A report on parent involvement in planning a randomised controlled trial in neonatology and lactation – insights for current and future research

Stage 1 – informal PPIStructure

The primary researcher (IL) is a neonatal clinician who has listened to parents of preterm babies during clinical work over ten years of practice, with a particular interest in lactation. IL also has personal experience of expressing breastmilk for prolonged periods and extensive involvement in breastfeeding support organisations and peer support networks as a service user and volunteer.

Qualitative literature and published PPI were used as another way to hear the voice of parents of very preterm infants [16, 17, 26, 28,29,30,31,32] and co-produced research prioritisation exercises [33] were consulted. All of these factors inform an understanding of parental experiences, priorities and values that can be used as a springboard for formal PPI planning.

Influence

This informal PPI foundation influenced decisions about the intervention area (relaxation recording), the population (mothers of very preterm babies), the method (RCT) and the broad outcome set (relating to expressed milk volume and breastfeeding experiences).

Stage 2 – online questionnaireStructure

An online questionnaire was advertised through social media channels of Bliss, a large charity supporting families with sick and preterm babies in the United Kingdom. Bliss provides parent and professional facing resources, individual support, national advocacy and supports research. The questionnaire was designed to elicit opinions from parents of preterm infants with experience of lactation on research priorities related to lactation, what outcomes are most important, potential practical and ethical difficulties of a trial and where parents obtain lactation information. All contributors were sent an executive summary of the findings of the questionnaire as feedback. No specific target number of respondents had been pre-specified however the questionnaire was closed after 1 week due to the overwhelming response.

Six hundred and seventy-five people responded to the questionnaire (respondents were not asked their gender but all had experience of lactation). Forty-seven percent had an infant born between 28 and 32 weeks of gestation (“very preterm”). A quarter (24%) had an infant born at less than 28 weeks’ gestation (“extremely preterm”). Half (51%) had provided breastmilk to their infant for 6 months or less. A quarter (23%) had provided breastmilk for more than 1 year. 95% of respondents reported white ethnicity. The sample included 30 mothers from an ethnic minority background and 52 who were under 25 years old at the time of the preterm birth.

Influence – choice of outcomes

The first key influence on the research design was in the choice of outcomes. When asked what questions they had about how to optimally express for and breastfeed a preterm baby, the most common question was how to express more milk and one in five respondents raised this in their free text response (21%). When asked about lactation problems, 27% reported that they had low milk supply and/or could not express as much milk as they wanted. This focus on milk volumes was seen in all groups, including younger mothers and those from ethnic minorities.

Respondents were asked to rank seven pre-defined research outcomes identified through the Stage 1 PPI process. Three outcomes were dominant; being able to express more milk (top priority for 32%), being able to breastfeed for longer (top priority for 25%) and feeling that milk supply was more secure (top priority for 19%).

These results gave a strong direction that the primary outcome of the study should be expressed milk yield, and that duration of breastmilk feeding was an important secondary outcome.

Although milk yield was dominant when asked about problems experienced and questions they would like answered, it was not a dominant response when asked what success or improvement meant in the area of lactation, there was very little consensus on this issue and no single type of response was given by more than 10% of respondents. Table 2 shows the diversity of responses (See Table 2).

Table 2 Examples of the diverse range of answers to questions about what ‘success’ means to parents in relation to breastmilk feeding in prematurity

The stress experienced in relation to lactation was evident from the language used by respondents. Frequently used descriptive words were “struggle” (used 87 times), “problem” (81 times), “difficult” (72 times), “stress” (60 times), “hard” (50 times) and “pressure” (24 times). Descriptions of stress and impact on mental health often revolved around milk volumes and milk supply.

“Constantly anxious about expressing enough milk”.

“[pumping] added to my post-natal trauma and depression. I will never forget how much of a failure. .. I felt.”

“I was heartbroken that I couldn’t do the one thing I thought I should have been able to do!”

This wide variety of issues raised, including mental health, expressing efficiency and duration of breastfeeding, informed the choice of secondary outcomes. As anxiety was the overwhelming emotion expressed in relation to expressing milk it was a clear choice for psychometric assessment. It was clearly important to many families to breastmilk feed beyond discharge, whereas discharge is often used as a surrogate marker of breastmilk feeding success due to the relative ease of measurement. However, there was no agreement on a single definition of successful breastfeeding for preterm babies; decision making around duration encompassed not only limitations experienced (such as milk supply or growth pattern), but also parental perceptions of the baby’s medical status and fragility, previous plans and the conflicts experienced between lactation plans and mental health. Therefore, the research design incorporated established public health goals, any breastmilk and exclusive breastmilk, at time points that were logistically or physiologically relevant. The first timepoint chosen was 36 weeks’ post-menstrual age (4 weeks before the estimated date of delivery), as a standardised timepoint commonly used in neonatal research. The second timepoint chosen was 4 months’ corrected age (4 months after the estimated date of delivery), as a time point representing the likely maximum milk supply required and an age where complementary feeds are indicated in almost all cases for very preterm babies [34]. It was also decided to ask participants for their breastmilk feeding goals at baseline to inform subgroup analysis.

Influence - addressing anxiety over milk yield in study design

A third of respondents (29%) reported at least one concern about the idea of a research study looking at expressing and breastfeeding in preterm babies. The predominant concern was that many mothers have feelings of guilt, judgement, inadequacy and anxiety related to lactation and that being in a study where volumes of expressed milk are monitored and questions asked about breastfeeding and mental health outcomes could exacerbate this, increasing a feeling of pressure or trauma.

Respondents made suggestions to mitigate these risks, which included:

Communicate sensitively, respectfully and tactfully

Ensure that families do not feel pressured over breastmilk outcomes but give gentle encouragement, reassurance, support and empowerment

Be explicit that many mothers have difficulty expressing and breastfeeding and this is ‘ok’

Minimise face to face questions as this could impose more pressure and guilt than filling in details electronically

Ensure mothers feel heard

These responses were used to draft the participant facing documents, using appropriate language to reduce any feeling of pressure and ensure that participants were aware that some people can find it hard to express milk. They also supported the decision to predominantly use a more impersonal electronic interface rather than face-to-face data collection.

Influence – minimising burden for participants

Some respondents reported a concern that the research study would be an extra burden to participants, taking time away from their baby and intruding into their lives at a stressful time. They recommended making the process as easy as possible to minimise this burden. Expressing milk was clearly already a considerable burden that was in direct competition with maternal needs, including factors that might be protective of their mental health such as sleep and time with family, so adding an extra component of yield measurement was a concern.

“. .. often a choice between sleeping and expressing which leads to horrendous guilt when choosing the former and risking health to prioritise the latter”.

“Expressing was just difficult due to the time spent between hospital/home/school run/eating, sleeping and showering!”

In response to these concerns, milk yield monitoring was minimised to three or four time points (depending on gestation at birth), rather than a daily log as used in some studies. Study design was built around electronic data entry, with participants receiving a personalised website link to enter data, in an attempt to make the process as quick and easy as possible. Text message and email reminders were automated for additional efficiency.

A tension exists between the desire for high accuracy assessment of milk yield, using weight, and the burden on participants, where reading the volume of milk from the container would impose less burden than using a weighing scale. A decision was made to prioritise the use of weights for accuracy while also minimising the number of timepoints that participants are asked to do this. However, it remained clear through the trial that weighing milk was difficult for some participants and for future trials, volume reporting may be preferable where medical grade containers with appropriate volume markings are available. It may also be preferable for research staff to measure milk yield rather than participants, although this could interfere with milk provision for the baby unless trained staff are available at all times.

An area of tension between scientific priorities and the burden on parents at a time of heightened stress was the timing of recruitment. PPI contributors raised concerns that recruiting in the first days after preterm birth would cause distress. However, there are strong physiological reasons to suspect that interventions to improve milk supply need to occur as early as possible after birth [35,36,37]. Many neonatal trials recruit participants in the first days after birth [38] and some in the first hours after birth where this is required due to a need for urgent action [39]. A decision was made to allow recruitment up to day four after birth to balance these factors.

Influence – additional data analysis

Contributors had many questions about milk expression, summarised in Table 3 in order of frequency raised. More detail is available in supplementary Table 1 (see Additional file 1). These questions influenced the choice of outcomes for a planned secondary analysis of the data gathered in the RCT, and is also a source of influence for planning future research (See Table 3).

Table 3 Questions for future research on breastmilk expression, from PPI contributors, in order of frequency mentionedInfluence - dissemination plans

The questionnaire asked for details about where parents found information about expressing for and breastfeeding their preterm baby. Multiple sources of information were reported, both formal and informal (Fig. 2). In response to predefined options, the majority (87%) of respondents said that they received information from neonatal unit staff and 40% used printed information found in the neonatal unit. 36% received information from maternity staff, such as a midwife, and 19% used social media for this purpose (See Fig. 2).

Fig. 2figure 2

Respondents’ sources of information on lactation for preterm babies. Multiple options could be chosen from a pre-specified list

Free text responses listed a wide variety of social media channels and websites used for information (more than 50). Respondents used Instagram, Twitter, Facebook, Pinterest, YouTube and apps as well as websites as sources of information.

An initial dissemination plan was made using these findings. This focused predominantly on dissemination to clinical staff, particularly neonatal nurses and neonatal infant feeding leads as the key professionals transmitting information to families. A partnership with national charity Best Beginnings is planned to include the trial results, along with wider information about optimal expressing, on printed cards that can be attached to breast pumps in neonatal units. Best Beginnings is a charity providing accessible resources to parents, including parents of sick and preterm infants. A short video is planned to disseminate findings directly to the public, which will utilise the list of social media settings provided by PPI contributors.

Influence - future research topics

The distinct area of how to successfully transition from expressing (with gavage feeding, where milk is put directly into the stomach via a tube) to direct breastfeeding was frequently raised. Specific questions for future research are listed in Table 4, in order of frequency raised by respondents. More detail is available in supplementary Table 2 (See Additional file 2 and Table 4).

Table 4 Questions for future research on direct breastfeeding in prematurity, from PPI contributors; in order of frequency raisedStage 3 – charity collaboratorStructure

UK national charity Bliss has supported the research project from an early stage. One staff member collaborated on the creation of the funding application, trial protocol and trial documents and supported dissemination of the online questionnaire to parents. A second staff member sits on the Trial Steering Committee. Bliss receives funding for their time, as recommended by an increasing number of research funders.

Influence

Input centred on prioritising the mental health and physical needs of parents throughout recruitment and participation in the study, for example, defining an appropriate response to high scores on mental health questionnaires. Bliss also facilitated the very high response rate to the online questionnaire described in Stage 2, through social media dissemination.

Stage 4 – online panelsStructure

A subset of 49 questionnaire respondents who had given consent for further contact and whose infants were under 32 weeks’ gestation at birth were invited in November 2020 to help review trial documents. This included all respondents who were under 25 years old at the birth of their baby or from ethnic minority groups. A formal role description was provided with the offer of associated payment in line with INVOLVE guidance [40].

Two panels were set up using videoconferencing software. Both were scheduled in an evening slot timed with the hope that contributors’ children might be asleep. Panellists were reassured that they could come and go according to their children’s needs. They were asked in advance whether they had any special requirements to maximise participation and sent documents to read in preparation.

Eleven mothers wanted to be involved in the online panels (22% of those invited) and were sent trial documents. Four contributors attended one of two online panels and two further mothers sent written feedback instead (12% of those invited). Two were from Black and other minority ethnic groups and one was under 25 years old. There was a range of gestational age experiences and one mother had experienced neonatal bereavement.

Initial activities were targeted at putting participants at ease, sharing their background and setting ground rules. Each document was discussed in turn and then pre-specified targeted questions were asked on particular terminology and issues that the researcher had experienced while writing the trial protocol. Finally, there was an opportunity for contributors to comment on the trial design and raise other points. Panellists were offered the opportunity to see final trial documents integrating their suggestions after the process was complete, as feedback.

It was also planned to discuss the documents in person one to one with parents currently in a neonatal unit to focus on demographics that were under-represented in the online questionnaire - for example mothers from non-European countries, teenage mothers and mothers from lower socioeconomic backgrounds. Unfortunately, this was not possible because of the SARS-CoV-2 pandemic restricting non-clinical presence in the hospital.

Influence – language and focus

Panellists’ feedback led to changes in language used and new documents being created, for example, a short introductory leaflet that would be less overwhelming than the full Participant Information Sheet. Free-text questions were added to the participant entered Case Report Forms to make them feel more accessible and encourage a relationship where the trial participants would feel seen as individuals. Resources were added to the end of each Case Report Form directing people to sources of support related to lactation, mental health and having a preterm baby.

Influence – choice of psychometric instrument

Panelists were shown a range of psychometric instruments covering both depression and post-traumatic stress (PTS) reactions. The anxiety instrument had already been chosen (the Spielberger State Trait Anxiety Index) and because a limited number of secondary outcomes could be included within the constraints of the sample size, one further instrument needed to be chosen. There was a clear preference for assessing PTS reactions and for the Post-Traumatic Check List for DSM-5 due to the tone of the language used in the questions. This preference for PTS assessment over depression was because contributors felt that PTS reactions are common after very preterm birth, poorly identified and are a priority for research.

Influence – additional training needs

One panellist focused on the need for specific bereavement training for research staff because of the risk of participants experiencing the loss of their baby, or one of their babies for mothers of multiples. All research staff were recommended to undertake bereavement training and some reported the ways in which they used this in talking to bereaved parents within the trial.

Stage 5 – RCT intervention modification exerciseStructure

All respondents who had given consent for further contact were invited in January 2021 to help choose and modify the relaxation recording to be used as the RCT intervention (approximately 250 invitations). The invitation emphasised that people from minority groups were particularly welcomed. A formal role description was provided with an offer of associated payment in line with INVOLVE guidance [40].

Sixty-two people responded to an invitation to be involved in the intervention modification and choice exercise (25% of those invited). Twelve were selected to include a range of gestational ages at birth and to include people from diverse backgrounds and numbers were limited due to PPI payment costs budgeted. Maternal age ranged from 25 to 48 years. One third of the group came from a minority ethnic background and two spoke English as a second language.

Two draft audio files were sent to contributors, both had been used in previous RCTs [41,42,43]. Contributors were asked for detailed feedback, both open ended and specific and to choose their preferred recording. Once modifications were complete, all contributors were sent feedback on the impact of their involvement.

Influence – intervention choice and modification

One recording was clearly preferred by the majority of contributors. The recording script was modified based on three key areas identified by PPI contributors, reducing a feeling of pressure and judgement on having a low milk supply; language and imagery appropriate for sick, immature babies; and language and imagery appropriate for women with traumatic pregnancy and birth experiences.

An example of contributor feedback (emphasis added):

“The mention of uterus stood out as a trigger word, even now so many years after I had my daughter. I don’t think I would want to be thinking specifically about my uterus as that was linked to trauma. .. There was also a mention of expressing being the number 1 priority, this felt a bit pressured in that moment.”

Summary of influence and impact

Key impacts of PPI for this trial were in defining the study outcomes, minimising participant burden and anxiety through trial design, modifying the intervention and informing a dissemination plan. The key impact for future work is providing research questions identified by parents.

Table 5 describes the nature of the influence of PPI on the research project using the Public and Patient Engagement Evaluation Tool (PPEET [44]) for inspiration. The GRIPP2 reporting checklist for PPI reports [45] is also provided as supplementary material (see Additional file 3 and Table 5).

Table 5 Areas where PPI influenced the research. Inspired by the Public and Patient Engagement Evaluation Tool (PPEET)

留言 (0)

沒有登入
gif