Introduction

Deaf children are most often born to hearing families [1] who have not yet experienced deafness and therefore require support with adapting their communication skills [2]. In the current study, ‘parent’ refers to the mother, father, or any primary caregiver and ‘deaf’ refers to all deaf and hard of hearing children, identified with a mild, moderate, severe, and/or profound level of deafness.

The quality of parent-child interaction (PCI) predicts deaf children´s language development [3,4]. A recent systematic review reported that the quality of parents’ linguistic input explained 31% of the variance in deaf children’s spoken language scores [5]. In many studies, deaf children’s spoken language development is 1–1.5 standard deviations lower than hearing peers [6]. Children with severe and profound degrees of deafness are the most delayed [7]. Much research suggests that communication is one of the major sources of stress when parenting a deaf child [8,9]. There is an internationally accepted guideline of working with families of deaf babies, known as ‘1-3-6 model’ where hearing screening occurs before the infant is 1 month old, a deafness diagnosis is complete by 3 months, and intervention is provided by 6 months [10]. Professionals (such as Teachers of Deaf children (ToDs), Speech and Language Therapists (SLTs), and Psychologists) maximise on this early start by offering family-centred support, education, and coaching. Professionals might suggest adaptations to a parent’s communication approach that includes greater focus on gaining the attention of the deaf infant, or on maintaining the joint attention between parent and child [11]. A parent may need support with ensuring their interactions are accessible and perceivable by the child; this might include the introduction of a signed language. There is evidence that by providing parents with the knowledge, skills and practice they need to enhance their communicative behaviours, positive changes in deaf children’s communication are observed [11,12].

Guidance for professionals working within early intervention programs for deaf children recommends the provision of tailored, individualised support to families [13]. An important first step to being able to offer this support is comprehensive assessment. Whilst there are assessments to track the deaf child’s communication skills, an assessment tool to appraise a parent’s strengths and needs in communicating with their deaf child does not yet exist. A recent systematic review of 61 international studies [14] identified which parent behaviours and assessment methods are most frequently used by researchers of PCI with deaf children aged 0–3 years. The parent skills identified were attention getting, joint engagement, parental sensitivity, and facilitation techniques that either enriched the deaf child’s language or ensured good access to the parent’s language input (ibid). In addition, the systemic review reported that many of these parent behaviours were associated with positive child language outcomes. Researchers tended to film mothers (rather than fathers), in labs for an average of 19 minutes. These videos were analysed frame by frame, with in-depth and time-consuming coding systems. Some of the studies used scales for measuring PCI instead of, or in addition to, coding. It is important to know the evidence-based parent behaviours positively associated with deaf children’s language development, however it is not straightforward to use these methods in clinical practice.

A recent survey of 190 UK-based professionals working with deaf infants aged 0–3 years [15] reported that PCI assessments overlapped with parent skills highlighted in the aforementioned systematic review [14]. In addition, a further 18 practice-based parent behaviours not reported in the systematic review were identified. In contrast to researchers, however, professionals often observed PCI live and made mental or paper-based notes, using their own knowledge and skills to analyse the interactions. Most professionals also stated that their assessments always or often led to collaborative goal planning with parents. In a sequential, focus group study, professionals shared how complex and multi-faceted assessing PCI in families’ homes can be [16]. Without any published guidance on how to conduct PCI assessments, it is difficult for professionals to know which parent behaviours are core, which to prioritise within their PCI observations, and which assessment approaches are best particularly when working within busy, complex family systems. Additionally, professionals do not have a formal way of showcasing progress to parents (or managers) through re-assessment. Therefore, an e-Delphi consensus study was designed to refine and gain expert international consensus on perceived best practice for an assessment of PCI where the child is deaf and aged 0–3 years. The objectives were to: 1) reach consensus on which parent behaviours to include in a new assessment and 2) develop best practice recommendations on how to approach assessing these behaviours.

The Delphi technique is an anonymous, iterative, and multistage method to synthesise expert opinion into a group consensus. It is carried out using questionnaires and feedback over a series of rounds [17]. The anonymity of participants aims to avoid a small number of experts dominating the discussion or the peer pressure to conform, and instead means full participation from all participants involved [18]. Delphi studies are used in health sciences to identify priorities [19], in policy making and developing practice guidelines [20], and where ‘best practice’ agreement is desired [21]. The current e-Delphi builds on previous research by the same authors [14–16].

Methods

Ethical approval was granted from City, University of London’s School of Health & Psychological Sciences Research Ethics Committee (ETH2122-0790). The reporting guideline for Delphi studies [22] has been used.

Feedback, review and round 2

Following round 1, participants received a report that summarised the 28 statements that had achieved consensus and the 41 statements that had not, showing the quantitative, group response per item. The report explained the reduction of items in the second e-Delphi round, i.e., the 28 statements that were not going to be presented in round 2 because participants had agreed they were essential and of highest importance, and therefore would form part of the proposed assessment tool. There were no eliminated or discarded statements, i.e., where 80% of participants rated a statement as a 1 (‘no importance’) or 2 (‘low importance’).

Statements that did not achieve high or low consensus in round 1 were then reviewed, with their wording scrutinised carefully to ensure that ambiguity was not a possible reason for items not reaching consensus (see data analysis and PPI sections below). Modifications to statements were made and re-presented for rating in round 2, as seen in other studies [22,30,31]. As before, optional open text boxes were offered for participants to justify answers. Fig 1 presents a flow chart illustrating the stages and outcomes of the Delphi process.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flow chart illustrating the stages of the e-Delphi process.

Note the denominator of total number of PBs and AAs changes from the ‘Preparatory Phase’ to the final ‘Included’ phase as there were some additional statements created in the ‘Review of R1’ phase.

https://doi.org/10.1371/journal.pone.0301722.g001

Results

Participants

Table 1 presents participant characteristics in terms of primary profession (many had dual roles), location, hearing status, and years of experience. Eighty-three (83) of the 95 registrants completed round 1. There was a relatively equal number of SLTs (n = 30) and academics (n = 28), with fewer QToDs (n = 22), NDCAMHS professionals (n = 2) and Psychologists (n = 1). Professionals were from thirteen different countries, with most participants contributing were from England (n = 50) and then the USA (n = 12). Most professionals were hearing (n = 78, 93%). On average, each panel member had nearly 20 years’ experience of working with deaf children. Seventy-two (72) of the 83 panel members were happy for their contribution to the study (and the assessment tool’s development) to be acknowledged and their names are listed in the acknowledgements section.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Participant characteristics.

https://doi.org/10.1371/journal.pone.0301722.t001

Round 1

Table 2 lists the statements where consensus was reached in round 1, their levels of consensus, the category the statement is in, and each statement’s median and interquartile range (IQR).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Round 1 ‐ consensus met (i.e., rated a ‘4’ or a ‘5’ by ≥ 80%).

https://doi.org/10.1371/journal.pone.0301722.t002

Parent behaviours.

In this first round, 21 of the 40 (53%) PB statements reached consensus, and percentages of agreement ranged from 80–99%, with IQRs <1. All eight statements within the parental sensitivity category gained consensus in this round, with seven of the eight statements in the top ten for highest percentages of agreement (94–99%). A parent being attuned or emotionally sensitive to the child’s emotions and behaviour had the highest figure of agreement overall (99%) with a parent being genuinely interested and involved also highly rated (98%).

All four statements related to joint engagement also reached consensus in round 1. Experts agreed on the importance of balanced turn taking (98%), maintaining joint engagement (98%), on the parent waiting for the child to initiate (94%), and the parent and child face watching one another (84%).

The remaining nine statements that reached consensus were linked to how a parent can increase a child’s access to language (1 of 4 statements achieved consensus) and how a parent can enrich their deaf child’s language (8 out of 19 statements achieved consensus). Experts agreed that a parent expanding on their child’s language (91%) and the parent interpreting the child’s behaviours with language (84%) were some of the most important skills to assess in PCI. Likewise, for access to language, 83% of participants felt it was important to assess the parent’s skills in physically positioning themselves at the child’s level.

There were 19 parent behaviours that did not reach consensus. None of the five statements on ‘attention getting’ reached consensus in round 1, and many participants gave comments as to why the statements did not work in their current form. Participants shared that a parent using auditory based attention getting strategies with an unaided deaf child would not be effective, and likewise a hearing parent using visual or tactile attention getting strategies might not be the most effective way of gaining the attention of a deaf child who is a good user of their hearing equipment. In addition, there were 3 parent behaviours related to accessible language and a further 11 parent behaviours linked to language enrichment that did not reach consensus.

Approaches in assessment.

Seven of the 26 (27%) statements related to assessment approach reached consensus in round 1. Participants agreed all four statements linked to goal setting were important and essential. These stated that assessments should lead to goal planning (80%), goals should be jointly discussed and agreed between the parent and professional (96%), the purpose of the parent-focused goal and impact on the child should be discussed between the parent and professional (98%) and parent-focused goals should be regularly reviewed and updated (96%).

Two of the four statements linked to developing an informed parent gained consensus. These were a parent being made aware of the purpose of the assessment (96%) and the parent having the opportunity to watch back and review a video of their PCI with a professional (91%). Lastly, one of the four statements on developing an empowered parent reached consensus. This referred to providing parents with the opportunity to assess themselves with encouraging support from a professional (88%).

Consensus was not reached on the 19 assessment approaches. These were three statements on setting up the assessment (n = 3), how to measure skills (n = 4), collaborating with parents (n = 3), multi-professional joint working (n = 2), cultural diversity (n = 1), and deaf infants with additional needs (n = 1).

Review and rewording.

Across both sections of round 1, there were no statements that were eliminated or discarded, i.e., where 80% of participants rated a statement as a 1 (‘no importance’) or 2 (‘low importance’). Therefore, all 19 PB statements and 19 AA statements were reviewed by the authorship team and the PPI group. There were seven PBs and six AAs where more divergent responses were present (i.e., IQRs >1). These are marked with an asterisk in the tables below. All of the statements, but particularly those with dissenting judgements, were carefully scrutinised using participants’ qualitative responses to aid the rewording of items. The review and rewording process resulted in less divergence in round 2. S1 Table displays the changes in wording for all 38 statements from their original round 1 presentation to how they were presented to participants for round 2.

Within the parent behaviours section, the main changes included creating different categories linked to child language use rather than parent skill. Participants had made it clear in their qualitative feedback in round 1 that some PBs were dependent on the language choice made by parents. For example, attention getting statements PB1, PB3, PB4 and PB5 were grouped together, and participants were informed that these would be most relevant where the deaf child benefitted from access to lip patterns and facial expressions. Similarly, attention getting statement PB2 and access to language statements PB21, PB22 and PB23 were grouped together, and participants were informed that these skills would be most relevant for deaf children who have access to sound and the potential to understand and use spoken language. Following the review of participants’ comments, it was decided that the remaining 11 reworded statements were relevant for all deaf children, regardless of listening skills and language choices. Two prominent reframes of statements within this section were PB30 ‘Parent reduces questions and increases on-topic comments’ and PB34 ‘Parent uses open questions’. Participants’ feedback had the following themes: PB30 said two things at once; it is impossible for parents not to ask questions; a professional might only want a parent to reduce questions if there is overuse, not all question types negatively impact language use; it is the balance between parent and child that is important, and balance between questions and comments is also key for parents. As they are linked, feedback on PB34 was similar but participants also made the case that open questions needed to be asked when the child was developmentally ready. It was decided that PB30 would be reworded and focused on question type: ‘Parent uses open questions in favour of closed questions’ and that PB34 would be more about the balance of a parent’s language use: ‘Parent balances open questioning with on-topic comments’. Disclaimers about the child’s developmental stages of cognition and language were added to PB28, PB29, PB32, and PB39. Lastly, one new parent behaviour was added to the ‘access to language’ category and this was about the parent’s use of child-directed speech or sign to pique the child’s interest. Therefore, for round 2, there were 20 parent behaviours to rate.

Within approaches to assessment, the categories and ‘essence’ of each statement mostly stayed the same, but statements were expanded to provide more clarification or better examples. Other adaptations included using softer, conditional terms such as ‘where appropriate’ and ‘could’ instead of ‘should’. This was to acknowledge participants’ feedback that professionals work with a highly heterogenous population and approaches in assessment would be context-bound. One statement that underwent extensive review was AA61, which related to families whose home language was not the officially recognised native language of the country in which they resided. This statement was split into three to gain consensus on whether: families should be observed in their home language (AA61a); professionals should acknowledge the culture of the family to prevent misinterpreting PCI assessment results (AA61b); and even where there may be cultural differences at play, all families could benefit from adapting their communication behaviours if unhelpful for the language development of the deaf child (AA61c). The team also added an optional question related to AA61abc with an open-text response to gain information on how participants have worked successfully in this area. From review of round 1 feedback and the preliminary work [14–16], it was decided that an additional free text box would also be added to AA50 (related to deaf-plus infants), as more expert insight would be beneficial for working with this population. Therefore, for round 2, there were two free text boxes and 21 statements on approaches in assessment.

Round 2

For round 2, 81 of the 83 (98%) participants returned. Table 3 lists the statements that reached consensus in round 2, the category the statement is in, the percentages of agreement, and each statement’s median and interquartile range (IQR). Consensus data from round 1 is also displayed, along with the change in percentage of agreement per item across the two rounds.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Round 2 –reworded statements achieving consensus (n = 24).

https://doi.org/10.1371/journal.pone.0301722.t003

Parent behaviours.

In the second and final round, a further 13 of the 20 (65%) reworded parent behaviour statements reached consensus, with agreement ranging from 80–95% and an IQR<1. This included six of the seven previously divergent PBs (i.e., IQR>1). Seven out of the 13 remaining statements within the language enrichment category, three out of the five remaining statements from the attention getting category, and three out of four statements remaining on access to language reached consensus. In this round, PB12 (‘parent waits and watches their child’s behaviours and gaze, using the child’s cues to tailor the language they will use’) gained the highest level of consensus (95%). This was a 23% increase from its agreement figure in round 1 and the only parent behaviour statement in round 2 that reached a consensus above 90%.

Many of the reworded parent behaviours discussed in the ‘Review and Rewording’ section above reached consensus. These were five of the seven statements re-grouped by child language (PB1, PB5, PB21, PB22, PB23), the new entry on child-directed language (‘PBnew’) and the reframed statements on questions and comments (PB30 and PB34).

Two of the largest increases in percentage agreement were for PB39 (‘Parent provides language input that is appropriate to child’s developmental stage’) and PB7 (‘Parent watches and waits when the child looks away, to allow the child to explore, to take a rest from interacting, to take a turn in initiating’). PB39 moved from 31% agreement (round 1) to 88% agreement (round 2), an increase of 57%, two points on the scale in terms of median score and had less divergence (IQR = 2 to IQR = 1). PB7 gained 28% of participants’ agreement (round 1) but achieved 83% agreement (round 2). The smallest change observed between round 1 and round 2, was PB28 (‘Where appropriate, parent uses mental state verbs i.e., ’like’, ’know’, ’think’). In round 1, the statement achieved 77% (median 4, IQR 4–5) and in round 2, the statement achieved 81% (median 4, IQR 4–5). Due to the slight decrease in the number of expert participants between round 1 and round 2, this is likely to mean that one more person rated the statement a 4 or 5.

There were seven out of 20 (35%) reworded parent behaviour statements that did not achieve consensus. Table 4 displays these ratings and the observable changes from round 1 to round 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. No consensus reached following rounds 1 and 2 (n = 17 statements).

https://doi.org/10.1371/journal.pone.0301722.t004

Five of these statements had an increase in the percentage of agreement between participants, with changes in agreement ranging from 6 to 29%. Two statements decreased in their agreement percentages, i.e., their agreement level lowered following rewording. These were PB4 (parent uses multiple strategies at one time to gain a child’s attention) and PB40 (parent uses touch as a tactile way of highlighting speech, tone, or rhythm in language). For PB4 (divergent in round 1 but not in round 2), participants argued that one mode at a time would be enough–auditory strategies, visual strategies, or tactile strategies. This would allow a parent to monitor which approach in attention-getting the child responds to best, so that the approach is child-led, and the child is not overwhelmed or over-stimulated by multiple methods. For PB40, many participants thought this was quite unnatural and intrusive for PCI and play, but rather a helpful strategy in word learning and/or syllable counting. Participants also felt that this strategy may be more appropriate for a therapist or teacher to use in direct work, because parents should not feel they need to be teachers during interactions with their infants.

A closer look at the low agreement statements highlighted two parent behaviour statements within the language enrichment category which were very near to achieving consensus: PB29 (parent talks ahead of actions or events; 77%), and PB32 (parent rephrases a child’s incorrect grammar; 76%). Both statements had increased in their percentage agreement figures (PB29 by 12% and PB32 by 7%) and in their spread of opinion when compared to round 1 results (IQRs of 3–4 in round 1 and IQRs of 4–5 in round 2). Qualitative feedback for PB29 was either that is not always possible to pre-warn a child of a change and that this parent behaviour is dependent on the age and stage of the child. For PB32, participants were keen to note that PCI should not include corrective grammar lessons as this can impact bonding and self-esteem. The remaining five parent behaviour statements that did not achieve consensus (PB4, PB3, PB33, PB37, and PB40) were considerably lower in percentage agreement (ranging from 62–13%) with a substantial gap from the statements bordering close to consensus, suggesting clear views from experts that the remaining five were less essential and less important items. Further, PB37 became divergent in round 2, i.e., an IQR>1. Through discussion, across the authorship team and in PPI meetings, it was agreed that statements PB29 and PB32 would be retained in the final list of statements and reworded to include caveats noted by participants in their qualitative feedback (see Table 5). Conversely the other five parent behaviours were not core and not included.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Reworded statements following the second and final e-Delphi round (n = 6).

https://doi.org/10.1371/journal.pone.0301722.t005

Approaches in assessment.

Eleven of the 21 (52%) reworded statements related to assessment approach reached consensus in round 2. This included three of the six previously divergent AAs (i.e., IQR>1). These were all three statements on cultural diversity (AA61a,b,c), the statement on deaf children with additional needs (AA50), the remaining two statements on developing an informed parent (AA43, AA56), the remaining three statements on empowering the parent (AA52, AA64, AA45), one of the three statements on collaborating with parents (AA48), and one of the two statements on multi-professional joint working (AA60).

The four statements with the highest percentages of agreement were also the four statements with the highest changes in agreement scores between round 1 and round 2. AA60 on joint working (professionals sharing information) had the highest number of agreement (95%) with a 37% increase from round 1 (58%). In their comments, participants agreed that with consent, information sharing is essential for reducing stress for families. The next statements with high levels on consensus were two items on cultural diversity. First AA61b (acknowledging the culture of the family) achieved 94% consensus and AA61a (observing PCI in the home language) achieved 93%, their changes in percentage agreement were 29% and 28% respectively. For these items, participants suggested that it would be best to involve professionals that share the culture of the family as they can inform the assessment, and make families feel more confident and comfortable using their own language. This would increase the validity of the observation. Some participants suggested that the PCI assessment could be done in the home language, and then if the parent speaks English, they could translate for the professional upon watching back. Another important point for professionals unable to work with bilingual co-workers or cultural brokers was that whilst it is important professionals understand the culture of the family, they should not make any general assumptions as variations exist. All participants’ responses to the open-text question on cultural diversity are provided in S2 File.

Another statement that achieved high levels of consensus was AA50 (flexible observations when working with deaf-plus children and their families), achieving 90% agreement which was an increase of 27% and one scale point from the results for this statement in round 1 (63%). Participants commented that parents could send their own videos to the professional as this would mean there was less pressure for the parent and child to ‘perform’ on the visit, or that the professional could observe highly interactive moments as well as moments where communication was more limited, where parents may need more support and guidance. Participants’ responses for the open-text question on working with the deaf-plus population are provided in S1 File.

There were 10 out of 21 (48%) reworded statements on assessment approach that did not gain consensus. These are displayed in Table 4. These were all three statements on setting up the assessment, all four statements on how to measure skills two statements on collaborating with parents, and one related to multi-professional joint working. For all of these statements, more participants increased their rating between round 1 and round 2, with changes in percentage of agreement figures ranging between 5 to 26%.

As with parent behaviours, statements that did not reach consensus within assessment approaches were reviewed. There were four statements that were very close to the consensus threshold of 80%, The first was AA49 (observing daily routines) with 78%. The percent agreed, median and IQRs for this statement remained high and stable between round 1 and 2. We did not gain much qualitative feedback from participants who rated this statement as a 1, 2, or 3 on the scale, other than this was harder to achieve for professionals working in clinical settings. Our PPI group of hearing parents felt this was an essential statement to include as so much communication centres around daily routines such as mealtime, feeding, bath time, nappy changing, and dressing, therefore a slight reword was made and this statement retained.

Statement AA47 (assessing in natural, familiar settings) had a percentage agreement of 76%. Statements AA44 (video recording length) and AA57 (giving all primary caregivers the opportunity to have PCI assessed) both achieved 75%. For all three statements, the percentage agreement increase was between 7–25% and all IQRs were narrower/less divergent in round 2 results. These statements were viewed as essential by both the authorship team and the PPI group and were reworded (see Table 5). Participant feedback was used to modify statement AA47 in order to give a parent more power, choice, and convenience when deciding on optimal settings for interaction. Similarly for AA57, participants shared this would be great but unrealistic with busy caseloads and hence ‘should’ was replaced with ‘could’. For statement AA44, many participants felt that three to five minutes was adequate and that 10 minutes of video recording was excessive. When these comments were presented to our PPI panel, parents of deaf children felt strongly that ‘up to 10 minutes’ should be the wording as they felt that a parent might need more than three to five minutes ‘to warm up’ and ignore the camera/the professional but agreed that more than 10 minutes of video recording would be excessive. Most importantly, parents felt video recording length should be discussed and agreed upon before the filming had begun. The remaining statements on approaches to assessment (AA42, AA41, AA54, AA58, AA55, AA59) were considerably lower in percentage agreement (ranging from 66–55%). AA54 and AA59 increased in divergence. Whilst they were not included in the final list of statements, it was noted that the majority of participants (i.e., over 50%) agreed they were important. Four of these statements were about measuring PCI, one about including siblings, and another about working with deaf adults.

Discussion

This study gained expert opinion on the core content and principles of a new assessment for PCI where the child is deaf and aged 0–3 years. Statements reviewed in the e-Delphi were based on a systematic review [14] and a studies of professional practice [15,16]. In addition, co-production work with a PPI team (17 hearing and deaf parents and professionals) was embedded throughout each study phase. Eighty-three experts (SLTs, QToDs, NDCAMHS professionals, Psychologists and academics) working internationally with deaf infants and their families agreed on the importance of 52 statements out of a possible 69 through two rounds of voting and feedback. A further six statements were included following data analysis and discussion between the authorship team and PPI group. Experts joined from all four nations of the United Kingdom, the USA and Canada, Malaysia and six European countries. On average, each expert participant had 20 years of experience in the field. Across the two rounds, there was excellent retention of expert participants, suggesting a good estimation of the degree of consensus in the final results (i.e., minority opinions did not leave the study).

The first objective of the study was to reach consensus on which parent behaviours to include in an assessment. All statements related to parental sensitivity achieved high levels of consensus. The importance of parental sensitivity in PCI with deaf infants is well evidenced [2,32,33]. In their large-scale, longitudinal study of 285 deaf children with cochlear implants, Cruz and team [34] found deaf children had 1.52 years less of a language delay when their parents had above average skills in maternal sensitivity and language stimulation. Pressman and colleagues [35] found that maternal sensitivity was not correlated with initial child language scores, but correlations were present in the follow up assessments 12 months later. They calculated that maternal sensitivity predicted expressive language and had a larger positive effect on the sample of deaf children compared to their hearing sample (ibid). It is appropriate therefore, that the top five behaviours with the highest agreement were related to the parent being attuned and adaptive to the child’s emotions and behaviour, the parent having a genuine interest in their child, and the parent responding to the child with on-topic behaviour or language.

All four statements on joint engagement also reached consensus, with balanced communicative turn taking, and maintaining joint engagement nearing total agreement. Many studies have shown the positive relationship between deaf children’s language scores and time spent in co-ordinated (or mutual) joint engagement between parent and child [2,36,37], i.e., the longer a parent and child are engaged, the better the child’s language will be over time. A particular behaviour linked to more successful instances of joint engagement was non-intrusiveness, i.e., following the child’s interest rather than directing them [2,37]. Parental sensitivity and initiating and maintaining joint engagement are therefore intertwined and reliant one another: a parent who is interested and attuned to their child’s behaviour, who shares the same interest as their child, will offer contingent comments and behaviours, and the dyad will be mutually engaged. The parent must also be aware of the balance in engagement, the space for the child to ‘take a turn’, verbally or non-verbally, in order to maintain the mutual participation of both child and parent. The longer the connection, the more bonding, enjoyment, and opportunities for language there can be. The fact that all sensitivity and joint engagement behaviours gained the highest levels of consensus fits with the existing literature; joint engagement and parental sensitivity are important to observe within PCI assessment.

Whilst most parent behaviours were viewed by experts as beneficial for developing spoken and/or sign languages, there were some that diverged across participants because of language choices made by parents. For example, a parent’s volume of voice or proximity to the child’s listening device would be more relevant for a deaf child who relies on their listening devices and is responding to and developing spoken language, than for a child who is developing sign language. Similarly, methods used to gain the child’s attention may differ. For children who are responding to and developing sign language, and/or benefiting from lip patterns and facial expressions, a visual way of gaining attention may be more effective. Parents of deaf children make more frequent use of visual and tactile strategies to gain their child’s attention when compared to hearing parents of hearing children [38]. In this current e-Delphi, tactile strategies for gaining attention or using multiple modes of attention getting at one time did not reach consensus. Participants chose PBs that were child-led, to see which methods (visual or auditory) the child responded to first. Rather than tactile strategies, the more passive attempts at getting a deaf child’s attention [39–43] were favoured by the e-Delphi’s expert panel. These include behaviours such as the parent actively waiting or pausing their communication until the child looks. A parent who displays this behaviour may well be following the more ‘non-intrusive’ patterns of gaining and maintaining attention mentioned above, where the parent follows the child’s lead and pace.

Many of the statements that achieved high levels of agreement were centred on the parent’s response to their deaf child, rather than the parent instigating anything new themselves. Expert participants seemed to value the statements that required a parent to step back, observe, receive the child’s action or utterance, and then respond. Statements achieving high agreement were the ones that promoted balance and a respect for the child’s place within the interaction, a respect for their gaze, actions, interests, and contributions. This responsive, child-centred ethos in interacting with deaf infants aged 0–3 years may also explain why participants did not highly rate behaviours such as PB33 (rephrasing grammar) or PB37 (modelling or highlighting mistakes in parent’s language). As many participants noted, the teaching of grammar, vocabulary, or speech production is not the job of a parent of a deaf child, particularly not at this age, and therefore it is not a vital behaviour to assess or promote in PCI work. Modelling language and exposure to grammar is important, hence our inclusion of PB33 in our final list of parent behaviours, however, parents should not be encouraging their deaf child to repeat, correct, or rephrase their utterances.

The second objective of the study was to develop best practice recommendations on how to conduct an assessment of PCI. Participants agreed that parents should be aware they are being assessed and have the power to choose where to be observed and what type of interaction they would like to have assessed (e.g., play, dressing, mealtime). Partnership working with parents, where the parent is seen as the expert of their child, forms the general principle of many policies and codes of practice in the UK such as the SEN Code of Practice [44] and The Best Start to Life [45]. Helpful, internationally accepted, guidance on working with families and deaf infants can be found within the position statement ‘Principles and Guidelines for Early Hearing Detection and Intervention Programs’ [10] and Moeller and team’s paper ‘Best practices in family-centered early intervention (FCEI) for children who are deaf or hard of hearing…’ [13]. Both of these documents also highlight the importance of parent/professional partnerships and recommend the role of the professional is to be a ‘supporter, partner, and coach’ ([10], p.25). The current study’s findings align with this philosophy; many participants agreed that PCI should be video recorded (at least once) and that parents should be given the opportunity to watch the interaction and assess themselves before receiving feedback from the supporting professional. Further, participants agreed that video feedback should be strength-based, i.e., parent and professionals identifying what is working well, but there could also be scope to sensitively highlight behaviours with potential to improve, as long as the overall review is positive and encouraging. The JCIH statement [10] recommends professionals use evidence-based practices and build on families’ strengths, fostering their confidence and competence in providing a range of jointly attended-to language opportunities with regular conversational turns throughout natural daily interactions.

A goal of FCEI [13] is also the development of trusting, respectful family-provider partnerships, characterised by honesty, shared tasks, and open communication. They also recommend a focus on facilitative family–child interactions, rather than child-directed therapies. These recommendations align with our participants’ agreement on collaborative conversations about goal setting, and sharing responsibility for the child’s developmental outcomes. The mention of ‘family-child’ interactions rather than one parent also speaks to our high agreement statements on giving all caregivers the opportunity to have their PCI assessed where appropriate and possible. For the first time, this consensus paper and the upcoming PCI assessment tool provides specific guidance on how to achieve an asset-based observation of parent-deaf child interaction, jointly reviewed by parent and professional.

Participants agreed that, for families who have a deaf-plus child, i.e., a deaf child with additional needs, more flexibility may be required for the assessment. For example, parents may find it is better suited to send the professional a video of interaction rather than video record a live visit, as the child may be more interactive without the external visitor. Likewise, participants agreed assessing less familiar daily routines should be considered and professionals should be prepared to watch less successful instances of communication (if that is the parent’s choice) in order to provide assistance with daily routines. Around 40% of deaf children are reported to have an additional disability [46] which can affect each family in different ways. As per expert comments in S1 File, it is therefore important to follow each family’s lead in how to assess PCI and to respect the parent as the expert of their child and their child’s conditions. Including deaf-plus children in studies that record every day routines is emerging [47–49], but these studies used audio-only recording software (LENA) which limits an analysis of PCI.

Experts concurred that families should be assessed in their home language with bilingual co-workers and/or interpreters helping to interpret the findings. Emerging research is now providing evidence that deaf children can learn two spoken languages [50,51] and that professionals should not discourage bilingualism to parents [52] providing the deaf child has good access to sound. Experts also agreed that cultural diversity and family context should be acknowledged when assessing and drawing conclusions from the PCI assessment, and that if there are behaviours that are unhelpful for the development of the deaf child, these should be raised and discussed sensitively using the video recording of PCI. Government data in England [53] have shown that deaf students who have English as an additional language (EAL) do less well at school than their deaf and hearing EAL peers and the need for better family support and education has been acknowledged. As many of our expert participants shared (see S2 File), a way of connecting with and accurately assessing and supporting parents who have EAL is through the use of video, bilingual co-workers, and using culturally based songs, toys, and games in the assessment.

In this study, expert participants agreed that parents should be given the opportunity to assess their own skills via a video review, but how this is done practically did not reach consensus. The two methods presented in our e-Delphi were the use of scales or a checklist. When using a scale, the professional might describe each parent behaviour before the parent reflects on their skill competency in that area. Scales were used in 34% of the research papers within the systematic review on assessing PCI with deaf infants aged 0–3 [14] and may be more sensitive to skill change than a present/not present checklist. When using a checklist, the professional may again describe the behaviour before the parent and professional discuss which skills were observed on the video. Checklists do not feature in PCI research in deafness but are used in therapies such as Palin PCI Therapy for Stammering [54]. A checklist has the potential to be less threatening than a scale as parents are in charge, simply selecting which behaviours they observed rather than reflecting on their competence in key skills. In the current e-Delphi, the use of scales achieved a slightly higher consensus (63%) than checklists (57%), but both failed to achieve the consensus threshold. When advising on Delphi studies, Hasson and colleagues state that consensus does not necessarily mean the correct answer has been found [55] and that results can help to structure discussion, raise items for debate, and help to streamline focus. Further discussion around professionals’ reservations with scales and checklists will help to better understand the responses and potential differences between professional groups. With regards to the next stages of the PCI tool’s development, how to measure parents’ skills will be decided upon in co-production stages with parents of deaf children and hearing and deaf professionals working in the field.

Many participants commented that the assessment tool was a worthwhile way of closing the gap between research and practice, but it was also the case that money, time, resources, and a heterogenous population [56] often led to huge variations in services being offered to families. A new PCI assessment might be more effective in allocating resources towards each family. However, we also acknowledge that organisational and contextual factors influence practice.

Professionals had mixed views about the inclusion of others in PCI assessments, specifically siblings and deaf professionals. Professionals who argued for sibling inclusion noted these points: increased ecological validity; highly informative; siblings are essential to modelling language; sibling communication is important for the deaf child; and regular individual special time may not be possible. Professionals who rated ‘1’, ‘2’ or ‘3’ provided no justification for their response, so limited insight can be gained here. We expected that if a parent has a deaf child at home along with another infant (perhaps under 5 and not in childcare), the observation might need to include siblings in order to reflect everyday language, everyday dynamics, and the natural environment. Despite its conditional wording i.e., ‘the observation may need to include siblings’, this statement was deemed less important and not essential to include in a PCI assessment.

There are three possible explanations for the mixed views around inclusion of deaf professionals. Firstly, the vast majority of panel participants were hearing. Secondly, many participants noted it was impossible to organise deaf professional inclusion because not all early intervention services employ deaf ToDs, deaf sign language instructors, deaf Language Specialists, deaf Family Support Workers or deaf mentors. Lastly, most participants were based in England, where there are no laws about the inclusion of deaf adults in early intervention. In the USA, including deaf adults in practice is federal law. If a larger sample had been recruited from that country, the statement may have achieved consensus. Both FCEI [13] and JCIH [10] recommend the inclusion of deaf adults in early intervention programs, not only for sign language instruction, but to offer families and deaf children support, guidance, and mentorship. There are multiple reported for families when deaf adults are involved in family-centred early intervention, such as reductions in stress and increased confidence [57], the opportunity to see possible successes for their children [58], learn a range of visual strategies to assist with language learning [59], and deaf adults being role models for families and deaf children [60]. If, once developed, the proposed tool is used internationally, discrepancies between countries on issues such as involving deaf adults should be included in the tool’s manual.

Conclusion

This study recruited 83 experts in the field of deafness to agree on the core content of a tool for PCI assessment for use with parents and deaf infants. A large number of parent behaviours reached consensus including parental sensitivity, positive affect, responsivity, maintaining joint engagement, and language facilitation techniques to ensure language is accessible and / or enriched. Some recommendations were also agreed upon for how a professional might approach an assessment of PCI. These included involving and educating the parent within the review of the assessment and being collaborative about next steps (i.e., discussing and jointly deciding upon areas to improve or adapt within their everyday routines). Developing e-Delphi statements into a PCI tool will support parents and professionals to effectively identify strengths within a parent’s interactions with their deaf child, leading to a more informed, empowered parent who has greater impact on their deaf child’s language.

Supporting information

Acknowledgments

We would like to thank all expert participants for their expertise and time in completing the e-Delphi and contributing towards the development of the assessment tool. Participants included: Alison Kendall, Dr Amy Lederberg, Alison Barley, Amanda Tivey, Amy Stephens, Anna Hughes, Ann East, Dr Alex Quittner, Dr Arlene Stredler Brown, Aya Oyama, Dr Bencie Woll, Dr Beatrijs Wille, Dr Patricia Elizabeth Spencer, Caroline Murphy, Dr Cathy Carotta, Charlotte Emery, Dr Emelie Cramér-Wolrath, Dr Daniel Holzinger, Dr Deborah James, Dr Derek Michael Houston, Dr Evelien Dirks, Dr Elaine Gale, Elizabeth Roche, Ellen Swann, Emma Mottram, Dr Gwen Carr, Helen O’Donnell, Hee Han Hui, Helene Elia, Dr Heidi J. Evans, Jacqueline Watton, Jane Fisher, Janice McWalter, Jayne Ramirez Inscoe, Jayne Langdale, Jane Gallacher, Jo Coote, Jean McAllister, Dr Joanna Hoskin, Julia Hollier, Julie Manning, Karen Reichmuth, Kathryn Reece, Katy Liriano, Kate Dixon, Dr Kerstin Watson Falkman, Kirsti Mackay, Lois Hatfield, Dr Lynne Sanford Koester, Dr Meghana Wadnerkar Kamble, Maria Nicastri, Marsha Locke, Mary Kean, Dr Mary Pat Moeller, Hilary Ann Dumbrill, Noel Kenely, Dr Patrizia Mancini, Dr Reinhild Glanemann, Dr Rachel O’Neill, Rachel Ward, Rhian Gardner, Sabrina Anderson, Sandra David, Sarah Allen, Dr Sophie Ambrose, Dr Susan Louise Tasker, Suzanne Harrigan, Teresa Quail, Vicki Lorriman Hughes, Wendy Martin, and Yvonne Seebens.

We would also like to give special mention and thanks to the research involvement group of deaf and hearing parents and professionals, who have been, and continue to be, vital in the development of this work: Alison Miles, Devi Krishnamoorthy, Dr Evelyne Mercure, Hannah Lumby, Husna Begum, Julie Hare, Julie Hughes, Juliet Viney, Karin Schamroth, Katerina Giachritsi, Lisa Smith, Martine Monksfield, Michelle Rayne, Paria Moghaddar, Sabina Iqbal, Tarryn Jacobs, Tina Wakefield, and Yasmena Waris.