SEARCH

SEARCH BY CITATION

Keywords:

  • conversation analysis;
  • patient participation;
  • doctor–patient interaction;
  • explanation;
  • diagnosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

In medical clinic visits, patients do more than convey information about their symptoms and problems so doctors can diagnose and treat them. Patients may also show how they have made sense of their health problems and may press doctors to interpret their problems in certain ways. Using conversation analysis, we analyse a practice patients use early in the medical visit to show that relatively benign or commonplace interpretations of their symptoms are implausible. In this practice, which we term pre-emptive resistance, patients raise candidate explanations for their symptoms and then report circumstances that undermine these explanations. By raising candidate explanations on their own and providing evidence against them, patients call for doctors to restrict the range of diagnostic hypotheses they might otherwise consider. However, the practice does not compel doctors to transparently indicate whether they will do so. Patients also display their ability to recognise and weigh the evidence for common, easily remedied causes of their symptoms. By presenting evidence against them, they show doctors the relevance of more serious diagnostic interpretations without pressing for them outright.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

A medical clinic visit can be seen as a sense-making event. There are various ways of understanding how the participants make sense of patients’ health problems. It can be viewed as a cognitive process wherein doctors, who have a specialised stock of knowledge about physiological processes, systematically gather information about patients’ presenting problems by soliciting information about their medical and social histories and current symptoms, performing physical exams and ordering laboratory tests. In this model, doctors develop and test hypotheses as information emerges in the investigation and finally arrive at diagnoses and plans for treatment. A widely used medical textbook describes the process in the following manner:

Appropriate medical care depends on the physician’s knowledge of the patient’s abnormalities … best described as a diagnosis. Each recognizable disease possesses distinguishing features that serve as clues. So the physician embarks on two parallel courses: (1) a search for clues to develop a list of problems, leading to (2) the generation and selection of hypotheses to reach a diagnosis. Clues are sought by taking a history, performing a physical examination, obtaining laboratory tests and imaging, and ordering other special procedures. The clues that are found suggest a list of problems from which is generated hypotheses to explain the cause of the problems in terms of diseases in a list called the differential diagnosis. The clues or facts obtained during the diagnostic examination are used to support or refute each hypothetical disease in the differential diagnosis and finally to arrive at the diagnosis (DeGowin and Brown, 2000: 2).

Social scientists, however, have argued that this sense-making process is more than a cognitive one; it has many social dimensions. Studies have shown that a number of socio-demographic factors, such as the participants’ gender and race, influence doctors’ interpretations of patients’ symptoms (Cooper-Patrick et al. 1999, Lorber and Moore 1997, Lurie et al. 1997, Rathore et al. 2000, Rathore and Krumholz 2004, Smedley et al. 2003). Moreover, patients as well as doctors try to make sense of illness; investigations of lay understandings of illness reveal that patients draw upon culturally prescribed models to make sense of their own health problems (Angel and Thoits 1987, Blaxter 1983, Hunt et al. 1989, Keller et al. 1989, Kleinman 1981, 1988). Some social scientists have also argued that it is important to examine social interactions between doctors and patients to see why certain interpretations of illness prevail. In some classic studies, sociologists have called particular attention to the socio-political dimensions of these interactions, such as the ways in which institutional identities and attending asymmetries in power and authority infuse the conversations and lead doctors to subordinate patients’‘life-world’ understandings of illness in favour of biomedical interpretations (Mishler 1984: 6; also see Cicourel 1983, Waitzkin 1979, 1991). Although diverse in nature, these social scientific studies have treated the sense-making process as a social process. Specifically, they have focused on the influence of external forces on sense-making in medicine and they have treated doctors’ and patients’ interpretations of illness as products, outcomes or results of these forces.

This article has emerged from a different stream of social scientific research on medical sense-making. Rather than focusing on the social forces that influence doctors’ and patients’ interpretations of illness, this body of research uses conversation analysis (CA) to investigate the concrete social practices the participants themselves use in interaction to display and negotiate interpretations of what patients’ symptoms mean. Numerous CA studies have revealed that the sense-making process in clinic visits is not only a social process but is one with its own orderly features. It is accomplished in and through the participants’ collaborative engagement in an array of actions, such as soliciting and providing information, which figure in broader activities such as testing diagnostic hypotheses. The ways in which these actions are performed can shape what information emerges in the visit, when it emerges and how it emerges, all of which has implications for what can be discovered about the symptoms in question. For example, when patients provide information to doctors during clinic visits they may do much more than convey information about how they are feeling: they often slant this information, meaning that they take positions vis à vis their problems and what is causing them to occur. Furthermore, when doctors and patients offer and respond to each others’ interpretations of symptoms, they can perform a diverse array of social moves in addition to showing how they have made sense of illness, including inviting or pressing the other participant to interpret the illness in a certain manner, casting certain interpretations as probable or improbable and positioning themselves as reliable and authoritative sources of medical knowledge.

Our investigation is part of a subset of CA research that focuses on the interactional resources that patients (and their representatives, such as parents of paediatric patients) employ during clinic visits to display their interpretations of illness and thereby draw their doctors’ attention to particular candidate explanations, cast themselves and their illnesses in a particular light, push for and resist particular interpretations, and in various ways suggest how doctors should view and handle the problems at hand. In this article we analyse a practice patients use to draw doctors’ attention to ways in which their problems could be interpreted and portray these interpretations as implausible, thereby suggesting to doctors that it would be relevant to pursue other diagnostic angles.

Background

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

Outpatient medical clinic visits generally proceed in phases that encompass different tasks or activities. Doctors collect data (information about patients’ problems and symptoms) through verbal and/or physical examination, analyse these data and present their diagnoses, and then offer advice about how the problems should be treated or managed (see Byrne and Long 1976). These activities occur over the following six phases: (i) opening; (ii) presenting complaint; (iii) examination; (iv) diagnosis; (v) treatment; (vi) closing (Heritage and Maynard 2006: 14). These phases offer different opportunities for patients’ participation in the sense-making enterprise and they impose various constraints upon that participation. In the following section, we provide a background for our study by reviewing the literature on phase-specific resources that patients employ to display their interpretations of illness and accomplish a variety of actions, including exerting pressure on doctors to interpret and treat their problems in particular ways.

Patients’ responses to diagnoses

The environment following the delivery of a diagnosis1 provides patients (and/or their representatives, such as parents of paediatric patients) with opportunities to display interpretations of their health problems. After the diagnostic informing the patient may indicate alignment with the doctor’s interpretation by providing an agreeing assessment (Peräkylä 2002, 2006). Doctors also treat patients’ minimal acknowledgments (yeh’), continuers and silence2 in this sequential location as an indication of alignment and willingness to move to the next phase of the visit: management or treatment of the condition (Peräkylä 1998, 2002, 2006, Heath 1992, Robinson 2003, Stivers 2005, 2006).

Patients also may also challenge doctors’ diagnostic interpretations in this location. They may disagree outright (Gill and Maynard 1995, Maynard 2003), or suggest diagnoses that are different from the ones that the doctors have offered (Peräkylä 2002). Patients may also resist doctors’ diagnoses in a tacit manner by reporting information about bodily states and symptoms that are inconsistent with these diagnoses. For example, patients may report that they have symptoms that are discrepant with the diagnosis (Peräkylä 2002, 2006) or that their symptoms have changed (Heath 1992), or they may describe a state of suffering that is incongruent with the condition that has been diagnosed (Heath 1992). By positioning these reports just after doctors’ diagnostic informings, patients invite their doctors to hear the reported circumstances against the contextual backdrop of the diagnostic news, and thus to hear the reports as diagnostically relevant – specifically, as evidence that the diagnoses are incorrect and should be reconsidered. When patients display disagreement with diagnoses in the post-informing environment they encourage doctors to delay the progression of the visit to reconsider their diagnoses, rather than move on to the management or treatment of the condition (Heath 1992, Peräkylä 2006).

Frontloading: patients’ displays of sense-making before doctors deliver diagnoses

Patients may also frontload the medical visit, meaning that they may present their own interpretations of their problems in early phases of the visit where doctors are, broadly speaking, collecting information. This environment provides patients with a number of opportunities not only to describe their problems but to show how they have made sense of them. By offering their own candidate explanations for illness early on, patients can draw doctors’ attention to potential causes for their illnesses and hint, suggest, or forthrightly ask doctors to consider their relevance during the visit (Gill and Maynard 2006).

Patients routinely offer causal explanations for illness as they present their problems and in the course of replying to doctors’ questions about them (Gill 1998, Gill and Maynard 2006). Some frontloading initiatives (for example, ‘Do you think it’s X?’) make it relevant for doctors to reveal their own sense-making in return – specifically, to respond by assessing the explanations. The design of other frontloading initiatives provides for a wider range of doctors’ responses. For example, patients may speculate about possible causes of their problems, they may sandwich their explanations within multiple-component turns or they may juxtapose symptom descriptions with reports of circumstances in ways that imply causal connections. Using these types of practices, patients can hint to doctors that they should investigate particular candidate causes while they are working up their problems, and do so without calling for them to assess these causal explanations then and there. That is, these initiatives do not firmly establish the conditional relevance of doctors’ immediate assessments, such that they would be ‘officially’ or ‘notably’ absent if they did not occur (Schegloff 1968, 1972: 76). Rather, they allow doctors to continue gathering information about patients’ symptoms with patients’ candidate explanations (possibly) in mind (Gill 1998, Gill and Maynard 2006).

When patients frontload the visit they also may be engaged in the more general work of showing that they have the types of problems that warrant medical visits (Heritage and Robinson 2006) and should therefore be taken seriously. Additionally, they may be working to cast themselves in a particular light – for example, as ‘reasonable’ patients who can be trusted to take an appropriate stance toward their bodies –‘neither too lax, nor hyper-vigilant’ (Halkowski 2006: 90). When patients provide their own interpretations of illness in the early phases of the visit they also may be angling for particular treatments. For example, in paediatric visits parents may offer candidate explanations at the beginning of the visit to suggest not only that their children’s problems are worthy of medical investigation but that they also worthy of treatment such as prescriptions for antibiotics (Stivers 2002).

In this article we reveal another dimension of frontloading in medical visits. We have found that in early phases of clinic visits, patients not only put forward interpretations of illness but also occasionally exhibit resistance to particular ways in which their illnesses could be interpreted. While intuitively it might seem that this would have to occur in response to doctors’ diagnostic informings, this is not always the case. Our analysis focuses on a practice patients use to discount commonplace interpretations of illness pre-emptively, during the phases of clinic visits where doctors are gathering information about their problems and symptoms. In this practice, the patients themselves raise candidate explanations for their problems and then report circumstances that call the explanations into question, showing doctors that these interpretations are unlikely to be worthy of investigation. By working to block certain avenues, the patients also suggest that it would be relevant for their doctors to look elsewhere for the source of their problems.

Data and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

Data for this study are videotapes of primary care medical consultations in the USA. We drew upon three sets of data recorded in three different decades in three different clinics: one set was collected in the late 1980s in an internal medicine clinic in a mid-western teaching hospital, the second was collected in the early 1990s in an ambulatory clinic located in a teaching hospital in a large eastern city and the third was collected in the early 2000s in a family practice clinic located in a mid-sized eastern city. The three collections include 50 consultations in total. All recordings were made with the participants’ informed consent. We use conversation analysis (CA) to analyse the data. CA has intellectual roots in the discipline of ethnomethodology (Garfinkel 1967) and shares with that discipline a fundamental interest in culturally available practices, understandings and other resources that members of society draw upon to achieve and reproduce a known-in-common social order. Conversation analysts describe how members employ these shared resources in interaction to accomplish a range of social actions and activities (see Clayman and Gill 2004, Heritage 1984, Pomerantz and Fehr 1997, Sacks 1992). Following this approach, we closely examined the recordings of the medical consultations along with detailed transcripts of the recordings and analysed how the participants used and responded to the practice of pre-emptive resistance during the consultations.

Organisation of the practice

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

The practice of pre-emptive resistance is organised in the following manner:

In an environment in the medical visit where the doctor is gathering information about the patient’s health problem and the patient is introducing or describing the problem or symptom:

  • 1
     Explanation: the patient raises a candidate explanation (X), showing that he or she has already considered the possibility that X causes the problem
  • 2
     Counter evidence: the patient resists the candidate explanation (X) by
    •   (i) reporting circumstances that serve as evidence that X is not the cause of the problem, or

    • (ii)
         implying there is no evidence that X caused the problem.

This is the base, or generic, practice. The patient may also add an upshot that explicitly rules out X:

  • 3
     Upshot: the patient asserts that X is not the cause of the problem.

Any particular example of this practice can be placed on a continuum in regard to how explicitly the patient resists the candidate cause that he or she has raised. The most explicit examples include an explanation, counter evidence in the (i) format and an upshot. Those that are the least explicit include an explanation and counter evidence in the (ii) format. In the article we consider examples across the spectrum.

We term this practice ‘pre-emptive’ because of its location in the medical interview: it occurs in advance of diagnostic informings, in environments where doctors are still gathering information about the nature of patients’ health problems. Patients make use of the opportunities the information-gathering environment affords: in response to doctors’ symptom-related queries they offer not only descriptions of their problems but also interpretations of what could potentially cause them and evidence that this is not what is causing them to occur. This positioning sets up an interpretive backdrop, enabling doctors to make sense of patients’ symptoms in light of the evidence they have provided.

An important dimension of the examples we consider in this article is that the patients display resistance to the possibility that their problems have ordinary, relatively benign causes.3 In doing so, they expose the relevance of alternative interpretations of their illnesses – that they are potentially more unusual or serious – yet without openly promoting those possibilities and therefore without displaying that they entertain worst-case scenarios. They present themselves as ‘reasonable’ patients (Halkowski 2006: 90) who have already used their own common-sense reasoning to consider the types of ordinary, mundane causes for their symptoms that doctors might also consider and who can recognise when evidence points away from mundane illness (Heritage and Robinson 2006, Sacks 1984).

However, the practice permits doctors to continue the medical workup rather than calling for them to actually reveal whether (or how) their own interpretations of the problem have been affected. Unlike other frontloading initiatives such as ‘Do you think it’s X?’ which establish the conditional relevance of an assessment (Gill 1998, Gill and Maynard 2006), the practice of pre-emptive resistance allows doctors to continue the primary task of this part of the medical interview: gathering information about the nature of patients’ problems. When doctors respond in this manner, it does not amount to an official absence, in sequential terms (Schegloff 1968, 1972). However, in these cases patients may do additional work to show that their problems are unusual, suggesting that they have received insufficient evidence of whether the doctors registered their concerns.

Analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

We will show four examples of the practice of pre-emptive resistance. In the first example, the patient, a 59 year-old woman, is visiting her new doctor for a check-up. Just prior to the extract shown below, the patient raised several health concerns including an ongoing problem with migraine headaches. She suggests she is puzzled about what causes the headaches, reporting, ‘they come and go so badly I just have to really wonder what triggers that’. Then she offered two additional concerns, problems with her bowels and fatigue. Taking up migraines first, her doctor began a series of questions by asking what medicines she takes for them.

In lines 1–2 (Extract 1a, below) the doctor continues gathering information about the problem, asking the patient how long she has had migraines. She responds that she has had them since she was three years old and then underscores that they have been a chronic problem (lines 3, 6–7). Our analysis will focus on lines 9–30 where the patient makes use of the opportunity that the environment of symptom description provides to present information about her problem. She indicates that she has considered two candidate explanations for the headaches, allergies to foods and to chemicals. She then reports a circumstance that serves as evidence against these explanations, discounting the possibility that the food and chemical allergies are the sole cause of her migraines and suggesting it is relevant for the doctor to look elsewhere for the cause.

(1a) [3: 175 (3:23)]
 1Dr:How long have you had this problem with these
 2migraines.
 3Pt:Oh- since I was three.
 4(0.2)
 5Dr:( ) Since age three.
 6Pt:Yeah. I’ve always had migraines. Al::ways had
 7migraine[s.  ]
 8Dr:     [Hm.]=
 9Pt.=And they’re never?. hh (.) I- I don’t know what
10triggers them. I kno::w that decaffeinated coffee.
11I knowthat. ((snaps fingers)) Boy I’ll get that just
12like that.
13Dr:>M hm?<
14Pt:.hh [Ah::       ] so I think there are some food=
15Dr:   [°>M hm?<°]
16Pt:=°allerg[ies and ] that° and I’ll also: (.) there=
17Dr:     [M hm. ]
18Pt:=are some chemical °a- (.) allergies if I get ahold
19of something that has biSULfites in it.°. hh But you
20see when I was three (0.4) there weren’t those kinds
21of things. [So what] triggered that .hh (1.2) at=
22Dr:     [M hm? ]
23Pt:=THAT time(0.4) I d- I don’t °know (n:: that).°
24(0.8)
25Pt:So it’s something that I’ve always had and I
26 [cer ]tainly can’t blame all those chemi[cals      ] =
27Dr: [Hm ]               [>M hm?<]
28Pt:=for it [because   ] I had it long before .hh=
29Dr:     [>M hm< ]
30Pt:=°we had all those chemicals an (1.5) (food)°
31Dr:Have you tried things for the migraines such as- (0.5)
32caffergot .h (0.5) ergotamine (.) things to- things
33to headoff that (0.3) thee ah (.) the pain from the
34headache.

After projecting and then aborting further description of her symptom (‘And they’re never?’ line 9), the patient broaches the issue of causality by inserting a claim that she doesn’t know what ‘triggers’ the migraines and then reporting ‘I kno::w that decaffeinated coffee’. By virtue of its positioning, this report invites the doctor to hear decaffeinated coffee as one such trigger (see Gill and Maynard 2006). She emphasises her certainty with ‘I knowthat’, snapping her fingers and reporting the immediacy of the reaction to the coffee (‘Boy I’ll get that just like that’ lines 11–12). She then provides a more general upshot of her interpretation, framing her reaction to coffee as evidence for ‘food allergies’ (‘so I think there are some food allergies’, lines 14 and 16). In this way she raises a candidate cause for her problem (food allergies) and exhibits that this is an obvious explanation for her to raise, given her experience with decaffeinated coffee.

However, via her initial claim that she does not know what triggers the migraines, she suggests that allergies are not the sole or primary cause. On the heels of this, the patient raises the possibility of a related candidate explanation, ‘chemical allergies’. As with ‘decaffeinated coffee’, she provides a specific example of a substance that gives her trouble, ‘bisulfites’ (lines 18–19). Thus, in a context of symptom description the patient raises the spectre of two related candidate causes for her migraines. Note that both of these interpretations of the problem posit a relatively benign condition that she could manage without the doctor’s help: if these substances were causing her migraines, she could potentially cut them out of her diet to avoid the headaches.4

In lines 19–21 the patient more explicitly challenges the possibility that food and chemical allergies are the only cause of her migraines by reporting circumstances that are inconsistent with that interpretation (Peräkylä 2002, 2006). She reports, ‘But you see when I was three (0.4) there weren’t those kinds of things’. She displays her reasoning that, given the absence of these substances when she was young, her migraines could not have been brought on by food and chemical allergies at that time (‘So what triggered that. hh (1.2) at THAT time (0.4) I d- I don’t °know’, lines 21 and 23). Having produced this counter evidence, she goes on to provide an upshot where she explicitly rules out the possibility that allergies could be the sole cause of her current headaches: ‘So it’s something that I’ve always had and I certainly can’t blame all those chemicals for it because I had it long before °we had all those chemicals an (1.5) (food)°’ (lines 25–26, 28, 30).

To summarise, in the environment of symptom description, the patient discounts two related candidate explanations for her migraine headaches. She does this by raising the explanations herself and then giving evidence that they are unlikely to be the only cause of her migraines. The migraines, she suggests, cannot be solely attributed to food and chemical allergies because they began before she had been exposed to these substances. Having raised these possibilities, the patient shows she is oriented to benign interpretations for her problem and that she is a ‘reasonable’ patient (Halkowski 2006: 90, Heritage and Robinson 2006). By providing evidence against them, she provides the doctor with the materials to let him see for himself that there is reason for concern, and that it is relevant for him to look elsewhere for the cause of her headaches – possibly to a cause that is relatively less obvious, less benign and less amenable to self-management.5

While the patient has shown that it is relevant to determine the cause of the headaches, and has indicated what types of causes are not relevant to consider, she does not compel the doctor to present his own diagnostic assessment then and there, in this phase of the interview (see Gill and Maynard 2006). Her contributions are designed as reports of information about her condition and are provided in the context of her response to the doctor’s symptom-related query in lines 1–2. They do not establish the conditional relevance (Schegloff 1968) of an assessment; the doctor may relevantly address the implications of the reports for a diagnosis but is not constrained to do so (see Drew 1984). The doctor can ask additional questions about her symptom, using the information the patient has provided as an interpretive backdrop for his ongoing diagnostic reasoning. The doctor takes this option, acknowledging the patient’s reports (lines 8, 13, 15, 17, 22, 27, 29) and then questioning her about her use of medications to abort the headaches (lines 31–34).

Although the patient has provided the doctor with this response option, she treats his focus on the management of her problem as heading in the wrong direction and does further work to direct his attention to how unusual the headaches are. After she describes the problems she has with medications the doctor presents her with the opportunity to report the headaches’ frequency. She characterises this as ‘a strange thing’ and provides an account for her characterisation:

(1b) [4:246 (5:21)]
 1Dr:How often do y[ou have-do you have headaches.
 2Pt:       [s
 3(0.2)
 4Pt:[UPWARDS ARROW]That’s a strange thing, because I can go:: I think
 5I ha::d up until about two months ago I musta
 6been: .hh three months (wisou-) without a headache.
 7(.)
 8Dr:Mm hm,
 9Pt:An then? Jus like that. I’ll get I:=will
10get them, and I will spend (.) two three weeks
11(.) having them.
12(0.5)
13Pt:[UPWARDS ARROW]Almost like clusters.
14Dr:Mm hm,

The patient suggests that the pattern of her headaches is odd, in that they will abate and then come back suddenly. She uses the term ‘clusters’ to describe them, a possible allusion to cluster headaches (an extremely intense form of migraine), although she ostensibly uses it to reference their pattern: for ‘two three weeks’ she will have one headache after another. The point we wish to underscore here is that she provides additional justification for concern and in doing so she orients to the possibility that the doctor had not found reason to be concerned about what causes the migraines.

The second example of the practice occurs later in the same clinic visit. The doctor invites the patient to talk about a medical condition (swollen feet), which is one of several concerns that the patient had noted on a health history form she completed before meeting the doctor. The patient raises a candidate explanation for the swollen feet, summer heat (this visit occurred during the summer), and provides evidence that it is unlikely in light of the symptom’s recent onset.

(2) [18:1211 (25:27)]
 1Dr:U::m (.) let’s see:. Couple of other things that
 2you’ve checked o:ff (0.7) .hh >you mentioned<
 3some::ah (1.5) ankle °swelling?°
 4Pt:Pch .h You know myfeet never used to swell at all.
 5In fact that was one of the things I always got iv: –
 6 (0.2) was admired by everybody.=How come you can take
 7your shoes o(h)ff and you(h)r feet never swell. .HH
 8You know. .hh And the la:st couple of months=an
 9course it’s summer.[An it’s] °hot.°[.hhh ] You know:.
10Dr:         [M hm ]   [M hm]
11Pt:A::hw (.) but my FEET have swelled.
12And I:[N:: ]E:Ver °had that b[efore.° ]
13Dr:    [M hm]         [M hm? ]
 : 
 : 
28Dr:Does it seem like everything is catching up with you?
29Pt:Ye::ah.
30Dr:hhhh ((Doctor is smiling.))
31Pt:[UPWARDS ARROW]All the things that people have had all these years
32and suddenly I get them and they fall on me you know.
33.hhh(0.8) Think I gotta start using my umbrella or
34something.
35(4.5)
36Dr:.hh Kay, then the other:- the other thing you
37mentioned was:: (.) you have (.) °pain with
38intercours:e.°

In response to the doctor’s introduction of ankle swelling as an item the patient had checked off on her health history form (lines 1–3), the patient reports a historical lack of foot swelling and its corroboration by admiring observers (‘You know myfeet never used to swell at all. In fact that was one of the things I always got iv:- (0.2) was admired by everybody.=How come you can take your shoes o(h)ff and you(h)r feet never swell’.).6 The patient thus casts foot swelling as an unusual or novel symptom for her, one that is worthy of the doctor’s attention. This becomes an important component in the patient’s subsequent reasoning about a potential cause of her foot swelling (summer heat).

Following this report, the patient begins an utterance (‘And the la:st couple of months’) which situates the onset of symptoms as recent and also projects a description or characterisation of those symptoms. However, rather than providing the projected description, the patient cuts herself off and reports, ‘an course it’s summer. An it’s °hot.° (lines 8–9). The positioning of this utterance (juxtaposed to a projected symptom description) invites the doctor to hear it as an explanation for her swollen feet (Gill and Maynard 2006). By initiating the report with ‘an course’, the patient displays that she has drawn upon common-sense knowledge to infer a causal connection, (that is, hot weather can cause swollen feet) and that she and the doctor have mutual awareness that this is a common cause of swollen feet. The doctor’s acknowledgments in line 10 may indicate recognition of this proposed causal connection and also may serve to ratify the patient’s orientation to this as shared common knowledge.

However, the patient then challenges the relevance of this candidate explanation by providing counter evidence. In lines 11–12 she reports circumstances that are inconsistent with what one would expect if summer heat were the correct explanation. She does this by reissuing her complaint using the contrast marker, ‘but’: ‘but my FEET have swelled’ and then repeating her claim that the symptom is atypical and novel (‘And I N::E:Ver °had that before.°’). This ‘negative observation’ (Peräkylä 2002: 232) serves as evidence against the candidate explanation: if the hot weather were indeed to blame, then in previous summers her feet would have been swollen.

By cutting herself off to acknowledge the hot weather after her projected symptom description (line 8), the patient orients to the possibility that the doctor might head down that diagnostic path as a result of her reference to the onset of symptoms during the recent summer months (line 8). By showing her attentiveness to the possibility that the recent heat could be causing the problem, and by exposing its logical inconsistency (that is, her feet never swelled in previous summers), the patient not only moves to potentially block that avenue of inquiry, she suggests that other diagnostic avenues are more relevant. At the same time, she displays her ability to consider (and weigh the evidence for) a relatively ordinary cause rather than automatically assuming that the swelling represents serious illness (see Pomerantz et al. 2007).

The doctor responds to the patient’s pre-emptive resistance in a more transparent manner than he did in Extract 1a, lines 31–34.7 He shows that he registers her concern (though does not share it) by attributing the problem to age (‘Does it seem like everything is catching up with you?’).8 That is, now that she is older the patient could be increasingly susceptible to such problems just like everyone else. He implies that the swelling is inevitable and therefore normal, rather than the sign of a serious underlying disorder. He smiles as he says this, and the patient aligns with his characterisation of her symptom by producing an agreement token in an ironic tone of voice (‘Ye::ah.’ line 29) and portraying it as ordinary (lines 31–34). This portrayal also normalises the sudden onset of her swollen feet by suggesting she was overdue for various common ailments – things that other people ‘have had all these years’.

Via gaze and body positioning, the doctor appears poised to continue questioning the patient about additional items on the health history form. Throughout this exchange he has oriented his gaze almost exclusively to the form, while holding it up off the desk with his pen in hand. After °hot.° (line 9), he jots on the form but immediately resumes his original posture. He puts his finger between the pages, and at line 28 (‘Does it seem like everything is catching up with you’) he turns the page, still holding his gaze on the form. In these ways, he shows an orientation to continuing his information-gathering activities rather than investigating what causes the ankle swelling. The patient initiates topic closing by producing a figurative expression (Drew and Holt 1998), ‘Think I gotta start using my umbrella or something’ (lines 33–34). This further characterises her situation as relatively trivial, in that it is just a matter of bad luck. The doctor then moves on, querying the patient about another item on the health history form (lines 36–38). The matter of swollen feet is dropped. Through their interaction, it has achieved the status of an ordinary problem that does not merit further concern.

The next two examples of the practice (Extracts 3 and 4a) also occur in one medical interview and they involve the same physical symptom, tenderness in the abdomen.9 In Extract 3, the patient introduces the symptom (lines 1–2), indicating the location with her hand. She immediately offers two candidate explanations, speculating that the discomfort may be originating from the gall bladder or the kidney. The doctor responds by asking for confirmation of the symptom’s location (lines 4 and 6) and when the patient confirms the location and emphasises the severity of the condition, the doctor questions the patient about its frequency (line 11). In this environment of symptom description the patient raises a third candidate explanation: that physical activity could be causing the abdominal tenderness. She resists this interpretation by implying that there is no evidence to support it.

(3) [13:677 (8:55)]
 1Pt:.hhh And then I get a lot of tenderness in this
 2area hh. And again. It’s probably (1.0)
 3[whether it’s] in the gall bladder?
 4Dr:[In the front ]
 5Pt:Kidney? [Er
 6Dr:       [Up in here.
 7Pt:Yeah. Like under the rib. Where I can’t get –
 8it’s get- (1.0) very sore.
 9(0.8)
10Pt:.hhh Ptch [And,        hhhh ]
11Dr:     [‘bout how often does] that come.
12Pt:Uh:: hhhh (1.0) This cn (1.5) m- be like at least
13once or twice a week. And I’ve been trying to see
14if I’ve been::: >you know,< lifting something or doing
15something. °.hhhh° (1.5)
16Dr:How long does it last when you g[et it. ]
17Pt:              [Ah::m ] (.) maybe
18a day or two.
19(2.2)
20Pt:Nuh- Enough tuh make me ner:vous.

The patient broaches the third candidate explanation by reporting that she has been monitoring her activities to see if she has been engaging in unusual physical activities (‘And I’ve been trying to see if I’ve been::: >you know,< lifting something or doing something. °.hhhh°’, lines 13–15). By juxtaposing this report to her symptom description and by using ‘And’ to link the two utterances (line 13), she invites the doctor to recognise that she is making a causal connection between the abdominal tenderness and activities such as lifting as well as the broader category of activities, ‘doing something’ (see Gill and Maynard 2006). She marks this as a common-sense category with the insertion of ‘>you know,<’ (line 14).

Having raised physical activity as a hypothetical cause for her symptom, the patient portrays it as unlikely. Through the formulation, ‘I’ve been trying to see’, she suggests that she has already searched her memory over some period of time to identify the occurrence of some relevant physical activity that might cause the problem. She audibly inhales (line 15), projecting continuation of her turn. At this moment the doctor, who had been writing in the patient’s chart, looks up at the patient. The doctor gazes at her and allows continuation while the patient looks directly at the doctor and remains silent. The patient shakes her head from side to side as the doctor returns her gaze to the medical chart and starts writing again. By remaining silent in the location where it would be relevant for her to report such activities and by shaking her head (as if at a loss), the patient implies that, having searched her memory, she has found nothing to report; she has not been engaging in activities that could cause the abdominal tenderness. This is the most tacit example of pre-emptive resistance in our data. The patient raises the explanation and then implies there is no evidence of engagement in physical activities – in a location where reporting such evidence would be relevant – thus suggesting that physical activities could not have caused her symptom.

After the silence the doctor turns back to the patient’s medical chart and asks her how long the symptom lasts (line 16). In so doing she treats the patient’s contribution as additional information about the nature of the symptom and she continues collecting more information, providing no outward indication of how she interprets the symptom. The patient responds to the doctor’s question (lines 17–18), and then does additional work to portray the symptom as unusual and troubling: she claims that the duration of the pain is ‘Enough tuh make me ner:vous’. (line 20).

Thus, in the environment of symptom description, the patient shows that she has already considered – and suggests that she has already dismissed – the possibility that her abdominal discomfort is caused by physical activity, a common and relatively benign cause of such discomfort and one she herself could potentially address by refraining from activities such as lifting. By raising the possibility, she orients to it as a line of inquiry the doctor is likely to pursue, and she works to pre-emptively block that path by implying that she found no evidence for it when she considered it herself. She makes it relevant for the doctor to look elsewhere, toward more serious possibilities such as the ones she originally mentioned (‘gall bladder’, ‘kidney’). In response, the doctor takes an option the patient has provided and continues gathering information about her symptom. Although her questions may be building up an evidential basis for a more serious explanation than ‘lifting’, she does not transparently indicate whether this is the case – that is, how or whether the patient's contributions have affected her interpretation of the problem. The patient’s effort to underscore her concern (line 20) treats this opacity as problematic, and bids for the doctor to recognise that the exclusion of a benign cause (physical activity such as lifting) justifies a concern that the problem may be more serious.

The final example of the practice of pre-emptive resistance (Extract 4a) occurs later in the same interview. Here, the patient more explicitly rules out the possibility that physical activity could be causing her abdominal discomfort. The doctor is performing the physical exam and has resumed questioning about the discomfort. In lines 121–122, she asks the patient how long she has had the symptom, seeks confirmation on how often it occurs (line 126) and asks whether the pain comes on ‘suddenly’ (line 128). The patient then raises and discounts the possibility that swimming could be causing her symptom.

(4a) [13:677]
121Dr:.hhh How long have you been getting this:: ah: sore –
122(1.0) ti:mes. (°Up here°)
123Pt:i- This: has been for about- s::ix months or so °I
124think.°
125(1.2)
126Dr:.hh >And you said< once or twice a week,
127Pt:Y::ah.
128Dr:Does it come on fairly suddenly?
129(1.0)
130Pt:Ye::ah.
131(0.8)
132Pt:It really- an- (.) when I- now I’ve been: swimming,
133(1.3) we:: go- we’ve- (.)°two or three times a week.°
134I haven’t been (1.0) swimming for the last oh five
135or six day:s.
136(2.0)
137Pt:A:n:::::::d (1.0) I ha- well I hadn’t had it before
138then. For:: (1.0) a week or so and that- so it isn’t=
139Dr:=Mh[m
140Pt:   [that that brings it on.
141Dr:>M hm<?
142(0.5)
143Dr:[Does ] food ever bring it on?
144Pt:[°But° ]
145Dr:[( )]
146Pt:[I don’t] kn:ow.

After confirming that the symptom does come on suddenly (line 130), the patient projects that she will continue describing the symptom (‘It really-’) and when it occurs (‘when I-’). However, she aborts these tacks and begins a narrative where she reports that she has ‘been swimming’ with a group on a regular basis (‘we:: go- we’ve- (.) °two or three times a week.°’). By reporting this just after confirming the symptom’s sudden onset, she invites the doctor to hear that she has connected swimming with the onset (see Gill and Maynard 2006). She reinforces this connection by presenting her swimming schedule in a way that corresponds to the symptom’s duration and frequency; that is, she presents both the swimming and the symptom as ongoing activities that happen a similar number of times per week. Thus, she shows that the swimming could hypothetically be responsible for the onset of the abdominal discomfort, and therefore that it is a possible cause of the discomfort.

The patient then resists this candidate interpretation by reporting counter evidence, circumstances that indicate swimming did not cause the symptom to occur (lines 134–138). The gist of this is that when she was engaging in her normal swimming routine, she did not experience the symptom. She first notes a change in her swimming schedule: she missed five or six days of swimming before this medical appointment (lines 134–135). She uses this as a temporal reference point in her next utterance (lines 137–138) where she reports that she did not experience the symptom ‘before then’– that is, during her normal swimming schedule, before the change of routine. Via this ‘negative observation’ (Peräkylä 2002: 232) she shows that her experience is inconsistent with what one would expect if swimming were indeed the cause of the symptom; if swimming were responsible, she would have experienced it when she was swimming regularly. Finally, the patient provides an upshot, explicitly ruling out swimming as the cause of the symptom (‘so it isn’t that that brings it on’, lines 138 and 140). The doctor acknowledges this (‘>M hm?<‘) and then asks her about a different type of candidate cause (‘Does food ever bring it on?’ line 143).

In firmly discounting swimming as a candidate explanation for the abdominal discomfort, the patient addresses a contingency that she had some hand in engendering: the doctor has not, as yet, revealed her own interpretation of the symptom or any diagnostic hypothesis (or hypotheses) she may be considering. As discussed earlier, the design of pre-emptive resistance provides for this contingency. Note, however, that the doctor’s question in line 128 was different from those in lines 121–122 and 126: it asked the patient to confirm whether the pain comes on suddenly, and therefore may have suggested to the patient that the doctor did have a specific diagnostic hypothesis in mind. In raising and then firmly ruling out the possibility that swimming caused the symptom, the patient orients to the possibility that the doctor is exploring whether the symptom occurs suddenly in conjunction with physical activity, despite the fact that she had already implied (in Extract 3) that there was no such connection. She works to block this potential line of inquiry. Evidence for this can be found in the patient’s lexical choices (‘brings it on’) and her emphasis on ‘that’ (line 140), through which she reframes the doctor’s prior questions as aimed at what makes the symptom come on, not just how it comes on (for example, suddenly or gradually). She more firmly steers the doctor away from the possibility that physical activity could be causing the symptom to occur and establishes an additional opportunity for the doctor to show how she is interpreting the problem.

By asking the patient, ‘Does food ever bring it on?’ the doctor transparently shows that she is shifting direction and is exploring a causal factor that is different from physical activity. In response, the patient claims not to know whether food produces the symptom (line 146) and in further talk (not shown here), describes her frustration about being sensitive to a number of foods. The doctor goes on to tentatively confirm the patient’s original (Extract 3) speculation that the pain could be originating from her gall bladder – specifically, that the gall bladder may be emitting stones after she eats and that they are getting caught and causing pain (see Gill and Maynard 2006: 142). In response, the patient indicates that she can live with the problem if she knows it does not represent serious illness:

(4b) [15:778(20:04)]
1Pt:W’ if- >ya know< even if (.) it’s- if I know:
2it’s not (.) uh major (1.0) something.=
3Dr:=Mm hm[:
4Pt:     [that it’s: (.) uh transitory [it    ]’s
5Dr:                [M hm]
6Pt:here onc:e in uh while.
7(.)
8Pt:I (mean) I (th) can cope with th[a:t
9Dr:             [Yeah:.

To summarise, in more explicitly ruling out the role of her own activities in causing her symptom, the patient addresses the interactional dilemma engendered by her original use of pre-emptive resistance (in Extract 3). In the first place, its design allowed the doctor to continue gathering information about the symptom rather than compelling her to indicate whether (or how) her own interpretation of the problem had been affected by what the patient reported. Secondly, the patient had implied that physical activity was not causing her symptom. When she received a possible clue that the doctor was testing a diagnostic hypothesis (Extract 4a, line 128) the patient took additional measures to show that she was attentive to a common cause of abdominal discomfort and to rule it out. This resulted in a relatively transparent indication, on the part of the doctor, that she was directing her inquiry elsewhere.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

It has been well established that the sense-making process in the medical visit is a social process that involves the participation of patients as well as doctors. Our study contributes to the growing body of CA literature revealing that patients may do more than convey information about their symptoms and problems so that doctors can diagnose and treat them; they also work within the different phases of medical visits to offer interpretations about what could be causing their health problems and they may press for and against particular interpretations. Our analysis of the practice of pre-emptive resistance shows that patients can take advantage of the opportunities afforded in the early, information-gathering phases of the clinic visit to show doctors that particular candidate explanations for health problems are implausible. They do this by raising the candidate explanations themselves and then presenting counter evidence, reports of circumstances that show these explanations are unlikely to be correct.

This practice allows patients to potentially manage some delicate actions in the medical interaction. It is a means by which they can indicate to doctors that it is relevant to narrow the range of diagnostic hypotheses they might consider during the visit. In our study, we analysed cases where patients pre-emptively resist the possibility that their symptoms have commonplace, relatively benign causes and expose the relevance of relatively unusual or serious diagnostic hypotheses. By doing this via reports of circumstances, patients let their doctors see for themselves that the circumstances run counter to what one would expect if the commonplace candidate explanations were correct, and that their symptoms could represent more unusual or serious illness. They show doctors that concern is justified rather than telling them so. This displays patients’ orientation to possible interactional risks of being heard to pressure doctors to consider the possibility that seemingly ‘run of the mill’ problems are unusual or serious, risks that include being perceived as hypochondriacal or not sufficiently oriented to whether symptoms are doctor-worthy (Heritage and Robinson 2006). The practice also provides a way for patients to manage this within the ordered phases of the interview. Patients show their awareness that doctors generate and test diagnostic hypotheses during the information-gathering phases and that this is an environment where they can indicate that their problems might not be run-of-the mill and therefore possibly influence the direction of medical inquiry. A matter for future research is whether this practice also may enable patients to circumvent disagreements over diagnoses and thus avoid pitting themselves against their doctors, avoid stalling the progression of the visit and the like.

The study advances the literature on patient agency and patients’ strategic employment of knowledge in medical visits. Patients use knowledge to which they are entitled in the service of interactional practices. In the examples of pre-emptive resistance we analyse, patients raise well-known common-sense explanations for their symptoms and then use reports of circumstances as counter evidence: migraines began before migraine triggers existed; feet did not swell in previous summers; no lifting coincided with abdominal tenderness; tenderness didn’t occur during a regular swimming routine. In doing do, patients make use of a category of knowledge that is legitimately theirs as lay members of society (Sacks 1984): first-hand knowledge about (i) circumstances and events in their own lives and (ii) the onset or occurrence of their own physical symptoms. The timing of (i) in relation to (ii) is also within their legitimate purview, and it represents an important empirical resource patients use to advance and block candidate explanations for illness in doctor–patient interaction. Just as doctors embark on ‘a search for clues’ and use them ‘to support or refute’ hypothetical diagnoses (DeGowin and Brown 2000: 2), it is clear that patients do so as well. However, this work is not just cognitive but something that is shared with doctors as an integral part of interactional practices, such as the practice of pre-emptive resistance.

This study also shows that patients make use of the structure of the medical interview when they use pre-emptive resistance to potentially influence the trajectory of the medical interview. Frontloading (presenting interpretations of illness during the information-gathering phases) is advantageous for patients not only because this environment offers them opportunities to put forward their own contributions, but also because it potentially allows doctors to take patients’ interpretations into account as they test diagnostic hypotheses. However, a distinctive feature of patients’ contributions in this sequential environment is that they do not necessarily require doctors to reveal whether they are doing so (Gill and Maynard 2006). In the case of pre-emptive resistance, a powerful resource patients have at their disposal to demonstrate that certain interpretations are unlikely to be correct (firsthand knowledge of symptoms and circumstances) is mobilised and presented in a certain manner (via reports of circumstances) in an environment where the primary activity is information-gathering. Doctors can treat these reports as ‘more information’. Doing this, rather than displaying recognition of the possible implications of the report for the diagnosis, may represent a form of ‘withholding’ (Drew 1984: 136), albeit a legitimate one in sequential terms. This may lead patients to feel their concerns have not been recognised and may engender further attempts to show that they are concerned. This dynamic is one of the more subtle and complex aspects of medical interaction and it is one that we suspect plays an important role in patients’ and doctors’ subjective impressions about the quality of their relationships and of medical visits in general. Future efforts to explore this connection are likely to be fruitful.

Footnotes
  • 1

     We reserve the term ‘diagnosis’ for doctors’ final specifications of conditions, diseases or injuries that cause patients’ problems and symptoms. We use the more general terms, ‘explanation’, ‘candidate explanation’ or ‘interpretation’ when referring to patients’ displays of sense-making in regard to illness and to doctors’ more general causal attributions. These encompass both overt and tacit attributions for symptoms, which are done by attributing symptoms to hypothetical conditions (for example, hormone deficiencies), to particular organs in the body (for example, the heart), and to various circumstances (for example, insufficient sleep). It also encompasses reports, speculations and inferences that something ‘brought on’ a symptom (Gill and Maynard 2006).

  • 2

     An exception is the ‘stoic’ response (Maynard 2003: 120) where non-vocal behaviour accompanying a silence – such as crossed arms and certain facial expressions – can be used to indicate resistance to a diagnosis.

  • 3

     See (Pomerantz et al. 2007) for an investigation of how patients show that more serious conditions are unlikely.

  • 4

     In lines 18–19 the patient suggests that she already avoids the consumption of a substance that triggers her allergies by framing any intake of ‘bisulfites’ as accidental: ‘if I get a hold of something that has biSULfites in it’.

  • 5

     As Halkowski (2006: 98) explains about the device ‘at first I thought “X” and then I realised “W”.’‘Rather than having looked for the dramatic or outrageous explanation for an event, we demonstrate ourselves to have looked for the most obvious and mundane account. Only if those fail do we broaden our search and include more dramatic hypotheses’. See also Sacks (1984), Jefferson (1986), Pollner (1987).

  • 6

     We did not have access to the written health history form to which the doctor refers in lines 1–3. The patient’s repair at line 4 (‘feet’, versus the doctor’s ‘ankle’ at line 3) suggests that the form included boxes for patients to check off and that she had checked off a box for ‘ankle swelling’.

  • 7

     In the interest of space we have omitted 11 lines where the patient acknowledges having had swollen feet many years ago during pregnancy and expresses dismay at having the problem now.

  • 8

     With ‘everything’ (line 28) the doctor is also likely to be invoking the sheer number of health issues the patient has already brought up in this visit. These include migraine headaches, dizziness, earaches, painful intercourse, concerns about HIV/AIDS, constipation, fatigue, chest pains and sweating at night. By responding in terms of ‘everything’, he manages to disagree with the patient’s suggestion about the cause of this specific problem (swollen feet) without actually issuing a disagreeing assessment.

  • 9

     See Gill (1998) and Gill and Maynard (2006) for analyses of additional features of the explanation-response sequences in Extract 3.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References

Earlier versions of this article were presented at the International Meeting on Conversation Analysis and Clinical Encounters (2007), International Conference on Conversation Analysis (2006), 14th World Congress of Applied Linguistics (2005), and Pacific Sociological Association (2004). We would like to thank Doug Maynard, John Heritage and the anonymous Sociology of Health & Illness reviewers for their very helpful comments.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Background
  5. Data and methods
  6. Organisation of the practice
  7. Analysis
  8. Discussion
  9. Acknowledgements
  10. References
  • Angel, R. and Thoits, P. (1987) The impact of culture on the cognitive structure of illness, Culture, Medicine and Psychiatry, 11, 4, 46594.
  • Blaxter, M. (1983) The causes of disease: women talking, Social Science and Medicine, 17, 2, 5969.
  • Byrne, P. and Long, B. (1976) Doctors Talking to Patients: A Study of the Verbal Behaviours of Doctors in the Consultation. London: Her Majesty’s Stationery Office.
  • Cicourel, A. (1983) Hearing is not believing: language and the structure of belief in medical communication. In Fisher, S. and Todd, J. (eds) The Social Organization of Doctor–Patient Communication. Washington, DC: Center for Applied Linguistics.
  • Clayman, S. and Gill, V. (2004) Conversation analysis. In Bryman, A. and Hardy, M. (eds) Handbook of Data Analysis. London: Sage.
  • Cooper-Patrick, L., Gallo, J., Gonzales, J., Vu, H., Powe, N., Nelson, C. and Ford, D. (1999) Race, gender, and partnership in the patient–physician relationship, Journal of the American Medical Association, 282, 6, 5839.
  • DeGowin, R. and Brown, D. (2000) DeGowin’s Diagnostic Examination, 7th edn. New York: McGraw Hill.
  • Drew, P. (1984) Speakers’ reportings in invitation sequences. in Atkinson, J.M. and Heritage, J. (eds) Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press.
  • Drew, P. and Holt, E. (1998) Figures of speech: figurative expressions and the management of topic transition in conversation, Language in Society, 27, 495522.
  • Garfinkel, H. (1967) Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
  • Gill, V. (1998) Doing attributions in medical interaction: patients’ explanations for illness and doctors’ responses, Social Psychology Quarterly, 61, 4, 34260.
  • Gill, V. and Maynard, D. (1995) On ‘labeling’ in actual interaction: delivering and receiving diagnoses of developmental disabilities, Social Problems, 42, 1, 1137.
  • Gill, V. and Maynard, D. (2006) Explaining illness: patients’ proposals and physicians’ responses. In Heritage, J. and Maynard, D. (eds) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press.
  • Halkowski, T. (2006) Realizing the illness: patients’ narratives of symptom discovery. In Heritage, J. and Maynard, D. (eds) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press.
  • Heath, C. (1992) The delivery and reception of diagnosis in the general-practice consultation. In Drew, P. and Heritage, J. (eds) Talk at Work: Interaction in Institutional Settings. Cambridge: Cambridge University Press.
  • Heritage, J. (1984) Garfinkel and Ethnomethodology. Cambridge: Polity Press.
  • Heritage, J. and Maynard, D. (2006) Introduction: analyzing interaction between doctors and patients in primary care encounters. In Heritage, J. and Maynard, D. (eds) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press.
  • Heritage, J. and Robinson, J. (2006) Accounting for the visit: giving reasons for seeking medical care. In Heritage, J. and Maynard, D. (eds) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press.
  • Hunt, L. Jordan, B. and Irwin, S. (1989) Views of what’s wrong: diagnosis and patients’ concepts of illness, Social Science and Medicine, 28, 9, 94556.
  • Jefferson, G. (1986) On the interactional unpackaging of a ‘gloss’, Language in Society, 14, 4, 43566.
  • Kleinman, A. (1981) Patients and Healers in the Context of Culture: An Exploration of the Borderland between Anthropology, Medicine, and Psychiatry. Berkeley, CA: University of California Press.
  • Kleinman, A. (1988) The Illness Narratives: Suffering, Healing, and the Human Condition. New York: Basic Books.
  • Keller, M., Leventhal, H., Prohaska, T. and Leventhal, E. (1989) Beliefs about aging and illness in a community sample, Research in Nursing and Health, 12, 4, 24755.
  • Lorber, J. and Moore, L. (1997) Gender and the Social Construction of Illness. Thousand Oaks, CA: Sage.
  • Lurie, N., Margolis, K., McGovern, P., Mink, P. and Slater, J. (1997) Why do patients of female physicians have higher rates of breast and cervical cancer screening? Journal of General Internal Medicine, 12, 1, 3443.
  • Maynard, D. (2003) Bad News, Good News: Conversational Order in Everyday Talk and Clinical Settings. Chicago, IL: University of Chicago Press.
  • Mishler, E. (1984) The Discourse of Medicine: Dialectics of Medical Interviews. Norwood, NJ: Ablex.
  • Peräkylä, A. (1998) Authority and accountability: the delivery of diagnosis in primary health care, Social Psychology Quarterly, 61, 4, 30120.
  • Peräkylä, A. (2002) Agency and authority: extended responses to diagnostic statements in primary care encounters, Research on Language and Social Interaction, 35, 2, 21947.
  • Peräkylä, A. (2006) Communicating and responding to diagnosis. In Heritage, J. and Maynard, D. (eds) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press.
  • Pollner, M. (1987) Mundane Reason. Cambridge: Cambridge University Press.
  • Pomerantz, A. and Fehr, B. (1997) Conversation analysis: an approach to the study of social action as sense making practices. In Van Dijk, T. (ed.) Discourse Studies: A Multidisciplinary Introduction. London: Sage.
  • Pomerantz, A., Gill, V. and Denvir, P. (2007) When patients present serious health conditions as unlikely: managing potentially conflicting issues and constraints. In Hepburn, A. and Wiggins, S. (eds) Discursive Research in Practice: New Approaches to Psychology and Interaction. Cambridge: Cambridge University Press.
  • Rathore, S. and Krumholz, H. (2004) Differences, disparities, and biases: clarifying racial variations in health care use, Annals of Internal Medicine, 12, 1, 6358.
  • Rathore, S., Berger, A., Weinfurt, K., Feinleib, M., Oetgen, W., Gersh, B. and Schulman, K. (2000) Race, sex, poverty, and the medical treatment of acute myocardial infarction in the elderly, Circulation, 102, 6428.
  • Robinson, J. (2003) An interactional structure of medical activities during acute visits and its implications for patients’ participation, Health Communication, 15, 1, 2759.
  • Sacks, H. (1984) On doing being ordinary. In Atkinson, J.M. and Heritage, J. (eds) Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press.
  • Sacks, H. (1992) Lectures on Conversation. Vols. 1 and 2. Oxford: Blackwell.
  • Schegloff, E. (1968) Sequencing in conversational openings, American Anthropologist, 70, 107595.
  • Schegloff, E. (1972) Notes on a conversational practice: formulating place. In Sudnow, D. (ed.) Studies in Social Interaction. New York: The Free Press.
  • Smedley, B., Stith, A. and Nelson, A. (eds) (2003) Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: Institute of Medicine and National Academies Press.
  • Stivers, T. (2002) Presenting the problem in pediatric encounters: ‘symptoms only’ vs ‘candidate diagnosis’ presentations, Health Communication, 14, 3, 299338.
  • Stivers, T. (2005) Parent resistance to physicians’ treatment recommendations: one resource for initiating a negotiation of the treatment decision, Health Communication, 18, 1, 4174.
  • Stivers, T. (2006) Treatment decisions: negotiations between doctors and parents in acute care encounters. in Heritage, J. and Maynard, D. (eds) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press.
  • Waitzkin, H. (1979) Medicine, superstructure, and micropolitics, Social Science and Medicine, 13A, 6019.
  • Waitzkin, H. (1991) The Politics of Medical Encounters. New Haven, CT: Yale University Press.