International Journal of Healthcare Simulation

Background

Speaking up is voicing a concern for the prevention of error and/or harm (physical and/or psychological) to healthcare staff and patients [1]. The definition of speaking up appears relatively straightforward, yet the reality for health professionals, it is not so simple. Miscommunication (including a speaker’s inability to speak up or having concerns ignored) is a significant contributor to patient harm [2]. There are a number of well-known barriers to speaking up, including hierarchy and fear of retribution [3] and poor organizational culture [4]. In addition, how a message will be received, including the anticipated response, or lack thereof, is a consistent challenge for clinicians when trying to speak up [5]. What is lacking in the healthcare speaking up literature is an empirical understanding of why receivers of a speaking up message respond as they do. To date, speaking up training tries to theoretically understand the receiver’s response, but little research on speaking has focused on the receiver’s perspective.

Some initial receiver-focused research within speaking up conversations is underway in the perioperative environment. Lemke and colleagues [6] studied the receiver behaviour of anaesthetists during anaesthetic induction, and Long et al. [7] in the general perioperative environment. More recently, theory-driven research, through the application of Communication Accommodation Theory (CAT [8]), has also examined the perceptions of health professional receivers from multiple disciplines and clinical environments (Barlow et al., 2022a). The current study used CAT to examine the behaviour of receivers.

CAT has been widely applied in healthcare communication research. For example, CAT has been used to study open disclosure conversations after medical error [9], nurse–doctor interactions [10], communication between medical specialties [11], and nurse–parent communication [12]. Healthcare is a rich intergroup environment, where salient professional identities can be defined by clinical discipline, seniority level, and clinical specialty or department. As an intergroup theory of communication, CAT is well suited to this environment, as it provides a framework to predict and understand intergroup interactions (e.g. between disciplines or departments) and their associated power differentials [8]. The theory posits that an individual’s communication goals, motivations and drivers of their communication behaviour, are influenced by their salient group identity (the most prominent identity within that interaction) [8]. In previous studies, the authors have demonstrated that speaking up conversations regarding patient care are intergroup interactions, influenced either by a discipline, e.g. nurse being spoken up to by a doctor, or seniority level, e.g. junior nurse being spoken up to by a senior nurse [13].

Social Identity Theory (SIT) explains how individuals within an interaction use behavioural strategies to establish, maintain or make themselves positively distinct from their speech partner, based on group membership [14]. Informed by SIT, CAT theorizes behavioural communication strategies people employ to manage their social relations with others [8]. When using accommodative strategies, similar to therapeutic communication, the person aims to adjust their language and tone to aid comprehension, build empathy and trust and minimize power differentials [12]. In contrast, nonaccommodative behaviour serves to maintain or extend the power differentials and does not facilitate comprehension. Nonaccommodative strategies serve to differentiate a speaker from their conversational partner [8].

Giles [8] explicates that there are five main communication strategies interactants use: approximation, interpretability, discourse management, emotional expression and interpersonal control. Approximation occurs through communication behaviours such as the adjustment of speech patterns, e.g. volume and pitch, rate and tone of speech and non-verbal behaviour to be similar or convergent (accommodative), or dissimilar or divergent (nonaccommodative) to their speech partner. The strategy of Interpretability refers to using language to help (or hinder) comprehension of the message by one’s communication partner. The speaker evaluates their speech partner’s ability to comprehend what they are saying and adjusts (or not) their language (including vocabulary, pronunciation, gestures) accordingly. An example of accommodative interpretability is when a speaker uses language and terminology to aid comprehension. Nonaccommodative interpretability refers to the use of language and jargon unique to a speaker’s health discipline, which may not be readily understood by a conversational partner from another discipline or specialty [10]. The strategy of Discourse Management refers to the interaction process, rather than the content of the conversation. When accommodative, it involves actively engaging the conversational partner, by encouraging participation through manoeuvres such as expanding the content, using open-ended questions and back-channelling, e.g. head nod, or saying mhmm. When nonaccommodative, active engagement from the other person is not encouraged or considered [12]. The strategy of accommodative Emotional Expression concerns appropriately meeting the emotional needs of a conversational partner, by actively acknowledging their needs and providing reassurance. It involves behaviours to ‘save face’, maintain relationships, give reassurance and show concern when appropriate. Emotional expression also includes the speaker regulating their own emotions and emotional expressions (warmth, happiness, appreciation) to maintain the social relationship. Nonaccommodative emotional expression occurs when the emotional needs of the other are ignored or not recognized, and/or a speaker does not regulate their own emotions and expressions to meet the other’s needs. The strategy of nonaccommodative Interpersonal Control refers to a conversational partner’s attempts to constrain a speaker in a specific role through the utilization of power differentials. When a speaker is accommodative, opportunities are created for their speech partner to move outside of their prescribed role [10]. An example of accommodative interpersonal control would be a senior clinician inviting a more junior clinician to voice their opinion (speak up) regarding the care of a patient.

The current study involved direct observation of receiver behaviour during a speaking up situation (discharge planning team) in a defined context using simulated (laboratory) conditions. The study had two research questions. First, what CAT strategies do receivers use to assist their understanding of a speaking up message to address the concerns of the speaker (a nurse)? Second, to investigate if the observed receiver communication behaviours differed across health disciplines (receiver groups).

Methods

Participants

One hundred and thirty-eight clinicians participated in the simulations. They consisted of nurses/midwives (n = 96), allied health, comprising of social workers, physiotherapists, radiographers, pharmacists and phlebotomists (n = 22), and medical officers (doctors) (n = 20). Most participants were from inpatient wards (n = 48) and critical care areas (n = 26). Years of experience ranged from being a new graduate or intern with less than 3 years’ experience (n = 44), through to more than 20 years of clinical experience (n = 15). See Table 1.

Table 1:

Participant characteristics (n = 138)

Characteristic	Nurse/midwife		Allied health		Medical officer
	N	%	n	%	n	%
	96	69.5	22	16.0	20	14.5
Clinical specialty
Critical care	18	18.8	1	4.5	7	35
Perioperative	13	13.5	0	0	2	10
Inpatient wards	35	36.5	8	36.4	5	25
Day stay areas	1	1.0	0	0	0	0
Antenatal areas	2	2.1	3	13.6	0	0
Birth suite	7	7.3	0	0	2	10
Outpatients	4	4.2	4	18.2	2	10
Interventional areas	5	5.2	3	13.6	0	0
Other	8	8.3	2	9.1	2	10
Missing	3	3.1	1	4.5	0	0
Years in profession
3 years or less	29	30.2	5	22.7	10	50
4–8 years	22	22.9	7	31.8	2	10
9–14 years	11	11.5	4	18.2	5	25
15–20 years	7	7.3	3	13.6	1	5
More than 20 years	11	11.5	2	9.1	2	10
Missing	16	16.7	1	4.5	0	0
Gender
Male	11	11.5	6	27.3	10	50
Female	85	88.5	16	72.7	10	50

Allied health: social work, physiotherapy, radiography, pharmacy and phlebotomy

Procedure

All participants received a standard e-mail reminding them of their attendance to the corporate speaking up training program, which included the Participant Information and Consent Form (PICF). Participants arrived with either their completed consent form or had an opportunity at the start of the training program to ask the lead investigator any questions regarding the activity and their involvement. Participants could choose to consent (or not) up until the commencement of the simulation activity. Participation in the research did not impact participants meeting the requirements for successful program completion. Participants self-enrolled into the speaking up program; therefore, it was unknown to the research team on any given day what clinical disciplines would be present and who would consent to participate. As a result, each simulation differed in group size and clinical discipline representation. Of the 22 simulations, at least two nurses/midwives were present in each simulation, medical officers were present in 11 and allied health in 10 simulations.

Within the simulation, participants were the discharge team, and upon entering the patient’s room, the junior bedside nurse, Mary (embedded simulated person), spoke up about her concern for the patient (manikin voiced by a simulation technician) going home without the appropriate and required home care in place. The pre-booked ambulance was on the way to transport the patient home, which contributed to a time pressure regarding decision-making. Owing to the need to get a timely decision, Mary had asked for her patient to be reviewed first by the discharge team, but her patient was, in fact, the last patient to be reviewed. The bedside nurse in all simulations spoke up using ‘hint and hope’, a commonly cited methodology used by nurses [16]. The two different speaking up messages were alternated across the 22 simulations. Both were accommodative but delivered with differing degrees of accommodation. One message was delivered by the nurse in a very polite manner, giving a lot of information (verbose), but not directly stating (instead hinting) the main concern. Alternatively, it was delivered in a less accommodating, less polite manner, which was more succinct and abrupt, and still did not clearly articulate the main concern (see Table 2 for examples). There was no clear nonaccommodative speaking up message, as it was deemed unrealistic that a more junior nurse, when faced with speaking up to a multidisciplinary team, would be openly confrontational. To investigate communication dynamics, the speaking up message was delivered to the team, rather than to a specific individual. Who responded, and how that response was framed, was completely up to the receivers.

Table 2:

Example of the initial speaking up message the speaker (actor) used within the simulations

What Mary was thinking	What Mary actually said
The patient lives alone with no home support. She is a high falls risk. Mary is thinking the patient should not go home today due to high risk of falling. The ambulance transport needs to be cancelled and a comprehensive discharge plan organized.	Message 1: Hint and hope – Accommodative, verbose Thank goodness, you’re here. This is Mrs Williams she’s due to go home today. The ambulance is coming within two hours, but the discharge has not been organized. So, I’m really worried that the appropriate care is not going to be in place. I really think we need a decision like now, whether she can go home or not. As I said, the discharge is not in place, so wondering what your thoughts are.
	Message 2: Hint and hope – Less accommodative, succinct/abrupt Look, I asked the ward receptionist if you could start the ward round here, not finish here. I’ve got the ambulance transport coming to pick up Mrs Williams and none of the discharge has been done and I don’t think she’s ready to go home. So, I just need you to do something.

After the simulations, participants undertook a structured debrief, as reported elsewhere [17]. The simulations were designed and delivered according to international standards in simulation and by educators trained in simulation methodology [18].

Results

Twenty-two simulations were run, recorded and analysed. The simulations ran for a mean of 4.52 minutes (range 3.0–7.24 minutes), SD 1.06. A total of 264 codes were identified from all receiver responses.

During the coding process, the coders (BW, MB) agreed that it was difficult to clearly analyse what aspects of receiver behaviour aligned to the strategies of discourse management and approximation. This difficulty arose because several receivers within an interaction asked short and direct questions, which meant it was not possible to accurately gauge approximation. For example, there was no opportunity to map how much receivers matched the speaker’s tone of voice. For discourse management, video angles did not always allow the receiver’s faces to be clearly visible. This meant that analysis of the receiver’s facial expressions or gestures that might signal encouragement for the speaker to continue the conversation was not able to be consistently analysed. As the interactions represented technical conversations, where obtaining relevant information was paramount, the strategy of interpretability was identified as the primary strategy rather than discourse management. When the receivers sought further clarification of the concern, it was apparent that they were not encouraging the speaker’s engagement in the conversational process, but rather required additional important information to help their interpretability of the concern. For these reasons the strategies of approximation and discourse management were removed from analysis. A key tenet of CAT is the influence of context. Where and when the conversation is occurring is theorized to influence the strategies individuals deploy. Therefore, it is not unexpected that receivers in our study used some strategies more than others. The three remaining CAT strategies of interpretability, interpersonal control and emotional expression were analysed by comparing the use of the communication strategies between the three receiver groups (nursing/midwifery, allied health, medical officers) and across speaker message types (accommodative/verbose, less accommodative/abrupt).

Across receiver groups, most receivers displayed accommodative behaviour (nonaccommodative behaviours n = 2) regardless of which of the two messages were being delivered. To compare the frequency of accommodative CAT strategies used per receiver group, the rate of codes for each CAT strategy was calculated as a proportion of the sample size for that receiver group (see Figure 1).

Figure 1:

Number of codes per CAT category in relation to sample size per receiver group

Discussion

The aim of the study was to investigate what CAT strategies receivers used to assist their understanding of a speaking up message and how receiver behaviour differed between health disciplines and across the two message types. Speaking up messages were delivered using the hint and hope methodology in either an accommodative, or less accommodative manner. How the speaking up message was delivered made some subtle observable differences in the receiver’s behaviour, particularly in the strategies of interpersonal control and emotional expression. Certainly, all receiver groups overwhelmingly displayed accommodative communication behaviours towards the speaker. Accommodative behaviours reflecting CAT strategies have been demonstrated to reduce the social distance between groups and thereby enhance communication outcomes [23]. However, the extent to which receivers accommodated towards the speaker and the CAT strategies they used, differed between the clinical disciplines. Additionally, the complexity of speaking up interactions was highlighted by receivers often using a combination of CAT strategies in their response.

Combining CAT strategies was a shared behaviour across all receiver groups, which appeared to both enhance the receiver’s understanding of the concern and meet the speaker’s needs. The use of emotional expression with either interpretability and/or interpersonal control were the most utilized by receivers. Uncertainty about the correct interpretation of the concern, fear of being wrong and/or fear of retribution are well-documented barriers to speaking up for clinicians [24]. The deployment of accommodative emotional expression by receivers (e.g. validating the speaker’s concern, demonstrating gratitude by thanking the speaker for speaking up, and demonstrating empathy) was important in building rapport, reducing the associated fear of speaking up and enhancing speaker engagement in the conversation. Once clarity of the concern was gained, receivers appropriately took control to move the conversation forward. Receivers often did this through offering suggestions and seeking consensus with the team and speaker.

Receiver identity (clinical discipline) influenced receiver behaviour. Interestingly, nurses/midwives demonstrated more interpersonal control when the speaker spoke in a less accommodative manner than an accommodative one. Within the debrief, this receiver group evaluated the less accommodative speaker as being frustrated and stressed. Attempting to take control of the situation was therefore deployed as a supportive measure for their ingroup member (nurse speaker). It was unexpected that nurses/midwives overall displayed the least emotional expression behaviours, given they are typically viewed as ‘the caring profession/s’ [25]. This behaviour was explored in the debriefs, where ‘listening to fix’ was identified as the largest barrier to effective receivership [17]. Nurses/midwives stated that ‘fixing’ the problem was their way of demonstrating emotional support to the speaker. The problem with this behaviour is two-fold. Firstly, we know from the literature that speaking up is hard, particularly for those in lower hierarchical disciplines [3]. Therefore, receiver support and emotional expression behaviours need to be explicit, rather than implicit to encourage participation in both the immediate and future speaking up interactions. Secondly, by asking task-based, ‘fixing’ questions, identification and understanding of the concern was delayed. If this behaviour was adopted in clinical situations where patient harm was imminent, this could potentially have dire consequences [2].

Regardless of the level of accommodation, the indirect nature of the messages (hint and hope) influenced interpretability of the concern. This forced receivers to implement strategies, often unsuccessfully, to try and achieve clarity. Speaking up mnemonics aim to help the speaker voice a succinct and clear concern [26]. There is, however, extensive research demonstrating that despite mnemonics and training, speaking up remains difficult, particularly in the presence of power status differentials between the speaker and receiver [27]. Like the junior nurse speaker in this study, the vague, hint and hope methodology is widespread amongst nurses, particularly when speaking to medical officers [16], and has also been noted in speaking up conversations during anaesthetic induction [6]. From our findings, it seems crucial that receivers are trained to receive a message. Receivers need to know how to manage self in the moment, to get curious and question further and to use open-ended questions that seek to understand, rather than to fix.

When observing allied health behaviour across the two message types, more emotional expression and fewer interpretability strategies were observed in the less accommodative simulations. This finding could reflect that when the speaker is less accommodating, allied health shift their behaviour from questioning to the provision of emotional support. This finding does align with our previous work, where more than any other receiver group, allied health was able to make positive attributions (accommodating emotional expression) about a nonaccommodative speaker [28]. The making of positive attributions, e.g. the speaker is concerned for patient safety, increased their perceptions of the acceptability of the message.

Medical officers were influenced the least by the speaker’s communication behaviour. This supports previous findings [29] where this receiver group evaluated both accommodative and nonaccommodative messages as equally acceptable [29]. It was expected that doctors would step forward as the first receiver, given their hierarchical positioning in the multidisciplinary team; however, this did not occur. Instead, medical officers stood back, listened and stepped into the conversation when a resolution was not readily forthcoming, which proved a successful tactic. The delayed engagement in the conversation may have been influenced by differences in communication training and role expectations between the disciplines. Clearly, how a message was heard and responded to, differed according to clinical discipline, and may account for why current speaking up training and standardized mnemonics are not always successful [30].

Regardless of discipline, it often took some time in the simulation for participants to realize they were being spoken up to. Participants were briefed prior to the simulated encounter; however, they were intentionally not informed as to whether they were going to be the speaker or receiver within the interaction. We acknowledge that delays in recognition and deployment of appropriate behaviour adjustments may have been influenced by the simulated context. Currently speaking up programs routinely only rehearse the act of speaking up and not receiving [31], which we believe is a key contributing factor. We argue that this lack of recognition or awareness of being spoken up to due to vague messaging readily occurs in real clinical practice. Receivers, therefore, may be being judged as ignoring the speaker and their concern, which may in part account for reports of inadequate receiver behaviour within the speaking up literature.

A key strength to this study was the application of a robust communication theory to analyse directly observed receiver behaviours. CAT highlighted how different disciplines have different approaches to receiving the same information. This finding has not previously been explicitly noted in the literature, and sheds light on an additional reason why speaking up remains difficult despite standardized mnemonics. Understanding differences between disciplines is important to increasing the efficacy of healthcare communication training.

There were limitations to the study. The speaker in all the simulations was a nurse. Whilst this allowed a consistent comparison of receiver behaviour, it needs to be noted that receiver behaviour may have been different if the message was delivered by another discipline. It must also be noted that for some clinicians, being a member of a patient discharge team was not part of their normal clinical role and may have affected their ability to engage. We do argue though that you do not have be a content expert to effectively listen and help facilitate a curious conversation. Participants knew they were being watched, and the fact that the clinical situation was not an emergency, high-stress, or high cognitive load event may have contributed to the accommodative stance of nearly all receivers. The literature from the speaker’s perspective has often reported poor or nonaccommodative receiver behaviour [32]. This requires further investigation. The situation however did allow for the elicitation and observation of unique communication behaviours by different receiver groups. To further enhance transferability, future studies should use CAT to observe receiver behaviour in high-stakes situations, where patient harm is imminent.

We compared the pattern of behaviour between receiver groups by the receiver being part of a group of clinicians who were the ‘discharge team’. This was a strength of the study, but also a limitation, as it meant that the strategies of approximation and discourse management could not be identified for individual receivers. Future studies should use CAT to observe and examine the speaking up interaction between a speaker and a single receiver. In addition, this study highlights the complexity of isolating communication strategies that may co-occur.

Conclusion

Speaking up is important for patient safety and effective clinical care; however, little is known about how people respond to speaking up messages. We found clinicians used the strategies of interpretability, emotional expression and interpersonal control to receive and respond to the speaking up message. How these strategies were deployed differed between the three discipline groups and with varying degrees of effectiveness in meeting the speaker’s needs. The findings have direct application to speaking up training for both speakers and receivers and demonstrate CAT could be used to help frame and inform recommended receiver strategies to enhance speaking up communication within and between disciplines. Clinicians being trained to respond effectively to speaking up messages will both improve patient outcomes and enhance the occurrence of clinicians voicing their concerns.