Restricted fluid bolus versus current practice in children with septic shock: the FiSh feasibility study and pilot RCT

Feasibility study

1. To review and explore, with input from parents/legal representatives, acceptability, potential barriers, information/documentation, decision-making and deferred consenting [or research without prior consent (RWPC)].

2. To review and explore, with input from parents/legal representatives, the potential patient-centred primary, and other important secondary, outcome measures.

Pilot trial with integrated perspectives study

1. To test the willingness of clinicians to screen, recruit and randomise eligible participants.

2. To estimate the recruitment rate.

3. To test, following randomisation, the delivery of and adherence to the intervention and demonstrate separation between the groups.

4. To test acceptability of the deferred consenting procedures and documentation.

5. To test follow-up for the identified, potential patient-centred primary and other important secondary outcome measures and adverse event (AE) reporting.

6. To inform the final selection of a patient-centred primary outcome measure.

7. To estimate the characteristics of the selected patient-centred primary outcome measure to inform sample size estimation.

8. To inform the content of, and time needed for, final data collection.

Progression criteria

The objectives would be reported against the following progression criteria:

1. Formal sign-up of sites (the required number of hub-and-spoke models) to achieve the overall sample size based on the observed recruitment rate, as determined by the pilot trial objectives concerning identification, recruitment rate, acceptability of deferred consenting procedures, selection of the patient-centred primary outcome measure and sample size estimation.

2. The recruitment rate achieved in the pilot trial (taking into account loss of participants as a result of declining deferred consent) being consistent with achieving the overall sample size required – as determined by the objectives concerning selection of the patient-centred primary outcome measure and sample size estimation – in an acceptable time frame from the sites giving formal sign-up in the first feasibility objective.

3. At least 80% of fluid bolus resuscitation in the external pilot study being delivered at the correct volume and timing ± 10%, as determined by the external pilot study objective concerning adherence.

4. The absolute total volume of fluid administered during the first hour and during the first 4 hours being lower (by ≥ 25%) in the more restrictive fluid bolus resuscitation group than in the current recommended fluid bolus resuscitation group, as determined by the external pilot study objective concerning separation.

Study design

The FiSh feasibility study was a qualitative semistructured interview study.

Research governance

The FiSh feasibility study was sponsored by Imperial College Healthcare NHS Trust and co-ordinated by the University of Liverpool and Intensive Care National Audit and Research Centre (ICNARC) Clinical Trials Unit (CTU). An ethics application was made to the North West – Liverpool Central Research Ethics Committee (REC) on 16 October 2015 and a favourable opinion was received on 15 December 2015 (REC reference number: 15/NW/0913). The protocol is available at www.journalslibrary.nihr.ac.uk/programmes/hta/1304105/#/ (accessed 23 May 2016). Local NHS permissions were obtained for the NHS trusts that participated as Participant Identification Centres (PICs) for recruitment routes 1 and 2 (see Recruitment).

Study management

The FiSh feasibility study was led by a qualitative researcher co-investigator (KW) with support from the Study Management Group (SMG), which included the chief investigator (DI) and other co-investigators (AC, DAH, NJ, MDL, PRM, SN, MJP and KMR). An experienced research associate (CBOH) was employed to conduct, transcribe and analyse the interviews.

Network support

To maintain the profile of the study, updates were provided at national meetings, such as the Paediatric Emergency Research in the UK and Ireland (PERUKI) annual meeting and the biannual Paediatric Intensive Care Society Study Group (PICS-SG) meetings.

Patient and public involvement

Two parents (NJ and AC) of children with experience of admission to hospital with severe infection were co-investigators and members of the SMG. They provided valuable input into the design and conduct of the study, including reviewing documents for parent interviews [e.g. draft pilot trial participant information sheets (PISs) and a list of potential outcomes, including outcome descriptions] and informing study recruitment approaches (i.e. identification of social media groups and charities). They were also involved in the review of study progress and findings.

Design and development of the protocol

The design and development of the protocol, including sample estimation, recruitment strategy and interview topic guide, were informed by previous trials conducted in paediatric emergency and critical care in the NHS and earlier research. 17–20 Previous relevant research was used to develop an interview topic guide (see Appendix 1) and draft PIS for the pilot trial (see Appendix 2). 18,19,21 The topic guide contained open-ended questions and prompts to help explore parents’/legal representatives’ views on the acceptability of the FiSh trial, including the draft pilot trial PIS and approach to consent. A separate set of questions was developed for bereaved parents/legal representatives.

Amendments to the study protocol

Following receipt of a favourable opinion of the study protocol from the REC on 15 December 2015, there were no amendments.

Recruitment

Recruitment and sampling procedure

Based on previous studies,19 it was anticipated that 15–25 parents/legal representatives would need to be recruited to reach data saturation. The study procedures for recruitment are summarised in Figure 1.

FIGURE 1.

Summary of study procedures for recruitment of participants.

Recruitment route 1: postal contact

Clinicians used hospital medical records to identify the 15 most recent parents/legal representatives (including up to five bereaved) of children who met the inclusion criteria. Those identified were sent a postal invitation, including a covering letter and FiSh feasibility study PIS describing how to register interest in taking part.

Recruitment route 2: advertising in paediatric intensive care units

Posters and copies of the FiSh feasibility study PIS were placed in family/relative waiting rooms near the PICU.

Recruitment route 3: advertising online, including social media

An advertisement was posted on Twitter (Twitter Inc., San Francisco, CA, USA; www.twitter.com) and Facebook (Facebook Inc., Menlo Park, CA, USA; www.facebook.com), which invited parents/legal representatives to register interest in participating in the study. Relevant charities and parent support groups were asked to place the advertisement on their website and social media.

Interviews

Participating NHS trust hospitals

Expressions of interest were received from seven NHS hospitals. Local research and development approval was obtained for four NHS hospitals in January 2016 for recruitment via routes 1 and 2.

Participants

A total of 58 parents registered interest, of whom 29 were screened (Figure 2); three were deemed ineligible and five did not confirm a date for interview. Data saturation was reached when 21 parents from 20 families were interviewed. 22,23,35 No individuals identified themselves as legal representatives and, therefore, this term is not used in the remainder of this chapter.

FIGURE 2.

The FiSh feasibility study recruitment process.

Parental perspectives

Sites

Sites were set up under a ‘hub-and-spoke’ model within three geographical regions: Bristol, London (North Thames) and Southampton. The ‘hubs’ were four regional hospitals with integral PICUs (two hospitals covered the same region), three of which also had an integrated ED and paediatric assessment unit [(PAU) or equivalent]. The ‘spokes’ were nine hospitals within the selected regions, which had an ED but did not have an integral PICU. The criteria for inclusion as a study site were:

• sites meet all responsibilities as stated in the pilot trial clinical trial site agreement

• sites identify and sign up a local PI

• sites identify a responsible FiSh trial research nurse (to be funded, or part-funded, centrally)

• sites agree to incorporate the FiSh trial into routine ED and PAU (or equivalent) activity, in particular, by highlighting the importance of screening at ED and PAU (or equivalent) presentation

• sites agree to adhere to randomisation allocation and to ensure adherence to the protocol

• sites agree, when possible, to recruit all eligible patients to the FiSh trial and to maintain a screening log.

The ‘hub’ sites were chosen during the grant application stage. For the nine ‘spoke’ sites, invitations for expressions of interest were sent to lead clinicians in emergency and critical care medicine at NHS hospitals within the three regions. Sites that expressed interest were asked to complete a site feasibility questionnaire. After reviewing the questionnaires, teleconferences were held with all potential hospitals to assess the engagement and support of the clinical team, the integration of the pilot trial into clinical practice (e.g. screening, randomising, delivering the intervention) and any potential challenges/barriers to running the pilot trial at their site. The final sites included in the study are listed in Table 2.

Recruitment

The trial procedures for recruitment and follow-up of participants are summarised in Figure 3.

FIGURE 3.

Summary of trial procedures for recruitment and follow-up of participants.

Treatment groups

Participants were randomly allocated to fluid bolus resuscitation using boluses of two different sizes over a 4-hour period (Figure 4). The period was divided up into 15-minute cycles, with one bolus of either 10 ml/kg or 20 ml/kg to be delivered within each cycle, at a rate left to the discretion of the treating clinician. There was no attempt to blind clinicians to the allocated treatment group as this was not felt to be pragmatic, given the ED setting and the need for clinical staff to prescribe, draw up and administer the fluid boluses. The maximum volume of fluid that could be given per bolus was either 500 ml (for those allocated to 10-ml/kg fluid boluses) or 1000 ml (for those allocated to 20-ml/kg fluid boluses). Other interventions, including referral to PICU, were not protocolised and were left to the discretion of the treating clinician. As a pragmatic pilot trial, FiSh also did not dictate the use of specific fluid products. At the end of each cycle, should the age-adjusted clinical signs of shock [defined as age-adjusted hypotension (Table 3) or prolonged CRT of ≥ 3 seconds] persist, then another bolus of the same size (i.e. according to the randomly allocated group) was given – again, at a rate left to the discretion of the treating clinician but within a 15-minute cycle.

FIGURE 4.

Fluid bolus resuscitation algorithm.

TABLE 3 - Age-adjusted hypotension shock criteria

SBP, systolic blood pressure.

Age group	SBP (mmHg)
0 days–1 week	< 60
1 week–1 year	< 70
1–2 years	< 75
2–5 years	< 80
5–12 years	< 85
12–16 years	< 90

The cycles repeated until either the end of the 4-hour resuscitation period or any of the hold criteria occurred. If a participant’s age-adjusted clinical signs of shock resolved or if a participant showed signs of fluid overload (defined as pulmonary oedema – either rales on auscultation or pulmonary oedema fluid in the endotracheal tube – or new or increasing hepatomegaly), the delivery of further fluid boluses was withheld. If, within the 4-hour resuscitation period, fluid boluses were indicated again, that is, age-adjusted clinical signs of shock were present without signs of fluid overload, then the cycles were recommenced with the allocated boluses until the end of the 4-hour resuscitation period. After the 4-hour resuscitation period, any further treatment was at the discretion of the treating clinician.

The maximum volume of fluid that could be given within the pilot trial protocol, regardless of allocation, was 120 ml/kg (excluding the 20-ml/kg fluid bolus pre randomisation). If > 120 ml/kg of fluid was required, then further treatment was at the discretion of the treating clinician.

Participants

Based on previous research and the pilot trial sample size,18 it was estimated that 100–150 parents/legal representatives (including both mothers and fathers of randomised children) would be recruited to the questionnaire element and 15–25 parents/legal representatives would be recruited to the interview element. In addition, three focus groups at the hub sites and up to 10 telephone interviews with staff who could not attend the focus groups were planned.

Interviews and focus groups

Data analysis

The analysis of qualitative data were broadly interpretive and iterative (see Table 19). 31,32 Informed by the constant comparative approach, the aim was to provide accurate representation of parental views on trial design and acceptability. 29,33–35 NVivo 10 software was used to assist in the organisation and coding of data. Data from voting handsets were cleaned and entered into IBM SPSS Statistics version 20.0 (IBM Corporation, Armonk, NY, USA). Descriptive statistics are presented with percentages.

Characteristics of participating sites

The characteristics of the 12 EDs recruiting patients to the FiSh pilot trial, compared with all EDs participating in the PERUKI research collaborative, are shown in Table 4. The characteristics were broadly similar. There was a lower percentage of paediatric hospitals participating in the FiSh pilot trial than in PERUKI, and a slightly lower proportion with PICU on site, but these characteristics are likely to be over-represented among PERUKI participants relative to the national picture.

Characteristics of patients at baseline

The treatment groups were, on the whole, well matched at baseline (Table 7), with slight differences in age and, consequently, weight. The median age of patients was 11 months for the 10-ml/kg fluid bolus group and 2 months for the 20-ml/kg fluid bolus group (Figure 12). The participants randomised to the 10-ml/kg fluid bolus group had a mean weight at baseline of 13.3 kg, compared with 9.1 kg for the 20-ml/kg fluid bolus group, which is in line with the slightly older age distribution in this group.

FIGURE 12.

Distribution of age, by treatment group. (a) 10-ml/kg fluid bolus group; and (b) 20-ml/kg fluid bolus group.

More than half the participants were male (61.5% in the 10-ml/kg fluid bolus group; 52.9% in the 20-ml/kg fluid bolus group). Most participants met the shock criteria because their CRT was ≥ 3 seconds (76.9% in the 10-ml/kg fluid bolus group; 88.2% in the 20-ml/kg fluid bolus group). The PIM2r score was balanced across treatment groups.

There were 37 participants (51.3% in the 10-ml/kg fluid bolus group; 50% in the 20-ml/kg fluid bolus group) recorded as having a confirmed infection. For 32 of these, the organism causing the infection was identified (Table 8). Of these, only 13 were bacterial infections in sterile sites. These included five Escherichia coli infections, two Enterococcus faecalis infections, three Streptococcus (two group A and one group B) infections, one Staphylococcus aureus infection, one Neisseria meningitidis infection and one Gram-positive coccal (unspecified) infection. There were 18 viral infections, 15 of which were respiratory pathogens (eight respiratory syncytial virus, five rhinovirus, one influenza A and one metapneumovirus), two of which were identified as enterovirus on blood polymerase chain reaction (PCR) analysis and one of which was identified as rotavirus in stools. One patient had falciparum malaria.

Characteristics

Of the 20 parents (19 mothers and one father) interviewed by telephone and 45 parents (34 mothers and 11 fathers) who completed a questionnaire, all had consented to the use of their child’s information in the pilot trial and none was bereaved. Parents interviewed were recruited from 10 out of the 12 (83%) recruiting sites and questionnaires were received from 9 out of the 12 (75%) recruiting sites. Parents were interviewed a median of 26.5 days after admission (range 14–96 days) and the median length of hospital stay for their child was 3 days (range 1–23 days). One parent had previous experience of their child participating in medical research; none had experience of RWPC. Telephone interviews with parents took between 30 and 60 minutes. All 45 parents completed the questionnaire on site, before their child was discharged from hospital.

Characteristics

Of the 27 participating site staff, 22 (81%) were involved in the clinical care of children [17/20 (85%) of focus group participants and 5/7 (71%) of telephone interview participants]. The staff sample included 14 (52%) nurses and 13 (48%) registrar- or consultant-grade doctors with a modal range of 1–2 years of experience in recruiting to paediatric clinical trials. Telephone interviews took between 30 and 60 minutes and focus groups lasted approximately 90 minutes.

Research without prior consent is acceptable

Before any discussion of the pilot trial during interviews, parents were prompted to describe their understanding of RWPC to explore explanations provided to them during the consent discussion, as well as their views on this method of consent. Although no parents had any previous knowledge of RWPC, the majority (n = 15/20, 75%) were able to appropriately describe this approach, explaining and supporting its use in emergency situations:

They just said that in these situations they just obviously stabilise the child as a priority and then discuss it afterwards because then there’s obviously no pressure to be involved and that’s sort of how it was explained. Um, so I hadn’t heard of it, no, but it doesn’t surprise me. In an emergency situation, you do what you’ve got to do, don’t you, and then deal with stuff afterwards.

P25, mother, telephone interview

Following the general discussion of RWPC, the focus shifted towards parents’ views on the use of this method in the pilot trial. Consistent with previous studies18,19,27 and the feasibility study findings described in Chapter 2, several parents (n = 7/20, 35%) were initially shocked or surprised to discover that their child had been enrolled into the pilot trial without prior consent. As shown in Table 18, 13 out of 45 (28%) parents who completed a questionnaire agreed with statement 2: ‘I was initially surprised to find out that my child had already been entered into FiSh’. However, following tailored explanation from site staff about why informed consent was not sought,30 all parents stated that they supported RWPC in this pilot trial:

I was initially surprised that obviously this had happened, because obviously the consent happened afterwards, but I wasn’t sort of concerned or anything, and then they explained everything really clearly, and the information leaflet was very good as well.

P12, mother, telephone interview

Indeed, the majority (n = 39/45, 86.7%) of parents who completed the questionnaire agreed with statement 6: ‘I was satisfied with the deferred consent process in FiSh’; no parents disagreed with this statement. Parents (n = 44/45, 97.8%) also agreed with statement 4, indicating that they had ‘understood why consent for my child’s participation in FiSh was sought after the treatment had been given’. Moreover, both questionnaire and interview responses indicated that, despite their children being entered into the FiSh pilot trial without prior consent, the majority of parents felt that their consent decision about the use of their child’s data was voluntary18,51,52 [42/45 (93.3%) agreed with statement 9: ‘I made this decision’; 42/45 (93.3%) disagreed with statement 10: ‘Someone took this decision away from me’; 40/45 (88.9%) disagreed with statement 11: ‘I was not in control of this decision’ and 40/45, 88.9% disagreed with statement 12: ‘The decision about research was inappropriately influenced by others’].

Importantly, however, support for RWPC in emergency care trials was felt by some to depend on the context of the specific study:

I think probably my concern was probably a strange concern, insomuch on this study, ‘cause I completely agree with it. It’s more, I think, a concern to give the impression that I think that consent after the fact is a good thing, because I don’t necessarily think it is. I think there are situations where I would strongly disagree with it. That’s really my only concern, that I would be in some way condoning that process as a process in itself, rather than it being a process specifically about, you know, this sort of subject.

P06, mother, telephone interview

More specifically, several parents stated that practitioners explained that rapid fluid bolus resuscitation is the standard treatment for children presenting to UK EDs, despite a lack of evidence for current recommended practice (i.e. 20 ml/kg). 3–5 This explanation, combined with a detailed description of fluid bolus therapy, helped to reassure parents that there were good reasons for RWPC in the FiSh pilot trial and that their child’s safety would not be compromised, as both treatment arms involved carefully titrated administrations of fluid:

I think because he was still getting the same, he was still getting the fluids and essentially still had the same amount but it was just over different sort of periods of time, I personally didn’t feel like it mattered ‘cause he was still getting the same thing.

P12, mother, telephone interview

Obviously the fluid is what he needed so just giving him a small amount wouldn’t have made a difference because they were keeping an eye on him, so it could have been added to.

P26, mother, telephone interview

Similarly, several parents indicated that they would have been concerned about site staff not seeking their prior consent if the trial had involved the administration of a new drug or any other ‘unknown’ clinical procedure, including surgery:

Um, if it’s a trial, I would want to know if it was medicine or drugs or if it was unknown surgery. So, unknown territory. The fluids, they’re going to get fluids anyway so for me, [RWPC] was not even vaguely a concern.

P40, mother, telephone interview

Finally, the majority (15/20, 75%) of parents interviewed said that their child’s survival probably affected their response to the pilot trial and willingness to provide consent. This was consistent with parental survey responses that explored parents’ reasons for consenting to the trial (see Table 19 and subsequent text):

I think that again it’d be down to outcome. If you have a positive outcome, I think the reaction would be completely different than if you had a negative outcome. I think then it could cause a lot of, kind of questioning on the parents’ part, had, you know, he been given a different drug or whatever, would the outcome be different. And actually it could, it could cause an awful lot of ongoing pain, wondering, you know, whether or not that was a factor that tipped the balance, you know.

P06, mother, telephone interview

Nevertheless, the remainder of parents interviewed stated that they would have supported the pilot trial and its method for consent even in less positive circumstances, suggesting that they would not have linked the trial to poor recovery. These views appear to have been influenced not only by the nature of the trial, but also by parental trust in doctors to prioritise their child’s life over research, as well as the sensitive manner of site research staff:

No, I’m pretty sure that if, if they started to do the FiSh and then something went completely wrong, I’m sure rather than stick to the trial, they’d put the patient first and think, right we need to do this, forget looking at the clock, this needs to be done.

P37, mother, telephone interview

The following conversation is from a telephone interview:

Hypothetically speaking, how would you have felt if they approached you about this study and your son had not recovered as well as he did?

Well the approach was really gentle and it wasn’t at all, um, you know, we weren’t being strong-armed into anything. So I don’t think, err, I know it’s a, would be a stressful and awful time anyway but I think it was done so delicately that I don’t think we would have been upset to be asked.

Do you think you would have had any concerns that the trial was linked to poor recovery?

No, I don’t think so. I mean, like I’ve said previously, if it was to do with different drug trials, then yeah 100%, but not for a you know, rehydration solution or whatever the trial is.

However, as the following quotation illustrates, some parents might question whether or not trial participation was to blame for a child’s poor recovery. This finding suggests that in such circumstances site staff should be prepared to address parents’ questions and concerns about their child’s poor recovery in the context of trial participation:

I understand how serious it was and I can’t thank them enough but I think everyone sometimes wants someone to blame and I think if it didn’t go the way that, that it was supposed to, I wouldn’t necessarily be blaming them but I would want an answer to that question.

P36, father, telephone interview

Responses to the pilot trial

Following the discussion of RWPC, the interview conversation shifted to the pilot trial. Parents were prompted for their views on several aspects of the trial, including their understanding of the aims, acceptability of the interventions involved and the reasons for consent (Table 19).

All parents interviewed had provided consent for the pilot trial. The parents cited their child’s trial participation as a way to help advance children’s emergency treatment to help other children in the future:

I kind of thought about other children and thought ‘do you know what, like it’s probably a good idea that they relook at this just to make sure that every child is getting the right treatment’.

P01, mother, telephone interview

The questionnaire included a series of statements to establish the reasons why parents had consented for their child’s information to be used in the pilot trial. Many parents ticked multiple reasons (mode, five reasons; mean, four reasons). In line with the qualitative findings, all parents indicated that a reason they provided consent for the use of their child’s information in the pilot trial was to help other children in the future (see Table 19), with 19 out of 25 (76%) indicating that this was the main reason for consent. A high proportion also indicated that they provided consent owing to a belief that medical research studies like the FiSh trial are important, with 3 out of 25 (12%) indicating that this was their main reason. Interestingly, half of parents in both interviews and questionnaires stated that they had participated because their child had recovered, which supports our qualitative findings that show that parents were happy to consent as their child had recovered by the time of consent.

Also of interest was that half of parents indicated that they consented to help their child, although this was a main reason for consent for only one parent. Only half of the parents interviewed appeared to understand the uncertainty around optimal fluid bolus resuscitation therapy and how the FiSh trial is needed because ‘we don’t know what is best’ (P33, mother, telephone interview). Interviews also revealed a potential lack of equipoise among some parents (7/20, 35%). This related to beliefs that one arm of fluid bolus resuscitation for septic shock was safer than the other, that the volume administered to their child was the correct amount given their recovery, or that both interventions were ‘the same essentially’ (P12, mother, telephone interview):

I really didn’t mind that. Obviously, they’re saying the small amount would be better for the child, so to me, you know, he was fine and he came round quick . . . so if someone had given him the larger amount I would have probably panicked anyway.

P26, mother, telephone interview

The operation was a success, you gave the right amount of fluid. Her, in our daughter’s case, we believe that that was the right amount because she didn’t need any more until the operation and I think that kind of kept her going.

P36, mother, telephone interview

In addition, the subject of the trial (i.e. septic shock) as well as its perceived severity appeared to have a significant impact on some parents. Specifically, 3 out of 20 (15%) mothers who took part in an interview, one of whom also completed a questionnaire, described their upset at finding out that their child had been treated for septic shock during the consent discussion. As the following quotations illustrate, clinical staff had not spoken to these parents using this term before research staff had approached them to discuss the pilot trial:

I had no idea my child had septic shock. When I was told about the study was the first time I found out, which was quite distressing.

P06, mother, questionnaire

We didn’t actually know he was being treated for septic shock . . . until they gave me the letter and I read it, I didn’t actually know that’s what he was being treated for.

P07, mother, telephone interview

I just heard sepsis and then I almost hit the floor.

P39, mother, telephone interview

One mother stated that ‘it was almost awkward for the nurse ‘cause she didn’t realise that she was breaking the news’ (P06, mother, telephone interview).

These parents recommended that, in future research, a member of the clinical team should have a discussion with parents about what has happened and their child’s condition, before a discussion about research occurs:

What could have been improved was to make sure the parents know what actually happened in the situation before speaking about a trial.

P26, mother, telephone interview

Approaching parents for consent

The majority of parents, including 41 out of 45 (91%) of those who completed the questionnaire, indicated that site staff had checked that it was a convenient time to discuss the pilot trial. During interviews, parents stated that they felt that they were approached appropriately within 24–48 hours of their child’s admission, ‘after sort of the initial, um, scary period’ (P01, mother, telephone interview), when the emergency situation had passed. Parents appeared to appreciate the timing of the discussion and that the research nurse had ‘made sure that it was the right time’ (P03, mother, telephone interview) to discuss research.

Two parents who completed a questionnaire indicated that the researcher had not checked the timing of the consent discussion, although this did not appear to affect their ability to understand the information or their satisfaction with the consent process. Although 38 out of 45 (84.4%) disagreed with statement 7 (see Table 18), indicating that they did not find it difficult to take in the pilot trial information, one parent agreed that it had been difficult to take in the information and six neither agreed nor disagreed with the statement. Furthermore, consistent with other studies in this area,38,40 many parents said that they did not ask questions or raise concerns with site staff. A few parents stressed that it was important for staff to tailor the amount of trial information provided to help meet individual needs and preferences:

You’ve got parents with very different levels of interest, education and cultures. I am somebody who likes to be as informed as possible and I’ll ask questions if I don’t understand. But, I think it’s a two-way process. Maybe the researcher has the option of going down the road of explaining to a parent if they want to know but perhaps doesn’t go down that road if the parent chooses to have sufficient information. Err, I think I would have appreciated a little more information in situ.

P38, mother, telephone interview

As per the findings in other critical care studies,19,27 parents stated that they would not wish to be approached for consent during resuscitation, as their child’s life was the priority. As the following quotation illustrates, parents questioned if they would want, or be able to take in, study information when their child was being resuscitated. One-third of parents indicated that they would have felt pressured to provide a decision about the pilot trial if approached by staff during resuscitation. They also suggested that, under such circumstances, parents would be more likely either to refuse to take part or to provide a decision that was uninformed as heightened emotions may affect parental capacity:

My focus had to be on my child. He needed me, right there and right then, it would have been, you know, noise in the background, and I probably would have nodded and not even taken it in.

P06, mother, telephone interview

I mean it’s such a stressful situation, you just need to be, but you just need to be as calm and relaxed as possible. By throwing that on to you I think people would just say no.

P13, mother, telephone interview

Interestingly, 5 out of 20 (25%) parents who were interviewed described how they were spoken to by site staff about the pilot trial during resuscitation. Of these, most parents (4/5, 80%) were unperturbed by this method of approach; however, this appeared to depend on the brevity of the conversation. One mother suggested that being approached by site staff during resuscitation was inappropriate and ‘just not necessary and just not a priority for the parent’ (P38, mother, telephone interview). She stated that she ‘would have been happy for them [site staff] to be there’, provided the explanation provided was simplified and kept brief:

I don’t think it’s necessary in the resus [resuscitation] department unless the parent asks. I think there’s so much going on so unless they ask or they overhear and they say, ‘oh, what’s that?’ Without brushing them off, you know, just a brief explanation is sufficient, which is what I had, which was perfect.

P40, mother, telephone interview

So, I was in A&E [accident and emergency] and there was a lady standing by the door sort of watching everything that was going on in the Resus [Resuscitation] Room. She introduced herself to me as somebody doing research and she said FiSh trial like the FiSh research project. And obviously I, you know, I remember because FiSh is obviously a fairly memorable acronym. And then she described it, but to be honest I remember her talking about fluids and giving probably too much information for me at that point in time.

P38, mother, telephone interview

Finally, during interviews parents were asked to reflect on who (doctor, nurse or researcher) approached them to discuss the pilot trial and whether or not their role affected the trial acceptability and their consent decision. The majority (70%) stated that they would not have a role preference so long as the practitioner ‘believed in the trial and its purpose’ (P13, mother, telephone interview) and was able to provide parents with ‘simple and clear’ (P40, mother, telephone interview) tailored explanations on key trial information. Nevertheless, several parents stated that they would appreciate a consultant’s involvement in the initial stages of consent. More specifically, they suggested that interacting with someone directly involved in their child’s care could promote parental reassurance and foster confidence in the pilot trial:

Well we had, I think the doctor had mentioned it and I think coming from the doctor that’s looking after your child, you feel a bit more confident in what they’re . . . because you’ve got confidence in them, you’re gonna have confidence in what they’re coming to do.

P35, mother, telephone interview

Consistency in who discussed the pilot trial and sensitivity of approach were also valued by parents. These factors appeared to positively affect pilot trial decision-making processes, reducing the likelihood of confusion and helping parents decide to take part:

I don’t remember who it was but it was definitely a different person and that’s a bit confusing in hospital ‘cause just as a consultant changes shift, then a nurse changes shift, so it seemed like the research team changed shift and I think if it had have been the same person it would have been easier to associate that individual with the research.

P38, mother, telephone interview

I would say yeah, their manner did play a part in us going ahead with it. I think if they’d have sort of come in straight faced and very sort of to the point, then we may have been a bit like, ‘oh God, what’s this?’. But because they sort of made it really easy to understand and, um, had a good rapport and sort of made us feel a bit more light-hearted about our actual stay at the hospital, um, it, it sort of made it easier to say yes.

P01, mother, telephone interview

Participant written information and terminology

Parents thought that the PIS was comprehensive: ‘I think it was quite informative from what I can remember . . . it pointed out probably everything that we needed to know at that point’ (P35, mother, telephone interview), and generally clear: ‘It was really, really easy to read . . . very straightforward’ (P13, mother, telephone interview), despite using certain medical terminology throughout: ‘I did think, you know because I’m not a clinical person, that like the language that it was using, it was explained quite well. I did actually understand what it was all about’ (P26, mother, telephone interview). These views were also reflected in questionnaire responses, with 42 out of 45 (93.3%) parents agreeing with statement 3: ‘The information I received about FiSh was clear and straightforward to understand’. No parents disagreed with the statement (see Table 18). The majority of parents (43/45, 95.6%) also indicated that they had enough opportunities to ask questions about the FiSh trial (statement 5, Table 18).

Although two parents reported that further explanation about their child’s treatment in the context of the pilot trial would have been appreciated – ‘Um, explain it more. I don’t really know what happened to my daughter in that sense, I don’t know what’s sort of gone on’ (P19, mother) – the large majority indicated that detailed verbal explanation of the pilot trial had been given by site staff. This discussion, in addition to the written PIS, was valued by parents and appeared to assist their understanding of the pilot trial:

I think the conversation definitely helped. I don’t think I would have fully sort of understood it just from the paperwork, but the doctor sort of went through it and made it, elaborated on some of the points.

P01, mother, telephone interview

Some parents suggested providing a shorter and more basic pamphlet which clearly summarises key aspects of the trial prioritised by parents, including trial safety and aims:

All I needed to know was that it wasn’t detrimental, it wasn’t harmful. You know, my son’s part of a study but actually at the end of the day he’s going to be getting exactly the same as somebody else, you know, not on the study. It’s just that you’re documenting it all and it’s going somewhere. Perhaps that’s, you know, I don’t need any more information than my son is safe.

P40, mother, telephone interview

Finally, as discussed, a minority of parents described significant distress at finding out that their child had been treated for septic shock during the consent process, which included delivery of the pilot trial PIS. In addition to recommending that a member of the clinical team should initially discuss the child’s condition with the child’s parents, parents agreed with suggested changes to PIS sentence structure and terminology specific to septic shock. For example, parents recommended emphasising ‘suspected’ infection or ‘suspected’ sepsis throughout all written information materials alongside a clinical conversation that briefs parents on how some words used (i.e. sepsis or septic shock) may be upsetting:

That leaflet is great but perhaps, like you say, the wording needs to change a little bit with regards [sic] to the septic shock bit.

P34, mother, telephone interview

Certainly, even just the word, rather than, if I’d been told about it right from the beginning that it’s an infection ‘cause you can tell that, you know, the immediate blood results and the high temperature and whatever. But until the blood results come back they don’t necessarily know, it could just be viral, couldn’t it? But I guess from a parental point of view, just knowing that your child is being treated immediately for suspected infection, at the moment we don’t know what it is. But then if that comes back as being more serious than an infection of whatever, this is actually the blood, then you know. I think parents need a build-up of information so it’s not so shocking when it all gets slammed into them.

P40, mother, telephone interview

Outcome measures

As part of an iterative process, findings from the FiSh feasibility study (see Chapter 2) were incorporated into the topic guide as interviewing progressed, in order to further explore parents’ perspectives about important outcome measures. Consistent with methods described in Chapter 2, parents were asked to think about their experience of their child being admitted for severe infection and then prompted to share their views on what effects they hoped the fluid bolus treatment had to help their child (i.e. what indicators were most important to suggest that their child was getting better).

On average, parents discussed seven outcomes. Appendix 3 shows the full list of outcomes identified in the analysis of parent descriptions when asked directly which indicators were most important to them. The top prioritised outcomes, in order of the how commonly they were described as being most important, were as follows:

• organ/physiological functioning, such as blood pressure, temperature and heart rate

• looking and behaving more like self, including becoming ‘more alert’ (P25, mother)

• machines as an indicator of improvement, including ‘the figures’ (P40, mother) and ‘the fewer machines that were operating’ (P38, mother)

• time spent on treatments

• long-term effects

• survival

• time spent in hospital.

FiSh pilot trial site initiation training

Site staff focus groups and interviews began with questions aimed at exploring staff views and experiences about the FiSh pilot trial site initiation training, including whether or not they felt that the training had sufficiently prepared them for the pilot trial. Over half (17/27, 63%) had been trained at the site initiation meeting, whereas 10 out of 27 (37%) had received training from a member of their own site team, often the PI or research nurse. Three site staff had received both forms of training. Without exception, site staff stated that the training had prepared them for recruitment and consent in the pilot trial. As the following quotation illustrates, site PIs stated that the training had provided them with the confidence and knowledge to disseminate the pilot trial training to the rest of their team:

Yeah, it was absolutely fine . . . it made me feel confident enough to go and do the training for the extended team.

P02, focus group 1

Site staff valued having access to the training slides after the site initiation meeting so that they could refer back to them at a later date. Access to training materials had assisted the dissemination of training more widely across relevant departments and teams:

The mixture of having the information to refer back to, so the protocols that you’ve got along with the consent inform – you’ve got to have the background knowledge, so you’ve got to sort of attend the training to know what you’re doing to fully understand why you’re doing it, but also to have the reminder bits of the bits that are important to explain to the families at the point at which you’re getting consent.

P01, focus group 1

I think having that resource was really good and everything you need to deliver training is in there.

P22, telephone interview

Some staff stated that the training had outlined clear pilot trial procedures, including simple recruitment criteria, which they felt were important in ensuring the engagement of all site staff, particularly those who work in the ED:

Who you recruit, who you randomise into it is . . . they’ve made it so kind of simple . . . it was very clear as to which cohort of patients they were looking for.

P21, telephone interview

Yeah, it was simple, wasn’t it, and that’s why it encourages, in A&E [accident and emergency] for instance, to get involved because it’s not complicated to pick the children out and to put them into the study.

P02, focus group 2

Others stated that, before the training, they had no previous experience of RWPC, which had made them feel anxious about being involved in the pilot trial. However, the inclusion of specific training on this approach to consent, which incorporated parents’ perspectives from the FiSh feasibility study, appeared to help prepare them for consent discussions with families in the pilot trial:

Um, I hadn’t really been involved in retrospective consent before, so it was very useful kind of hearing them kind of talk about the kind of background of retrospective consent and why they were doing it, and all that kind of stuff. Because it’s really helped when I’ve gone and consented people, explaining why we’re doing it now and we didn’t do it yesterday type of thing. Um, so that probably was the more useful bit out of it, if that makes sense.

P21, telephone interview

The training package was very well put together and it had lots of relevance. Um, you know, supporting evidence for the rationale and lots of interesting sort of qualitative stuff about the methodology.

P22, telephone interview

I think with this particular study, what was quite concerning for most was the consenting part, obviously, because everybody was feeling how do we sort of speak to parents when they’re ready, so like we’ve asked them when their child has already been entered into the study in a way. And I think, having had the sort of the initiation day and sort of discussions, I think that gave me sort of a sound understanding of it.

P26, telephone interview

Screening pilot trial patients

Site staff described the methods used to help raise awareness of the pilot trial to ensure that every patient admitted to the ED or PAU (or equivalent) needing a fluid bolus was screened for the pilot trial. This included placing FiSh trial posters, stickers and memos where fluid was stored to remind staff to screen, as well as using computer systems to regularly search for potentially eligible recently admitted patients. In most sites, all clinical staff were described as being involved in screening. In some sites, a bleep system was successfully used, ‘so the A&E [accident and emergency] team know that in the event that there’s a child coming in who’s requiring fluids and sort of fits into the sort of FiSh criteria’ (P26, telephone interview), to inform research nurses that a consent discussion would be needed. Others stated that there were a few designated research nurses responsible for monitoring patients for eligibility to the pilot trial, as well as other studies being run at their site. One site used a rota system to ensure that someone was watching and checking computer systems for potentially eligible patients during the day:

We have a rota, and it’s Monday to Friday. There’s somebody watching from 8 till 6-ish, 8-ish, it depends. Some days it’s later than others.

P13, focus group 2

Site research staff stated that potentially eligible patients were sometimes missed at weekends, and that they would use computer systems to identify them the following week and retrospectively complete screening forms. During focus groups, staff members were asked to use the voting handsets to indicate whether or not the screening process could be improved. A total of 13 out of 20 (65%) clinicians indicated the process could be improved, whereas 7 out of 20 (35%) did not. Staff who took part in a telephone interview also stated that the screening process worked well, with few suggestions for improvement.

Suggestions for improvements to the screening process to help ensure that patients are not missed in the FiSh trial included ensuring 24/7 research cover in the ED, making sure that all staff were trained (including new staff on rotation) and ensuring that study posters are placed in the optimal location:

I mean I think it would be fantastic if the research cover could be a bit integrated into the areas that we’re recruiting because I do think that it raises awareness.

P16, focus group 2

So just in terms of improving it for the future, ‘cause we’ve said that there are multiple clinicians and nurses involved and they change every 6 months. Training the clinicians.

P20, focus group 3

We have got a big poster down there but maybe we need to think . . . of where it is. It’s on the back wall.

P10, focus group 2

Randomising pilot trial patients

The sealed-envelope method of randomisation was viewed as being simple and easy to use. During focus groups, staff members were asked to use the voting handsets to indicate whether or not they felt that the randomisation process could be improved. The majority (14/20, 70%) indicated that the randomisation process could not be improved. This view was shared by three out of seven (42.8%) staff who took part in a telephone interview.

The system is very much in place. I do like in ED the system where we use the sealed envelopes because it’s quick and easy.

P23, telephone interview

Problems identified with the process included an envelope being opened in error when a child did not meet the inclusion criteria or an envelope being opened too early. Staff stated that these issues occurred at the start of the pilot trial when some site members who ‘didn’t really understand the process’ (P25, telephone interview) were not completely familiar with the pilot trial protocol.

In one focus group, site staff stated that the treatment allocation had been misread. They suggested that, for the FiSh trial, the allocation should be in larger, bold text on the case record form, so that there is no doubt which treatment should be given:

They [ED clinician] were very confident they had done the right thing, but from the paperwork, it obviously was indicating they had not done the right thing, but they didn’t realise at all, because they felt, ‘we’ve recruited them to FiSh, isn’t that great?’ And we’ve given the second bolus and then we reassessed them and all of this, and it’s just such a great documentation but just unfortunately the wrong bolus.

P19, focus group 3

One staff member described how their localised protocol for sepsis treatment differed from the pilot protocol, which led to two patients being randomised in error:

I think we’ve been slightly overenthusiastic, if that makes sense. So, our local protocol is currently very heavily weighted towards lactates in sepsis, and if you’ve got high lactates, give fluid. Um, so we’ve had two children that, as per FiSh randomisation criteria, haven’t fitted needing a second fluid bolus. But as per our local sepsis protocol, have fitted being given a second bolus, i.e., their lactate hasn’t improved. And people have I guess kind of rightly in the face of sepsis more than 20 kilo[gram]s let’s randomise them into FiSh. But actually their cap [capillary] refill and their blood pressures have been fine. So I think, that’s kind of local protocol fighting the FiSh protocol. It’s not the fault of FiSh, it’s just it clashing with our local protocol.

P21, telephone interview

Protocol adherence

During the focus groups, staff members were asked to use the voting handsets to indicate whether or not they had experienced any difficulties adhering to the pilot trial protocol. Of these, 6 out of 20 (30%) staff indicated that they had experienced protocol adherence difficulties, whereas 5 out of 20 (25%) had not. However, 9 out of 20 (45%) staff opted not to respond to the question using their handsets, although insight was gained into protocol deviations at these sites through group discussion. Some staff (3/7, 43%) who took part in telephone interviews described some difficulty with protocol adherence. In addition to the randomisation issues described above, protocol deviations described by staff related to a lack of equipoise and difficulties in administering fluid boluses within 15-minute cycles.

Case report form

Many staff stated that it was difficult to complete the CRF in ‘real time’ while they were focused on resuscitating a child. Many felt that this had led to missing or incomplete CRF data, which were difficult for research nurses to complete retrospectively:

If they’re [ED staff] not recording things at the time that those boxes are wanting, then it makes it difficult for us [research nurses] to know what to put in those boxes.

P24, telephone interview

Many stated that research support in the ED was needed to help complete the CRF as, although the form was ‘really straightforward’ (P24, telephone interview), it was difficult and ‘time-consuming’ (P23, telephone interview) to complete while treating a critically ill child. As one ED clinician described, randomising and treating participants over the weekend without research support was stressful:

Both of our patients were randomised at the weekends, so there was no research input whatsoever it was all done off the back of the nurses that we had trained. Um, and so the only thing that they said is that it’s just a very, very scary experience and that, you know, they’re still trying to concentrate on the child. But they also want to get the trial right, so it’s a very stressful experience.

P25, telephone interview

Some staff suggested that additional training and experience of completing the CRF would help to improve the quality of data entry, whereas others felt that changes should be made to the recording of CRT to reflect a score rather than a number of seconds:

You put your capillary refill time in like point whatever of a second. That isn’t how people measure it. So that immediately put people off as a sort of first hurdle. And when you’ve sort of fallen at the first hurdle, I think people lost trust with it.

P27, telephone interview

Suggestions were made to help improve the format of the CRF, including having specific sections to complete in an emergency situation, as well as a clearer way of recording the time that a fluid bolus was administered:

It looks quite daunting, so maybe if there was a way of trimming it down so that only the bare essentials that are needed from the emergency, and then the rest could be picked up retrospectively from notes, from the research team. Which is kind of how we had to work the first one.

P25, telephone interview

One comment would be just to allow for a time box at the top of each of those sections so that, you know, I know it says 15 minutes but what time was that? Was it actually 15 minutes? And then that would probably help you with the data collecting as well. Um, I don’t, I didn’t feel like a flow from reassessing at the end of the 15 minutes to whether a trial bolus was given or not, I didn’t think that was a natural progression.

P19, focus group 3

Additional issues to consider for the FiSh trial protocol

Experiences of research without prior consent in the pilot trial

Staff involved in other NIHR-funded critical care research stated that they were ‘less daunted’ by the use of RWPC in the pilot trial, as their previous experiences of such consent discussions with parents ‘had all gone without a hitch’ (P21, telephone interview). Staff at one site stated that they did not have any initial reservations about the use of RWPC in the pilot trial:

It was going to be quite an interesting experience and we were all quite excited about it.

P25, telephone interview

However, the majority of staff involved did not have prior experience of RWPC and described their initial anxieties about this consent method before the pilot trial began. As one PI described, the site team took their time to consider the consent process before agreeing to participate in the pilot trial:

We had to go through the process quite a lot, that was sort of the sticking point for the trial for quite a long time until they understood why it’s working that way.

P25, telephone interview

Many staff were concerned about how parents would react to being told that their child had been entered into the pilot trial without their prior consent, particularly if their child was still ‘really sick’ at the time of the conversation or ‘had died’ (P01, focus group 1). As described previously, the FiSh feasibility study findings, which showed how parents support the use of RWPC in the FiSh trial, appeared to have alleviated some initial staff anxieties:

I think I, along with the majority of people, was quite anxious about the honesty with the parents to begin with, was quite anxious about how they’d react, and I think the biggest thing that helped me was hearing that there’d been views from parents that they actually were happy about it.

P20, focus group 3

However, it was often the experience of conducting the pilot trial that helped to fully address any remaining anxieties about the use of this approach to consent in the pilot trial. Staff stated that parents had responded positively to the consent discussion and appeared to understand the reasons why informed consent could not have been sought prospectively. Furthermore, staff confidence in conducting the pilot trial appeared to increase over time:

And obviously the first deferred consent . . . you know it’s the see one, do one, type thing and then I think you realise, ‘oh, actually it is OK’.

P02, focus group 1

Both of them [parents of two recruited patients] were really, really understanding and recognised why we hadn’t asked them at the time.

P18, focus group 3

But as time has gone on, people have bought into it more and more and more.

P02, focus group 1

All staff supported the use of RWPC in the FiSh trial and thought that seeking prospective informed consent in the ED or PAU (or equivalent) when a child was being resuscitated would not be acceptable to families.

Staff viewed the consent form as simple and easy to use. One focus group participant suggested that the form should be one sided so that ‘nothing could be missed’ (P18, focus group 3). Finally, staff stated that child assent had most often not been sought as children were ‘too young’ (P27, telephone interview), or ‘didn’t have the capacity’ (P19, focus group 3) to be involved in pilot trial consent discussions with parents. Only one site member stated that ‘in retrospect, thinking about it, we should have thought to go back and get assent from that child once she’d been woken up’ (P18, focus group 3).

Summary of parents’ perspectives in the pilot trial

The integrated perspectives study findings add to the feasibility work (see Chapter 2) to show that the trial protocol was feasible and acceptable to parents. Consistent with the feasibility study findings and with associated guidance on RWPC,18,19,26,27 some parents of children randomised to the FiSh trial were initially surprised about the concept of RWPC. However, following tailored explanations from site staff about why prospective consent was not sought, all parents stated that they supported the RWPC approach in the pilot. Again, parents strongly valued the advancement of medical research and stated that they provided consent as a way to help other families and children in the future. 17,19,28,30,36,37

The use of the feasibility study findings to develop site staff initiation training on how to explain the pilot trial, including RWPC, appeared to assist parents’ understanding in the pilot trial. Importantly, concerns and misunderstandings about key aspects of the trial observed in parents who took part in the feasibility study were not observed in the pilot trial sample. Parents of children who took part in the pilot trial recalled site staff explanations about the nature of fluid bolus resuscitation and how this was standard care, not a new or novel treatment for septic shock, and about the lack of evidence for current recommended practice (i.e. 20 ml/kg). These verbal explanations, rather than written trial information, were prioritised by parents and helped to reassure them that there were good reasons for conducting this research without seeking prior informed consent, and that their child’s safety would not be compromised, as both treatment arms involved carefully titrated administrations of fluid. Nevertheless, researchers should bear in mind that parents may struggle to voice their questions or concerns during research discussions. 38,40 In line with previous studies in this area, these findings should be incorporated into staff training to encourage site staff involved in the FiSh trial to use open-ended questions and provide ample opportunities for questions,38,53 in order to identify and respond to parental concerns. 40,54

Importantly, trial acceptability and parents’ willingness to consent to their child’s information being included in the pilot trial appeared to hinge, in part, on the nature of the trial and their child’s recovery. 19,43 Not all children who experience septic shock will make a full recovery; therefore, site staff should be prepared to respond to parental questions or concerns related to the possibility that trial participation may have resulted in harm. 26,27 In line with our previous findings,19 parents viewed this study, involving variations in routinely used fluid bolus resuscitation therapy, as safer than trials of a new drug or surgery, and therefore as an acceptable study in which to use RWPC. Of added interest, several parents indicated that they consented in order to help their child, and some revealed a preference for one fluid bolus volume over another. This may indicate that parents did not understand that the optimal fluid bolus resuscitation volume for septic shock is not known, and, therefore, any direct benefit could not be assured. We recommend highlighting this in staff training programmes and including it as a key message to convey in consent for a full trial, if planned.

The professional background of the member of site staff approaching families for consent in the FiSh trial did not appear to affect significantly trial acceptability and willingness to provide consent – provided that adequate explanation was offered and familiarity was demonstrated with the child’s clinical case. Parental accounts of the pilot trial consent process supports existing guidance for site staff to check appropriate timing with the clinical team27 before approaching parents within 24–48 hours, when the acute situation has resolved. 17,19,30,38,41 Indeed, discussions during resuscitation were, for the most part, viewed as inappropriate unless conversations were kept brief, or parents had explicit trial-related questions.

Importantly, some parents became distressed when first hearing that their child had been treated for septic shock during the pilot trial discussion. We recommend that site staff check that a member of the clinical team has spoken to parents about their child’s condition and treatment before discussing the FiSh trial. To assist understanding and parent–practitioner communication, parents emphasised the need for clear and concise written information materials alongside verbal tailored explanation. 38,40 Similar to the findings in other clinical trials,19,39 parents recommended amending language used in both written and verbal explanations, and provided insight on what parents regard as appropriate language specifically in the context of septic shock. This recommendation may be related to a recent increase in media coverage on sepsis in children, with the term being associated with a high risk of mortality in children with severe infection. As a result, terminology such as ‘suspected’ sepsis or ‘suspected’ septic shock may not have been used sufficiently in the PIS and pilot trial verbal explanations. This small change to patient information could reduce parents’ initial anxieties relating to the severity of their child’s illness.

Finally, parents provided their views about what outcomes are most important in children with severe infection. In order of importance, parents reported that improvement in organ and physiological functioning, looking and behaving more like their normal selves, improvement in clinical figures and less need for machines could be used as indicators that fluid bolus resuscitation therapy was effective. Interestingly, these parents, who were interviewed shortly after their children’s treatment for septic shock, prioritised shorter-term rather than long-term outcomes, which were potentially most salient to them at the time of being interviewed. This is in contrast to the prioritisation of longer-term outcomes, such as disabilities, by parents in the feasibility study who were interviewed approximately 1 year after their children’s hospital admission. Child survival as an outcome was rarely discussed by parents in the pilot trial sample, all of whom had children who had survived. Such findings may suggest that parents’ prioritisation of outcomes may be influenced by their experience of their child’s illness, the survival of their child and the point at which they are asked about outcomes of importance in the course of their child’s illness. These combined qualitative findings should be incorporated into any future priority setting or core outcome set development work to ensure the selection and measurement of outcomes that are patient centred.

Overall, these findings highlight the value of doing pre-trial research to inform approaches to recruitment and consent in challenging trials. 55 Incorporation of these findings into staff initiation training, approaches to consent and information materials for the full FiSh trial, if recommended, will help parents to make informed decisions about the use of their children’s information.

Summary of site staff perspectives in the pilot trial

Overall, findings demonstrated that site staff supported the pilot trial and were willing to screen, recruit and randomise eligible patients to the FiSh trial. The study also provides new evidence in addition to the existing literature18–20,27 that can be used by practitioners involved in the design, ethics approval and conduct of children’s emergency care trials. Staff valued the FiSh pilot trial site initiation training, which sufficiently prepared them for recruitment and approach to consent. Access to site initiation training materials after the site initiation meeting helped PIs disseminate training across relevant departments and teams, a model that can be replicated in the FiSh trial. Similar to previous studies,18,20 confidence in conducting the pilot trial appeared to increase over the recruitment period, when staff became more familiar with the screening, randomisation and consenting processes.

Some staff without prior experience of RWPC were initially anxious about being involved in the pilot trial. Concerns were mainly related to how parents would react to being informed that their child had already been included in the pilot trial without their prior consent, potentially affecting the parent–practitioner relationship. 20 The inclusion of FiSh feasibility study findings in site initiation training, which showed parental support for the FiSh trial and its use of RWPC, helped to prepare staff for pilot trial consent discussions with parents. However, in staff without previous experience of this method, it was often the direct experience of conducting the pilot trial, and observing parents’ positive responses to the consent discussion, that provided staff with full confidence about the acceptability of this approach to consent in the FiSh trial. Consistent with previous findings,18,20 the views of practitioners with prior experience of RWPC should be considered in the design and conduct of future paediatric critical care trials, which should not focus solely on the beliefs of those without prior experience of this method of consent. Furthermore, future research would benefit from audio-recording RWPC discussions20,56 to improve future training by providing insights into how practitioners explain this to parents and to identify parental views and concerns. Suggestions for improvements to the screening process for the FiSh trial included increasing research support in the ED, ensuring that all relevant staff are fully trained when the medical staff rotation occurs every 6 months, and ensuring that posters and study materials are placed in the optimal location.

The pilot trial recruitment process and inclusion criteria were described as being simple to follow in an emergency situation. The majority of staff stated that the pilot trial randomisation process worked well, and, although some protocol deviations had occurred, these often occurred in the first few patients recruited at their site. Suggestions were made to improve the CRF for the FiSh trial. In addition to research support in the ED to help complete the CRF, staff suggested that the form should be reformatted, so that sections to be completed ‘in real time’ during resuscitation are clarified, or separated from sections that could be completed retrospectively from medical notes. As shown in other studies, staff can find recruitment difficult, with too many competing demands on their time,57 and this is particularly pertinent in a paediatric emergency care situation. Additional training would be needed on randomisation and CRF completion in a full FiSh trial, such as a simulation of the screening, randomisation and allocation procedures in the FiSh trial site initiation training. A step-by-step guide, or video, might help to improve the quality of CRF completion and, therefore, the quality of data obtained.

Staff found the administration of fluid bolus within the 15-minute time frame to be challenging. Although some stated that they had changed their practice and had been able to administer fluid boluses more quickly to adhere to the trial protocol, there were issues with medical devices (i.e. slow pumps and small cannulas) used to deliver fluid boluses. Such medical devices would need to be reviewed in preparation for the definitive FiSh trial.

Finally, telephone interviews and focus groups with staff indicated a lack of equipoise among some ED clinicians who had not adhered to the pilot trial protocol when a child had been randomised to a 20-ml/kg fluid bolus allocation because of concerns about fluid overload. Some staff also described a ‘clinical creep’ of a tendency to use restricted volume fluid bolus resuscitation without scientific evidence. Studies have consistently shown that trial recruitment can be negatively influenced by staff perceptions and experiences, including the decision not to recruit eligible participants because of a lack of equipoise with regard to particular treatments or individuals. 58–61 Although it appeared that only a small proportion of FiSh trial site staff were not in clinical equipoise, it has been suggested by other studies that this number would probably increase in the light of increasing use of one’s own preferred method of treatment and in the absence of forthcoming evidence. 61 Moreover, previous findings suggest that the role and importance of equipoise may not always be fully understood by all. 60 Taken together, these findings highlight the crucial need to conduct the definitive FiSh trial, incorporating explicit training on clinical equipoise, to identify the optimal fluid resuscitation volume for children with septic shock, to achieve the best possible outcomes and to inform clinical guidelines and practice in the NHS.

Overall, clear communication and ensuring that participants understand trial information and discussions is a challenge for practitioners in many trials. 62,63 This demonstrates the significant value and impact of our findings for future trials by providing recommendations for the enhancement of staff training in preparation for the conduct of ethically sound research in the paediatric critical care setting.

Feasibility study

This was a qualitative interview study involving 21 parents (18 mothers and 3 fathers, and 7 of whom were bereaved) with children admitted to a UK ED with presumed septic shock in the previous 3 years. The study provided insight into the acceptability of the FiSh trial to parents. The majority of parents interviewed were unfamiliar with RWPC, yet supported its use in the FiSh trial. Parents were concerned about the change from currently recommended treatment, but were reassured by an explanation of the current evidence base, the rationale for fluid bolus therapy and the monitoring procedures that would be in place. Parents made recommendations about both the timing of consent and the patient information materials. Bereaved parents suggested that site staff should not attempt to obtain consent immediately after a child’s death, but supported a personalised postal ‘opt-out’ approach to RWPC for the FiSh trial. All parents interviewed in this phase of the project would have provided consent for the use of their child’s data in the FiSh trial, citing altruism as their main motive.

Parents also gave views on outcome measures that they considered to be most important in children with severe infection. Parents commonly discussed the following outcomes: long-term effects, including health and development; the child looking and behaving more normally; improvement in organ and physiological function; less need for medications and machines; and survival.

Pilot trial

The pilot trial was conducted in 13 hospitals from July 2016 to April 2017. Inclusion criteria were children presenting to ED with clinical suspicion of infection and shock after receiving 20 ml/kg of fluid. Children were randomised to 10- or 20-ml/kg fluid boluses every 15 minutes for 4 hours if in shock. Exclusion criteria were prior receipt of > 20 ml/kg of fluid, conditions requiring fluid restriction and full active resuscitation not being within the goals of care. A number of potential end points were collected. Seventy-five children were randomised: 40 into the 10-ml/kg fluid bolus group and 35 into the 20-ml/kg fluid bolus group. Two children were withdrawn as their parents were unable to be approached to provide consent, but no parents declined consent. Baseline characteristics were broadly similar, with some imbalance in age and weight. The majority of children (59% in the 10-ml/kg fluid bolus group and 74% in the 20-ml/kg fluid bolus group) required only a single trial bolus before shock resolved. The mean total volume of fluid received (in ml/kg) was significantly lower in the 10-ml/kg fluid bolus group than in the 20-ml/kg fluid bolus group. Fluid boluses were delivered in accordance with the protocol in 79% of children in the 10-ml/kg fluid bolus group and 55% in the 20-ml/kg fluid bolus group. There were no deaths or SAEs. Overall, 29% were admitted to PICU. Length of hospital stay, transfers to a PICU, length of stay in a PICU and days alive and not in a PICU up to 30 days post presentation to an ED did not differ significantly between the groups.

Integrated perspectives study

This was a questionnaire, interview and focus group study involving all hospitals participating in the pilot trial. Forty-five parents completed a questionnaire, 20 families and 7 site staff (3 doctors and 4 nurses) were interviewed and 20 site staff (10 doctors and 10 nurses) participated in three focus groups. One-third of parents were surprised to discover that their child had been entered into the pilot trial without prior consent. Nevertheless, all parents supported the trial because they were reassured by explanations about why prior informed consent could not be sought, the nature of fluid bolus resuscitation and lack of evidence for current recommended practice. Parents suggested improvements to how the trial should be introduced, including how ‘suspected’ sepsis should be emphasised. Parents prioritised shorter-term physiological outcomes, such as reduced heart rate, over longer-term health outcomes, such as disabilities. This is in contrast to the prioritisation of longer-term outcomes by parents in the feasibility study who were interviewed approximately 1 year after their child’s hospital admission. Child survival as an outcome was rarely discussed by parents of children who took part in the pilot study, none of whom was bereaved. Such findings may suggest that parents’ prioritisation of outcomes may be influenced by their child’s illness experience, survival and the point at which they are asked about outcomes of importance in the course of their child’s illness. These combined qualitative findings should be incorporated into any future patient-centred outcome priority-setting exercise or core outcome set development work to ensure that the selection and measurement of outcomes for sepsis research are patient centred. Site staff made some suggestions to improve the protocol and the training package. There was a lack of equipoise among some parents and some clinical staff who favoured a 10-ml/kg fluid bolus volume. However, overall, both clinical and research staff supported the FiSh trial, with their confidence in delivering the protocol increasing over time.

Feasibility study

Overall, the findings supported the proposed FiSh pilot trial, including the use of RWPC. Although the sample size was small, data saturation was achieved and, importantly, both bereaved and non-bereaved parents were included. However, as the majority of parents were recruited via social media, our sample may have comprised parents with a pre-existing interest in research and may not be representative of the parents of children who might be recruited in a trial. Additional qualitative research was therefore conducted within the FiSh pilot trial to confirm the findings in the feasibility study and to understand the perspectives of site staff.

Pilot randomised controlled trial

The pilot trial was successfully delivered, with screening logs showing that > 90% of eligible patients were recruited. Twelve patients were ineligible as a result of receiving > 20 ml/kg of fluid; the opportunity may have been missed to randomise these patients earlier. It is also possible that these may have represented a subset of sicker patients. Despite the logistical challenge posed by setting up 13 study sites for a short pilot, recruitment was very close to the target of one child per site per month. The process of RWPC worked smoothly, although it was not tested in the emotional aftermath of a child’s death, as, thankfully, no patients died during the course of the study. The intervention was by and large delivered successfully, with good separation being achieved in terms of volume of fluid delivered. Potential outcomes were successfully collected as part of the trial data collection, with more data coming from the linkage with PICANet. Some participants had missing data from the PICU stay because of referral to a PICU that was not in the pilot trial. It would be vital to have local approvals for all PICUs for a larger trial.

However, there were significant logistical problems in setting up some of the smaller study sites, two of which were discovered during this process to have inadequate research infrastructure to support the study. This necessitated the recruitment of another two sites in the middle of the pilot trial. Although the overall recruitment rate was just under the anticipated recruitment rate of one patient per site per month, with a final rate of 0.9 patients per site per month, recruitment was driven mainly by three study sites (Southampton, Chelsea and Westminster, and Bristol), which between them recruited 38 participants. The remainder of the study sites struggled to recruit more than a handful of participants each.

In the event of a larger study looking at fluid bolus volumes being undertaken in the UK, it would be important to consider those sites that might be best suited for such a study. These would probably be large university hospitals with a children’s ED, an integrated PICU/high-dependency unit and good pre-existing research infrastructure, rather than small district hospitals.

Most importantly, the population was not as unwell as expected. The mean PIM2r score (a PICU severity of illness score giving a population risk of mortality) was 2.1%. There were 51 out of 73 (70%) participants who required only one study fluid bolus and only 8 out of 73 (11%) participants required more than two study boluses. Only 21 out of 73 patients (29%) were transferred to a PICU. All the participants survived. Furthermore, despite many guidelines suggesting that elective intubation and ventilation should be considered when ≥ 40 ml/kg of fluid has been administered in this clinical setting, because of the risk of pulmonary oedema, out of the 46 children who received ≥ 40 ml/kg (±10%) of fluid, including the 20 ml/kg pre randomisation, 16 were transferred to a PICU. Thus, not only were the participants less unwell than expected, but many of those who were apparently unwell enough to merit referral to a PICU, or discussion with a PICU retrieval team, improved and remained under the care of general paediatricians. As the intervention was not blinded, it is possible that potential outcome measures, such as admission to a PICU, may be subject to bias either directly through knowledge of the allocated treatment (performance bias) or through the decision to admit to a PICU being influenced by treatment factors including the volume of fluid received. It was not considered practical to blind clinicians to the allocated treatment in the ED setting.

Overall, the pilot trial was successfully delivered. However, there were more protocol deviations in the 20-ml/kg fluid bolus group than in the 10-ml/kg fluid bolus group and qualitative data implied that some clinicians may lack equipoise. The randomised population was less unwell than expected and only a minority of participants received more than one fluid bolus or ended up in a PICU. Thus, the trial as currently conceived is not feasible.

Possible reasons for the patients being less unwell than expected in this study are explored below.

Integrated perspectives study

This study involved parents from 10 of the 12 hospital sites that randomised participants to the FiSh pilot trial and staff from 9 of the 12 hospitals. Although the sample size was relatively small, data saturation was reached. However, opportunities to sample fathers, bereaved parents and parents who declined to consent were limited, as only one father consented to be interviewed, all children recruited into the pilot trial survived and no parents declined consent. This limits our understanding of these groups and means that firm recommendations cannot be made on how to conduct RWPC when a child has died. However, our qualitative data were strengthened by the integrated perspectives mixed-methods approach, providing an overall understanding by investigating the experiences and opinions of both parents and site staff.

Although attempts were made to include children and young people in the discussion about the pilot trial, child assent was not taken by any of the participating sites, because children were either too young or too sick before discharge for a discussion about research. Thus, children’s views on the FiSh trial remain unknown.

Parent and site staff perspectives: research without prior consent

Our findings add to the existing literature on RWPC in paediatric emergency and critical care settings: they provide an important contribution and demonstrate the value of using qualitative methods in considering the family-centred parameters needed to inform the design and conduct of challenging trials. 53,64

Other studies looking at fluid bolus therapy

Current guidelines recommend aggressive fluid resuscitation in children with severe sepsis and septic shock. The latest ACCM-PALS guidelines65 recommend rapid fluid bolus resuscitation (of 20 ml/kg, up to a total of 60 ml/kg over the first 15 minutes) during the first hour of emergency assessment and treatment.

The literature supporting the ACCM-PALS recommendations consists of retrospective studies involving small numbers of children. 6–8 There are no data from prospective clinical trials. However, there is now increasing evidence that fluid overload in critical illness is deleterious. In Africa, the recent FEAST9 multicentre RCT reported a 30% increase in mortality associated with fluid bolus resuscitation. A recent systematic review66 identified only one RCT, other than FEAST, that investigated different size fluid bolus therapy in children presenting to EDs in septic shock. 13 This was a single-centre study in India, randomising 147 children to 40 ml/kg of fluid over 15 minutes followed by dopamine, compared with 20 ml/kg of fluid over 20 minutes up to a maximum of 60 ml/kg over 1 hour followed by dopamine. No difference was found in any outcome measure between the groups, although the 40-ml/kg fluid bolus group had a higher incidence of hepatomegaly than the 20-ml/kg fluid bolus group, indicating fluid overload.

Thus, at the time of writing, the optimum strategy for fluid bolus resuscitation in children with septic shock remains unknown in the developed world. One paediatric fluid resuscitation RCT, other than the FiSh trial, is currently in progress. The SQUEEZE14 trial is a Canadian RCT that is comparing fluid with early inotrope therapy. Results are eagerly awaited. In PICU, many observational studies67–70 have recently shown that fluid overload is associated with morbidity and mortality in a variety of clinical settings.

Our data also suggest that the FEAST9 trial, along with the PICU studies, may have already affected clinician equipoise in the UK, with a preference for less generous fluid boluses than those recommended in the current guidelines already entering routine clinical practice. Further work is thus required to confirm equipoise still exists before a large-scale trial of fluid bolus volumes is undertaken in the UK.

The impact of vaccination

Only 32 out of 73 (44%) participants in this study had a confirmed bacterial infection in a sterile site or a viral infection. Of these infections, 18 were viral, the majority of which were respiratory and likely, therefore, to have minimal cardiovascular dysfunction. Thirteen participants had confirmed bacterial infection in sterile sites, only one of which (a single case of meningococcal sepsis) was preventable by a vaccine. Such results would until recently have been impossible to imagine – that, over a 9-month period in the winter in 12 UK EDs, so few children would present with confirmed bacterial sepsis. A few years ago, the incidence of meningococcal and pneumococcal sepsis, in particular, was much higher (about 600 cases of group B meningococcus were reported in 201371). Although more recent absolute figures are not available, the incidence is now much lower, following the introduction of group B meningococcal vaccination in September 2015. Since then, there has been a 50% incidence rate ratio reduction in group B meningococcal cases in vaccine-eligible children. 72 The incidence of group C meningococcal infection was already very low following the introduction of the vaccine in 1999. In relation to pneumococcal disease, in 2006, a seven-valent vaccine, pneumococcal conjugate vaccine (PCV7), was added to routine childhood immunisation, and in 2010 a PCV containing polysaccharide from 13 common capsular types (PCV13) replaced PCV7. Before the introduction of PCV7, the mean annual incidence of invasive pneumococcal disease was 44.4 per 100,000 children < 2 years of age, with a mean case fatality rate of 3.5%. Following the introduction of PCV7 and now PCV13, the mean annual incidence of invasive pneumococcal disease is currently at around 10–12 per 100,000 children. The impact of the comprehensive childhood vaccination now offered in the UK on vaccine-preventable infections cannot, thus, be understated.

Furthermore, there has been a recent trend in many hospitals to implement ‘sepsis bundles’ to help staff recognise and manage early sepsis. In the UK, evidence-based guidelines for recognition and early management of children with sepsis have been produced and disseminated by the National Institute for Health and Care Excellence (NICE). 73 Thus, it is also possible that some children presented early to the ED, were recognised early, treated promptly and responded to treatment. This would reduce the severity of illness in the population as a whole and also reduce the need for fluid bolus therapy. Unfortunately, data that might determine whether or not this was the case were not collected as part of the FiSh feasibility study; hence, this possibility must remain a speculation.

Overall, it is most likely that vaccination, and possibly improved recognition and management of early sepsis, is the cause of both the unexpected low severity of illness of the population of children randomised into the FiSh pilot trial and the difficulties many sites had in recruiting more than a handful of participants.

Progression criterion 1

• Formal sign-up of sites (the required number of hub-and-spoke models) to achieve the overall sample size based on the observed recruitment rate, as determined by the pilot trial objectives concerning identification, recruitment rate, acceptability of deferred consenting procedures, selection of the patient-centred primary outcome measure and sample size estimation.

• Response: qualitative work showed that there was buy-in for the FiSh trial, but as a result of the full trial not being deemed feasible there has been no formal sign-up of sites.

Progression criterion 2

• The recruitment rate achieved in the pilot trial (taking into account the loss of participants as a result of declining deferred consent) was consistent with achieving the overall sample size required (as determined by the objectives concerning selection of the patient-centred primary outcome measure and sample size estimation) in an acceptable time frame from the sites giving formal sign-up in the first feasibility objective.

• Response: the overall recruitment rate in the pilot trial was very close to anticipated, although there was considerable variation across sites. However, the severity of illness of the participants recruited was much lower than expected and, if the inclusion criteria were to be changed to restrict to a more severely ill cohort, then the number of eligible children would inevitably be reduced and the feasibility of completing recruitment in an acceptable time frame may be brought into question.

Progression criterion 3

• At least 80% of fluid bolus resuscitation in the external pilot study was delivered at the correct volume and timing, ± 10%, as determined by the external pilot study objective concerning adherence.

• Response: adherence to both volume and timing was close to achieving the 80% threshold in the 10-ml/kg fluid bolus group but was substantially worse in the 20-ml/kg fluid bolus group, suggesting that, despite current recommendations, rapid infusion of 20-ml/kg, fluid boluses does not represent current usual practice in the UK.

Progression criterion 4

• The absolute total volume of fluid administered during the first hour and during the first 4 hours is lower (by ≥ 25%) in the more restrictive fluid bolus resuscitation group than in the current recommended fluid bolus resuscitation group, as determined by the external pilot study objective concerning separation.

• Response: the absolute total volume of fluid administered during the first hour was only 7% lower in the 10-ml/kg fluid bolus group than in the 20-ml/kg fluid bolus group, and by the end of the 4-hour intervention period it was 23% lower. However, once imbalance in children’s weights between the groups was taken into account, this criterion was achieved, with the volume of fluid (in ml/kg) being 35% lower in the first hour and 44% lower over the entire 4-hour period.

Recommendation 1: the definitive FiSh trial, with its current design, should not be conducted

The full FiSh trial, with the current design, is not feasible, for the reasons given above. However, the optimal volume for fluid bolus resuscitation in the developed world remains unknown and infection is still an important cause of death in young children, even in the post-vaccine era. Thus, it may be that a different trial, with different inclusion criteria, or a different intervention, would be feasible and would answer this question. Possibilities might include adjusting the inclusion criteria to randomise patients to the trial only after receiving 40-ml/kg fluid bolus resuscitation, to select a more unwell population. Disadvantages of this approach would be the lower number of eligible patients in each centre and the difficulty in achieving separation. Alternatively, an intervention that compared usual management with a fluid-sparing/early inotrope arm, such as that currently being tested in the SQUEEZE trial,14 might have more chance of success.

Recommendation 2: further observational and epidemiological research should be conducted

Further observational and epidemiological research is required, both in EDs and in PICUs, to determine both current management and if there is still a population of children in the UK in whom a study looking at fluid bolus volumes could usefully be conducted. This would require the prospective collection of data on children presenting to EDs with suspected sepsis. In order to succeed, such a project would need a collaborative approach involving paediatric EDs, Acute Paediatric Units and PICUs. In relation to PICUs, it should be possible to perform an audit, linking to – and using – PICANet data collection systems, to determine the number of children on PICUs with severe sepsis and septic shock, where they are being admitted from (ED or ward) and their comorbidities.

Recommendation 3: further understanding of the moving clinical landscape is needed

Further qualitative work is required to determine whether or not clinical equipoise still exists among front-line professionals dealing with acutely ill children in relation to the research question. Our data seem to suggest that the equipoise, which previously existed, may already be lost in favour of clinicians using smaller, more limited, fluid bolus volume resuscitation than that currently ‘officially’ recommended. This may be a consequence of the FEAST9 trial or of smaller observational studies demonstrating a link between fluid bolus administration or fluid overload and poor outcomes in critically ill children.