help for NVP were dependent on the trial site, the catchment area and the health service provision therein. Ethnic composition, health behaviour and antidepressant use were factors at each site. Trial sites that recruited more women were characterised by primary care providers that tended not to prescribe trial drugs as first- or second-line antiemetics. Sites that were unsuccessful were those that had to work hard to change the culture of prescribing antiemetics within secondary care. Relationships with A&E departments varied across trial sites; A&E departments in which research was not a priority was a factor. The availability and engagement of GCP-trained staff also varied across sites. Staff confidence and morale varied depending on recruitment rate, recruitment success, which ranged from 11% to 63% of eligible patients.