Elsevier

The Lancet

Volume 353, Issue 9147, 9 January 1999, Pages 108-111
The Lancet

Early Report
Concordancing: use of language-based research in medical communication

https://doi.org/10.1016/S0140-6736(98)02469-6Get rights and content

Summary

Background

The available literature on medical communication reports almost exclusively on observational, qualitative studies. We aimed to apply a novel approach to the analysis of doctor-patient consultation by means of computer concordancing. This methodology, established in linguistic research but rarely applied to professional language, allows both the quantitative and qualitative study of language.

Methods

We analysed the language of 40 doctors and their patients during 373 complete primary-care consultations. We examined the use of jargon by doctors, the language of power and absence of power, and ways in which language was used to diminish the potential threat of the presenting disorder.

Findings

There was no evidence that the doctors used medical jargon. Some doctors used language associated with social power, and some patients used language associated with absence of power. There was substantial evidence that the doctors used language to express emotions (eg, anxiety), to diminish threats (eg, words such as “little”), and to reassure patients.

Interpretation

Concordancing is a valuable resource to study the consultation. The finding that doctors do not use jargon suggests that they are aware of the need to avoid it, but it does not follow that they are easily understood by patients. The use of some elements of the language of power may imply that consultations may be less democratic than is appropriate. The language of emotion and diminution is a major part of the primary-care doctor's repertoire and denotes a therapeutic use of language.

Introduction

The 1991 Toronto consensus statement1 remains the best-known statement about the importance of medical communication skills. The statement offered evidence that this subject could be taught and could improve the well-being of patients. In addition, it called for “new research methods” since the “Complementary development of alternative methods will allow the identification and description of aspects of clinical communication not already recognised or studied. Qualitative methods need to be encouraged, to complement the now standard quantitative and interaction analysis measures”.1

Quantitative studies make the presumption that interaction can be reduced to the counting of behaviours. Much of the assessment literature, for instance, deconstructs consultations into discrete points, defined in behavioural (and therefore observable) terms that boost test reliability—for example, did the doctor greet the patient; were open questions asked? Of this approach, Marinker2 remarks: “coming to an understanding with people is now being taught as ‘behavioural skills’, as though Marshall McLuhan had never related the medium to the message”. The difficulties of quantitative studies are best illustrated by the frequency with which behavioural marking schedules introduce solely subjective terms, such as: “Performs the history taking and the review of symptoms properly3 or “Gives appropriate advice”.4 This shortcoming might matter less if the measurable behaviours generated were not so easily rote-learned and performed—a machine can be programmed to “greet the patient”. Quantitative studies also beg the question of how much of each criterion is desirable: is a lot of eye-contact better than a little? The answer can only be: it all depends.

This particular difficulty was first developed within medicine by Stiles,5 who drew attention to the poor correlations between process (what the doctor said) and outcome (what effect it had on the patient). The advice such studies offer (for example, to emphasise and repeat the most important points) has been fundamental to teacher training for generations, where on the whole it is considered too obvious to research. This observation suggests that only the most self-evident of these behavioural criteria are completely generalisable.

These difficulties suggest that in quantitative analysis the ideal is to identify variables that: are measurable; are so embedded in the process of the consultation that they cannot easily be performed; and are not self-evident. We aimed to identify such parameters through the application of concordancing, a recent method for the analysis of language that has not been applied in medicine. We explore the potential of concordancing through its application to a database of consultations between primary-care physicians and their patients.

A basic concept in the study of language is that meanings cannot be completely quantified: if words were like numbers, it would be hard to understand why we bother with both. Any quantitative analysis must, therefore, take place in a qualitative context, such as the Toronto statement called for (there is in fact a healthy qualitative tradition6, 7, 8 not discussed in the statement).

We report three types of inquiry to show possible relations between quantitative and qualitative work. The first is a quantitative study of doctors' use of jargon. In the second, a complex qualitative concept of doctors' power in the consultation is given a partial, quantitative operational definition. In the third, a further complex qualitative concept is given a quantitative definition in the context of diminishing threat, but here we emphasise the manner in which qualitative and quantitative research can support each other.

Section snippets

Consultations

We used concordancing software to analyse a UK database of 373 consultations, collected in 1993–95, with 40 native speaker (English), primary-care doctors. Our aim was to study the performance of doctors who might represent good practice. All the doctors had some connection with medical education, either because they held an academic appointment (n=11), or because they were involved in some other way in undergraduate or postgraduate training (n=29). 28 of the doctors were men, and 12 were

A quantitative study: the use of medical jargon

Our first inquiry explores the repeated finding in descriptive studies that doctors use jargon with patients. Jargon can be defined as the use of (a) technical terms, (b) beyond the hearer's understanding, (c) unnecessary (there is an alternative term available), and (d) without explanation. Use of jargon is perceived as “bad” because it inhibits the patient's understanding and threatens the patient with the unknown. But jargon seems also to be regarded as a proxy measure of attitude: only

Discussion

We looked at three types of application for concordancing research. In the first, a purely quantitative study of jargon yielded a straightforward finding. In the second, on power, we argued that setting up language-oriented operational definitions, from the patient's language rather than the doctor's, can help our understanding of complex qualitative issues. In the third, illustrated by threats, the relation between quantitative statements and their qualitative context is shown in action.

Our

References (16)

  • R Shuy

    The medical interview: problems in communication

    Primary Care

    (1976)
  • M Simpson et al.

    Doctor-patient communication: the Toronto consensus statement

    BMJ

    (1991)
  • M Marinker

    Myth, paradox and the hidden curriculum

    Med Education

    (1997)
  • J Cox et al.

    An instrument for assessment of videotapes of general practitioners' performance

    BMJ

    (1993)
  • HF Kraan et al.

    The Maastricht history taking and advice checklist: studies of instrumental utility

    (1987)
  • WB Stiles

    Evaluating medical interview process components: null correlations with outcomes may be misleading

    Med Care

    (1981)
  • C West

    Routine complications: troubles with talk between doctors and patients

    (1983)
There are more references available in the full text version of this article.

Cited by (52)

View all citing articles on Scopus
View full text