Objectives

The objectives were to seek agreement between parents of children with neurodisability and HPs on which interventions and which outcomes are considered to be the most appropriate for young people with neurodisability and EDSD.

Methods

Study design

As the aim was to establish consensus, an iterative online Delphi survey was employed. A questionnaire was sent to parents and HPs in two rounds; parents and HPs received the same questionnaire in each round. Both survey rounds were open for 3 weeks, with a week between the two rounds to allow for data analysis. In the first round, respondents were asked to rate the importance of individual intervention categories and outcomes for children with neurodisability who have EDSD. In the second round, respondents were shown the results of the first round in the form of bar charts of both parent and HP ratings of the importance of each intervention to be included as part of a package for children with EDSD, and the importance of each outcome. Respondents were then asked to re-rate the importance of each intervention and outcome in the light of the results from round 1. No items were removed from the survey between round 1 and round 2.

As an incentive, respondents were informed that at the end of the second round of the Delphi survey they would be entered into a prize draw to win one of five £100 vouchers available for each stakeholder group (parents and HPs).

Questionnaire development

The round 1 questionnaire was developed by listing the interventions and outcomes identified from the synthesis of evidence gathered through the systematic and mapping reviews (see Chapters 3 and 4, respectively), the national survey (see Chapter 7) and the second round of focus groups (see Chapter 9). The research team developed the structure and format of the questionnaire, drawing on previous experience of Delphi surveys and with reference to methodological recommendations.¹⁰²

The questionnaire comprised (1) demographic characteristics, (2) questions about outcomes that could be used to measure progress in eating, drinking and swallowing and whether or not interventions were effective, and (3) questions about the inclusion of specific interventions that could be delivered at home by parents of young children with neurodisability and EDSD. The majority of questions used fixed-choice rating options, although participants could suggest any additional outcomes or interventions not listed and/or comment through free-text boxes. Respondents were invited to rate how important they thought it was to include each of the 25 discrete interventions as part of a package of treatment for EDSD and their perception of how important they thought each of the 22 outcomes were. Ratings were made on a 9-point scale in which categories were labelled 0–3, ‘not important’; 4–6, ‘important but not essential’; and 7–9, ‘essential’. Respondents were able to tick ‘unable to score’ if they were not able to comment based on their knowledge or experience, or if the item was not relevant or applicable to them. Electronic, web-accessible versions were hosted by Newcastle University using Qualtrics (Provo, UT, USA).

The PAG reviewed the draft survey documents and offered advice about wording and layout. The final online questionnaire was piloted by a small number of parents, HPs and researchers. Respondents were able to use a ‘back’ button to review or change their answers as required and a progress bar told them how far through the questionnaire they were (see Report Supplementary Material 2 for the first round and Report Supplementary Material 3 for the second round).

Procedure

Parents and HPs who took part in the national survey (see Chapter 7) and who had expressed an interest in completing another survey were contacted via e-mail with an invitation to take part in the Delphi survey, and a hyperlink to the online questionnaire. To maximise participation, both respondents and non-respondents from round 1 were invited to take part in round 2.

Analysis

Consensus was conservatively defined as ≥ 67% of both stakeholder groups separately rating that it was essential (rated 7–9 at round 2) to assess an outcome domain or to incorporate intervention in the treatment package.¹⁰²

Results

Response

Invitations to participate were sent to 196 parents and 175 HPs (Figure 11). Of these responses, in each round there were:

FIGURE 11

Flow diagram of Delphi survey recruitment.

• parents –
  • round 1: n = 81/196 (41%)
  • round 2: n = 61/196 (31%)
  • with 52 parents taking part in both rounds.
• HPs –
  • round 1: n = 76/175 (43%)
  • round 2: n = 61/175 (35%)
  • with 51 HPs taking part in both rounds.

The round 1 HPs comprised:

• speech and language therapists (n = 29)
• dietitians (n = 5)
• occupational therapists (n = 20)
• physiotherapists (n = 5)
• clinical psychologist (n = 1)
• paediatricians (n = 12)
• gastroenterologist (n = 1)
• nurses (n = 2)
• other (n = 1).

The round 2 HPs comprised:

• speech and language therapists (n = 28)
• dietitians (n = 3)
• occupational therapists (n = 14)
• physiotherapists (n = 2)
• clinical psychologist (n = 1)
• paediatricians (n = 10)
• gastroenterologist (n = 1)
• nurses (n = 2).

Sample characteristics

Table 20 shows the characteristics of the respondents who completed the first and second rounds of the survey, and the non-respondents. In the first round, the respondents comprised a similar proportion of parents and HPs (49% and 51%, respectively). Respondents were predominantly from England. The majority of respondents were aged 41–50 years (49% of parents; 33% of HPs) and most respondents were female (94% of parents; 92% of HPs). The majority of respondents identified as white British (96% of parents; 92% of HPs). Half of the parents had a child with non-physical EDSD (49%) and half had a child with mixed or physical EDSD (51%). The majority of HPs worked with children with mixed EDSD (75%).

TABLE 20

Characteristics of Delphi survey respondents for rounds 1 and 2, and non-respondents

In the second round, a similar proportion overall of parents and HPs responded (50% and 50%, respectively). Respondents were predominantly from England. The majority of parents were aged 41–50 years (53%) and the majority of HPs were aged 51–60 years (34%). The majority of respondents were female (95% of parents; 94% of HPs) and most identified as white British (97% of parents; 90% of HPs). Around half of the parents had a child with non-physical EDSD (52%) and half had a child with mixed or physical EDSD (48%). Most HPs worked with children with mixed EDSD (74%).

The characteristics of respondents who completed both rounds of the Delphi survey and those who completed the second round only were very similar apart from age, for which the majority of those completing both rounds were aged 41–50 years (46%), whereas the majority of those completing the second round only were aged 31–40 years (63%). See Appendix 14 for full details of the characteristics of respondents who completed both rounds of the Delphi survey and those who completed round 2 only.

Interventions

Table 21 shows the proportion of parents and HPs who rated interventions as essential in the first and second round of the Delphi survey. The percentage of respondents giving a rating of ‘essential’ (score of 7–9) increased from round 1 to round 2 for most interventions, for both parents and HPs. Consensus was achieved for 17 of the 25 interventions at round 1, increasing to 19 out of the 25 interventions at round 2. There was substantial agreement between the ratings of participants who completed both rounds of the survey and the ratings of participants who completed round 2 only (data not shown).

TABLE 21

Parents’ and HPs’ rating of interventions as essential in rounds 1 and 2 of the Delphi survey

Figure 12 shows the interventions that were viewed as essential by ≥ 67% of parents and HPs in the second round of the Delphi survey (see Appendix 15 for the interventions viewed as essential by ≥ 67% of parents and HPs in the first round). In the second round, parents and HPs identified that a large number of interventions (n = 19) were an ‘essential’ part of an intervention package for young children with neurodisability and EDSD. These comprised interventions directly targeting EDSD, such as Modifications (Environment, Equipment, Food or drink, Placement of food and Positioning), improved mealtime communication (Enhancing communication strategies, Responding to the child’s cues for feeding and Pace of feeding) and desensitisation strategies (Graded exposure to foods or textures and Oral and sensory desensitisation). Other interventions not specifically targeting the child’s EDSD were also viewed as essential, such as Sharing information about the nature of the child’s difficulties, Medication and Parental support. None of the 25 interventions was viewed as ‘not important’ to include in an intervention package by either parents or HPs. See Appendices 16 and 17 for ratings of ‘not important’, ‘important but not essential’ and ‘essential’ for interventions by parents and HPs in the first and second rounds of the Delphi survey.

FIGURE 12

Interventions viewed as essential by ≥ 67% of parents and HPs in round 2 of the Delphi survey.

Outcomes

Table 22 shows parents’ and HPs’ agreement on outcomes rated as essential in the first and second rounds of the Delphi survey. In keeping with the findings about interventions, the percentage of respondents rating the outcome as ‘essential’ typically increased between the first and the second rounds. Figure 13 shows the outcomes viewed as essential by ≥ 67% of both parents and HPs in the first and second rounds of the Delphi survey. These are presented together given that the items for which there was consensus did not change between rounds. The 10 outcomes viewed as essential included child-focused outcomes, such as Physical health, Safety, Oral motor control and Quality of life. They also included family-focused outcomes, such as Quality of life of family, Parents’ understanding of the child’s difficulties and Parents’ mental health. None of the 22 outcomes was viewed as ‘not important’ by parents or HPs. See Appendices 18 and 19 for ratings of ‘not important’, ‘important but not essential’ and ‘essential’ on outcomes by parents and HPs in the first and second rounds of the Delphi survey.

TABLE 22

Parents’ and HPs’ agreement on outcomes rated as essential in rounds 1 and 2 of the Delphi survey

FIGURE 13

Outcomes viewed as essential by over 67% of parents and HPs in round 2 of the Delphi survey.

Summary of the Delphi survey findings

The Delphi survey established consensus between parents and HPs on the ‘essential’ interventions to include in an intervention package for young children with neurodisability and EDSD, and the important outcomes. The Delphi survey showed that for both parents and HPs 19 interventions were viewed as ‘essential’ to include in an intervention package and 10 outcomes were viewed as ‘essential’; these focused on both the child and the wider family. Schedule of meals and Visual supports were viewed as ‘essential’ interventions for inclusion in an intervention package by HPs only, and Oral motor exercises and Psychological support for children were viewed as ‘essential’ interventions to include by parents only. Training to self-feed and Changing behaviour at mealtimes were not viewed as ‘essential’ to include in an intervention package. None of the interventions or outcomes included in the Delphi survey was viewed as ‘not important’ to include in an intervention package, reflecting the complexity of neurodisability and highlighting the need to tailor interventions to meet the needs of individual children and their families.

Strengths and limitations of the Delphi survey

We used a wide range of recruitment sources that were also used in the national survey (see Chapter 7), and a reasonable number of respondents to that survey showed willingness to contribute further. This meant that we had information to enable a comparison of characteristics of those who participated and those who did not, and, therefore, the Delphi survey invitation was sent only to previous survey respondents. The overall response for the surveys was typical and acceptable (≈ 40%) for the Delphi approach, and there was little difference in the characteristics of respondents between round 1 and round 2 (see Appendix 14). The response ratings about ‘essential’ interventions and outcomes were very high; therefore, a higher response would have been unlikely to lead to different results. Our consensus definition of 67% was conservative; however, most interventions and outcomes were rated ‘essential’ by a much higher proportion than this. Nonetheless, if we had used a higher percentage to define consensus, our findings would have differed slightly. Contacting non-respondents from round 1 in round 2 increased the responses in round 2 and, therefore, improved precision; although this is not always carried out in Delphi surveys, there was substantial agreement between the ratings from participants who completed both rounds of the survey and those who completed round 2 only.

Patient and public involvement in the Delphi survey

The Delphi survey questionnaires and information sheets were developed by the research team, including the parent co-investigators. These draft documents were reviewed by the PAG who offered advice around content, wording and layout. These documents were further amended to take account of the discussions in the second round of focus groups and these amended versions were reviewed again by the PAG. The PAG advised on the ordering of statements relating to interventions and outcomes for the Delphi survey to ensure that these were presented in an order that made sense to the reader.

How did the Delphi survey findings inform the next study stage?

The Delphi survey findings informed decisions around which interventions and outcomes to take forward for further discussion at the stakeholder consultation workshops (see Chapter 11), and for consideration when designing future research and pragmatic trials.

The Delphi survey identified that a large number of interventions were viewed as essential to consider for improving EDSD in young children with neurodisability. The research team used this information and that gathered previously to make pragmatic decisions about which interventions to include in the toolkit so that they could be selected for use depending on the child and family’s individual needs.

The Delphi survey also showed that a large number of outcomes were valued by parents and HPs, focusing both on the child and the wider family. These essential outcomes were presented at the stakeholder consultation workshops (see Chapter 11) to facilitate further discussion regarding the important primary and secondary outcomes for use in future research and trials.