of Family Medicine and General Practice

com-Abstract Objective: To describe the communication or verbal behavior of physician-patient-companion


Introduction
The cornerstone of general practice is the consultation about which much has been written, from psychology to sociology, psychiatry and anthropology, but focuses on an encounter between two people: the patient and the physician. In practice, a third person (companion, who is usually a family member) frequently accompanies a patient during medical encounter triadic consultation), but nevertheless reviews about the presence of a com- panion of the patient in consultation are rather scarce in our environment. The subjective experience of disease is built by patient in the family context and is expressed in the medical consultation, often, with the presence of a companion of the patient [1,2].
A second adult-usually parents or husband or wife accompanying the patient consultation is always significant and deserves the attention of the doctor. Although many consultations occur only on a patient, others of them involve companions with the patient in the office [3]. However, conventionally, physician training focuses on an encounter between two people: the patient and the physician (dyadic consultations). But, in practice, a third person (companion) frequently accompanies a patient during medical encounter [4].
If little is known about the presence of the patient's companion in the family medicine consultation, the less we know about the dynamics of communication in triadic interviews, which has been tried to measure on the basis of verbal participation questionnaires [5], as well in relation with some aspects of patient-centered communication [6], as well Using the Rotter interaction analysis [7], or by studying different aspects of the bio psycho social model in the clinical interview with the patient and his family [8], or by measuring whether family history and family problems were being taken into account during the visit [9]. Physicians use a wide range of family interviewing approaches, and, despite these attempts; clearly, much research is needed in this area [10,11].
In this context, we carried out a qualitative and quantitative study with the objective of understanding, describing, and measuring the verbal behaviour of medical-patient-companion in triadic consultations, and of physician-patient in dyadic consultations, and to assess their similarities and differences with the hypothesis That the communication or verbal behaviour among the actors present in the consultation is different in the consultations with companion (triadic) vs. In consultations without a patient's companion (dyadic), and these differences could have important implications in clinical management and interpersonal relationship.

Designing and variables
A qualitative, observational, narrative study was conducted during the months of November and December 2016, through the audio recording of the consultation, and verbal content analysis of the interviews.
A proposed Open University classification, though based on similar principles to Bales and Flanders, was used. It was chosen from other existing ones because it is a much simpler system and consist an abbreviated sys-tem devised originally to study management skills and behaviour by the Huthwaite Research Group [12]. Six categories were proposed to classify verbal behaviour, as follows: 1) Proposing: a behaviour that advances a new concept or suggests a course of action; 2) Supporting or agreeing: a behaviour that includes a conscious and direct statement of support or agreement with another person or their concepts; 3) Disagreement: a behaviour that involves a conscious and direct statement of difference of opinion, or criticism of the concepts of another person; 4) Giving information: a behaviour that offers facts, opinions or clarifications to other individuals; 5) Seeking information: a behaviour that seeks facts, opinions or clarifications of another individual or individuals; 6) Building: a behaviour that extends or develops a proposal that has been made by another person.
Other variables that were collected: age (del acompañante y del paciente), sex (the companion and the patient), and time in minutes of the consultation.
The study is descriptive in its approach since it aims at describing an existing phenomenon and it is qualitative in nature although the study uses a quantitative method for data collection.
In all cases the doctor was the same professional, a family doctor who remains in the same consultation for over 25 years.
The location was a family medicine office, in the Health Centre Santa Maria de Benquerencia, Toledo, Spain, which has a list of 2,000 patients. Patients of both sexes over 14-years-old were included (In Spain family doctors attend patients over 14-years-old).
Companion was defined as any person who accompanied the patient in the consulting room or that consult instead the patient. Patients were included only one time, thus, were excluded the repeated consultations of same patient, including only the first visit.
Also excluded were interviews in which the patient was not present (his or her companion was alone), when there were more than 1 companion with the patient (since it made the verbal analysis very complicated by the interference of one another), emergency consultations, and the phrases of courtesy of initial and final greetings were not included in the content analysis of the interview.

Sample
A non-random sampling, intentional -of conven- ience -was carried out by the investigators. A suitable sample number was considered when saturation occurred, i.e. no new data were obtained. Sampling adequacy evidenced by saturation means that sufficient data to account for all aspects of the phenomenon have been obtained. By definition, saturating data ensures replication in categories; replication verifies, and ensures comprehension and completeness [13]. Other studies of the same subject have included similar sample sizes [6,14].
Because of the study is initially qualitative, the sample is not intended to generalize results, but to delve into them, to understand them. The criterion of maximizing the diversity in obtaining the sample was taken into account, and all types of interviews were included, with the widest possible situations.

Ethic aspects
The informed consent of all patients and companions for using of data in research was obtained.

Analysis
The interview was recorded in audio, and later transcribed to Microsoft ® Word. With the written text of the interview, their content was subsequently analyzed, classifying the interaction in the doctor-patient interview or in the physician-patient-companion group, according to the 6 proposed categories [15,16].
Once the qualitative study is completed, the results of the number of behaviours in the total of triadic and dyadic consultations are presented in a quantitative way, without probability equations because the sample was not probalistic and so it cannot be apply any theory of this discipline. But the sample is reasonably no different from the one randomly selected. Besides, it would be very strange if there were individuals over represented or absent in the sample studied. So, with all cautions, quantitative results could be extrapolated to the population of family medicine consultations [12].

Control of validity and reliability of the study
Internal validity (Credibility): A qualitative study has internal validity when it is credible. We present a description and interpretation of a human experience or phenomenon such that people who live that experience immediately recognize the descriptions and interpretations as their own.

External validity (Applicability):
Although it is not an objective of the qualitative studies, that do not pretend to generalize findings, to the extent that the researchers have respected the criterion of maximizing the diversity in obtaining the sample and make a detailed description of the Context and participants, the findings may be applicable in similar contexts.

Technique to control bias
Triangulation: (it is to get different perspectives of the phenomenon studied).

Among different evaluators:
The written transcripts of voice recordings of the interviews were read by the research team to reach agreements on the categories that were used. The process was as follows: 1) Each researcher made several individualized readings of each interview, obtaining the frequency of each of the 6 categories, until obtaining a file of categories of each journal.
2) A group agreement of categories was made for each interview -assigning phrases of original data to new categories -and thus forming a file of definitive categories for each interview. To facilitate this process, we worked from the categories of a researcher, which were compared with those of the rest in group work. And finally 3) The results were interpreted.
Triangulation of data: 10 primary health care interviews were randomly selected with simulated patients [17,18], available on social media of YouTube ™ , the videos for the purpose of coding were transcribed and the same methodology was applied of codifying them in the 6 categories, to value agreements or disagreements between actual consultation interviews and videos of Internet. In this way different perspectives of the studied phenomenon were obtained using different data of investigation [19,20].

Results
From November 23 to December 2, 2016, 33 unaccompanied consultations (dyadic consultations: physician-patient), and 19 consultations with companion (triadic consultations: physician-patient and companion) were recorded in audio. When analyzing the content of the verbal dialogue, it was observed that since the interviews numbers 7-8 of each group, no new results appeared, so the sample was considered saturated with 20 interviews (10 interviews in each group).

Triadic consultations
In the 10 triadic consultations, 254 verbal phrases were included that indicated behavioural styles in the doctor, 216 in the patients, and 99 in the companions (total 569). In the triadic interviews, "Giving information" (38%), "Searching for information" (22%) and "Proposing" (20%) predominated in the doctor. In patients and companions were predominant, "Giving information" the doctor and 198 of the patients were included. In dyadic interviews predominated, in both physicians and patients, "giving information" (37% in physicians and 50% in patients) ( Table 1). The mean age of the patients was 46 years (Range: 32-62). There were 5 females and 5 males. The mean duration of dyadic consultations was 6 minutes (range: 3'-11 ').
So, in the consultation with a companion (triadic) there are more behaviours of "Supporting or agreeing" on the part of the patient and the companion, and in the doctor more of "Proposing" than in the dyadic consultations (Table 1). Triadic consultations were of a slightly longer duration (8 minutes vs. 6 minutes on average). Table 2 and Table 3 present some "verbatims" (literal phrases) in relation to the behaviours of each actor in triadic and dyadic consultations. Table 4 presents the quantitative data of the triadic consultations (physician + patient + companion) globally vs. dyadic consultations (doctor + patient).

Dyadic consultations
In the 10 dyadic consultations 210 verbal phrases of   Table 5 presents the comparison of verbal behaviours in patients between triadic and dyadic consultations.    The conceptualization and physician training focuses on an encounter between two people: the patient and the physician. In practice, a third person (companion) frequently accompanies a patient during medical encounter. However, there is a high prevalence of the presence of companion (25% of the interviews are with companions) [1,[22][23][24][25]. Besides, the presence of the companion of the patient in consultation is a metaphor from the patient [26]. The companion shapes the experience obtained by the patient's physician. A second adult accompanying the patient in the consultation is always significant and deserves the attention of the doctor [2]. In the consultation there are always three sites: the doctor, the patient and the family (physically present or not). This triadic relationship can be drawn as shown in (Figure 3). But previous research on communication in the interview has focused primarily on "dyadic" consultations between physician and patient [27,28], and attempts at Investigation of the dynamics of communication in consultations in which the interview is with "triads" in family medicine are scarce and partial [5][6][7][8][9][10][11]29].
The group communication dynamics that are de- Table 6 presents the comparison of the physician's behavior in triadic and dyadic consultations. There is a greater frequency of "Proposing" behaviors, and a weak decrease in "Supporting" in triadic consultations.

Results of data triangulation
When performing triangulation with videos of queries taken from the Internet, which included six dyadic and three triadic queries, it was found that it was feasible to apply the same coding system. The results indicated: • In triadic consultations, patients have more behaviour to supporting or agreeing than in dyadic • The companions (in the triadic ones) act giving information • The doctor in the triadic consultations has more behaviour to give information It should be borne in mind, that they are simulated interviews, and that there is no effect of continuous care, which does occur in the interviews of the consultation, which may cause "less information" to be sought in the latter, being known by the doctor.
When comparing the results of the interviews of the actual sample of the query, with Internet videos of simulated queries, it is observed that the Internet results validated those obtained in the sample of the actual consultation.

Discussion
The clinical interview is an essential competence of the family doctor and communication a key piece in the doctor-patient relationship. Health care and its outcomes depend on how the professional and the consultant communicate, since profitable communication is a major component in health recovery [21].   However, we found a slightly greater number of verbal behaviours in triadic interview patients (216 verbal behaviours of patients in triadic consultations vs. 210 in dyadic consultations) ( Table 4 and Table 5).
And it may be interpreted that longer consultations are associated with more adequate diagnoses, at least in psychological problems [41]. Older patients have been reported to pose fewer topics in all content areas (personal medical habits, psychosocial sphere, and physician-patient relationship) in triads than in dyads. Patients may be less assertive and expressive in triads than in dyads. That is, patients can be excluded in conversations at visits where a third person is present [42].
The doctor-patient discourse is characterized by the asymmetry that emerges in the encounter, the distribution of turns in the conversation and the type of language strategies that are used. While the doctor has at his disposal more linguistic alternatives to participate in the speech, the patient has more restricted possibilities of using the same linguistic resources in that event. The physician, in his role, searches for information about the patient, makes a diagnosis, proposes and evaluates a treatment. The patient tends to give personal information about his psycho-social identity and to initiate, according to the need and possibility, subjects that he considers relevant. Our study reproduces this scheme, finding more behaviours of "Giving information" and "Supporting or agreeing" on the part of patients and companions, and of "Proposing" by the doctor.
Michael Balint suggested that the patient begins the consultation by offering one or more problems and issues to the doctor. The doctor respond to these offers, indicating his acceptance or rejection of them, until some kind of compromise is worked out. In consequence, while the matters that are discussed in the consultation reflect the problems presented by the patient, they include only those aspects that the doctor indicates that are allowable [43]. In our study, the verbal behaviours of the physician in triadic and dyadic consultations are similar, veloped in the bipartite and tripartite meetings are not identical. From a sociological point of view, these tripartite interactions could influence the type of topics covered in the event. For example, could be affected the level of involvement the patient might use in the consultation. Although the communication skills described for the physician-patient relationship might be equally relevant in any communication with accompanying family members or volunteer friends or caregivers, they may need to be applied in a slightly different way.
We find that the presence of a companion can improve medical proposals or interventions (Table 1, Table  2, Table 3, Figure 1 and Figure 2), which is in line with other authors [30][31][32][33], and this underlines the importance of encouraging companions to participate in the consultation [34], but we also find that the participation of the companion may not imply that it is a source of more information (Table 4), which has been previously reported [14,[35][36][37], and in any case at the cost of a longer duration of the consultation (8' vs. 6'), time greater than that which has been reported for the hospital setting (4 minutes and 17 seconds for each patient on the ward and 20 seconds for his or her relatives) [38], but less than other authors in general medicine (10-12 minutes) [39].
Although our results may support the fact that the participation of the companion may not imply that it is a source of more information, since the patients in triadic consultations presented the behaviour of "Give information" in 43%, and in the dyadic interviews in 50%, it should be noted that in our study the presence of continued care for years by the doctor, may be the reason that it was not necessary to look for more information, since much of it was already known.
The absence of significant differences in duration of the consultations in a specialized outpatient clinic has also been reported, suggesting that patients occupy proportionally less discourse space (measured by the number of words produced) in accompanied interactions than in unaccompanied interactions [40]. The content of the interviews was not collected: Only the class or behavioural style of the interaction process. The conceptualization of the disease is another point of interest that can vary between participants, doctor, patient and companion. A different understanding of the origin of a disease, for example, can cause communication problems and lead to misunderstandings.

Silences have not been collected:
We could interpret the silence that the patients assume, in some segments of the conversation, is a reflection of the emotional load they feel with the situation they are living. But when picking up the audio recording it was not possible to know when the patient's voice was silenced.

Non-verbal communications have not been collected:
Verbal and nonverbal messages should be studied as inseparable phenomena they happen together [46].

Conclusions
Although the concepts of measuring and understanding are sometimes opposed [47], we try to do both in this study. Thus, in short, we find that the triadic vs. dyadic consultations 1. They are discreetly longer lasting.
2. There are more agreements and more initiatives are proposed. except for a higher frequency of "Proposing" behaviours and a weak reduction of "Supporting" in triadic consultations (Table 6).
Effective and empathic management of triadic communication that avoids unnecessary disruption and frustration requires specific communication skills (e.g.: rules and order of conversation). All this said emphasizes that triadic communication in medical encounters can be useful, but it is difficult. Our results suggest some preliminary strategies for health professionals to increase the utility of physician-patient-companion communication (Box 1) [44,45].

Limitations of the study
Coding difficulties: There may be difficulties in sorting, since in one sentence more than one category can be observed. A disagreement phrase can also be a proposal, and a proposition can also give information. Many phrases can be considered either as giving or requesting information, but once the technique has been learned, and biases have been avoided through the triangulation of researchers, it can provide useful data on the behavior of individuals and groups [12].

Difficulties of transcription from audio to text:
Doctor, patient and companion -contribute to the discourse, and thus present overlaps, juxtapositions and narrations together, and these markers of joint production are difficult to transcribe and codify.

Difficulties in the role of patient and companion:
In the triadic interviews, the companion, not infrequently consults for herself or herself, and there is an inversion of roles: the one who was a companion becomes a patient, and vice versa. This situation complicates the coding of behavioural styles.
The question of representativeness of the sample: How "typical" was the behaviour of the participant? Will the "typical" activity be found if the sample were larger? Is the activity collected subject to fluctuations? Interviews were recorded on normal consultation days, communication was not subject to fluctuations, and we thought that by maximizing the diversity of the participants, they represented the patients and companions usual of the consultation. It may be thought that the size of the sample is small, but in qualitative studies this usually is small, and sample size was given by the saturation of the data, and was similar to that of other studies of the same subject [6,14]. Once the qualitative study was completed, the results of the number of behaviours in the total of triadic and dyadic consultations were presented in a quantitative way, without probability equations because the sample was not probalistic and so it cannot be apply any theory of this discipline. But the sample is reasona-