Different temperature thresholds for antipyretic intervention in critically ill children with fever due to infection: the FEVER feasibility RCT

The FEVER qualitative study

To review, with input from parents/legal representatives:

• the acceptability of the selection of temperature thresholds and options for analgesia for a definitive FEVER RCT

• potential barriers to recruitment, the proposed process of decision-making and deferred consenting, and co-develop information and documentation for a definitive FEVER RCT

• the selection of important, relevant, patient-centred, primary and secondary outcomes for a definitive FEVER RCT.

To review and explore, with input from clinicians:

• the acceptability of temperature thresholds and options for analgesia for a definitive FEVER RCT

• potential barriers to recruitment, deferred consenting and associated training needs for a definitive FEVER RCT.

The FEVER observational study

• Estimate the size of the potentially eligible population for the definitive FEVER RCT.

• Confirm, using empirical data, the temperature threshold(s) currently employed for a standard approach for antipyretic intervention in NHS PICUs.

• Estimate the characteristics [e.g. mean and standard deviation (SD)] of selected important, relevant, patient-centred primary outcome measure(s).

The FEVER pilot randomised control trial with integrated-perspectives study

• Test the willingness of clinicians to screen, recruit and randomise eligible critically ill children.

• Estimate the recruitment rate of critically ill children.

• Test the acceptability of the deferred consenting procedure and participant information.

• Test, following randomisation, the delivery of and adherence to the selected temperature thresholds (intervention and control) for antipyretic intervention and to demonstrate separation between the randomised groups in peak temperature measurement over the first 48 hours following randomisation.

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

• Inform the final selection of a patient-centred primary outcome measure.

Research governance

An ethics application was made to the North West – Liverpool East Research Ethics Committee on 16 October 2015 and received a favourable opinion on 21 December 2016 (reference number 16/NW/0826). The protocol is available at www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/ (accessed 16 November 2018). Local NHS permissions were obtained for the NHS trusts for recruitment routes 2 and 3 (see Recruitment and sampling procedures).

Study management

The FEVER qualitative study was led by a co-investigator (KW). An experienced research associate (ED) was employed to organise, conduct and analyse the interviews and focus groups. The SMG was responsible for overseeing day-to-day management of the entire FEVER feasibility study, including the FEVER qualitative study.

Network support

To maintain the profile of the FEVER feasibility study, updates were provided at national meetings, such as the biannual Paediatric Intensive Care Society Study Group meetings.

Patient and public involvement

Two parents (CF and JW) and one young adult (BF) with experience of severe infection and admission to hospital 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. the draft pilot trial participant information sheets) 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. 16–20 Relevant research was used to develop a parent interview topic guide (see www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/; accessed 23 January 2019), two participant information sheets for the pilot trial (bereaved and non-bereaved) (see www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/; accessed 16 November 2018),17,18,20,21 a list of potential outcome measures17 to inform discussions with parents/legal representatives during interviews (see www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/; accessed 23 January 2019) and a staff focus group topic guide (see www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/; accessed 23 January 2019).

The interview and focus group topic guides contained open-ended questions and prompts to help explore parent/legal representative and staff views on the acceptability of a definitive FEVER RCT, including the draft FEVER pilot RCT participant information sheets and approach to consent. A separate section of questions was developed for bereaved parents/legal representatives.

Amendments to the study protocol

Minor amendments to the protocol, patient information sheet, schedule of events and statement of activities were approved by the Health Research Authority (HRA) on 10 January 2017.

A substantial amendment to the qualitative study protocol was submitted to the North West – Liverpool East Research Ethics Committee. The amendment requested the inclusion of an incentive description in the study recruitment advertisements and placement of the study advertisements in newspapers. A favourable opinion was received on 1 March 2017.

Participants

Based on previous studies,17,20 it was anticipated that 15–25 parents/legal representatives would be recruited to reach data saturation (i.e. the point at which no new major themes are discovered in analysis). We aimed to conduct a focus group with site staff (nurses and doctors) working in the four PICUs and associated retrieval services involved in the subsequent pilot RCT. We anticipated that a minimum of 4 and a maximum of 10 clinicians would attend each of the four focus groups (16–40 in total).

Recruitment and sampling procedure

For the focus groups, an e-mail invitation and participant information sheet was sent to lead clinicians (FEVER feasibility study co-investigators) in paediatric emergency and paediatric critical care medicine at the four NHS hospitals taking part in the pilot RCT. Co-investigators disseminated focus group invitations to all relevant staff and arranged a location and time for the focus group.

Parents were recruited through four routes (described in the following sections) to maximise the potential sample within the 6-month active recruitment period and to encourage diversity within the sample.

Interviews

Focus groups

Transcription

Digital audio-recordings were transcribed verbatim by a professional transcription company (Voicescript Ltd, Bristol, UK) in accordance with the Data Protection Act 1998. 23 Transcripts were anonymised and checked for accuracy. All identifiable information, such as names (e.g. of patients, family members or hospital their child was admitted to), were removed.

Data analysis

Qualitative data analysis of the interviews and focus groups was interpretive and iterative (Table 1). Utilising a thematic analysis approach, the aim was to provide accurate representation of parental views on trial design and acceptability. Thematic analysis is a method for identifying, analysing and reporting patterns (or themes) within data. Analysis was informed by the work of Braun and Clarke24 and their guide to thematic analysis. This approach allows for themes to be identified at a semantic level (i.e. surface meanings or summaries) or at a latent level (i.e. interpretive – theorising the significance of the patterns and their broader meanings and implications). 28 NVivo 10 (QSR International, Warrington, UK) software was used to assist in the organisation and coding of data. Elizabeth Deja (a psychologist) led the analysis with assistance from Kerry Woolfall (a sociologist).

Participants

Parents/legal representatives

A total of 46 parents registered interest, of whom 34 were screened (Figure 1). Eight were deemed ineligible and one did not confirm a date for interview. No individuals identified themselves as legal representatives; therefore, this term is not used in the remainder of this chapter.

FIGURE 1.

Recruitment to the FEVER qualitative study.

Parent characteristics

The sample included 25 parents: 20 mothers (4 bereaved) and 5 fathers (2 bereaved). Bereaved parents were interviewed at a mean of 19 months (SD 9.8 months) since their child’s admission (range 12–38 months). Children were a median age of 1.96 [interquartile range (IQR) 4.21] years (range 0.0–15.58 years) and had a median hospital length of stay of 12.5 (IQR 33.25) days (range 0.5–140 days) owing to streptococcus (n = 2), meningitis (n = 2), pneumonia (n = 1) and sepsis (n = 2). One bereaved father stopped the interview before diagnosis information was collected. Non-bereaved parents were interviewed a mean of 15.63 months (SD 6.5 months) since admission (range 6–29 months). The children had a median age of 1.5 years (IQR 3.87 years) (range 0.0–15.58 years) and a median hospital stay of 30.67 days (IQR 37.35 days) (range 7–140 days) for streptococcus (n = 2), meningitis (n = 4), sepsis (n = 6), chest infections (pneumonia, n = 5, and bronchitis, n = 2), toxic shock (n = 1), varicella (n = 2) and unknown infection (n = 2).

A total of 9 out of 24 parents (36%) had previous experience of trial participation either directly (themselves) or indirectly (through their child’s participation). Five parents (20%) had previous knowledge of RWPC. Ten out of 24 parents (41.7%) could not recall what temperature their child reached during their admission to hospital with a fever and severe infection. Of those who could recall, 7 out of 24 (29%) stated that their child’s temperature had exceeded 38.0 °C, whereas 5 out of 24 (21%) stated that their child’s temperature had exceeded 40 °C. A total of 15 out of 24 parents (63%) reported giving their child paracetamol at home, whereas the children of 13 out of the 24 parents (54%) were known to have been administered paracetamol by staff in hospital. Interviews took, on average (median), 47.59 minutes (SD 19.26 minutes) (range 15–105 minutes).

Site staff perspectives

A total of 56 staff took part in six focus groups across the four sites, with a median of 49.72 minutes (SD 11.14 minutes) in length (range 30.59–59.48 minutes). Staff mainly self-identified as nurses [20 junior nurses (36%) and 25 senior nurses (45%)]; there were also eight senior doctors (14%) and three junior doctors (5%). All were involved in the clinical care of children.

Parent perspectives

The FEVER pilot randomised controlled trial participant information sheet

All parents considered the participant information sheet (see www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/; accessed 16 November 2018) to be ‘fairly clear’ (P10, mother, non-bereaved), ‘make sense’ (P02, mother, bereaved) and written in comprehensible language:

I think they would understand it. It’s not been put in medical jargon, it’s simple for them to read.

P06, mother, bereaved

Parents identified parts of the participant information sheet that required clarification, including if not treating a temperature could cause a seizure, the importance of explaining RWPC at an early point in the recruitment discussion, that all other treatments would be given and some formatting suggestions (Box 2).

BOX 2 - Examples of points for clarification raised by parents

• Does not treating temperature increase the chance of seizure?

• How long will they be in the trial for?

• What if my child has already had paracetamol?

• It needs to be clear from the start that the child has already been entered in the study.

• Needs highlighting that the child is receiving all other treatment.

• Would prefer it if the headings were not in red.

Given the emotive situation in which the participant information sheet was going to be read, many suggested that the draft participant information sheet was too long and would benefit from all key information being summarised on the first page. Parents said that this format would provide them with the essential information needed at the time of the consent discussion with the option to read the rest of the information sheet at a later time. Consequently, during later interviews, we explored views on preferred information for the ‘important things you need to know’ section on the first page of the participant information sheet. As a result, the FEVER pilot RCT participant information sheet summary section was revised using information parents prioritised when considering FEVER information (Figure 2).

FIGURE 2.

Important things that you need to know: revised participant information sheet.

Site staff perspectives

Approach to consent in the FEVER randomised controlled trial

As shown in Figure 3, 22% (11/49, seven missing) of staff used their handsets to indicate that the use of RWPC was very acceptable or acceptable (25/49, 51%) in the proposed FEVER RCT. Although 14% (7/49) did not have any strong feelings (selected neutral), 12% (6/49) indicated that RWPC was unacceptable in the proposed FEVER RCT.

FIGURE 3.

Staff views on the acceptability of RWPC in the FEVER RCT.

Some members of staff discussed their support for the study and its use of RWPC, saying that there is a need for scientific evidence to address this research question and that they did not view informed consent as feasible owing to parental incapacity in an emergency situation:

. . . by giving paracetamol early, it might be actually prolonging their hospital stay and making them sicker for longer with consequences of actually getting other illnesses in the intensive care . . .

They can’t give consent . . . can they?

. . . they can’t actually give informed consent, they can barely say their name.

Others found it reassuring that although they would not be seeking informed consent, it’s not ‘cloak and dagger stuff’ (R12, FG4), as they anticipated that parents would be informed about trial participation and consent sought fairly shortly after randomisation:

I would anticipate that people do consent quite quickly, I don’t think it would need to be 4 days before you did consent, it would be able to be done within a reasonable . . . I don’t know exactly what time frame but I don’t think the parents would be in the dark about it for very long, so I think that it would be acceptable.

P01, FG6

As the following quotation shows, some staff supported RWPC in this trial, as they stated that the most effective temperature threshold to use in children presenting with a fever and suspected infection is not known, therefore children would not knowingly be denied an effective treatment:

Because we’re not denying a treatment that has a proven benefit, I would find that less acceptable, but because we don’t know whether it helps or not I think it’s acceptable to try and find out.

P02, FG6

Those who used their handsets to indicate a ‘neutral’ response or indicated that RWPC was unacceptable in the FEVER RCT went on to describe a range of reasons for their views. Most commonly, staff were concerned about RWPC in a trial that would require them to not to treat a child’s high temperature, therefore being ‘inactive’ in a potentially life-threatening situation. As the following quotation illustrates, some staff with previous experience of RWPC in paediatric RCTs viewed ‘withholding’ an emergency intervention without prior consent as less acceptable than administering a ‘needed’ intervention:

When you have to do something that’s an emergency, you know, it’s needed, it’s an emergency, you have to intervene and you have to do that without consent, that’s very different to withholding something.

P02, FG4

Staff were also concerned that parents may notice their child’s temperature and respond negatively to finding out that they were not being treated with an antipyretic owing to participation in a trial without their prior consent:

The thing is what won’t it [trial participation] be really obvious to the parents watching.

P03, FG4

I just wonder whether not treating people, which is what this says, is so completely contradictory to parents . . . that it just won’t sit well.

P05, FG3

But if we’re now saying, ‘oh we, we’re not ignoring it but we’re just not treating it,’ I don’t know whether you’d get a good response in an already very highly emotive environment.

P05, FG1

Staff concerns about the impact of research without prior consent on parent–clinician relationships

Many members of staff spoke of how RWPC in the FEVER RCT could negatively impact on rapport, relationship building and communication with parents: ‘You know, potential barrier to communication or, you know, working relationship’ (M3, FG1). One nurse spoke of how the ‘whole ethos of paediatric nursing is you’re meant to do it in partnership with the, you know, child and family’ (P05, FG1), and how RWPC goes against this ethos.

Others were uncomfortable about changing their usual practice of administrating antipyretics for a fever, as they were unsure of the clinical impact that would have on a child. Some stated that they would not give consent for their own child to be in the study, so they were not willing to put someone else’s in the FEVER RCT without prior consent. Nevertheless, they noted their conflicting desire for the study to be conducted to inform future clinical practice:

I think for me it’s a personal thing, I wouldn’t put my daughter into that trial to start off with, so I wouldn’t get, er, uninformed consent from another parent, simple as. There’s no way I would let my daughter’s temperature go that high before giving them paracetamol.

P09, FG4

I suppose the thing is it’s a bit rich isn’t it that, because we’re all happy to benefit from research. But we’re not all . . . that happy to perhaps participate.

P02, FG4

Towards the end of one focus group, provisional findings from the parent interviews were described to staff, including details of how parents viewed the use of RWPC and the higher threshold as acceptable. Although a few queried the representativeness of the interview sample, this information appeared to appease staff concerns, stating that if parents are happy with the study then staff would find taking part in FEVER more acceptable:

If that’s what parents are saying then, how can you dispute what the parents are saying.

P09, FG4

Key findings to inform the FEVER pilot randomised controlled trial

This qualitative study provides insight into the acceptability of the FEVER RCT by exploring the views of parents with relevant experience and site staff at the four pilot RCT sites.

This research found that parents supported the proposed trial and the use of RWPC. All stated that they would (hypothetically) consent to the use of their child’s information in the trial. Parents’ views on trial acceptability appeared to be influenced by a number of factors, including that all other treatments for suspected infection are given, the nature of the intervention (e.g. antipyretics being non-invasive), trust of medical staff to act in the best interests of their child, a desire to help other children in the future and a belief that the trial question is important.

The majority of parents were not concerned about the proposed restrictive temperature threshold of 37.5 °C. In contrast, many voiced specific concerns about the acceptability of the higher temperature threshold, including if not treating a child’s temperature would increase the likelihood of seizures. Parents stated that they would still consent to a trial that did not treat a fever with an antipyretic until 40 °C, as they trusted staff to monitor their child and act in their best interests. However, owing to concerns about 40 ^°C being a high and ‘scary number’, the majority suggested that the FEVER RCT would be more acceptable if the permissive temperature threshold was slightly lower (e.g. 39.9 °C or 39.5 °C).

Focus groups with site staff revealed concerns about 37.5 °C being too low a temperature threshold at which to administer an antipyretic, and the majority echoed parents’ perspectives in describing a permissive threshold of 40 °C as being too high. Staff were concerned about the acceptability of the permissive threshold to parents and about not using paracetamol for analgesia in the less unwell, spontaneously breathing patients who may be in pain or discomfort.

Many staff viewed RWPC as acceptable in the proposed trial, although for some the acceptability of RWPC was dependent on the permissive temperature threshold selected for the trial. Parents also viewed RWPC as acceptable in FEVER. Many described prospective informed consent as being inappropriate in emotive critical care situations and suggested that the RWPC process should be carefully managed by site staff in order to limit any potential negative impact on parents’ trust in staff. Parents’ views on RWPC were in line with previous research and guidance. 18 They recommended that staff in the proposed FEVER RCT should inform parents that their child has been randomised to the trial at the earliest possible time. Parents also recommended that staff should use their judgement about when is appropriate to approach each family, as well as tailor the level of information provided to meet individual needs. Bereaved parents recommended that communication about a child’s involvement in the proposed FEVER RCT after death should be personalised, honest and compassionate. Parents identified parts of the participant information sheet that required clarification, including if not treating a temperature could cause a seizure, that all other treatments would be given and some formatting suggestions, such as all prioritised information being on the first page.

Suggestions on how staff concerns could be addressed to assist ‘buy-in’ included restricting inclusion criteria to mechanically ventilated children; involving clinical staff as well as research staff in site training; providing information about the scientific evidence underpinning the trial; and clarifying key issues, such as what cooling methods could be used.

Qualitative study findings were used to develop the pilot RCT protocol, including the selection of temperature thresholds, inclusion criteria to require that patients are mechanically ventilated, participant information materials and site staff training package. To help address staff concerns about how parents may react to trial discussions, parent perspectives would be communicated in site training, highlighting parental acceptability of RWPC, temperature thresholds, parents’ questions and concerns about the study and, importantly, suggestions on how staff could address such questions.

Study design

The FEVER observational study was an epidemiological cohort study of fever attributable to infection in critically ill children following an unplanned admission to a PICU.

Research governance

Approval was granted by the Greater Manchester West Research Ethics Committee on 11 January 2017 (reference number 17/NW/0026) and the HRA on 17 January 2017 (Integrated Research Application System reference number 209929). The protocol is available at www.journalslibrary.nihr.ac.uk/programmes/hta/154401/#/ (accessed 16 November 2018).

The FEVER observational study was registered on the ClinicalTrials.gov database. Registration was confirmed on 20 January 2017 (reference number NCT03028818). The study was adopted in the National Institute for Health Research (NIHR) Clinical Research Network (CRN) Portfolio as a non-consenting study on 15 February 2017 and issued the NIHR CRN Portfolio number 33374.

Local confirmation of capacity and capability was obtained from each participating hospital trust to collect data for the study. For NHS England sites, the statement of activities was used as the study agreement. For sites in Northern Ireland, Wales and Scotland, an adapted clinical trial site agreement, based on the model agreement for non-commercial research in the health service, was signed by each participating hospital trust and the sponsor (ICNARC).

Data collection was embedded in the Paediatric Intensive Care Audit Network (PICANet), which has approval to collect patient-identifiable and personal data without consent. No patient-identifiable data were collected or used for the FEVER observational study. A release of data and a customised data collection request was made to the Healthcare Quality Improvement Partnership for access to unidentifiable routine PICANet data and to collect the additional data required for this study. Healthcare Quality Improvement Partnership approval was received for this purpose on 1 March 2017.

Study management

The study co-ordinator was responsible for day-to-day management of the FEVER observational study, with support from the ICNARC CTU and the SMG.

Amendments to the protocol

After receiving approval of the protocol from the HRA on 17 January 2017, three non-substantial amendments were submitted and categorised:

• Minor amendment 1 (approved 3 February 2017) consisted of the addition of 16 research sites.

• Minor amendment 2 (approved 21 February 2017) consisted of the addition of two research sites, and changes in principal investigator (PI) responsibilities in four participating sites.

• Minor amendment 3 (approved 17 March 2017) consisted of changes in PI responsibilities for a participating site.

NHS support costs

In order to enable screening and data collection for the FEVER observational study, resources equivalent to 0.09 whole-time equivalent of a Band 6 clinical/research nurse per PICU for a period of 6 months were agreed prior to the submission of the research grant for the FEVER feasibility study.

Participants: sites

Twenty-two PICUs (Table 3) were selected from those actively participating in PICANet. These sites represented a variety of UK PICU configurations, including general or combined intensive care units in general academic medical centres or within stand-alone children’s hospitals. In order to be selected to participate in the FEVER observational study, sites had to meet the following inclusion criteria:

• Meet all responsibilities as stated in the statement of activities (for sites in England) or the adapted clinical trial site agreement (for sites in Scotland, Wales and Northern Ireland).

• Identify and sign up a local PI.

• Identify a responsible FEVER observational study research nurse (to be funded or part-funded, centrally).

• Agree to collect additional FEVER data on all patients eligible for data collection.

• Agree to enter study data for all eligible patients via data collection screens and send securely to ICNARC.

Site initiation

Site initiation meetings were held via teleconference with all participating sites prior to the commencement of data collection (between 1 and 8 March 2017). The purpose of these meetings was to present the background and rationale for the study and to discuss data collection and submission. The PI from each participating site attended the meeting.

Following participation in the site initiation meetings, screening and data collection was commenced at participating sites once the statement of activities (sites in England) or clinical trial site agreement (sites in Scotland, Wales and Northern Ireland) had been signed and necessary approvals were in place. The scheduled start of data collection for the FEVER observational study was 1 March 2017. Sites initiated after this date collected data retrospectively on participants admitted from this date.

Investigator site file

An electronic investigator site file was provided to all participating sites. This contained all essential documents for the conduct of the study and included the approved protocol, all relevant approvals (e.g. local confirmation of capacity and capability), a copy of the statement of activities/clinical trial site agreement, copies of the approved participant information leaflet and family poster and all standard operating procedures (e.g. for entering data onto the customised data collection screen).

Communication and motivation

The study co-ordinator, with support from the research assistant, maintained close contact with the research team at participating sites by e-mail and telephone throughout the study. A teleconference was held mid-way through data collection, on 14 June 2017, with research teams at participating sites. The purpose of this was to provide a forum for site teams to ask questions and to discuss interim data submission at 3 months to allow analysis to inform the pilot RCT. PICANet provided regular status reports, which displayed the number of participants eligible for data collection, the data collection completion and the outstanding number of associated validation queries. These reports were used to identify specific sites that were falling behind on data collection.

Study population

Outcomes

• Potentially eligible population for a definitive FEVER trial.

• Temperature thresholds employed in usual care.

• The distribution of potential outcome measures for a definitive FEVER trial.

Data collection

Data submission

Participating sites downloaded a data extract from PICANet, consisting of the data collected specifically for the FEVER observational study and the specified PICANet routine data. The study data set was anonymous and submitted to ICNARC for analysis (Figure 4).

FIGURE 4.

The PICANet processes.

Sample size

Based on PICANet data, a sample size of approximately 1100 children was deemed to be sufficient to permit the calculation of 30-day mortality (anticipated to be 6% based on data from PICANet) with a 95% CI of ±1.4% based on the following assumptions:

• There were 21,326 emergency/unplanned admissions to the 27 UK PICUs during the period of 1 January 2011 to 31 December 2012. This corresponded to an anticipated overall sample size for the FEVER observational study in a target number of 20 UK PICUs over a 6-month period of approximately 3960 admissions (33 per PICU per month).

• There were 11,007 (51.6%) emergency/unplanned admissions to the 27 UK PICUs during the same period for patients who were admitted with a primary or secondary diagnosis consistent with a probable infection. Of these, 946 admissions (8.6%) met one or more of the proposed FEVER exclusion criteria, corresponding to approximately 1900 admissions (16 per PICU per month) who required the additional FEVER data collection on temperature and FEVER management.

• Of admissions with probable infection, 1583 (14.4%) were re-admissions of the same child. Based on data from the UK Paediatric Intensive Care Outcome Study (UK PICOS),29 approximately 70% of children admitted with infection have a peak temperature of ≥ 37.5 °C (estimated to be the likely temperature threshold).

Interim analysis

An interim analysis of data from the FEVER observational study was undertaken in July 2017 to inform the final design of the pilot RCT.

Study populations

All potentially eligible participants for the definitive FEVER trial were included in the analyses. Analyses were carried out based on the following populations:

1. All patients not meeting any stage 1 exclusion criteria and with confirmed or suspected infection (stage 2 criteria).

2. As population 1, restricted to patients receiving any mode of mechanical ventilation (invasive, non-invasive or high-flow oxygen) on days 1–2 of PICU admission.

3. As population 1, restricted to patients receiving invasive ventilation on days 1–2 of PICU admission.

Analysis methods

Statistical analysis

Statistical analyses were conducted in accordance with a prespecified statistical analysis plan written prior to the final data lock. The final analyses were conducted using Stata^®/SE version 14.2 (StataCorp LP, College Station, TX, USA) and R version 3.4.3 (The R Foundation for Statistical Computing, Vienna, Austria).

Missing data

All analyses were undertaken in the complete-case population. There was no imputation of missing data.

Participants

From 1 March 2017 to 31 August 2017, a total of 4126 children with unplanned PICU admissions were recorded from 22 hospitals. A total of 4008 patients (97.1%) had completed data entry relating to stage 1 criteria, and, of these, 737 (18.4%) met one or more of the stage 1 exclusions. Of the 3141 patients (78%) meeting stage 1 criteria, 1263 (40%) had no infection and a further 25 (0.8%) had missing data. The remaining 1853 patients (59.3%) meeting stage 2 criteria were identified as the potentially eligible population for a definitive FEVER RCT. We were unable to successfully collect PICANet data from 12 patients (0.6%); therefore, 1841 patients (99.4%) were included in the analysis for population 1, 1532 patients (82.7%) for population 2 and 1156 patients (62.4%) for population 3 (Figure 5). The variation across sites in the number and percentage of patients meeting stage 1 and 2 criteria is presented in Appendix 1.

FIGURE 5.

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) diagram for the study.

Baseline characteristics

Participant demographics and case mix are presented in Table 4. Age distribution, temperature at first face-to-face contact and highest temperature recorded on day 1 appear similar across the three populations. The proportions of antipyretic interventions on day 1 also appear consistent across groups; however, the use of intravenous opiates on day 1 is greatest in those receiving invasive mechanical ventilation (population 3) (74% of invasively mechanically ventilated patients received intravenous opiates). Infection was confirmed during the PICU stay for approximately two-thirds of participants in each of the populations, with viral infections and the lungs being the most common type and site of infection, respectively. Again, this was consistent across groups, accounting for > 60% (both type and site) of infections in each of the three populations.

Outcomes

Temperature thresholds

Figure 6 shows the proportion of participants in each population receiving an antipyretic intervention on day 1 with a maximum temperature above or below a given temperature ranging from 36.0 °C to 39.5 °C. In all participants (population 1), 55% of participants with a maximum temperature of > 36 °C received antipyretics, and this increased to 62% of participants at > 37 °C, 82% of participants at > 38 °C and 91% of participants at > 39 °C. The proportion of participants receiving antipyretics at the given temperatures on day 1 remained consistent when narrowing the inclusion criteria across populations 2 and 3 (Figure 6); however, there was variability across sites in temperature thresholds employed for antipyretic interventions on day 1 (Figure 7).

FIGURE 6.

The proportion of participants receiving antipyretic intervention on day 1 with a temperature above or below a given threshold. (a) Population 1; (b) population 2; and (c) population 3.

FIGURE 7.

Variation across sites in the proportion of participants receiving antipyretic intervention on day 1 among participants with a temperature of ≥ 37.5 °C (a, c and e) and ≥ 38 °C (b, d and f). (a and b) Population 1; (c and d) population 2; and (e and f) population 3.

For all participants not receiving an antipyretic intervention on day 1 but receiving subsequent antipyretic intervention, around 15% had a maximum temperature of ≤ 37 °C on the first day they received antipyretics (Figure 8). This increased to 68% of participants with a maximum temperature of ≤ 38 °C and 94% of participants with a temperature of ≤ 39 °C. The proportion of participants treated with antipyretics at these thresholds appeared to remain consistent across populations 2 and 3, with the mean temperature on the first day of receiving antipyretics being 37.8 °C (SD 0.8 °C) in population 1, 37.8 °C (SD 0.8 °C) in population 2 and 37.8 °C (SD 0.7 °C) in population 3. Again, there was considerable variability across sites in terms of temperature thresholds for antipyretic intervention on days 2–5 (Figure 9); however, the majority of participants in populations 1–3 with a temperature of ≥ 37.0 °C received an antipyretic intervention. The subgroup analyses of temperature thresholds employed by confirmed versus suspected and site/type of infection can be found in Appendix 2.

FIGURE 8.

Cumulative distribution of maximum daily temperatures of participants not receiving any antipyretics on day 1 and receiving subsequent antipyretic intervention. (a) Population 1; (b) population 2; and (c) population 3.

FIGURE 9.

Variation across sites in the median value of maximum temperatures on the first day on which antipyretics were received. (a) Population 1; (b) population 2; and (c) population 3.

Potential outcome measures

The distributions of potential outcome measures were consistent across the ≥ 37.5 °C and ≥ 38 °C temperature thresholds (Table 6). Overall, PICU mortality (population 1) was around 5%, and this was higher among mechanically ventilated patients, at around 5.5%, and even higher among invasively ventilated patients, at around 6.5%. Similarly, PICU length of stay increased with narrowing inclusion criteria for all participants and those who survived to PICU discharge. Although the median duration of organ support among those receiving the support was similar across all populations, the proportions of patients receiving mechanical ventilation and cardiovascular support and the mean duration across all patients (including those not receiving the support) increased with narrowing inclusion criteria. Use of renal support was low in all populations. Consistent with the results on mortality, length of stay and organ support, the mean number of days alive and free from PICU and mechanical ventilation reduced across populations 1–3 with the narrowing of inclusion criteria.

Summary of findings informing the FEVER pilot randomised controlled trial

An interim analysis was conducted in July 2017 on data collected from 21 out of 22 sites from between March and May 2017 to inform the design of the pilot RCT. This showed that the proposal to narrow the inclusion criteria to mechanically ventilated patients from the FEVER qualitative study was feasible in terms of recruitment. Data from the interim and full analyses also confirmed the lower temperature threshold for the usual care group of the pilot RCT at ≥ 37.5 °C, as this temperature falls within usual care across UK PICUs.

Study management

The study co-ordinator at the ICNARC CTU was responsible for the day-to-day management of the FEVER pilot RCT and Kerry Woolfall was responsible for day-to-day management of the integrated-perspectives study. Both aspects were completed with support from the ICNARC CTU and the SMG.

Patient and public involvement

Engagement with patients was vital to the successful conduct of the FEVER pilot RCT with integrated-perspectives study. A parent of a child who was admitted to a PICU with severe infection provided oversight on the FEVER pilot RCT TSC; three members were co-investigators (JW, CF and BF) and members of the SMG. They provided input into the conduct of the FEVER pilot RCT, including reviewing literature to be given to patients and their families (e.g. participant information sheets). In addition to this, patients and members of the public were engaged through the FEVER qualitative study in which the result informed the pilot RCT design and trial materials.

One of the main aspects of the integrated-perspectives study was to obtain input from parents of children who were enrolled in the FEVER pilot RCT.

Design and development of the protocol

The FEVER pilot RCT protocol was designed to inform key parameters and to test the design and possible conduct of the proposed definitive FEVER RCT. The protocol design was informed by the perspectives of parents and clinicians, as described in Chapter 2, and the observational data, as reported in Chapter 3.

Amendments to the FEVER pilot randomised controlled trial protocol

Following receipt of approval from the HRA on 11 August 2017, one substantial and two minor amendments were submitted and categorised.

Substantial amendment 1 (approved on 30 October 2017) consisted of the following:

• the addition of the inclusion criterion ‘referral requiring PICU admission to a participating unit’, which clarified for transport services that randomisation could only take place during transport for eligible patients being transported to a participating unit

• the removal of the exclusion criterion ‘sickle cell disease’ as the SMG felt there to be no clinical reason for those with sickle cell disease not to take part in the trial

• the addition of the exclusion criterion ‘previously recruited into the FEVER pilot RCT’

• amendment of the existing exclusion criteria ‘severe rhabdomyolysis’ to ‘rhabdomyolysis’ and ‘receiving or requiring mechanical ventilation’ to ‘new requirement for mechanical ventilation’

• minor administrative changes to the protocol that added clarity to the trial processes.

Minor amendment 1 (approved on 22 November 2017) consisted of the addition of a researcher contact details card to the participant materials given to the parent/legal guardian who agreed to take part in a telephone interview. This was to help assist contact and arrangement of interviews.

Minor amendment 2 (approved on 28 December 2017) consisted of a minor administrative change to the site staff questionnaire to mirror the focus group topic guide.

NHS support costs

The NHS support costs were agreed prior to the submission of the research grant to include screening to identify eligible patients and obtaining informed consent from parents. This equated to 0.15 whole-time equivalent of a Band 6 research nurse for 4 months in each of the four participating PICUs.

Sites

Four sites agreed to participate in the FEVER pilot RCT (Table 7). The sites agreed to participate prior to the submission of the grant application, and for study continuity, including participation in the FEVER qualitative study and the FEVER observational study. Each FEVER pilot RCT site was obliged to:

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

• identify and sign up a local PI

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

• agree to incorporate the FEVER pilot RCT into routine transport team and PICU activity, particularly highlighting the importance of screening at first contact

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

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

Participants

Recruitment

The trial procedures for recruitment and follow-up of participants are summarised in Figure 10.

FIGURE 10.

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

Inclusion criteria

Patients were eligible for inclusion in the FEVER pilot RCT if they met all of the following criteria:

• unplanned PICU admission

• aged ≥ 28 days and < 16 years

• referral requiring PICU admission to a participating unit

• fever of ≥ 37.5 °C in the first 48 hours following contact with the paediatric retrieval service/PICU

• new requirement for mechanical ventilation

• treating clinician presumes the cause of the fever is an infective process.

Exclusion criteria

Patients were excluded if they met any of the following criteria:

• acute encephalopathy, including convulsive status epilepticus

• postcardiopulmonary bypass or known/suspected cardiomyopathy/myocarditis

• rhabdomyolysis (defined as serum creatine kinase concentration at least 10 times the upper limit of normal)

• malignant hyperthermia, neuroleptic malignant syndrome or drug-induced hyperthermia

• receiving palliative care, or death perceived as imminent

• previously recruited to the FEVER pilot RCT.

Interventions

The trial incorporated a pragmatic approach to temperature control in both trial groups. An example treatment algorithm is illustrated in Figure 11. Treatments to reduce temperature were only permitted in response to a temperature threshold of ≥ 39.5 °C in the permissive group and a temperature threshold of ≥ 37.5 °C in the restrictive group. The treatment strategies for the restrictive and permissive group commenced from randomisation and were sustained until PICU discharge or death while receiving mechanical ventilation (defined as invasive, non-invasive or high-flow oxygen). During periods when the participant did not receive mechanical ventilation, the trial processes were not applicable. If the participant subsequently recommenced mechanical ventilation, the trial processes were reapplied.

FIGURE 11.

Example treatment algorithm.

Outcomes

• The proportion of eligible children recorded in the trial screening logs who were recruited to the FEVER pilot RCT and the reported reasons for non-recruitment.

• Recruitment rate of eligible, critically ill children.

• The proportion of recruited children whose parents subsequently declined to give consent or who withdrew their child from the FEVER pilot RCT having initially given consent.

• Adherence to the temperature thresholds for antipyretic intervention in both the intervention and standard care groups.

• Separation between the treatment groups in maximum temperature measurement over 48 hours following randomisation.

• A comparison between treatment groups and assessment of the completeness of follow-up for the following potential outcome measures:

  • Mortality at discharge from a PICU (‘PICU mortality’).

  • Mortality at ultimate discharge from an acute hospital (‘hospital mortality’).

  • Mortality at 30 days post randomisation (‘30-day mortality’).

  • Length of stay in a PICU (calculated in days from the date of randomisation to the date of discharge from or death in a PICU).

  • Duration of organ support (calculated as the number of calendar days on which the organ support was received in a PICU up to day 28) –

    • Number of days of mechanical ventilation.

    • Number of days of cardiovascular support.

    • Number of days of renal support.

  • Number of days alive and free from a PICU to day 28.

  • Number of days alive and free from mechanical ventilation to day 28.

Safety monitoring

Participants were monitored for AEs occurring between randomisation and PICU discharge. Specified AEs were:

• seizures

• rhabdomyolysis (defined as serum creatine kinase concentration at least 10 times the upper limit of normal)

• cerebral oedema.

Unspecified AEs were defined as an unfavourable symptom or disease temporally associated with the instigation of the FEVER pilot RCT treatments, whether or not it was related to the trial treatment, that was not deemed to be a direct result of the participant’s medical condition and/or standard emergency or critical care treatment. All AEs were recorded in the eCRF and reported as part of routine reporting throughout the FEVER pilot RCT to the DMEC. AEs that were assessed to be serious (i.e. prolonging hospitalisation or resulting in persistent or significant disability/incapacity), life-threatening or fatal – collectively termed serious adverse events (SAEs) – were reported to the ICNARC CTU. SAEs that were unspecified and considered to be possibly, probably or definitely related to the trial treatment were reported to the Research Ethics Committee within 15 calendar days of the event being reported.

Data collection

A secure, dedicated eCRF, hosted by ICNARC, was set up for data to be entered by staff at participating sites. The eCRF was accessible only to authorised users with access being centrally controlled by the study co-ordinator or data manager (access was granted after cross-checking the site delegation of study duties log). Each individual was provided with a unique username and password and had access to data only for participants recruited at their site. The data set for the FEVER pilot RCT included the minimum data required to confirm patient eligibility, describe the patient population, monitor and describe delivery of the intervention, assess outcomes and enable identification of the patient through PICANet to collect applicable routinely collected data.

Sample size

The FEVER pilot RCT was set up without a defined primary outcome and, hence, without a usual power calculation to determine sample size. Instead, the sample was determined to be adequate to estimate critical parameters to be tested to a necessary degree of precision.

Data from 1537 children admitted to a PICU with a reason for admission associated with infection and a temperature of ≥ 37.5 °C in UK PICUs indicate a mean peak temperature in the first 24 hours following PICU admission of 38.5 °C with a SD of 0.7 °C (unpublished data from the UK PICOS study;29 Kathryn M Rowan and David A Harrison, ICNARC, 2016, personal communication). Based on this, a sample of 84 children was anticipated to be sufficient to give 90% power to demonstrate a separation of 0.5 °C in mean peak temperature between the higher temperature threshold group and standard care group. The FEVER pilot RCT would therefore need to recruit 100 children to allow for up to an anticipated 16% withdrawal following deferred consent. 32

Allowing for a rate of non-recruitment of 30% of eligible children, recruitment was anticipated to be completed with four PICUs/retrieval teams recruiting for 4 months (anticipated recruitment rate of 6.25 children per PICU per month).

Statistical analysis

Outcome definitions

The following potential outcome measures for the definitive RCT were compared between the treatment groups:

1. PICU mortality

2. hospital mortality

3. 30-day mortality

4. length of stay in a PICU (calculated in days from the date of randomisation to the date of discharge from PICU or death in a PICU)

5. duration of organ support (calculated as the number of calendar days on which the organ support was received in a PICU up to day 28) –

   1. – number of days of mechanical ventilation

   2. – number of days of cardiovascular support

   3. – number of days of renal support

6. number of days alive and free from PICU to day 28

7. number of days alive and free from mechanical ventilation to day 28.

Participants who died between randomisation and day 28 were assigned a value of zero days alive and free from PICU/mechanical ventilation to day 28. For participants who survived to day 28, the number of days alive and free from PICU (mechanical ventilation) were calculated as the number of calendar days between the day of randomisation and day 28, inclusive, on which the participant was not in a PICU (not receiving mechanical ventilation in a PICU) at any time during the calendar day.

Analysis methods

Mortality outcomes are reported as the number and percentage in each group. The risk ratio and absolute risk reduction were calculated unadjusted, with 95% CIs based on approximate standard errors.

Length of stay in a PICU is reported as the mean (SD) and median (IQR) in each treatment group, overall and stratified by survival to PICU discharge.

Duration of organ support is reported as the number (%) receiving the organ support, the median (IQR) duration of organ support among participants receiving the organ support and the mean (SD) duration of organ support among all participants (with participants not receiving the support assigned a value of zero) in each group. The difference in means was calculated unadjusted, with 95% CIs based on Student’s t-distribution assuming equal variance.

The numbers of days alive and free from a PICU/mechanical ventilation are reported as the mean (SD) in each group and the difference in means with 95% CI, calculated as for the duration of organ support.

In addition, histograms of the continuous outcomes by treatment group are presented.

Missing data

To assess the follow-up procedures, the number (%) of participants with complete follow-up data for each of the potential outcome measures for the definitive RCT, overall and by treatment group, is reported.

All analyses were undertaken in the complete-case population. There was no imputation of missing data.

Harms

The number and percentage of participants experiencing each prespecified AE (plus any other AEs as reported) while in PICU were summarised for each treatment group. Numbers of SAEs, severity and reported relatedness to treatment are also reported.

Additional analyses

If the results of the FEVER observational study suggested a potential narrowing of the inclusion criteria (e.g. a higher temperature threshold or restricting to participants receiving invasive ventilation), then it was planned to repeat the FEVER pilot RCT analyses on the subgroup of patients likely to be eligible for the definitive FEVER RCT.

Participants

Based on previous research and the FEVER pilot RCT sample size,17,18 it was estimated that approximately 50 parents/legal representatives (including both mothers and fathers of randomised children) would be recruited to the questionnaire element and 15–25 (depending on data saturation) parents/legal representatives would be recruited to the interview element. In addition, four focus groups with site staff (approximately 6–8 per group) and an online questionnaire for those who could not attend focus groups were planned.

Recruitment and consent procedure

Procedures

Transcription

Digital audio-recordings were transcribed verbatim by a professional transcription company (Voicescript Ltd, Bristol, UK) in accordance with the Data Protection Act 1998. 23 Transcripts were anonymised and checked for accuracy. All identifiable information, such as names (e.g. of patients, family members or the hospital their child was admitted to), were removed.

Data analysis

Qualitative analysis of interview and focus group data was interpretive and iterative (Table 8). Utilising a thematic analysis approach, the aim was to provide accurate representation of parental views on trial design and acceptability. Thematic analysis is a method of identifying, analysing and reporting patterns (or themes) within data. Analysis was informed by the work of Braun and Clarke24 and their guide to thematic analysis. This approach allows for themes to be identified at a semantic level (i.e. surface meanings or summaries) or at a latent level (i.e. interpretive – theorising the significance of the patterns and their broader meanings and implications). 28 NVivo 10 software was used to assist in the organisation and coding of data. ED (a psychologist) and Kerry Woolfall (a sociologist) analysed and synthesised the data. Quantitative data from the parent and staff questionnaire were cleaned and entered into IBM SPSS Statistics version 20.0. Descriptive statistics are presented with percentages.

Sites

All sites were open to recruitment within 3 months of receiving full HRA approval (Table 9). Initial local document packs were sent in June 2017 and the final packs were sent on 17 August 2017. The median time from receiving the final local document packs to the pilot trial opening at sites (i.e. start of screening) was 51.5 days (range 39–83 days) and to the first participant recruited was 62 days (range 41–84 days). Delays in opening were predominantly attributable to the time taken to obtain local confirmation of capacity and capability.

Participants

Between 25 September 2017 and 19 December 2017, a total of 154 patients were screened and deemed to meet the inclusion criteria (Figure 12). Of these, 15 patients (9.7%) met one or more of the exclusion criteria. Of those patients deemed to be eligible, 26 (18.7%) were missed and a further 12 (8.6%) were not randomised owing to local clinical decisions. Overall, 101 eligible children (72.7%) were randomised into the pilot RCT. One patient randomised to the permissive group was identified immediately post randomisation as a duplicate randomisation and removed from the study. Therefore, 72.5% (100/138, 95% CI 64.5% to 79.2%) of eligible patients were appropriately randomised into the study.

FIGURE 12.

The CONSORT flow diagram. a, Two patients were under the care of social workers, two had a lack of understanding of study information and one was a site decision. b, Includes 139 patients eligible for randomisation minus one duplicate patient. c, Includes 101 randomised patients minus one duplicate patient.

The overall recruitment rate of 11.1 participants per site per month (95% CI 9.0 to 13.5 participants) was greater than the pre-trial estimate of 6.25 (Figure 13); however, recruitment rates varied across sites (Figure 14). The number of patients eligible, screened and recruited per month at each site can be found in Table 10.

FIGURE 13.

Cumulative recruitment over time compared with pre-trial expected recruitment.

FIGURE 14.

Overall recruitment rate at each site. GOSH, Great Ormond Street Hospital.

Consent

Consent to continue in the study was refused in 8 out of 100 patients (8.0%, 95% CI 4.1% to 15.0%). Of these, seven had been allocated to the permissive treatment group (accounting for 14.3% of patients in this group, 95% CI 7.1% to 26.7%). In addition, seven parents/legal representatives (five in the permissive group and two in the restrictive group) refused consent to be treated in accordance with the study protocol but consented for ongoing data collection, and a further three patients treated in the permissive group had consent withdrawn so that an antipyretic intervention could be given.

In one hospital, 7 out of the 10 patients randomised to the permissive group either refused or withdrew consent to the intervention, which accounted for almost half (46.7%) of the total number across all four sites. The consent procedure was also frequently carried out earlier than stated in the protocol (24–48 hours post randomisation), with the consent procedure being completed on the same day as randomisation for 29 out of the 91 patients (31.9%) to have a face-to-face approach. Eight out of 29 patients (27.6%) refused or withdrew consent if the consent procedure was completed on the day of randomisation, whereas 10 out of 62 (16.1%) refused or withdrew consent on the subsequent days after randomisation.

In addition, five patients were withdrawn from the study centrally as the parents/legal representatives were unable to be approached for consent. Therefore, 87 out of the 100 correctly randomised patients were included in the analysis of objectives 4–6 (87.0%, 95% CI 79.9% to 92.2%).

Baseline patient characteristics

Overall, the two treatment groups appear well matched at baseline (Table 11), with only a small difference between the groups in terms of mean age and a balance across groups in the proportion of participants aged < 1 year. Over half of the participants were male (63.2% in the permissive group and 55.1% in the restrictive group) and bronchiolitis was the primary diagnosis (50% in the permissive group and 49.0% in the restrictive group). Illness severity was balanced across the two groups in terms of PIM2r and PIM3 scores, with approximately 90% of participants invasively ventilated (92.1% in the permissive group and 89.8% in the restrictive group). Baseline temperature was the same in both groups; however, a greater proportion of participants in the restrictive temperature group received pre-randomisation paracetamol (Table 12).

Temperatures

Highest temperatures

Figure 15 shows the distribution of the highest temperatures for participants over the first 6 days of the intervention based on individual patient measurements presented as both box plots and medians and IQRs. On average, participants in the restrictive group had reached a temperature of 37.5 °C by 24 hours and participants in the permissive group had a temperature of > 37.5 °C until 42 hours. The permissive group spent a greater proportion of time below the allocated treatment threshold over both the first 48 hours and the first 6 days of the intervention (Figure 16). Overall, the proportion of time spent below the allocated temperature target was a median of 1.0 hours (IQR 0.9–1.0 hours) in the permissive group compared with a median of 0.7 hours (IQR 0.5–0.8 hours) in the restrictive group.

FIGURE 15.

Highest temperature, by 6-hour time period and treatment group. (a) Box plot; and (b) median and IQR.

FIGURE 16.

Distribution of time spent at each temperature, by treatment group. (a) During the first 48 hours; and (b) during the first 6 days.

Separation between the groups

Over the first 48 hours, on average, participants in the permissive group had a mean peak temperature of 38.8 °C (95% CI 38.6 to 39.1 °C) compared with 38.4 °C (95% CI 38.2 to 38.6 °C) in the restrictive group. This resulted in a separation in mean peak temperature over the first 48 hours of 0.5 °C (95% CI 0.2 to 0.8 °C) between the groups. The separation of around 0.5 °C between treatment groups was maintained up to around 84 hours following randomisation (Figure 17).

FIGURE 17.

Difference between groups in the mean of the highest temperature at each time point with 95% CI.

Interventions

Receipt of antipyretic interventions

The proportion of participants receiving antipyretic interventions while mechanically ventilated over the first 6 days of the intervention was greater in the restrictive group (Figure 18) than in the permissive group. This was consistent across both paracetamol and external/other cooling (Table 13).

FIGURE 18.

Proportion of participants who received any antipyretic intervention while mechanically ventilated.

TABLE 13 - Percentage of participants receiving antipyretic interventions while mechanically ventilated

Number of hours from randomisation.

Number of participants receiving mechanical ventilation in each time period.

Hoursa	Treatment group
	Permissive				Restrictive
	n b	Antipyretic intervention (%)			n b	Antipyretic intervention (%)
		Paracetamol	NSAID	External and other		Paracetamol	NSAID	External and other
Baseline	38	7.9	0.0	7.9	49	18.4	2.0	6.1
0–6	38	13.2	0.0	13.2	49	51.0	0.0	24.5
6–12	37	10.8	2.7	21.6	49	34.7	0.0	22.4
12–18	36	8.3	0.0	5.6	47	29.8	0.0	23.4
18–24	35	11.4	0.0	2.9	46	34.8	0.0	21.7
24–30	35	8.6	2.9	5.7	45	35.6	0.0	13.3
30–36	35	11.4	2.9	2.9	43	18.6	0.0	11.6
36–42	32	9.4	3.1	6.3	40	20.0	2.5	2.5
42–48	29	13.8	0.0	3.4	39	15.4	2.6	15.4
48–54	27	14.8	0.0	7.4	38	15.8	0.0	21.1
54–60	25	8.0	0.0	0.0	37	10.8	0.0	2.7
60–66	24	16.7	0.0	0.0	36	13.9	0.0	8.3
66–72	23	8.7	0.0	13.0	34	29.4	2.9	11.8
72–78	23	4.3	0.0	4.3	36	19.4	0.0	11.1
78–84	22	9.1	0.0	0.0	34	20.6	0.0	8.8
84–90	21	9.5	0.0	4.8	32	21.9	0.0	9.4
90–96	19	15.8	0.0	5.3	28	25.0	0.0	14.3
96–102	18	11.1	0.0	11.1	26	23.1	0.0	19.2
102–108	19	15.8	0.0	5.3	24	29.2	0.0	12.5
108–114	18	0.0	0.0	0.0	23	30.4	0.0	8.7
114–120	17	17.6	0.0	0.0	20	25.0	0.0	15.0
120–126	17	11.8	0.0	0.0	18	27.8	0.0	11.1
126–132	17	17.6	0.0	0.0	16	6.2	0.0	12.5
132–138	15	20.0	0.0	0.0	16	31.2	0.0	18.8
138–144	15	13.3	0.0	0.0	16	12.5	0.0	12.5

Use of opiates

As shown in Figure 19 and Table 14, the vast majority of participants received opiates and the proportions were similar across groups.

FIGURE 19.

Proportion of participants receiving opiates while mechanically ventilated.

TABLE 14 - Percentage of participants receiving opiates while mechanically ventilated

Number of hours from randomisation.

Number of participants receiving mechanical ventilation in each time period.

Hoursa	Treatment group
	Permissive			Restrictive
	n b	Opiates (%)		n b	Opiates (%)
		Intravenous opiate infusion	Opiate bolus		Intravenous opiate infusion	Opiate bolus
0–6	38	84.2	28.9	49	85.7	28.6
6–12	37	83.8	18.9	49	81.6	26.5
12–18	36	75.0	19.4	47	80.9	21.3
18–24	35	77.1	22.9	46	82.6	10.9
24–30	35	74.3	8.6	45	84.4	17.8
30–36	35	65.7	17.1	43	83.7	7.0
36–42	32	65.6	12.5	40	80.0	12.5
42–48	29	69.0	13.8	39	79.5	15.4
48–54	27	66.7	11.1	38	78.9	7.9
54–60	25	68.0	16.0	37	73.0	10.8
60–66	24	66.7	16.7	36	69.4	13.9
66–72	23	69.6	13.0	34	70.6	8.8
72–78	23	69.6	17.4	36	63.9	13.9
78–84	22	68.2	13.6	34	61.8	8.8
84–90	21	66.7	28.6	32	59.4	15.6
90–96	19	63.2	10.5	28	64.3	10.7
96–102	18	66.7	5.6	26	50.0	0.0
102–108	19	68.4	15.8	24	54.2	8.3
108–114	18	72.2	16.7	23	56.5	13.0
114–120	17	58.8	17.6	20	65.0	20.0
120–126	17	52.9	17.6	18	66.7	0.0
126–132	17	52.9	29.4	16	62.5	6.2
132–138	15	46.7	13.3	16	62.5	6.2
138–144	15	40.0	6.7	16	50.0	6.2

Protocol adherence

Protocol deviations were reported in 39 out of 628 6-hour time periods (6.2%) in the permissive group and 60 out of 810 time periods (7.4%) in the restrictive temperature group. Overall, 39% of participants in the permissive group and 55% of participants in the restrictive group experienced at least one protocol deviation (Table 15). The number of participants who received antipyretic interventions when below the allocated temperature threshold was higher in the permissive group (14, 36.8%) than in the restrictive group (11, 22.4%). In a total of 37 out of 589 6-hour periods (6.3%), these deviations showed that participants in the permissive group were receiving mechanical ventilation and had a documented temperature measurement below the threshold, compared with 20 out of 504 6-hour periods in the restrictive group (4.0%). In contrast, fewer participants did not receive the recommended antipyretic interventions above the allocated temperature threshold in the permissive group (2, 5.3%) than in the restrictive group (20, 40.8%). In 2 out of 39 6-hour periods (5.1%), these deviations showed that participants in the permissive group were receiving mechanical ventilation and had a documented temperature measurement above the threshold, compared with 40 out of 306 6-hour periods in the restrictive group (13.1%).

For the permissive group, reasons for receiving antipyretic interventions during the 6-hour period when the temperature was below the threshold included paracetamol for pain or discomfort when receiving non-invasive or high-flow oxygen or being weaned from invasive ventilation (16 occurrences); external or other cooling given in error (10 occurrences); lack of trial communication (five occurrences); other staff errors (four occurrences); and withdrawal of consent for the intervention (one occurrence). No clear explanation from the site was provided for the remaining deviation. When restricting the analyses to protocol deviations solely while the participants were receiving invasive ventilation, the proportion of 6-hour time periods when antipyretic interventions were received while the temperature was below the threshold in the permissive group reduces from 6.3% to 3.1% (15 occurrences out of 479 6-hour time periods). This is lower than the corresponding proportion (3.5%, 15 occurrences out of 424 6-hour time periods) in the restrictive group.

For the restrictive group, the circumstances when antipyretic interventions were not received during a 6-hour period with a maximum temperature at or above the threshold were paracetamol having been given in the previous or subsequent time period (19 instances), paracetamol having been given in both the previous and subsequent time periods (six instances), a single-period spike in temperature (six instances), the temperature dropped below the threshold by the subsequent time period (five instances), the participant was no longer mechanically ventilated by the subsequent time period (one instance) and external/other cooling was given in the subsequent time period (one instance). No clear explanation from the site was provided for the remaining deviations. Only two participants had two consecutive 6-hour time periods in which antipyretic interventions were not received when the temperature was at or above the threshold. Of these, one participant received paracetamol in the subsequent time period and the other participant’s temperature subsequently dropped below the threshold.

Potential outcome measures

For participants included in the analyses, the completeness of outcome data was high, with one participant excluded from all mortality outcomes owing to withdrawal of consent and four participants excluded from hospital mortality owing to withdrawal of consent (n = 1) or because they remained in hospital at the time of the data lock (n = 3) (Table 16). As anticipated for a small pilot trial, there were no significant differences between the groups in any of the outcomes (Table 17). Four participants remained in hospital. Histograms plotted for continuous measures are presented in Appendix 3.

Safety monitoring

There were no reported SAEs. In total, three AEs were reported. One seizure was reported in the permissive group (an expected AE) that was deemed unlikely to be related to the trial intervention. There were two AEs in the restrictive group: one seizure that was deemed unrelated to the trial intervention and one case of rhabdomyolysis (an expected AE) that was deemed possibly related to the trial intervention (Table 18).

Additional analyses

As specified in the methods, because the FEVER observational study suggested that it may be beneficial to narrow the inclusion criteria to include invasively ventilated patients only, rather than all those with mechanical ventilation, the FEVER pilot RCT analysis was re-run on this group of patients only (see Appendix 4).

Characteristics

Of the 60 parents (49 mothers and 11 fathers) who took part, eight parents (13%) of seven children reported that they had not provided consent for their child’s continued involvement in the FEVER pilot RCT. One parent interviewed was bereaved at the point of interview. Parents were recruited for interviews and questionnaires from all four sites. Seven of the parents interviewed (7/19, 37%) said that they had previous experience of participating in medical research. No parents reported prior experience of RWPC. Parents were interviewed a median of 31 days after randomisation (range 9–70 days). Their children had received treatment for respiratory illness (e.g. bronchiolitis and respiratory syncytial virus) (18/19, 94%), cancer (1/19, 5%) and septic shock (1/19, 5%). Interviews took between 20 and 50 minutes (mean 32 minutes).

Characteristics

Of the 97 staff who provided information about their role, 75 out of 97 (77%) were nurses, 45 out of 97 (60%) were senior-level staff. Fourteen out of 97 (14%) were doctors, 12 (86%) of whom were consultant-level doctors. The majority of staff (79/98, 81%) were involved in the clinical care of children (44/48, 92% of online questionnaire participants; and 35/48, 73% of focus group participants). Focus groups took an average of 52.5 minutes (range 22.52–106.00 minutes).

Research without prior consent is acceptable in the proposed FEVER randomised controlled trial

Questionnaires and interviews to explore parents’ views on RWPC were used in the FEVER pilot RCT and proposed FEVER RCT. As shown in Table 19, 17 parents (35%) who completed a questionnaire indicated that they were initially surprised to find out that their child had already been entered into the FEVER pilot RCT. However, the majority (89%) were satisfied with the RWPC process for this study.

Interview data supported questionnaire findings. Parents described their initial surprise at finding out that their child was involved in a trial without their prior consent. However, many were supportive of RWPC for this trial, as well as other PICU trials, as they understood the reasons why informed consent could not be sought in a critical care situation:

I think it’s not ideal to find out they are in a study, you know, without your prior consent, but I didn’t have any issues with that particular study and I understand the reasons why they’re doing it, and I think it’s a, you know . . . a good reason and for a good, good cause.

P79, interview, mother, permissive

I don’t think there’s any other way better to go about it.

P79, interview, mother, restrictive

Parents described how RWPC was acceptable as the FEVER pilot RCT discussion took place when the situation had calmed down and the emergency situation had passed. Some emphasised the importance of clinicians broaching the trial discussion at the first appropriate opportunity:

No, I think it was better the way they did it because I had calmed down, he was in intensive care, everything was settled down and then they spoke with . . . Yeah, I do actually think it’s better that way around.

P49, interview, mother, permissive

As long as it’s sort of discussed at the first suitable opportunity, erm, then that’s fine . . . Yeah, as soon as it’s appropriate to.

P82, interview, mother, restrictive

Nevertheless, there were some exceptions. Four (4/48, 9%) questionnaire and two interview participants indicated that they were not satisfied with RWPC in the FEVER pilot RCT. Only one out of the eight parents who declined consent and took part in an interview (n = 2) or questionnaire (n = 6) indicated that they were unhappy regarding RWPC; however, the use of RWPC was not the reason they had declined consent.

Interestingly, four parents did not believe that their child had been entered into the trial without their prior consent as they believed that they had provided informed consent prior to randomisation:

Because at that point, um, she hadn’t been assigned to it [a threshold] and it was randomly. I mean ‘cause we made that decision at the time, she may have been in the 39 [°C group].

P53, interview, father, restrictive

They come to me and asked us first . . .

Did they?

. . . before putting him in the trial. Yes, they did, sorry.

Yeah.

They didn’t just put him in. They, they come and asked me first if they could do the trial.

This finding suggests that in a few cases either the consent process was not followed correctly, randomisation to a temperature threshold had not been explained clearly at the point of broaching the trial or the parent did not have the capacity at the time to understand the RWPC process. This finding was subsequently discussed in three staff focus groups. All staff confirmed that RWPC had been used for all randomised participants without exception.

One of the mothers who believed they (the parents) had consented before randomisation explained that although they would have been unhappy that their child had been entered into the FEVER pilot RCT without their consent, they would still have provided consent for the use of their child’s information and to remain in the FEVER pilot RCT as they wished to help others:

I wouldn’t have been ha-, like happy at first that the would go ahead and do it without asking permission but then I would still say, well you can still, you can still do. It . . . it was just one them things if this could help then obviously I am gonna do it, but it’s just not knowing what they were doing that I wouldn’t have been at first.

P85, interview, mother, permissive

Another mother described how her dissatisfaction with RWPC was exacerbated in the FEVER pilot RCT as she had noticed a sticker on her child’s bed indicating trial participation before staff had broached the FEVER pilot RCT:

I walked onto the ward that morning and I saw the, um, little sticker on the side of his bed that said that, the higher temperature and it said FEVER RCT and I had no idea what it meant. I think he came round later and said that, er, and I was a bit shocked that they were asking me for consent for him to be on the trial, even though he was already on the trial.

P84, questionnaire, mother, permissive, declined consent

Others appeared to be unhappy with any alternative to informed consent for research, rather than specific concerns about RWPC in this pilot RCT:

Well I personally don’t agree with, um, somebody else deciding that for me . . . before they come and speak to me about it.

P84, interview, mother, permissive, declined consent

Parents’ experiences of the FEVER pilot randomised controlled trial consent process

The majority of parents, including 96% of those who completed the questionnaire, indicated that site staff had checked that it was a convenient time to discuss the FEVER pilot RCT (Table 20). Most parents (46/48, 96%) also indicated that the information they received about the FEVER pilot RCT was clear and straightforward to understand. This was corroborated during interviews with parents describing how staff had approached them appropriately, with well-timed, clear, comprehensive study information:

It was all, all fine and I understood what they were saying and was happy to sort of go ahead with it, with the trial. If it, if I was, if it wasn’t explained to me too well, I probably wouldn’t have bothered doing it, so I don’t think there was any problems with it being explained to me.

P77, interview, father, restrictive

I feel like it was well-explained to me.

P79, interview, mother, permissive

I didn’t think it [trial information] was missing anything.

P78, interview, mother, restrictive

Patient information materials were also described as being ‘clear and understandable’ (P80, interview, mother, permissive), ‘jargon free’ (P30, interview, mother, restrictive) and the ‘perfect’ length (P85, interview, mother, permissive):

I mean we were given so much information on various other things and that’s one thing I’ll always say, you know, everything was clear and in plain English. Um, yeah, it was easy to follow all the literature.

P79, interview, father, permissive

One mother suggested that in the proposed FEVER RCT participant information sheet it would be useful to include information about trial outcome measures to help fully understand what the trial aims are:

The outcome measures would be good so people could understand, they understand the reasons why and then understand what it is that is actually being looked at as part of the reasons.

P80, interview, mother, permissive

Parents’ views on trial acceptability

Following discussion of the consent processes, the focus of the interviews shifted to parents’ views on the acceptability of the proposed FEVER RCT, including the temperature thresholds.

All parents interviewed, including those who declined their child’s continued participation, supported the proposed trial:

I think it’s a brilliant idea, so I’m all, I’m all for it.

P80, interview mother, permissive

It sounded quite interesting.

P84, interview, mother, permissive, declined consent

This was underpinned by the perception that ‘it was an interesting trial’ (P30, interview, mother, restrictive) that had been ‘clearly explained’ (P74, interview, mother, permissive), and involved a non-invasive treatment (paracetamol), which parents were familiar with:

It’s just how are they gonna give the paracetamol, when they’re gonna give it. I mean if it was more severe, um, more of an invasive study, um, I might have been a bit, I might have had to query it a little bit more but I was happy with, with everything.

P53, interview, father, restrictive

Many also considered the proposed FEVER RCT to be acceptable and had consented to the FEVER pilot RCT as they did not perceive it as causing harm to their child:

No, I just thought, you know, what’s, what’s the harm? There’s no harm in taking like, letting her into the study.

P78, interview, mother, restrictive

Parents of children allocated to the restrictive temperature threshold found the trial very acceptable as they viewed giving paracetamol at this temperature as being their normal practice at home:

I think it probably helped, the fact that it’s something that I would do myself anyway, so I felt comfortable that they were giving him paracetamol at 37 degrees [°C], which I would do anyway. I think that was the main thing. It wasn’t something that was out of my comfort.

P82, interview, father, restrictive

Parents also viewed the permissive threshold to be acceptable; however, this acceptability was contingent on the perception that their child was not in any discomfort. Interestingly, no parents referred to the temperature their child reached when discussing acceptability:

The only thing would be if she wasn’t on any other kind of pain relief, um, but there’s other things to manage, her discomfort.

P73, interview, father, permissive

Why parents consented to the FEVER pilot randomised controlled trial

The questionnaire included a series of statements to establish why parents had consented for their child’s information to be used and/or provided consent to continue involvement in the FEVER pilot RCT (Table 21). Parents often indicated multiple reasons for providing consent (mode four reasons, mean two reasons). In line with qualitative findings, the main reason why parents provided consent was to help other children in the future:

At the end of the day it will benefit other children in the future so, I wanted to obviously help [. . .] if we can find the information to help other children in the future, that would be good.

P35, interview, mother, restrictive

A high proportion of parents (32/41, 78%) indicated that they provided consent to help other children in the future and because of a belief that medical research studies like FEVER are important (32/41, 78%). After this desire to help other children and future research, parents were motivated by the belief that trial participation could help their own child (30/41, 73%). ‘Other’ reasons for consent included two parents who felt that their child had benefited from additional monitoring owing to trial participation and two parents who referred to how their child was comfortable or recovering well. Another parent described wanting to ‘give back to medical service as NHS done so much to help my child’ (P59, questionnaire, mother, restrictive threshold).

Although 25 out of 41 parents (60%) reported providing consent because they trusted the doctor or nurse who explained the FEVER pilot RCT, no parents identified this as their main reason for providing consent. During interviews, parents also spoke of their trust in the person who explained the study. However, they also appeared to place their trust in the knowledge, skills and values of clinicians looking after their child. Parents repeatedly emphasised the importance of their trust in doctors and nurses to place the needs of their child above the needs of the study:

She’s in a hospital. I mean them people know better than me, so I understand that they would never put a child in harm’s way.

P81, interview, father, restrictive

Yeah. I mean we were trying to take the view that we were, we’re in their hands and they know what they’re doing so, um, I’ve, I’ve always felt that, you know, it wouldn’t . . . you know, they’re not gonna put any undue risk on the child so yeah, I was happy with that level set. Um, and it’s set at that level for a reason I’m sure.

P73, interview, father, permissive

Parental acceptability of the permissive threshold was underpinned by trust in staff to act in their child’s best interests

During interviews, parents described how they expected staff to act in their child’s best interests, which included not adhering to the protocol by administering an antipyretic if at any point clinical staff felt that it was needed. The parents who were interviewed appeared to understand and value their ability to withdraw or decline consent for their child’s continued involvement in the FEVER pilot RCT:

I know if anything did happen, yous can stop at any time. Stop it if they saw it was getting out of hand and he, and I felt like it, what like it, it wasn’t helping, that I would stop it. Erm, but I knew he was in like capable hands, obviously being in the, intensive care unit that he was going to be absolutely fine, they wouldn’t let him go to the stage of him getting poorly.

P85, interview, father, restrictive

I was happy enough for him to undergo the trial but if at any point the nurses thought he could do with the Calpol, or I thought he could, then I wanted the trial to stop it could do.

P49, interview, mother, permissive

Some parents referred to how the trial was acceptable as ‘my child is comfortable’ (P49, questionnaire, mother, permissive). Indeed, two mothers described how they found the trial acceptable and gave full consent, but later chose to withdraw their child from the study when they were being weaned from the ventilator as they were concerned about their child being in pain or distress:

The only time we eventually pulled him from the trial and gave him paracetamol was when he was awake and he was a lot more distressed, and that was harder for me to watch, especially, especially the way he was, and I said, ‘look, if it’s gonna help, I’d rather you gave him it,’ but when he was sedated and he was ventilated and everything, he did get a temperature and, like I say, he brought it back down himself.

P49, interview, mother, permissive, consented then withdrew

He was withdrawn because, um, he didn’t actually have a temperature at all and he was in a little bit of pain once he’d been extubated. So they wanted to give him some paracetamol but he didn’t have any temperature any way.

P76, interview, mother, permissive, consented then withdrew

Why parents declined consent to the FEVER pilot randomised controlled trial

We gained insight into the reasons why eight parents (8/18, 44% of total decliners) refused consent for their child to continue to take part in the FEVER pilot RCT. Of these, six parents completed a questionnaire and two took part in an interview.

One parent reported declining owing to her child no longer being ventilated and a belief that her child was ‘therefore not part of the research requirements’ (P67, questionnaire, mother, restrictive, decliner). Of the other seven parents, six had children who were allocated to the permissive threshold and one was in the restrictive threshold. Two were parents of the same child: the mother (P84) took part in an interview and the father (P72) completed a questionnaire.

As shown in Table 22, only one parent declined consent to use data already collected, as well as consent for their child to continue in the trial. Parents’ reasons for declining consent included concerns about any negative impact on their child owing to the child’s pre-existing medical condition, concerns about their child being in pain or discomfort, increased workload, risk of epileptic seizures owing to family history, incapacity to make an informed decision and a wish to have the answer to the research question before trial participation.

Importantly, interview and questionnaire data suggest that parents who declined some element of their child’s participation often described their support for the proposed FEVER RCT. No parents were unhappy about their child’s involvement in the FEVER pilot RCT:

From being on other wards, as soon as babies or children even spike a little bit of a temperature they seem to be quite quick on the, quick let’s give them paracetamol. We can say that first hand where we’ve always been like a kind of, a hold off [. . .] I think sometimes paracetamol can hide other things going on as well.

P83, interview, mother, permissive, decliner

As described in Chapter 2, Parent perspectives, one mother (P84, interview, mother, decliner) was dissatisfied with RWPC in any trial. However, the parents who declined appeared to be satisfied with the consent process and provided similar questionnaire responses to those who had consented to the FEVER pilot RCT. This was corroborated during the interviews, with both mothers saying that consent discussions were clear and well timed:

And did you think the timing was appropriate?

Yeah, it was absolutely fine. The bloke that I spoke to so lovely about it. But, because we did feel a bit bad afterwards saying no, because he was so nice, but obviously, you know, at the end of the day he respected our decision.

Parent-centred outcomes

As part of an iterative process, findings from the FEVER qualitative study (see Chapter 2) were incorporated into the topic guide as interviewing progressed to further explore parents’ perspectives about important outcome measures. Parents were asked to think about their experience of their child being admitted to hospital with a suspected infection and then prompted to share their views on what effects they hoped the FEVER treatment would do to help their child.

At this point in the interview, the majority of parents’ responses focused on a reduction or reduced fluctuation of their child’s temperature. Parents also prioritised stopping any discomfort and/or pain, getting the child ‘back to themselves’ and reducing time in hospital. Others found this question difficult to understand: ‘I don’t know really’ (P79, interview, mother, permissive) and ‘I have no idea [laughter]’ (P76, interview, mother, permissive). All parents were then asked ‘What would you be looking for as an indicator that your child was getting better?’. The most common responses were:

• time spent on mechanical ventilation

• looking and/or behaving like normal self

• child not in discomfort and/or pain

• improved mood.

The interviewer then repeated back the outcome measures identified to the individual, as well as any other outcomes that had been referred to and noted by the researcher during interview discussions. Parents were then asked to rank their identified outcomes in order of importance for a proposed FEVER RCT. The top-prioritised outcomes in order of how commonly they were described as being most important were:

• time spent on mechanical ventilation

• child not in discomfort and/or pain

• reduction of, or reduced, fluctuation of child’s temperature

• looking and/or behaving like normal self

• time spent in hospital

• number of re-admissions to hospital.

As no parents had mentioned survival as an outcome measure, the researcher queried the inclusion of survival in the list after parents had prioritised outcomes. All except one parent stipulated that survival is the most important outcome measure for the proposed FEVER RCT and gave a range of reasons why she had not initially mentioned it (Box 6). These reasons included death not being something she wished to consider or that she had not considered owing to her child’s condition (bronchiolitis) and, therefore, she thought that it would not be a salient outcome measure for this trial:

I hadn’t thought about it, but also, yeah, don’t think it’s necessarily the best outcome measure that you could have.

P80, interview, mother, permissive

Therefore, other than survival, six prioritised outcomes were identified from the combined analysis of outcome measure discussions, initial prioritisation and ranking. These were:

• length of time on breathing support (e.g. ventilator, high-flow oxygen)

• not in pain and/or discomfort

• how quickly vital statistics are back to normal (including temperature)

• looking and/or behaving like normal self

• length of time in hospital

• number of re-admissions to hospital.

Interestingly, this list includes only short-term outcomes, in contrast to the outcomes list created from the FEVER qualitative study (see Chapter 2), which included longer-term outcomes. This finding may indicate that parents’ prioritisation of outcomes may be influenced by their experience of their child’s illness, survival and the point at which they are asked about outcomes of importance in the course of their child’s illness.

Combining outcomes-related data from both the qualitative and integrated-perspective studies suggests that, apart from survival, parents would prioritise seven outcomes for measurement in the proposed FEVER RCT (Box 7).

Site staff perspectives

Restrictive temperature threshold

In questionnaires and focus groups, staff were asked to rate how acceptable they viewed the restrictive temperature of 37.5 °C. As shown in Table 23, most (95%) indicated that the restrictive temperature was acceptable (66% considered it very acceptable and 29% considered it acceptable). Only five nurses did not view this threshold as acceptable, two of whom described how they were unhappy with the lower threshold as it was not their normal practice to treat a fever at this low temperature:

I understand this was for the trial, however it was mentioned more than once from myself and others that we would not normally treat with paracetamol a temperature of 37.5 [°C].

P49, questionnaire free-text response

So I’ve put not acceptable . . . because actually I wouldn’t normally give it to them at 37.5, so why, all of a sudden, am I overmedicating someone. That’s more uncomfortable to me, than waiting till like 39 [°C].

P01, FG5

Overall, these findings are in stark contrast to those reported in the FEVER qualitative study (see Chapter 2), in which the majority of participants had concerns about unnecessarily treating children with an antipyretic. Although this pilot RCT staff sample was larger, the majority of participants had also taken part in the FEVER qualitative study focus groups. As the following quotation shows, experience of conducting the FEVER pilot RCT appeared to change staff views about the acceptability of the restrictive temperature threshold. Parents’ positive responses to their child’s participation in the restrictive group of the FEVER pilot RCT appeared to influence staff views on the acceptability of administering an antipyretic at a lower temperature than their usual practice:

Everybody that was in the lower end of it, I found were like happy to take part.

P01, FG4

Bedside clinical staff acknowledged that they may not have normally given paracetamol at such a low temperature; however, they described how, despite initial reservations, they did follow the FEVER pilot RCT trial protocol and did not find it unacceptable to treat children at 37.5 °C:

I wasn’t like, oh I shouldn’t be doing this. But, erm, but, yeah, I mean I suppose if we, because we were following sort of like a protocol I mean, yeah, it didn’t feel, it didn’t feel wrong if that’s what you’re asking?

P03, FG4

One participant described how it was acceptable to treat at such a low temperature as a child may have a fever or be in pain:

I think it could be acceptable if you’re like looking at different things. Like they might be in pain or there’s a temperature there. If you look at like the whole, you might give them paracetamol if they’re 37.5 [°C].

P02, FG5

Permissive temperature threshold

Staff had mixed views about the acceptability of the permissive temperature threshold. Approximately half (42/79, 53%) indicated that the 39.5 °C threshold was acceptable, whereas just under half (37/79, 47%) did not (Table 24). There were no notable differences between the views of staff who completed the questionnaire and the views of those who took part in the focus group. However, staff trained by their colleagues were more likely to find the permissive threshold not acceptable, or very unacceptable, than staff who received training from the FEVER study team, which included the chief investigator. This may have been linked to the training issues highlighted in The FEVER pilot randomised controlled trial site training, as all staff who rated the site training as poor had been trained by their colleagues. These staff appeared to have remained unclear about the scientific rationale for the study, which negatively influenced their views on the acceptability of the trial.

As described in Chapter 2, FEVER qualitative study findings highlighted parent and staff concerns about a permissive threshold of 40 °C. These findings informed the selection of a FEVER pilot RCT permissive temperature threshold of 39.5 °C and limiting inclusion criteria to children already receiving ventilation. In focus groups, many staff who were positive about the permissive threshold in the FEVER pilot RCT described how they valued how staff and parent views had been listened to. Changes made to the protocol positively influenced staff perceptions about the acceptability of the study. Some research nurses responsible for training site staff spoke of how being able to describe how parent and staff views had informed changes to the study had helped engage colleagues in the FEVER pilot RCT:

I know from giving the training, having the feedback that things have been changed according to what parents had said changed the kind of response of the people that we were giving training to. You know, there was an initial kind of barrier up and then once you explained certain parts of it had been changed, staff did relax a little bit into it and go ‘right OK,’ and engage a bit more in the training so whether that helped.

P01, FG1

As the following interview excerpt shows, prior concerns about not being able to use an antipyretic to treat pain had been addressed by the FEVER team limiting the pilot RCT sample to children receiving mechanical ventilation. Staff described how this inclusion criterion had changed their views on trial acceptability, as children would already be receiving other medication, such as an opioid, for pain relief:

It was mainly that they all had an infusion running of kind of something else, an opioid and . . .

And they would have had that anyway?

Because they were ventilated, yeah.

Yeah, OK. Did that make quite a difference [to staff views on trial acceptability] then, that vent [ventilation] and infusion?

Yeah.

You didn’t see the sort of effects and it wasn’t like a self-ventilating patient that you might see sort of crying and being quite distressed.

P02, FG4

Others stated that their previous concerns about high temperatures causing harm or discomfort were not observed. One nurse anticipated that experience of conducting the proposed FEVER RCT would help address staff concerns in the future as ‘seeing is believing’ (P03, FG1):

I think also what’ll happen is as some patients are randomised into the higher temperature and people are, see that they’re actually manageable and it doesn’t cause them any, um, harm, then people generally get more accepting of it can be one or the other category. It’s kind of seeing is believing.

P03, FG1

Some staff were surprised at how positive parents in the FEVER pilot RCT had been about their child’s participation in the permissive threshold group. A few attributed these positive responses to staff explanations about the use of alternative medication for pain. Others described how some parents had an interest in the trial question and the potential to inform the future treatment of fever in children with a suspected infection:

I think a lot of them were happier once they understood there were alternatives we could give for pain.

P04, FG2

I think I definitely found that a lot of families were really interested in it, and like something like temperature, something that they actually understand. And I think, if they understood that we’re doing something because maybe the outcome is, is better, I think then families were a lot more on board.

P04, FG5

In contrast, staff who did not find the permissive threshold acceptable were unhappy about not administering paracetamol when they thought that a child was uncomfortable or in pain. These concerns meant that some staff did not adhere to the protocol and administered paracetamol before a child’s temperature had reached 39.5 °C:

I often looked after patients that I felt were in pain and may benefit [from] paracetamol for comfort. I found it unacceptable that I felt I should not use paracetamol in these circumstances when it is our first line drug for pain. I found doctors often reluctant to prescribe paracetamol for comfort, I am a nurse, it is extremely important to me that my patients are comfortable. I continued to request paracetamol for pain and comfort.

P39, questionnaire response

I’ve never known a trial before where I feel like potentially we’re making our patients more uncomfortable.

P01, FG2

Think the thing I was most uncomfortable with, er, was the, the higher level . . . I thought it was too high because at that stage the patient I was looking after was, was very distressed and very uncomfortable from what I remember I gave paracetamol because I didn’t fi-, I didn’t think it was fair on the child to leave them that hot and that distressed.

P01, FG3

Many were concerned about not giving paracetamol for pain or discomfort when a child was conscious. A few staff said that alternative medications for pain, such as morphine, seemed ‘extreme’ (P01 and P03, FG2). Suggestions were made to revise the proposed FEVER RCT protocol to exclude patients not on mechanical ventilation (e.g. high-flow nasal oxygen) or those close to being extubated when sedation is being weaned:

I think people/nurses have found it difficult to not be able to give there paracetamol when the child is upset.

P22, questionnaire free-text response

We’re much happy to, happier to be compliant, erm, if the child was intubated and ventilated and knocked out . . . most of our problems came in really, I would say, in classifying Optiflow [Fisher & Paykel Healthcare Limited, Panmure, New Zealand] and, you know, high-flow nasal cannular as mechanical ventilation. I think that obviously there are huge problems in . . . because a child is awake.

P05, FG2

Do not include high flow.

P63, questionnaire free-text response

And then at the point that you’re weaning their ventilation . . . you’re weaning . . . their sedation to wake them up. So you’re reducing their level of comfort so you would automatically normally give paracetamol . . . I know it sounds dramatic but a little bit like collateral damage. Like for the greater good that they’re gonna be in more discomfort while they’ve got, they’re in intensive care to find out these results sort of thing. But, yeah, [if] they were on morphine that would be easier.

P03, FG2

Staff in one focus group described how they ‘do not sedate anybody that much unless you are very sick’ (P04, FG3). They discussed the benefits of using paracetamol to ensure that a child is comfortable and how its use can prevent the need for other drugs. As the quotation below suggests, this practice appeared to apply to children on invasive ventilation, which meant that staff had concerns about changing their usual practice in FEVER, leading to protocol deviations when children in the permissive threshold looked uncomfortable or in pain:

As paracetamol is such, erm, a good drug and we use it for lots of situations, it’s not just temperature control . . . You can reduce the amount of other drugs you give, so you actually, you know, you could aim people towards getting them extubated because they’re, they’re breathing better and, you know, ‘cause you can reduce the amount of morphine and use paracetamol as a, as an adjunct with those drugs. And actually taking that away almost, especially with somebody like him because the [FEVER patient] pa-, I remember this case as well, erm, it, it kind of tied your arms behind your back because you’re like that wou-, is what you would use. That’s [paracetamol] we would usually use, you know?

P04, FG3

These members of staff described a clinical decision to withdraw a pilot RCT participant without input from parents owing to staff concerns about how the child looked:

It was easy until that point and, erm, I jus-, I felt, I don’t, you know, I didn’t just make the decision on my own, I think I spoke to other people. The parents weren’t there, erm, I don’t know if I should have got their consent or not, but I just felt, erm, it was difficult to stick to it looking at how the child was . . . So I made that decision, well, along with other people.

P01, FG3

In each focus group, staff provided a few examples of parents who had withdrawn their child from the study because had been unhappy or concerned about their child’s well-being in the permissive temperature threshold group. These examples appeared to be uncommon. In most cases, parents had informed staff that they would have consented for their child to remain in the trial if they had been randomised to the lower threshold:

I knew that one of the parents, erm, I think withdrew because their child was starting to get to like 39 [°C] and was uncomfortable, and they didn’t feel happy keeping their child involved in the study. So, again, if it was slightly lower threshold for giving it you’d probably maintain like a higher participant rate.

P01, FG3

His view was that he would be happy to do the study, but he felt 39.5 [°C] was too high, erm, that the study didn’t need to be done and that we already knew exactly what was going on . . .

P04, FG4

Others described how waiting for a child’s temperature to rise had caused parents stress, leading them to withdraw their child from the FEVER pilot RCT:

Sometimes parents were uneasy about watching the temperature climb even after consenting to the trial. A few families took their child out of the trial because it was causing more stress to them.

P76, questionnaire free-text response

Parents’ perspectives

The findings suggest that the majority of parents who took part in an interview or completed a questionnaire supported the proposed FEVER RCT. Administering an antipyretic for fever was something that parents understood and were familiar with and, as a result, parents were interested in the trial question and felt that the proposed trial was important. The non-invasive nature of the intervention appeared to influence parents’ views on trial acceptability, which they did not perceive to pose a risk to their child’s well-being. As shown in previous studies, parents strongly valued the advancement of medical research and commonly stated that they provided consent as a way to help other families and children in the future. 16,20,36–39

Descriptions of the permissive temperature threshold as a ‘scary number’ by parents in the FEVER qualitative study were not observed in data from parents of children who took part in the FEVER pilot RCT. Interestingly, no parents referred to the temperature their child reached when discussing trial acceptability. Nevertheless, parents only viewed the permissive threshold as being acceptable if their child was not in pain or discomfort. Parents trusted clinicians to act in the best interests of their child and expected them to deviate from the protocol by administering an antipyretic if at any point they felt that it was needed. Both parents interviewed who had withdrawn their child from the trial did so because of concerns about pain or discomfort. Similar concerns were also described by some parents whose children were in the permissive threshold group and had declined their child’s continued participation in the trial. Others had declined because of reasons such as their child’s pre-existing medical condition, risk of epileptic seizures owing to family history or incapacity to make an informed decision owing to the situation.

Importantly, interview and questionnaire data suggest that parents who declined some element of their child’s participation in the FEVER pilot RCT still supported the proposed FEVER RCT. Parents valued clinicians’ willingness to give an antipyretic, as well as the option to withdraw or decline continued participation in the trial when their child was perceived to be uncomfortable or in pain. No parents who took part in this study element expressed anger or dissatisfaction about their child’s involvement in the FEVER pilot RCT.

Using information that parents’ prioritised in the FEVER qualitative study to inform the participant information sheet and staff training appeared to address potential parental concerns about the trial. For example, anticipated concerns from the FEVER qualitative study, such as if not treating a fever could cause a seizure, were not voiced by parents in the FEVER pilot RCT. Parents described the participant information sheet as being clear, comprehensive and an appropriate length to read in a critical care situation.

As shown in previous studies,17,18 some parents were initially surprised to find out that their children had been entered into a trial without their prior consent. However, following a tailored explanation from site staff about why prospective consent was not sought, the majority of parents were satisfied and supported this approach to consent for this trial, as well as other PICU trials. Parents described how RWPC was acceptable, as the FEVER pilot RCT discussion took place when the emergency situation had passed. Some emphasised the importance of clinicians broaching the trial discussion at the first appropriate opportunity, which is consistent with existing guidance. 30 There was a minority of exceptions. A few parents were unhappy that their child had been randomised to FEVER without their prior consent. However, during interviews, parents described how their concerns were not specific to the proposed FEVER RCT, as they would not be happy with any alternative to informed consent for research.

Some parents stated that they had provided informed consent for the FEVER pilot RCT. Although it was found that most did support the use of RWPC, this finding suggests that there were instances in which either the RWPC process was not followed by staff, the trial consent process had not been clearly communicated or parents did not have the capacity to understand information provided in the critical care situation. This finding is perhaps not surprising as at the point of discussing the trial the child’s clinical care may not have been affected for the purposes of the trial. Indeed, if a child’s temperature had not reached a ‘standard care’ threshold (e.g. < 37.5 °C) when staff broached the trial, then parental decision-making was not dissimilar to a prospective informed consent decision as nothing had happened at that point in time, apart from online randomisation. This interpretation, coupled with concerns voiced about children being in pain or discomfort in the permissive threshold group, may help to explain the higher decliner rate in this group.

Outcomes

Combined data on parent-centred outcomes from the qualitative and integrated-perspective studies suggest that, apart from survival, the following outcomes should be prioritised for measurement in the proposed FEVER RCT: length of time on mechanical ventilation, long-term effects of illness on child (physical and developmental), looking and/or behaving like their normal self (examples included improved mood, improved communication, being more like themselves and more alert, sitting up and starting to eat and drink), not being in discomfort and/or pain, number of days spent in the PICU and hospital, vital signs back to normal (e.g. heart rate, breathing rate and temperature) and effect on family (e.g. emotional, functional and financial). Interestingly, as also shown in the FiSh feasibility study,17 parents who were interviewed shortly after their child’s admission in the FEVER pilot RCT prioritised shorter-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 FEVER qualitative study, who were interviewed a mean of 19 months after their child’s hospital admission. These findings add further evidence to suggest that parents’ prioritisation of outcomes may be influenced by their experience of their child’s illness, 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 findings should be considered in any future priority setting or design of core outcome set development work.

Site staff perspectives

Overall, the majority of clinicians viewed the trial and its use of RWPC to be acceptable as well as practically possible to conduct. Many staff confirmed their support for the proposed FEVER RCT, often stating that this trial was needed to provide an answer to an important research question.

Staff viewed the restrictive temperature threshold to be acceptable. This finding was in stark contrast to those reported in the FEVER qualitative study (see Chapter 2), in which the majority of participants had concerns about unnecessarily treating children with an antipyretic. Although the FEVER pilot RCT staff sample was larger, the majority of participants had also taken part in the FEVER qualitative study. As shown in previous critical care research, staff experience of conducting a trial can help address negative preconceptions about how parents may respond to the trial and its use of RWPC. As described by one nurse (P06, FG1), ‘seeing is believing’. Parents’ positive responses to their child’s participation in the restrictive group of the FEVER pilot RCT appeared to influence staff views on the acceptability of administering an antipyretic at a lower temperature than their usual practice. In the permissive group, staff described how parents appeared to value staff explanations about the use of alternative medication for pain. In line with the parent perspective findings, staff described how some parents were interested in the trial question and supported the study as they wished to help inform future treatments for children with a fever and suspected infection.

Those who were positive about both temperature thresholds appeared to value how parent and staff perspectives had been incorporated into the FEVER pilot RCT design. Changes made by the FEVER team, such as the selection of a FEVER pilot RCT permissive temperature threshold of 39.5 °C rather than 40 °C, and amending inclusion criteria to require that patients are mechanically ventilated, had positively influenced staff views on trial acceptability. Some research nurses responsible for the dissemination of training felt that being able to show how FEVER qualitative study findings had informed the protocol had helped to facilitate engagement of colleagues in the FEVER pilot RCT.

Nevertheless, polarised staff views about the acceptability of the permissive temperature threshold were found. Approximately half of the sample indicated that the 39.5 °C threshold was acceptable, whereas just under half did not. Those who raised concerns did not find the permissive threshold acceptable when a child was perceived to be uncomfortable or in pain. This had resulted in staff administering paracetamol or actively cooling a child before their temperature had reached the permissive threshold. As previously described, parents valued staff decisions to administer paracetamol when they felt it was needed. To help address this issue, suggestions were made to exclude patients not on mechanical ventilation (e.g. those on high-flow nasal oxygen) and to end trial participation when sedation is being weaned in preparation for extubation.

The majority of clinicians positively rated site training, which they felt prepared them for recruitment and consent in the FEVER pilot RCT. An important finding was that staff who viewed the FEVER pilot RCT training to be poor were often trained by their colleagues rather than the FEVER team at the site initiation meeting. It was also found that staff trained by their colleagues were more likely to find the permissive threshold unacceptable when compared with staff trained at the site initiation meeting. These staff remained unclear about the scientific rationale for the study, which negatively influenced their views on the acceptability of the trial. Although half of participants did not experience problems adhering to the protocol, others raised uncertainties or misunderstandings about what constituted active cooling, which led to protocol deviations.

As shown in Chapter 5, the FEVER pilot RCT recruitment target was reached ahead of time, yet the findings of this research suggest that in a future trial, recruitment rates could be further increased. Collaborative working between retrieval teams and clinical and research staff assisted the identification of eligible patients at sites. However, the FEVER pilot RCT was set up during a very busy winter period, which led to difficulties in ensuring that all clinical staff and retrieval teams were able to attend training. One retrieval team was unable to engage in training and limited research support meant that screening did not always taking place at evenings and weekends; therefore, eligible patients were missed. Nurses played an important role in identifying patients yet were unable to randomise until a GCP-trained doctor was present. Many nurses were enthusiastic about the study and wanted to play a more active role in randomisation in a future trial.

When combined, these findings suggest the need to focus on the content, timing and dissemination of training in the proposed FEVER RCT. Suggestions to help improve recruitment and trial conduct for the proposed trial include additional research support, GCP training for senior nurses so they can randomise patients and a comprehensive training package that can be disseminated to all staff. This research identified that training content in the proposed FEVER RCT should include clarification on what constitutes active cooling, the scientific rationale and existing evidence that informed the study question, presentation of the findings on parents’ and staff perspectives about trial acceptability, and any changes made to the FEVER RCT protocol based on these findings.

Overall, the findings of this research highlight the value of doing pre-trial research to inform the design of challenging clinical trials. 40 Incorporation of these findings into staff training, and the protocol for the proposed FEVER RCT, should help assist staff buy in to improve recruitment, protocol adherence and, importantly, ensure that the trial is patient centred.

The FEVER qualitative study

The FEVER qualitative study provided insight into the acceptability of a definitive FEVER RCT by exploring the views of parents with relevant experience and site staff at the four pilot RCT sites.

Interviews were conducted with 25 parents – 20 mothers (four bereaved) and five fathers (two bereaved) – of children admitted to a UK PICU in the preceding 3 years. Overall, parents supported the conduct of a definitive FEVER RCT. Despite describing support for the study, concerns were raised regarding the acceptability of the permissive temperature threshold, with parents voicing specific concerns regarding pain and discomfort, and if not treating a child’s temperature would increase the likelihood of seizures. Owing to these concerns, the majority of parents suggested that a permissive temperature threshold of 39.5 °C would be more acceptable than 40 °C. In line with previous research and guidance,18 RWPC was considered to be an appropriate methodology by parents with all stating that they would consent for the use of their child’s information in the study.

Parents’ views on which potential outcome measures they perceived to be the most important were sought. Parents prioritised the following outcomes: (1) long-term morbidity, (2) looking and behaving more normally, (3) length of time on breathing support, (4) time in a PICU and hospital and (5) how quickly vital statistics are back to normal. Parents’ views on outcomes were incorporated into potential outcome measures collected in the pilot RCT. Interestingly, parents rarely prioritised survival as an outcome. This finding was also observed in the recent FiSh feasibility study. 17 As it is surprising that all parents would not prioritise survival as the most important outcome, this finding was further explored with parents of children involved in the pilot RCT.

Fifty-six clinical staff from the four sites involved in the pilot RCT took part in six focus groups. Site staff had concerns regarding the temperature thresholds, suggesting that 37.5 °C for the restrictive temperature group was too low for the administration of antipyretic intervention. Staff also echoed parents’ concerns regarding the permissive temperature threshold and were concerned about not using paracetamol for analgesia in the less unwell, spontaneously breathing patients, who may be in pain or discomfort. Many staff found RWPC acceptable for the proposed trial; however, concerns were raised regarding acceptability to parents of participants randomised to the permissive group.

Overall, the clinical trial protocol was considered acceptable by most parents and staff.

Qualitative study findings were used to develop the pilot RCT protocol, including reducing the higher temperature threshold, narrowing inclusion criteria to require that patients were mechanically ventilated, revising participant information materials and developing a site staff training package. To help address staff concerns about how parents may react to trial discussions, parent perspectives were communicated in site training, highlighting parental acceptability of RWPC, temperature thresholds, parents’ questions and concerns about the study and, importantly, suggestions on how staff could address such questions.

The FEVER observational study

The FEVER observational study, conducted in 22 PICUs from March to August 2017, identified a potentially eligible population of > 3000 cases per year or 10.9 (95% CI 10.3 to 11.5) eligible cases per site per month. The high number of potential patients permitted consideration of adjustments of study design by testing the impact of more stringent inclusion criteria. Importantly, mandating invasive mechanical ventilation rather than including non-invasive ventilation and high-flow humidified oxygen reduced the eligible patient population to around 2000 per year or 7.6 (95% CI 7.1 to 8.1) per site per month.

Wide variation in the temperature thresholds associated with antipyretic interventions was observed both across participants and across sites; however, the majority of critically ill children with a maximum temperature of ≥ 37.5 °C received antipyretic interventions on the first 5 days in a PICU. These findings remained consistent when the inclusion criteria were narrowed to include only mechanically ventilated or invasively mechanically ventilated patients.

The PICU mortality in the potentially eligible population was around 5%, increasing to 5.5% and 6.5% in mechanically ventilated and invasively mechanically ventilated patients, respectively. PICU length of stay and duration of mechanical ventilation and cardiovascular support also increased with narrowing inclusion criteria, whereas the number of days alive and free from a PICU/mechanical ventilation reduced.

Integrated-perspectives study

A total of 60 parents (49 mothers and 11 fathers) of 57 children recruited to the FEVER pilot RCT took part in the integrated-perspectives study, of whom 41 completed a questionnaire, 12 took part in an interview and 7 took part in both parts. Most parents supported the proposed FEVER RCT and thought that the proposed trial was important. No parents referred to the temperature their child reached when discussing trial acceptability. However, parents only viewed the permissive threshold as being acceptable if their child was not in pain or discomfort. Concerns about pain or discomfort were cited as reasons for non-consent and withdrawal among parents of children randomised to the permissive group. Nevertheless, both interview and questionnaire data from parents who declined their child’s participation in different elements of the pilot RCT still supported the conduct of the proposed FEVER RCT. One-third of parents were surprised that their child had been entered into the pilot RCT; however, the vast majority (89%) were satisfied with the RWPC approach and supported its use.

Parents in the integrated-perspectives study prioritised the following outcomes: (1) length of time on breathing support, (2) not in pain and/or discomfort, (3) how quickly vital statistics are back to normal, (4) looking and/or behaving like normal self, (5) length of time in hospital and (6) number of re-admissions to hospital. However, when prompted by the researcher all except one parent stated that survival is the most important outcome measure for the proposed FEVER RCT.

Ninety-eight members of staff at the four FEVER pilot RCT sites took part in the integrated-perspectives study, with half completing an online survey and half attending a focus group. Overall, staff supported the proposed FEVER RCT and thought that the research question was important. Staff found the restrictive threshold of 37.5 °C acceptable; however, only just over half indicated that the permissive threshold of 39.5 °C was acceptable. Staff trained by the study team, which included the chief investigator, were more likely to find the permissive threshold acceptable than staff trained by the trial team or who received ‘secondary’ training by site staff. When questioned, 46% of staff admitted difficulty in adhering to the protocol. The majority of difficulties related to willingness of staff to not treat pain and discomfort, resulting in staff administering paracetamol or actively cooling a child before their temperature had reached the permissive threshold.

Barriers to delivering the definitive FEVER randomised controlled trial

Synthesising the results from the multiple studies identifies a number of barriers to delivering the definitive FEVER RCT, but also informs how these barriers may be overcome (Table 25).

A major concern is the acceptability of the temperature threshold in the permissive group. Both parents and clinicians raised concerns regarding the permissive temperature threshold in the qualitative elements of the study, suggesting that it is only acceptable if the child is not in pain or discomfort. This is likely to have resulted in higher rates of both non-consent and withdrawal in this treatment group and a higher proportion of patients being treated with antipyretics when below the temperature threshold. The findings suggest that this could be addressed in the definitive RCT by restricting the inclusion criteria to only those patients who were receiving invasive mechanical ventilation (comprising 80% of ventilated patients with a temperature of ≥ 37.5 °C in the FEVER observational study and 90% of patients recruited to the FEVER pilot RCT) and ceasing the intervention when beginning to wean the patient from invasive ventilation.

The integrated-perspectives study identified a lack of acceptability of the permissive threshold among some site staff, particularly those not trained directly by the FEVER trial team, which may therefore have stemmed from a lack of understanding of the rationale for the study. This is likely to have been transmitted to parents when seeking consent, contributing to the higher rates of non-consent. This may be addressed through improved training of the site teams, making use of data from the feasibility studies to underpin the variation in usual practice, the importance of the research question and acceptability of the trial to parents. In addition, further resources should be made available for use during secondary training by local site staff, such as a video from the chief investigator explaining the rationale for the study, example study and consent explanations and answers to frequently asked questions.

Despite advice to not approach parents about the study until it was considered an appropriate time, many parents were approached on the first day their child was in a PICU, and the rates of non-consent and withdrawal were much higher when the consent procedure was completed on the first day. The site training (including resources for secondary training) must reinforce the rationale for not approaching the parents too early. In addition, the high visibility of the study within the PICU (e.g. stickers within the bed space) prompted some parents to ask questions to the bedside nurses before the site team had the opportunity to properly explain the study at an appropriate time. The need to ensure that the clinical team are aware of the treatment allocation and adhere to the assigned threshold needs to be balanced against the additional concerns that this visibility may create when the child’s condition has not stabilised.

Non-consent and withdrawal were particular issues in one of the four study sites and, although this may just have been bad luck (e.g. the parents would have not provided/withdrawn consent regardless of the study being conducted and approach to consent), the site team considered that this could have been attributable in part to their local approach to sedation. Tailoring of the site training to specific local needs – in this case, any adjustment required to their usual approach to sedation and analgesia to facilitate delivery of the study – may have helped to mitigate these issues.

A notable proportion (13%) of recruited patients could not be included in the analysis either because of not being able to approach parents/legal representatives to obtain consent or as a result of a complete refusal of consent for the study. Although the first of these situations will reduce the power of the study, the second is of greater concern as increased refusals in the permissive group may introduce a bias in the study results, particularly if these refusals relate to the patient’s ongoing clinical trajectory in the time period between randomisation and the approach for consent. As well as damaging the scientific integrity of the study, excluding participants who have been randomised and received the intervention raises ethical concerns as, for example, if consent was refused following an AE caused by the intervention, then potential safety signals would be missed. Therefore, there is a strong rationale to seek approval for the definitive RCT to use data from the clinical record and other routine sources (including PICANet and death registrations) without explicit consent, while ensuring that parents/legal representatives are fully informed of any processing of data whenever possible and retain the option to opt out. We would seek consent for conducting longer term quality-of-life follow-up with the participant.

Although non-adherence to the higher temperature threshold in the permissive group happened in only 6% of the 6-hourly monitoring time periods, these affected over one-third of patients at some point during their stay. Around half of the non-adherence was owing to paracetamol being used for analgesia among patients receiving non-invasive ventilation or high-flow oxygen, or who were being weaned from invasive mechanical ventilation, and, therefore, the proposed change to include only invasively mechanically ventilated patients would address this. The remaining instances of non-adherence were mainly attributable to misunderstandings of what did or did not constitute ‘external cooling’ or ‘other cooling’, which may be improved by closer attention in site training. Cooling was pragmatically defined in the study as ‘any action taken with the express intention to reduce body temperature’. These measures, which may be as simple as removing a blanket, were also not reliably documented in the patient record, making retrospective collection and validation of study data for these items difficult.

Lack of understanding, and lack of documentation, of cooling measures may also have contributed to the apparently high rate of non-adherence in the restrictive group, in which 41% of patients had at least one 6-hour time period during which their temperature was > 37.5 °C and they did not receive any antipyretic intervention. However, this definition of non-adherence was very strict, with over half (65%) of occurrences of non-adherence being for patients who received paracetamol during one 6-hour time period and then did not receive a subsequent dose during the next 6-hour time period, often as their temperature was already dropping and a number of participants (15%) with a brief, transient spike in temperature. Indeed, only 11 participants (four in the permissive group and seven in the restrictive group) had two or more consecutive periods of non-adherence. The fact that, for most participants in the restrictive group, temperatures dropped below the threshold relatively quickly and the prespecified separation in temperature was achieved indicates that adherence was not as bad as the headline figure of 41% would suggest. For a large trial, it would be necessary to refine the adherence monitoring to identify the most important non-adherence events, for example, when antipyretic interventions were not given for an extended period of time.

Primary outcome measure

With regard to the choice of a primary outcome measure for a definitive RCT, opinions of parents differed between those in the FEVER qualitative study whose experience of a PICU had been some time previously (and who favoured outcomes related to longer-term recovery) and those in the integrated-perspectives study whose experience was more recent (and who favoured outcomes related to time on machines). Combined data on parent-centred outcomes from the qualitative and integrated-perspectives studies suggest that, apart from survival, the following outcomes should be prioritised for measurement in the proposed FEVER RCT: length of time on mechanical ventilation, long-term effects of illness on child (physical and developmental), looking and/or behaving like their normal self, not in discomfort and/or pain, number of days spent in the PICU and hospital, vital signs back to normal and effect on family.

Mortality was 5.0–6.5% across the populations considered, and, therefore, there would not be a sufficient number of events to power a trial to detect a feasible reduction in mortality as a primary outcome. Longer-term functional and developmental outcomes are difficult to assess in this population owing to the wide age range and the pre-existing impairments of many patients. Consequently, the recommended primary outcome for the trial (consistent with the top-prioritised outcome from parents across the qualitative and integrated-perspectives studies) would be either number of days alive and free of mechanical ventilation or time to liberation from mechanical ventilation (accounting for a competing risk of death).

Table 26 illustrates alternative scenarios for the sample size of a definitive trial based on the number of days alive and free of mechanical ventilation. The calculations assume 28-day mortality in the restrictive group of 6% and a mean duration of mechanical ventilation of 7 days among survivors, based on data from patients receiving invasive mechanical ventilation with a temperature of ≥ 37.5 °C in the FEVER observational study. This corresponds to a mean of (28 – 7) × (1 − 0.06) = 19.74 days alive and free of mechanical ventilation. A SD of 8 days was assumed for all calculations, based on the FEVER observational study. This results in required sample sizes of between 750 and 3000 patients, depending on the assumptions regarding effect size and power. Based on an average rate of eligible patients of 8.5 per site per month (from the FEVER observational study, excluding the four smallest sites, which would not participate in a trial), with recruitment of 60% of eligible patients (73% achieved in the FEVER pilot RCT) and allowing for 13% withdrawal (as observed in the FEVER pilot RCT), 10 sites would achieve these sample sizes after recruiting for between 17 and 69 months and 15 sites would achieve these sample sizes after recruiting for between 12 and 46 months.

The FEVER qualitative study

The FEVER qualitative study recruited rapidly and achieved saturation with opinions from both parents and clinical staff. Overall, the findings supported the proposed FEVER RCT including the use of RWPC and were used to inform the methods and conduct of the FEVER pilot RCT. However, the majority of parents were recruited via social media advertising, with no parents being recruited directly from a PICU, and, therefore, the findings may be biased towards parents with a pre-existing interest in research and may not be representative of the population recruited into the trial.

The FEVER observational study

The FEVER observational study had the great strength of over-recruiting sites above the original plan, with 22 out of the 27 UK PICUs participating. No previous prospective study of UK PICUs has achieved this level of national coverage. This provides strong support for the importance of the question of fever thresholds to PICU staff. This study also confirmed that data linkage to PICANet is a highly efficient approach to study design. Despite only recruiting in spring and summer months, this research was able to identify high numbers of patients who would be suitable for inclusion in a clinical trial. However, that fact that data collection took place over a short period of time means that it did not capture the full range of seasonal variability that may happen in emergency admissions, and, therefore, the results may not be fully representative of the population of a future RCT. Finally, although the linked observational study design enabled a large number of patients to be recruited, this made it impossible to distinguish the reasons why particular interventions were being delivered, for example, whether paracetamol was given as an antipyretic or for pain relief, limiting the ability to draw definitive conclusions regarding thresholds for antipyretic management in usual practice.

The FEVER pilot randomised controlled trial

The FEVER pilot RCT was the first attempt to compare fever thresholds in critically ill children globally and was delivered successfully. In particular, the FEVER pilot RCT recruited within a little over two-thirds of the planned recruitment period despite some delays in site approvals. The inclusion and exclusion criteria were easily followed by site staff and a high completeness of outcome data was achieved. The between-group separation in maximum temperature in the first 48 hours of 0.5 °C was greater than many clinicians expected and compares favourably with other studies in this field. 10,41 No safety concerns were identified. Although recruitment ahead of schedule is a strength of the pilot RCT, the fact that the recruitment period was conducted over a different period to the FEVER observational study is an important limitation. Seasonal differences in admissions (i.e. ≈50% of admissions for bronchiolitis in the pilot RCT) may have resulted in differences between the potential trial populations included in the observational and pilot studies, in which 49% of mechanically ventilated patients (population 3) were < 1 year old in the observational study, compared with 64% being 1 year old in the pilot RCT. Therefore, this may misrepresent the true population of a definitive RCT that would recruit year round. However, despite this, the prognostic variables of interest appear to be consistent across the study populations. Another important weakness is the proportion of parents who declined consent to continue in the study, especially in the permissive group. If this were to be reproduced in a full clinical trial, this would threaten the validity of any result. Finally, the relatively high level of non-adherence observed in both groups is a weakness of the pilot RCT, although this may be somewhat mitigated through restricting the inclusion criteria and intervention period to both recruit patients and deliver the intervention only when invasively ventilated.

Integrated-perspectives study

The findings are strengthened by the involvement of parents with direct experience of their child being enrolled into the FEVER pilot RCT, as well as staff involved in recruitment and consent procedures and the clinical care of randomised children. The recruitment targets were either reached or exceeded, involving 60 parents of 57 out of 100 patients. Insight was gained into the views of 8 out of 18 parents (44%) who had declined their child’s continued participation in one or more aspect of the pilot RCT. In particular, the interviews with parents who declined consent and nursing staff who found the protocol challenging to follow provided valuable information to assist with refining the study process for a definitive RCT. However, it is unknown whether or not the predominantly positive views of the declining parents who took part in an interview or questionnaire were shared by other parents who declined the FEVER pilot RCT and also declined participation in the integrated-perspectives study. In addition, the insight into the consent process was limited by the subjective recollections of parents and practitioners rather than insight gained through observations or the analysis of audio-recoded consent discussions. Nevertheless, the findings were strengthened by the integrated-perspectives mixed-methods approach, providing a multiperspective understanding of parent and site staff opinions and experiences of RWPC, specifically in the context of the FEVER pilot RCT.

A definitive trial of permissive versus restrictive temperature thresholds for antipyretic intervention using the FEVER protocol tested here should not be conducted

Our studies identified major concerns with respect to the feasibility of a definitive trial using the same protocol as the FEVER pilot RCT, particularly with regard to the acceptability of the intervention among less sick children and the use of paracetamol as analgesia for patients who were not invasively ventilated. As a result, our recommendation is to not attempt a definitive trial with this protocol.

A definitive trial of permissive versus restrictive temperature thresholds for antipyretic intervention using a modified protocol should be conducted

The thorough, mixed-methods approach to evaluating feasibility conducted here has confirmed both the continuing importance of the research question to parents and to clinical staff and the size of the population with the potential to benefit. Approaches to temperature management vary widely but are largely very restrictive. The unmet need of evidence to guide this very common intervention is therefore confirmed.

We recommend that a definitive trial is undertaken to answer this important question with some key protocol amendments, as outlined previously, that will increase the acceptability of the trial protocol to parents and bedside staff and improve the retention of children in the trial.

The definitive trial should include an internal pilot phase to ensure that these amendments have achieved the desired effect.

Further international work is required to agree important outcome measures and/or develop new outcome measures for clinical trials among critically ill children

Although we were able to identify outcomes that were important to parents, the mixed-methods approach that we undertook highlighted the difficulty of selecting a primary outcome that is measurable, free from bias and with which it would be feasible to detect a plausible treatment effect. We recommend further research to achieve international consensus on a core outcome measurement set42 for paediatric intensive care and, if required, to develop new measures and instruments, or further validate existing measures and instruments, to address any gaps identified during this process.

Highest temperatures

FIGURE 31.

Highest temperature, by 6-hour period and treatment group. (a) Box plot; and (b) median and IQRs.

FIGURE 32.

Distribution of time spent at each temperature, by treatment group. (a) During the first 48 hours; and (b) during the first 6 days.

Antipyretics

FIGURE 34.

Proportion of patients who received antipyretic interventions while receiving invasive mechanical ventilation.

Opiates

FIGURE 35.

Proportion of patients receiving opiates while mechanically ventilated.