Correcting medical decisions: a study in nurses’ patient advocacy in (Finnish) hospital ward rounds

During daily hospital ward rounds, medical teams, led by doctors, assess the progress of an individual patient’s health. It is widely reported in the research literature that nurses play a relatively passive role during these rounds, because although they may have valuable information about the patient’s condition and progress, and indeed their role includes advocacy on behalf of their patients, nurses nevertheless can experience difficulties in participating during case constructions. Here we report an instance from a (gastro-surgical) ward round in a Finnish hospital, in which nurses played a key role in reversing a consultant’s initial decision to discharge a patient. They did so not by directly challenging the consultant’s opinion, but by employing indirect means to introduce their discrepant perspective: they provide descriptions and ask questions that draw attention to information that results in the doctor coming to a different assessment than theirs of the patient’s condition, and a different decision about what should be done (the patient was not discharged from hospital). The encounter reported here is taken from a corpus of ward round discussions in a Finnish hospital. The method of our study is Conversation Analysis.


Introduction
This paper arises from research into ward rounds in a large Finnish hospital. In the bedside consultation that is the focus of this report, it emerged that the nurses present at the bedside had information about the patient which was initially overlooked by the consultant (known as a specialist, in US medicine). We explore the difficulties that nurses may experience in participating in the decision-making process during the ward round. Put simply, they have first-hand knowledge about a patient's medical condition, and are able to assess a patient's progress over a period of time; what, then, when nurses' assessments are discrepant with a consultant's assessment? It becomes evident in this particular gastro-surgical ward round that the nurses regard the consultant's assessment of the patient and his consequent decision to discharge as incorrect. However, the difficulty nurses face in this situation is what might be described as their restricted entitlement to correct a consultant. In common with most organisations (Lane 2006), medical organisations, including hospitals, are structured hierarchically. Nurses are lower than consultants in the medical hierarchy, and it is known that those lower in an organisational hierarchy can have difficulty correcting the (perceived) mistakes of their superiors.
The ward round is an integral and important part of the daily routine of most hospitals; though this study is of ward rounds in a Finnish hospital, they are a familiar feature of hospitals in many other countries. The multi-professional consultations 'play a crucial part in reviewing and planning a patient's care. They are an opportunity to inform and involve patients, and for joint learning for healthcare staff' (Royal College of Physicians 2015). Hill (2003) considers the ward round as a pivotal part of the daily decision-making process about patients' care and treatment. Ward nurses are members of those multi-professional teams (Caronia and Saglietti 2018). While hospital ward rounds are conducted by doctors, nurses may input valuable information. They have considerable direct knowledge of a patient's condition, progress and responsiveness to treatment, having observed and assessed a patient over some period. Nevertheless, nurses report difficulties in participating in the decision-making discussions, and in raising relevant patient issues during the ward round (e.g. Manias and Street 2001: 442, 447). Weber et al. (2007: 346) even describe ward rounds as a 'dyadic interaction between patient and physician with only minor contributions from nurses'. Nurses may experience restrictions in their autonomy which results in them being relatively passive in opening discussion during the rounds (Leino et al. 2017). In particular, nurses tend not to open discussion about the planning and evaluation of a patient's treatment, about issues such as disease progression, discharge and follow-up treatment (Leino et al. 2017: 5-6). However, nurses possess knowledge of a patient's particular medical circumstances that can be valuable in deciding on treatment and disposition. Taking internal medicine as an example, Weber et al. (2007: 346) concluded that as the majority of the patients within this branch of medicine are elderly, it is particularly important to exploit nurses' knowledge in such matters as the final placement of the patient, or the feasibility of a therapeutic regimen. Interprofessional collaboration, which is characterised by communication, shared decision-making and integration of the knowledge and expertise of each professional (see e.g. D' Amour et al. 2005), is officially encouraged in this medical context (Weber et al. 2007: 344, Leino et al. 2017, see also Royal College of Physicians 2015). However, it is evident from their reports (e.g. Manias and Street 2001) that collaboration and shared decision-making pose difficulties for nurses at the grass-roots level of daily ward rounds.
Medical organisations are structured hierarchically. Nurses do not have epistemic authority, deontic authority or rights in decision-making that are equivalent to those possessed by doctors, let alone by consultants/specialists. Nurses have considerable relevant medical expertise, experience and insight, but they might be regarded as having less epistemic authority (Heritage 2012) than doctors, as regards medical knowledge. They do not have the same deontic authority as doctors to decide about patients' treatment; hence also they do not have the same rights as doctors to decide 'what should be done' concerning patients' treatment in hospital (see also Stevanovic and Per€ akyla 2012). These are the principal respects in which nurses are lower than doctors in the medical hierarchy, and which account for the reported passivity of nurses in ward round or multi-professional discussions. This has been reported in other hierarchical organisations such as aircraft cockpits; nurses' difficulty in redressing (what is perceived to be) a superior's mistake is one shared by aircraft flight personnel such as co-pilots when they realise that the captain might have made a mistake. Fatal aircraft accidents have been attributed to co-pilots reluctance to correct a captain's error or to draw attention to that error, or a captain's failure to accept input from junior crew members (Helmreich et al. 1999: 20; on the Kegworth air disaster see Department of Health report and archived 2012). For nurses there is an additional complication, which is that their role is to act as an advocate for and on behalf of their patients, a role that includes informing others about their patient's condition and unmet needs and to protect their patients (Hanks 2005and Hanks 2008, Josse-Eklund et al. 2014. Given the prominence of this aspect of their role, nurses are subject to conflict between their role in the medical hierarchy and their countervailing role as patient advocate. How they resolve that conflict, in circumstances where they may perceive that a decision is being made that is contrary to the (medical) interests of their patient, is the subject of this report.
The nurses in the episode we focus on here faced the practical dilemma of how to redress the consultant's perceived error without being entitled to correct him. A recent study showed that nurses have subtle means of contributing to medical case construction in interprofessional discussions (Caronia and Saglietti 2018). However, the literature on nurses' general difficulties of having their voice heard during the round (e.g. Leino et al. 2017, Liu, Manias and Gerdtz 2012, Manias and Street 2001 suggests that correcting their superiors is not an easy task. In this paper, we show how the nurses in this encounter managed to intervene to redress the consultant's assessment and decision concerning a particular patient, and thereby to prevent an outcome that they regarded as contrary to the patient's safety. They managed to do so without directly or explicitly correcting the consultant. They adopted a strategy of progressively revealing information indicating that the consultant's decision was inappropriate, based on his incomplete knowledge of the patient's condition and an erroneous assessment of his progress. Our study contributes to the emerging body of knowledge on the conversational details of the joint construction of the medical case by nurses and doctors (e.g. Caronia and Saglietti 2018). By so doing, this paper responds to the calls for research on the participation and involvement of nurses during rounds (see Leino et al. 2017: 2).
Decision-making in medical interaction has been studied across a variety of contexts, from neonatal critical care to, for example, paediatrics, oncology and neurology (Collins et al. 2005, Shaw et al. 2016, Stivers 2005, Toerien et al. 2013. In these studies, decision-making is a process between the medical professional and the patient (or the caregiver). Decision-making between medical professionals has received much less attention in conversation analytic studies. This study explores the process of interprofessional decision-making, and thereby also responds to the calls for extending investigation to multiparty situations where doctors may share tasks with other health professionals (see Barnes 2019: 310). Our study contributes to a growing body of conversation analytic work that offers a 'more nuanced account of medical authority than is typically portrayed in the sociological literature' (see Toerien et al. 2013: 873). It contributes also to our understanding of interactional asymmetries in institutional interactions (Drew and Heritage 1992:47-53). Such asymmetries are generally considered to involve those between professionals and lay persons (clients etc.); here, though, we consider the asymmetry between professionals, in the context of how treatment decisions are reached. Participants' orientations to such asymmetrieshere exhibited in nurses' indirect means of correcting the consultantare evidence also of their orientations to their roles and identities, as displayed in the constraints on nurses' contributions to ward round discussions.

Data and method
The case we examine here is from a corpus of video recordings of authentic ward rounds in a gastro-intestinal surgical ward in a large Finnish hospital (recorded 2016-2017). These recordings were transcribed in detail using the conventions used in Conversation Analysis (see Jefferson 2004). The full transcript and a glossary of the transcription conventions can be obtained as supplementary material. All names and other identifying information have been omitted in order to ensure complete anonymity. Figure 1 below may help to give the reader an idea of the configuration of participants in this multidisciplinary ward round of bedside consultations.
The patient is an elderly man who has a history of urological problems, and has recently undergone surgery resulting in having been fitted with an ostomy bag or pouch. The medical team consisted of the consultant (specialist) gastroenterologist, who was entitled to make the decisions about the further procedures concerning the patient; the registrar (i.e. junior doctor specialising in internal medicine); the ward sister (senior nurse overseeing an entire ward of patients); and the nurse. A trainee nurse was present but did not feature in the interaction that is our focus here. Typically, a bedside consultation lasts 3-6 minutes, during which the consultant asks the patient about his/her condition and feelings, checks his/her medical record and decides what will be the next step in his/her treatment. Bedside consultations of 29 patients were video-recorded and transcribed; a total of two consultants, three registrars, one ward sister and 13 nurses participated in the study. Ethical approval was granted by the local health authority's research committee, and informed consent was obtained from all participants.
The method for the study is Conversation Analysis (hereafter CA; for an overview, see Sidnell and Stivers 2013). Through a micro-level CA approach, it is possible to determine the actions in which people are engaged, the linguistic and embodied resources and practices mobilised to perform these actions and the sequential patterns that are systematically associated with the management and consequences of those resources and practices. For example, CA research has shown how a change of a single word ('anything else you want to ask' vs. ' something else you want to ask') can be consequential for eliciting patients' hitherto unspoken agendas (Heritage et al. 2007). We will use CA to demonstrate how the nurses in a particular encounter managed to reverse a consultant's assessment of a patient's progress and his decision about the discharge of the patient. We demonstrate that they did so through mobilising interactional resources associated with managing repair/correction in an implicit manner, at an unofficial level of the interaction (Jefferson 2018); these resources and linguistic practices reflect nurses' subordinate role or position in the multi-professional medical team, while nevertheless enabling them to play a key role in decision-making concerning the patient's future treatment.
We should note that focusing on a single episode in interaction is justified methodologically in CA research when explicating a practice, pattern or phenomenon which, although not rare per se, nevertheless is observed in rather special or unusual circumstances, or in a context in which that practice has a special significance or serves some special and distinctive Figure 1 The medical team around the patient's bed: the consultant (C) at the left of the bed, the nurse (N) on the right, the ward sister (WS) further away, the registrar by the computer.
(organisational) task. Examples of the analysis of single cases of conversational phenomena include Schegloff (1987), in which he explicates the way in which the coherence of some episode in interaction derives from the overarching action being conducted, and not, as might be supposed by the 'topic' of conversation; Drew (1989), who investigated a single greetings episode in order to explore general social rather than cognitive aspects of memory attributions (in claims about 'remembering'); and a study by Drew and Penn (2016) of a single brief episode in which a woman with aphasia is attempting to tell her therapist something about her pastthe study focuses on sources of misunderstanding and incomprehension in talk.
It will be evident from the glosses of these 'single instance' studies that they serve to depict and explicate through the details of a single episode, quite systematic and general phenomena that are the constitutive practices which together give the episode its particular interactional character (on which see Schegloff 1987: 101-3). We have selected the encounter that is the focus of this analysis because it is the only instance in our corpus of 29 ward rounds in which a consultant's assessment and decision is revoked; there are only three cases in our entire corpus in which a nurse questions a doctor's assessment or decision, and only in this instance do we have evidence that a decision was changed as a result of a nurse's intervention. While it is singular in that respect, nevertheless the interactional means through which the nurse(s) achieved this reversal are methodic practices employed by the nurse(s) that are widespread in interaction (Jefferson 2018, Schegloff et al. 1977. Hence, we explore through the means of this single encounter how these indirect, methodic practices for correction are deployed to reverse a decision made by a superior (see Pomerantz 1990 on the analysis of single episodes to illuminate the use of methodic and quite systematic interactional practices). How this occurred and how the nurse(s) handled the awkwardness associated with questioning the judgement of a medical superior goes to the heart of the role conflict to which we referred above. Our analysis aims to illuminate how a nurse manages the rather unusual or even rare circumstances in which the interactional norms associated with hierarchy conflict with the imperative to ensure accurate and safe decision-making on the ward (nurses' advocacy role).

Findings
The encounter that is the focus of our analysis is an excerpt from a routine ward round by a senior consultant in the hospital. Having first discussed the patient's readings with other members of a medical team, he approached the patient's bed. At this point, the consultant (C) was joined by three nursing staffthe ward sister (WS), a nurse (N) and a trainee nurse (TN). Only the nurse and the ward sister played a significant role in the unfolding interaction with the patient and doctor; the nurse was assigned to this patient and therefore was principally responsible for his care and monitoring his condition. The physical configuration of the participants can be viewed on material accompanying this paper [ Figure 1: Configuration of participants], consisting of an anonymised still of the configuration of the participants.
In order to provide the context for a more detailed analysis of the interaction, it will help to give an overview of the excerpt. The starting point is the consultant's decision that the patient has made sufficient progress to be discharged home from hospital that day, a decision arrived at after looking at the patient's record since admission. However, at the end of this episode, the consultant decided that the patient should not be discharged but instead should be moved to a different (surgical) ward. Evidently the consultant's assessment of the patient's condition and his readiness for discharge changed quite significantly between these two decision points. The consultant's initial decision was reversed through a series of (heuristic) phases that can be summarised as follows: We should emphasise that these are not stages in any formal sense, for instance in the overall organisation of a certain form of interaction (on the stages of the primary care medical consultation, and a critique, see Robinson 2003). We have identified these phases in this particular interaction only as a heuristic device to represent the series of interactional moves through which the initial decision was replaced with another quite contrary decision. Closer examination will show how the initial decision was implicitly challenged and how a quite different decision came to be made. It will reveal particularly the role that the nurses, both the ward sister and especially the nurse, played in the process of changing that initial decision.

First phase
Having reviewed the patient's case notes with the registrar away from the bed, the consultant moved to the foot of the patient's bed, and announced the result of his reviewa decision to discharge the patient (lines 21-22 The consultant's three-part summary (Jefferson 1990) in lines 13-14 (kidneys started functioning, you can pee, the ostomy works) indicated a positive assessment of the patient's progress. The consultant did not (physically) examine the patient; he relied therefore on the review with the registrar, and on the patient's collaborative positive assessment (feeling better) in lines 4-10. Having accounted for and dispensed with the patient's understated concern indicating a possible problem (almost too well) (line 16), the consultant delivered his decision that the patient should be discharged (lines 21-22). The interrogative format of the delivery of this decision (would it be) is perhaps closest to the 'suggestion' format for treatment recommendations identified by Stivers et al. (2018): the decision is offered to the patient almost as an option (line 21-22), which the patient accepts with the token aha (line 26).
The decision to discharge the patient that day was marked as a conclusion from or an outcome of his prior assessment, in the turn-ending 'then' (sitte, line 22). There are, however, two points to notice in this interchange. First, the patient responded to the consultant's decision by partially repeating it (today, line 24), then producing the particle aha (line 26). Partial repeats can be associated with repair initiation, even surprise (e.g. Thompson, Fox and Couper-Kuhlen 2015: 60-4). The particle aha is used to receipt informings that redirect ongoing talk (Koivisto 2016). The patient had not perhaps expected to be discharged that day, his lack of conviction in responding to C's decision is consistent with his reservation that the ostomy has started to work almost too well (lines 15-16). Second, the ward sister and nurse exchanged glances (line 21, Figure 3), after the ward sister has pointed to something on the patient's bed (presumably some soiling of the sheets), a point that turns into a sweep with her finger (line 19, Figure 2). Their smiles and the nurse's facial gesture convey a 'knowing' or collusive quality. The ward sister's gesture and their mutual glances and expressions were not in the consultant's sight line so not evident to C; it is apparent from the video that they expressed to one another, covertly and non-vocally, their scepticism concerning C's assessment of the patient's situation. Second phase Immediately after the patient acknowledged the consultant's suggestion, the nurse intervened. Up to this point, the discussion had been between consultant and patient, but now N joined the conversation: Extract 2. The nurse began with the discourse particle (tota 'well', line 27), conveying hesitancy (Hakulinen et al. 2004: 979). N did not continue her turn, but this is enough to indicate that from her point of view there is more to the matter; tota functions as a 'my-side alert' (Heritage 2015: 98). C invites N to continue through both verbal and embodied means; he repeats the particle and turns towards N (line 29). Having been granted 'license to speak', N takes a longer turn (lines 30-41) outlining the difficulties the patient has experienced using the ostomy.
It is now clear that the nurse did not share the consultant's positive view of the patient's condition. However, she did not openly contradict anything the consultant has said, but was indirect in her approach in disagreeing with him. First, her initial turn consisted of only one particle; she could have stated the problems right away, but instead indicated only that she wishes to take a turn (line 27, tota 'well'). Only after being explicitly prompted by the consultant (line 29) did she proceed to describe the problems (lines 30-38). Second, she did not refer to or explicitly question or challenge C's decision to discharge the patient that day. She volunteered information (Caronia and Saglietti 2018: 4) simply by describing problems she has encountered while taking care of the patient. Yet it is evident in her choice of linguistic formulations that the consultant's positive, upbeat assessment of the patient's condition and therefore readiness to be discharged was not in her view correct. She referred to her (and her colleagues') actions with a construction implying necessity (line 34, ollaan jouduttu 'we had to put back'), and to the patient's actions with a modal verb (line 35, ei pysty 'cannot') conveying the physical restrictions of the patient's capabilities. The trajectory of C's questions was to show that the problems revealed through N's account in lines 30-41 occurred only during this most recent period in hospital.
Extract 3b. 47 C: onks siin€ a nyt tullu u-uus ongelma sit t€ ass€ a has it now become a n-new problem now here during 48 sairaalahoidoj jakson #aikana# the hospital period To the consultant's question about the ostomy, the patient reported that, previously when at home, he has not had difficulties having managed to empty it regularly (lines 59-62), which the consultant acknowledged (lines 69 and 72).
Extract 3c. From the consultant's questions, the picture began to emerge of the patient's ability to manage the ostomy on his own when at home. Responding to N's evidence that the ostomy arrangement is not working effectively (lines 34-41)which implicitly conveyed that the patient may not yet be ready to be dischargedthe exchanges in this phase conveyed instead that the patient might be capable of managing these difficulties at home.

Fourth phase
After the patient has portrayed himself as capable of managing the ostomy at home, the consultant turns towards the nurses. The ward sister (WS) uses the opportunity to join the conversation, asking the patient about his living circumstances (line 76).  The ward sister does not usually question patients during ward rounds; generally consultants interview patients about their condition and feelings (Leino et al. 2017). WS's interpolation here, asking whether the patient lives alone (line 76), clearly served to draw attention to the fact that he would have to manage everything independently if discharged from the hospital, a difficulty that was further highlighted when the WS elicits the information that not even occasional help would be available (line 79).
Similar to the earlier intervention by the nurse, WS's questions revealed information that implicitly contradicted C's original positive assessment of the patient's condition and capabilities. However, this contradiction was not stated explicitly; the indirectness that characterised N's intervention is similarly present in WS's turns. On the surface, she simply asked a question and a follow-up question about the patient's living circumstances, without further commenting on the answers. However, the implication of her questions that the patient would be on his own after discharge from the hospital conveyed scepticism that the patient would be able to manage the ostomy, and his daily life, independently. This kind of strategic use of questions has been shown to be the means by which nurses 'influence and manage interprofessional interactions in a manner that is diplomatic and polite' (Arber 2008(Arber : 1331, see also Caronia and Saglietti 2018: 6).

Fifth phase
By implying his ability to cope on his own at homefor example, mentioning help from his neighbourthe patient displayed his recognition of the scepticism implicit in WS's questions. In his response in lines 81-82 and 91, he attempted to transform and push back against the negative implications of the WS's question, by demonstrating that he would be able to cope. At this point, N intervened once again, further describing problems but without being explicit about the consequences of these problems. Whereas before the issue was whether or not the patient can manage his daily life at home, now N moves back to specifically medical matters, by describing the problems with the patient's ostomy.
Extract 5. N begins her turn with the particle kyl, implying some contrast to the prior talk (Hakulinen 2001). Even though the patient portrayed himself as capable and functional, N introduced a problem that is beyond the control of the patienthis digestion. She reported that the patient was processing neither medication nor food nor drink, as is evident from all his intake appearing unprocessed in the ostomy (lines 92-96). To emphasise her point, she used extreme case formulations (Pomerantz 1986): the pronoun kaikki ('all') refers to a maximal quantity, and the clitic particle ki ('even') highlights that even the berry sauce is unaltered.
Thus, in her second intervention, the nurse shifted the focus of her observations from practical matters associated with managing the ostomy (leakage and emptying) to the more significant medical problem that the patient is no longer processing food and medication. Her second intervention was nevertheless as indirect as before: she did not say anything about the discharge, but simply reported what she has seen, leaving it to C to draw the necessary conclusions.
This intervention by N initiated a transition that is reflected in the participants' interactional space (see Mondada 2009); the consultant turned away from the patient's bed to walk back toward the computer (line 99).

136
C: ni t€ aytys pikkasen varovaisemmin ottaa t€ a€ a kotiutus = ni so we'd have to take a bit more carefully this discharge from hospital= 137 me voitas j€ arjest€ a€ a sill€ a tavalla et ton kirurgisen sairaalan we could arrange so that via that surgical 138 kautta, hospital, 139 (0.5) 140 C: meil on ((kadun nimi)) ((kaupunginosan nimi)) on toi samaa we have in ((street name)) ((part of town)) it's the same 141 yksikk€ o€ a. The consultant referred to taking the patient's discharge from hospital a 'bit more carefully' (line 136), thereby not making explicit that he will not be discharged. He then refers to measures that might alleviate the problem the patient is experiencing with his ostomy (lines 140-142), and finally constructs this in such a way that it is something happening 'today' (line 143). The eventual decision is not to discharge the patient that day, but instead to move him to another ward (lines 110-112). This changed decision is represented to the patient as consistent with the initial decision (see Drew 2003 on constructing consistency between quite different versions).
Discussion: the role of nurses in correcting a consultant's decision In this episode, a decision made during a hospital ward round to discharge a patient on the same day was corrected and changed; it came to be agreed that the patient will not be discharged, but will instead be transferred to a different (surgical) ward. The initial decision (to discharge) was made by the consultant. The intervention by nursing staff had a decisive impact on the decision outcome; through information to which they drew attention, they implicitly 'challenged' the consultant's decision which was subsequently reversed. Through communicating in this indirect way, nurses brought their knowledge of the patient's condition and progress to bear on the outcome and safety of the eventual decision.
It is evident that the patient had experienced difficulty with leakage from the ostomy, a problem that C assessed initially as minor and temporary. It is clear that the nurses were doubtful about that assessment. WS pointed to and showed gesturally the extent of the soiling of the patient's bed. She and N shared a collusive moment of mutual gaze and recognition that the problem was more significant than C allowed. This collusion between them, their mutual scepticism, suggests that they can in some respects be considered asand act asa team (Goffman 1961); they share an assessment of the patient based on their first-hand observation of the patient's condition, over several days. At any rate, there is direct evidence that the nurses did not share or agree with the assessment on which C based his decision (second phase).
While the nurses played an active role in correcting the assessment of the patient's condition and thereby changing or 'correcting' the decision about his treatment, they did so through means that can best be described as indirect. They did not explicitly or directly correct the consultant or his decision; they did not explicitly question (i.e. doubt) the initial decision, nor did they say that C's assessment is incorrect. Indeed, they avoid correcting the otherthough they initiated a sequence at the end of which he 'corrected' himself (i.e. came to a different decision) (Schegloff, Jefferson and Sacks 1977). The nurses' participation in this ward round consultation was characterised by several forms of indirectness, including an 'incomplete' turn suggesting that from N's perspective there is more to tell. Only when prompted by the consultant to continue did N reveal information from which a different, and discrepant, conclusion or assessment may be drawn. Part of the indirectness of this practice, that is providing discrepant information, is that N did not draw any conclusion from this information, nor did she make explicit that it is discrepant with C's assessment. She left it to the other to draw conclusions from her account of the patient's problems with his ostomy.
Other interventions by the nurses that worked similarly to guide C towards a different assessment of the patient's condition are WS's questions to the patient about his living alone. The consultant had been leading the patient in the direction of agreeing that he was (had been) able to cope with the ostomy at home, to which WS responded with questions that revealed that the patient lives alone with no homecare (lines 76-81). The patient resisted that by emphasising that he had other help. At this point, when an assessment seems poised in favour of the patient's ability to care for himself independently, N intervened again (lines 92-96) to provide information that left it to C to recognise the negative implication for his assessment of the patient's readiness to be discharged (again, N and WS act as a team). She did not explicitly contradict or question the consultant's assessment, nor overtly disagree with him. By just 'reporting the facts', she employs a practice, 'reporting' (Drew 1984), that left it to him to recognise and act upon the implication of what she has reported, resulting in his changing his decision.
A feature of these interventions that further contributed to their indirectness is that they manage to move the consultant towards arriving at a different decision in what might be regarded as a stepwise manner (Jefferson 2018). It became clear only towards the end of this sequence that the ultimate reason for the nurses' concern about discharging the patient is evidence that he was unable to process his medication and food, and therefore that his condition had most likely deteriorated. Whatever information was discussed before the consultant saw or spoke to the patient did not alert him to this dysfunctionality. The nurses could 'see' this dysfunctionality from what 'comes through the ostomy unaltered' (line 93). However, N did not introduce this matter at the first opportunity. She responded initially to the difficulties experienced with the ostomy bag. In some respects, this was fitted to the patient's own concerns, to the reservations he appeared to have with the consultant's initial assessment and his surprise with the consultant's decision that he could go home today (e.g. P's repeat of 'today'). N's stepwise approach consisted, then, in first addressing the matter of the problems with the ostomy bag. Only when she was not successful in forestalling the decision to discharge the patient did N introduce, again indirectly, the more significant underlying probleman inability to process food and medication.
There is a further aspect of this stepwise approach which, though more speculative, is worth mentioning. The nurse's account of the difficulties with the ostomy (lines 34-41) was, as noted above, somewhat fitted to or in line with the patient's own expressed concerns. In that respect, they are not 'news' to the patient, who did not, therefore, need to be guarded from this information. The later information in lines 92-96 about the 'medicine coming through the ostomy unaltered' is more serious in its medical implications about the patient's condition. By giving only this information, through to the 'berry sauce came out the same colour' (line 96), and without making any assessment, the nurse managed simultaneously to avoid alarming the patient, and to leave it to the consultant to determine the implications for the further treatment of the patient.

Conclusion
The indirectness of the nursing staff's interventions in this interaction with a consultant is evidence for their orientation to their position and role in the medical hierarchy. Their role is hierarchically asymmetric with that of the consultant, given that the consultant has stronger deontic authority, that is has the right to determine other's future actions (Stevanovic and Per€ akyl€ a 2012: 297). This asymmetry presents barriers to nurses' participation in the ward round, barriers consisting of constraints in participation rights, associated with decision-making in multi-professional interactions. Nurses' orientation to their asymmetric hierarchical role and the associated constraints in participation rights are manifest in the indirectness of the practices through which they intervened in such a way as to avoid directly contradicting or correcting the consultant, leaving it to him to recognise and act upon the implications of their accounts.
While directly correcting another participant is generally dispreferred in interaction (Schegloff et al. 1977), the norms against correction are all stronger in hierarchical, multi-professional teams. The principal finding here is that the nurses in this episode had the communicative means to be relatively active in participating in decision-making. Those communicative means involved avoiding explicitly challenging or directly correcting the consultant's emerging assessment/decision (see also Schoeb et al. 2018). The fact that the nurses resisted implicitly and indirectly can be seen as evidence of their professional competence, and their role as advocates on behalf of patients: they mobilise their knowledge to influence the decision about the future action but do so in a way that takes into account the professional positions of the participants. While resisting the doctor's decision, the nurses oriented to their institutional roles and the constraints associated to those roles, thereby constructing their professional competence as nurses.