1 Introduction

The aim of this article is to present some preliminary findings obtained during my research in the field of doctor-patient interaction [1]. It concentrates on the role of questions and answers, especially how they contribute to asymmetry [2], the characteristic feature of the relationship between doctors and their patients (Mishler 1984, Todd and Fisher 1983, Müllerová 2002). Moreover, it discusses some differences and obscurities in previous research (e.g. Korsh, Gozzi, Francis 1968, West 1983, Ainsworth-Vaughn 1998), and offers a possible explanation. Although more studies touching the topic have been published in recent years (e.g. Gwyn 2002, Humphreys 2002), my work, as it will be explained below, differs both in the material and methodology.

2 Corpus

The corpus under investigation is a collection of authentic materials recorded by Maria Györffy (2001) in consulting rooms all over Great Britain and the United States. It is divided into 13 units, each of them based on a different medical speciality. For the purposes of my analysis I have chosen five of them, namely Internal Medicine, Obsterics and Gynaecology, Paediatrics, Oto-rhino-laryngology and Orthopaedics. Each speciality comprises ten dialogues, which means there are 50 dialogues under examination, the total number of turns being 725. The advantage of such corpus is that it gives an opportunity for so called "communicative comparison" [3], i.e. a comparative analysis of D-P dialogues in different medical specialities, which, as I hope, could reveal a more complex picture of doctor-patient inteviews.

3 Method

According to Rolf Wynn, there are two major approaches to doctor-patient relationship research: "...the sociolinguistic approach (of which ethnomethodological conversation analysis is quite typical) and the medical approach (of which Bales' Interaction Process Analysis and Byrne & Long's method are typical)." [4] In his book The Linguistics of Doctor-Patient Communication. An analysis of the methodology of doctor-patient communication research (1995), Wynn comments on the strengths and weaknesses of both of them, and concludes his analysis by suggesting a combination of the two methodologies. I take his point of view and try to combine the qualitative aspects of conversation analysis (e.g. the interpretation of questions and answers) with the quantitative aspects of the medical (psychological) approach.

4 Analysis

From 50 dialogues consisting of 725 turns I have excerpted 374 questions. In accordance with the research which has been done before (West 1993, Humphreys 2002), I understand question as "...any utterance requiring a response within the context of the interaction, regardless of form (interrogative, declarative, etc); any formal question; and any utterance that receives a response as though it were a question" [5] and I categorize questions into three groups, namely yes/no questions (Y/N), either/or questions (E/O), and open questions. Furthermore, I distinguish doctor-initiated and patient-initiated questions, and classify them according to their place (phase) [6] within the dialogue. My findings are offered in the form of tables and figures.

Tables 1 and 2 present absolute and relative frequencies of the distribution of questions between doctors and patients. They also give information about the number (percentage) of questions in particular medical specialities with regard to question types and phases of the D-P talk. Tables 3, 4 and 5 offer results of the F-test, which reveals how significant the distinctions between the medical specialities in this collection are [7]. I have also calculated correlations [8] between particular variables (participants, medical specialities, question types and phases). However, the only significant correlation (r=0.33) is between the participants and phases. For example, patients usually raise their questions during the phase of examination and treatment. There are no patient-initiated questions asked during the history-taking phase. The other correlations are not as important.

Table 1

Abs. frequency	Participant		Question type			Phase			Total
Medical speciality	D	P	Y/N	E/O	Open	Hist	Exam	Treat
Internal medicine	99	1	43	28	29	24	76	0	100
Gynaecology	65	1	36	8	22	0	64	2	66
Paediatrics	88	4	45	8	39	13	79	0	92
ORL	55	8	50	7	6	1	61	1	63
Orthopaedics	47	6	25	1	27	0	47	6	53
Total	354	20	199	52	123	38	327	9	374

Table 2

Relat. frequency	Participant		Question type			Phase			Total
Medical speciality	D	P	Y/N	E/O	Open	Hist	Exam	Treat
Internal medicine	99%	1%	43%	28%	29%	24%	76%	0%	27%
Gynaecology	98%	2%	55%	12%	33%	0%	97%	3%	17%
Paediatrics	96%	4%	49%	9%	42%	14%	86%	0%	25%
ORL	87%	13%	79%	11%	10%	2%	96%	2%	17%
Orthopaedics	89%	11%	47%	2%	51%	0%	89%	11%	14%
Total	95%	5%	53%	14%	33%	10%	87%	3%	100%

Table 3

F	Participant
	Internal	Gynaecol.	Paediatr.	ORL	Orthop.
Internal m.		2×10-8	10-11	4×10-25	2×10-22
Gynaecology			2×10-5	8×10-14	3×10-12
Paediatrics				2×10-5	2×10-4
ORL					0.73

Table 4

F	Question type
	Internal	Gynaecol.	Paediatr.	ORL	Orthop.
Internal m.		0.22	0.2	0.02	0.14
Gynaecology			0.73	4×10-3	0.52
Paediatrics				10-3	0.72
ORL					10-3

Table 5

F	Phase
	Internal	Gynaecol.	Paediatr.	ORL	Orthop.
Internal m.		7×10-13	0.05	10-11	0.02
Gynaecology			10-8	0.76	4×10-6
Paediatrics				10-7	0.48
ORL					2×10-6

5 Findings

The importance of questions and answers in D-P dialogues is undeniable. They "introduce, develop and dissolve topics" [9], and help to reach the correct diagnoses and treatment. Statistics I have tried to elaborate show that as many as 649 turns (90 %) out of 725 are formed solely by questions or answers. Out of 374 questions, 354 (95%) are initiated by doctors, only 20 (5%) are initiated by patients. 199 (53%) belong to Y/N questions, 52 (14%) to E/O questions, and 123 (33%) to open questions (Figure 1). 38 questions (10%) appear during the history-taking phase, 327 (87%) during the examination phase, and 9 (3%) during the treatment phase (Figure 2).

Figure 1	Figure 2

Considering the findings of previous researchers (Ainsworth-Vaughn 1998, West 1983), these results are somewhat surprising. The greatest difference is in the distribution of questions and answers between doctors and patients. As West points out, only 9% of all questions in her material are patient-initiated. The relative frequency of patient-initiated questions in the corpus studied by Ainsworth-Vaughn is much higher (40%). Thus the very first problem which needs to be solved is to find an explanation for this divergence.

One might suspect that the distinction could be closely connected with the context of the particular medical speciality under examination. Patients, as nonprofessionals, are not, of course, acquainted with any of these disciplines, they lack the knowledge of medical terminology, and hardly understand the process of examination and following treatment. However, for some reasons, one may still believe that certain medical branches (e.g. Paediatrics or ORL) are easier to become familiar with than to understand, for instance, Internal Medicine or Orthopaedics [10]. As a result, patients are more confident and would rather ask more questions when visiting a paediatrician or an oto-rhino-laryngologist than when seeing an internist.

Such hypothesis would be acceptable if there was no evidence against. Unfortunately, both West and Ainsworth-Vaughn worked with collections of dialogues betweeen patients and their family practitioners, i.e. within the same medical speciality. Still, their findings were considerably different. Furthermore, the correlation (r=0.2) I have calculated in my material proves there is no direct correspondence between the medical branch and participants, and even some results of the F-test (e.g. the comparison of ORL and Orthopaedics with regard to participants, F=0.73) show that we cannot take the proposed assumption for granted.

It might also appear worth investigating whether the asymmetrical findings in the distribution of questions and answers could not be sought in the differences in sex, race and age of doctors and patients who are studied. However, according to West (1983): "Neither sex nor race (of physician or patient) seemed to influence the distribution of questions between parties." [11] She also claims that the very same conclusion works for the category of age. Thus the explanation must be looked for somewhere else.

The interpretation of D-P questions and answers in the corpus I have worked with prove that most questions patients ask take place when they really feel worried about something. Usually they must undergo an operation (examples 1 and 2) and they want doctors to explain the operation in detail. They want information about what it involves, whether it is dangerous or not, what possible complications they may expect. Simply, patients wish to hear that the surgical intervention will help them and there is nothing to fear. It doesn't matter whether such dialogues take place in consulting rooms of ORL practitioners or orthopaedic specialists. It doesn't make a big difference of which age, sex or race physicians and their patients are. The most important thing which should be taken into consideration when solving the problem of the distribution of D-P questions is the level of patient's anxiety about his/her health problems and their treatment.

Example 1:

D: You've got a deviated nasal septum. This part of your nose is cartilage, and instead of being straight it's twisted and the twist is blocking you on the left side. I'm pleased to say we can fix it for you. We can put it right with an operation to straighten up your nose, as there are nomodicenes or tablets really that will help.
P: Is it a big operation?
D: No, not too big. It's quite common. If you agree, we'll bring you into hospital the day before the operation. You can usually go home the day after your operation, or possibly the second day after that. We do it under a general anaesthetic. It's done through your nostrils, there's no cuts on your face.
P: No black eyes?
D: Not for this operation. When you wake up from anaesthetic, you'll probably have a bandage up both nostrils overnight so, you see, you'll have to breath through your mouth that night. Would you like the operation?
P: Will it work?
D: Yes, we can say that we can make things a lot better than they are now.

Example 2:

D: Mrs Wallis, I have the results of your X-rays. These show you have severe osteoarthritis of your left hip. This is due to a congenital dislocation of the hip which you've had since birth. I think the best treatment for you would be an operation to replace your left hip.
P: Tell me, Doctor, is that a major operation?
D: Yes, it is undoubtedly. But you are having so much trouble I do not think there is any other alternative.
P: I'm very worried about this, Doctor. What does the operation involve?
D: It is a major operation which would require you to be in hospital for about two weeks. You'll come to hospital a day or two before surgery so that we can examine you and check that you are fit for an anaesthetic. The operation itself involves quite a long cut on the outside of your thigh and then the worn part of your hip will be cut away and replaced with a metal and plastic joint. This should make you more comfortable, and your hip less stiff. But of course, as with any operation, there's a small risk of complications.
P: What are those complications, Doctor?

5.1 The role of doctor-initiated questions

As it was mentioned above, the doctor-initiated questions are much more numerous than those initiated by patients. Out of 354 doctor-initiated questions, 188 (53%) could be classified as Y/N questions, 52 (15%) as E/O questions, and 114 (32%) as open questions (Figure 3). 38 (11%) take place during the history-taking phase, 315 (89%) during the examination phase, and only 1 (0%) during the treatment phase (Figure 4).

Figure 3	Figure 4

Unlike patients (see below), doctors freely take advantage of using all the question types. Because the correlation between the question type and the participant is quite low (0.03), there is no significant correspondence between these two attributes of the consultation. What is significant, however, is the fact that one hundred percent of doctor-initiated questions are asked during the history-taking and examination section. Since doctors have on the average only eight minutes to "establish rapport, discover the reason for a patient's visit, verbally and physically examine the patient, discuss the patient's condition, establish a treatment plan, and terminate the exchange" [12], these two parts are solely reserved for them to gather information, to concentrate and diagnose responsibly.

We may see that there is only one doctor-initiated question asked during the treatment phase in my corpus (example 1). This question appears after a long stretch of doctor's talk about the possible treatment, and invites a female patient to raise any questions concerning her problems. As the correlation proves (r=0.33), there is a connection between participants and phases in their dialogue. The treatment section is used by doctors to explain the process of treatment or therapy, and by patients to ask some additional questions about their diagnosis and following cure (example 2).

Example 1:

D: Would you like to ask me any questions...?
P: Doctor, I wonder how I got these warts.

Example 2:

P: Is it a big operation?
D: No, not too big. It's quite common.

In addition to what has been said about the role of doctor-initiated questions during the treatment section, some recent studies mention a specific sequence of doctor's talk, which is repeatedly appearing throughout each medical consultation, especially during the treatment phase (e.g. Humphreys 2002). The sequence consists of a rhetorical question, an answer, and final intepretation. As Humphreys points out: "This Question-Answer-Interpretation sequence places the patient in a position where they could, if they wished, question not only the treatment offered, but the thinking behind it. Therefore, this strategy is significant in balancing the asymmetry between doctors and patients." [13] Because the interpretation phase is also strongly rooted in the corpus I have been working with (example 3), I may confirm her findings.

Example 3:

D: You have been exposed to the genital wart virus through sexual contact with somebody who has genital warts or has the virus in his skin and the genital area. Unfortunately, I cannot tell you how long you have had the virus in your skin or who you could have caught it from, as it does not have to be your present partner but could have been from a partner several years ago.... The warts will disappear with treatment, but unfortunately, I can't give you any guarantee that they will not return.

5.2 The role of patient-initiated questions

The distribution of patient-initiated questions in my collection is as follows. 11 of them (55%) belong to Y/N questions, no question (0%) could be classified as E/O questions, and 9 (45%) belong to open questions (Figure 5). There are no patient-initiated questions (0%) taking place during the history-taking phase, 12 questions (60%) take place during the phase of examination, and 8 questions (40%) take place during the phase of treatment (Figure 6).

Figure 5	Figure 6

There are three important points which we should take into consideration when interpreting the patient-initiated questions. The first is the fact that no questions are asked during the history-taking phase. At the beginning of the consultation, the interview is usually in doctors' hands. Physicians aim to acquire as much information as possible, and there is no place for patients to raise their questions, even if they wanted. To keep this form, physicians make only short pauses not to give patients too much time to ask. They also avoid giving them any explicit invitation to do so. According to Humphreys: "This may indicate a higher degree of consensus to the conventional roles of doctor and patient." [14]

Another interesting point to be noticed is that there are no E/O questions on the side of patients. As the sample of patient-initiated questions is very low, it is difficult to say why. In my opinion, the most probable explanation could be deduced from the type of information patients look for. Patients usually raise their questions when they do not understand what doctors would like to know, what they actually ask about (example 1), when they need some advice (example 2), or when they desire to be given direct replies to their worries and anxieties (example 3). In all these situations Y/N questions (example 3) or open questions (examples 1 and 2) are preferred. Most patients, as lay people, do not have enough medical knowledge to offer their carers two or more alternatives, and therefore it is easier and more natural for them to use the other two question types.

Example 1:

D: What is it /the phlegm/ like?
P: What do you mean?

Example 2:

P: Well, what should I do in the meantime, Doctor?
D: Well, the important thing is to make sure that he has sufficient fluids...

Example 3:

P: Will my hearing get better, Doctor?
D: Yes, it will.

The last point to be mentioned here is in a sharp contrast with the findings of Korsch, Gozzi, Francis (1968) [15] and West and Page (both in Todd and Fischer 1983), whose results show that there is a tendency to ignore patient-initiated questions. For example, according to West, only 87% of questions raised by patients in her corpus were answered. My findings differ rapidly because all questions asked by patients in this material are answered. Here I am again about to agree with Joanne Humphreys and her explanation that: "Recent social changes have altered the balance of power between doctors and patients through increased openess on the part of the medical profession and greater access to information for patients." [16] The same results of her research as mine give some evidence for such argument. However, it is necessary to add that the shape of doctor-patient interaction has not been changed in all its aspects, and that the asymmetrical relationship has been preserved.

6 Conclusion

To sum up the principal points of the present article, let me repeat the most important findings in relation to asymmetry:

1. 354 (95%) questions are initiated by doctors in this collection - asymmetry
2. Doctors use all question types - asymmetry
3. There is only one doctor-initiated question asked during the treatment phase - symmetry
4. Doctors tend to offer interpretations of their decisions - symmetry
5. Only 20 (5%) questions are initiated by patients in this material - asymmetry
6. Patients usually ask questions when they desire direct replies to their worries - asymmetry
7. No patient-initiated questions are asked during the history-taking phase - asymmetry
8. There are no E/O questions on the side of patients - asymmetry
9. All questions asked by patients are answered by their doctors - symmetry

What has been said might suggest that the asymmetrical relationship between physicians and patients has been preserved so far. However, there is a tendency to "reduce hierarchies and renegotiate roles" [17].