Embed Size (px)
Citation preview
www.elsevier.com/locate/pateducou
Patient Education and Counseling 65 (2007) 51–57
Using linguistic analysis to explore medicine counter assistants’
communication during consultations for nonprescription medicines
Mark Garner a, Margaret C. Watson b,*
a School of Language & Literature, University of Aberdeen, King’s College, Aberdeen AB24 3UB, Scotland, UKb Department of General Practice and Primary Care, University of Aberdeen, Westburn Road, Aberdeen AB25 2AY, UK
Received 1 February 2006; received in revised form 26 April 2006; accepted 1 May 2006
Abstract
Objective: The purpose of this study was to undertake linguistic analysis of consultations between medicine counter assistants (MCAs) and
customers for the supply of nonprescription medicines.
Methods: The linguistic corpus comprised 168 recorded and transcribed consultations. Utterances were tagged as information eliciting,
information or advice giving and other. The analysis focused on triads, beginning and ending with an MCA utterance (MCA1! customer’s
response!MCA2). The use of the mnemonic, WWHAM (Who for, What symptoms, How long, Any medicine tried, other Medication taken)
was also analysed.
Results: A total of 505 triads were identified. Of the 773 utterances, 61% were information eliciting, 13% were information giving, 14% were
advice giving and 11% were ‘‘other’’ or unclassifiable. No consultation involved WWHAM in its entirety.
Conclusion: MCAs do not appear to have been made sufficiently aware of the ways in which their exchanges with customers during
consultations for nonprescription medicines are crucially different from natural conversation.
Practice implications: In order to help customers decide upon an appropriate nonprescription medicine, the MCA has the role of both
informing and advising. Training should focus on informing and advising as distinct functions, and the potential problems caused by
combining them.
# 2006 Elsevier Ireland Ltd. All rights reserved.
Keywords: Communication; Counseling; Linguistics; Community pharmacy services; Non-prescription drugs
1. Introduction
The reclassification of medicines is a global method to
restrict government drug budgets and increase public access
to medicines. In the UK, over 70 medicines previously
available only on prescription (i.e. Prescription Only
Medicines (POMs)) have been reclassified to Pharmacy
(P) or General Sales List (GSL) medicines [1]. The supply of
these nonprescription medicines (NPMs) from community
pharmacies is controlled by the Medicines Act 1968 and by
the Royal Pharmaceutical Society of Great Britain (RPSGB)
[2]. Whilst the greater availability of medicines has been
* Corresponding author. Tel.: +44 1224 553785; fax: +44 1224 550683.
E-mail address: [email protected] (M.C. Watson).
0738-3991/$ – see front matter # 2006 Elsevier Ireland Ltd. All rights reserved
doi:10.1016/j.pec.2006.05.004
welcomed by many organisations, there are concerns about
the inappropriate supply of NPMs [3–9].
The majority of consultations for NPMs involve medicine
counter assistants (MCAs), who are members of pharmacy
support staff who have undertaken accredited training [10].
In the late 1980s, when medicine reclassification began, a
questioning framework (WWHAM [11,12]) was developed
to assist MCAs during consultations for NPMs. WWHAM
represents: Who is the medicine for? What are the
symptoms? How long have the symptoms been present?
Any medicine tried already? and What Medication is
currently used? MCAs or other members of pharmacy
support staff need to elicit relevant information from the
customer and communicate with the customer, in order to
derive an appropriate treatment recommendation. Whilst
such a recommendation requires adequate knowledge on the
.
M. Garner, M.C. Watson / Patient Education and Counseling 65 (2007) 51–5752
part of the MCA, there is evidence that sub-optimal
communication is associated with many of these consulta-
tions [13], and that this is a significant factor in the
inappropriate supply of these medicines [14]. This is the
focus of the analysis reported here.
1.1. Communication skills
The past couple of decades have witnessed a growing
interest in the role of communication in health care.
Communication skills training has become a component of
many medical and health-care curricula worldwide, and
interdisciplinary research involving various social and
human sciences has been undertaken in many aspects of
this rich field. Parrott [15], for example, identifies more than
250 publications relating to communication and health care,
well over 90% of which have appeared since 1990.
The considerable contribution that communication theory,
particularly when informed by linguistics, could make to
health research has not yet been fully realized [16]. This may
be due partly to the tendency of linguistics as a discipline to
inadequate engagement with normative, corrective, and
therapeutic issues [17,18], and partly to the widespread
assumption by practitioners in many professions centrally
concerned with human interaction, that communication is not
problematic [19]. Yet applied linguistic research, for example
involving the various forms of increasingly sophisticated
discourse analysis, has a significant and unique contribution to
make to a deeper understanding and improved practice in the
delivery of health care at all levels.
This paper reports a linguistic analysis of interactions
between MCAs and pharmacy customers in consultations
involving the supply of NPMs.
The framework for the analysis was drawn from
operational communication research [20]. This branch of
applied linguistics aims to enhance the operational effective-
ness of organisations such as the emergency services and
transport authorities through developing and implementing
efficient communication practices, specifically in relation to
language use. Its outcomes are a series of recommendations
for standardised language procedures relating to practices to
be employed or avoided within the given context, which are
enshrined in, for example, training programmes and com-
munication manuals. The operational communication frame-
work is relevant to health-care communication research
insofar as both are concerned with professional interactive
practices in contexts involving focused, definable goals
related to the achievement of some specific public good.
Linguistic analysis is fundamental to this approach. It
provides the descriptive foundations for understanding and
evaluating current practice within the particular professional
sphere of operations. It also furnishes the specific elements
of recommendations for enhanced practice. Experience has
shown [21] that explicit recommendations concerning the
linguistic forms to be used or avoided are more effective in
developing communication skills than the sort of general
advice (‘be clear and concise’, ‘avoid closed questions’) that
is often found in training programmes and manuals.
Contemporary discourse analysis offers a variety of
methods for contextually based language description; the
choice of which to use is determined by the aims of the
research and the demands of the specific professional sphere
under study. The research reported here applied a relatively
simple form of context-sensitive conversation analysis for
reasons that are explained in the following section [22].
1.2. Framework for this study
Operational communication research is an inherently
collaborative undertaking. It requires inputs from both
linguists and practitioners, for whom communication is a
tool of the trade. It uses a four-phase methodology, although
this paper reports only on the linguistic analysis conducted in
the first phase. The other phases, however, are presented
briefly in order to describe the framework within which the
analysis was conducted. The study reported here, the analysis
of interactions between MCAs and customers, relating to the
supply of NPMs, is phase 1 of an on-going, larger project
aimed at developing improved communication training for
MCAs (phases 2–4). In the following discussion, ‘exchange’
refers to the totality of the language used within an interaction,
from the first word spoken to the last. An ‘utterance’ is one
speaker’s uninterrupted turn within the exchange.
1.2.1. Description and analysis of linguistic usage
Like all forms of discourse, communication by health-
care practitioners is embedded in the context, and can be
properly described only in conjunction with that context
[23]. It differs from natural conversation, however, in that its
communicative functions are defined by professional
requirements, which are often (as in the case of the present
study) explicitly codified. It is thus an example of
‘institutional talk’, the characteristics of which are [24]:
1. I
nstitutional interaction involves an orientation by at leastone of the participants to some core goal, task or activity
(or set of them) conventionally associated with the
institution (e.g. community pharmacy) in question. In
short, institutional talk is normally informed by goal
orientations of a relatively restricted conventional form.
2. I
nstitutional interaction may often involve special andparticular constraints on what one or both of the
participants will treat as allowable contributions to the
business at hand.
3. I
nstitutional talk may be associated with conceptualframeworks and organisational cultural values and proce-
dures that are particular to specific institutional contexts.
The analysis of operational discourse is typically
conducted on two levels: general behaviour and specific
procedures. The former level includes the protocols for
conducting interactions—who speaks to whom, the broad
M. Garner, M.C. Watson / Patient Education and Counseling 65 (2007) 51–57 53
discourse functions (or purposes of the interaction), and the
information to be transacted. The latter level includes the
sequencing of utterances in an exchange and the language by
which broad discourse functions are realized at utterance
level (such as the framing of questions, the marking of
functions, and the vocabulary used). In the present study, the
general communicative practices were fixed in advance. The
participants (MCA and customer) and the discourse
functions are assumed by the research topic, and the
information to be transacted has been defined by the
pharmacy profession (WWHAM). The linguistic analysis
thus concentrated on specific procedures, in particular the
sequencing and functions of utterances.
1.2.2. Communication planning
In this phase, improved communication practices and
procedures are designed in the light of the results of the
analysis. The stipulations for linguistic usage are couched in
varying degrees of strictness appropriate to the requirements
of the communicative context and the practitioners. They
may vary from firm rules reinforced by sanctions, to general
advice and recommendations for best practice.
1.2.3. Evaluation and modification
The designed practices and procedures are tested accord-
ing to a range of criteria appropriate to the setting, evaluated,
and modified as required. Evaluation takes into account both
linguistic issues (such as ambiguity, brevity, and cohesion)
and practical operational issues (such as the ease of learning
and using the stipulated or recommended forms). In contexts
in which information communication technology such as
radios, computers and telephones are employed, technologi-
cal issues also help to determine the evaluation.
1.2.4. Implementation
Once the final versions of the new practices and procedures
are agreed upon, they are formulated in a series of comm-
unication guidelines or a manual, and introduced to the body
of practitioners through an appropriate form of training.
2. Method
2.1. Procedure and participants
In 2004, a non-participant observation study was
conducted with 42 staff from 21 community pharmacies in
Grampian, Scotland. Each member of support staff was asked
to consent to 60 minutes of their working day (in relation to the
supply of advice and/or NPMs) being audiotaped using a clip-
on microphone. The study was approved by Grampian Local
Research Ethics Committee. A poster was placed in the
pharmacy window and at the medicines counter during
observation days, informing customers about the study. In
addition, MCAs were asked to inform each customer that they
were being taped and to seek their consent for this to continue
during the consultation. No customer requested the micro-
phone to be switched off during their consultation.
2.2. The corpus
Because the focus of the study was on the communicative
practices of MCAs themselves, a small number of interactions
which also involved the intervention of a pharmacist were not
included in the analysis. Also excluded were those sections of
any given interaction that were not related specifically to
NPMs (e.g., discussions about the weather).
2.2.1. Broad communicative functions
The functions of the exchanges are very limited. They are
to ensure that:
(a) th
e MCA elicits sufficient information to provide thecustomer with the appropriate NPM, and
(b) th
e customer is made aware of essential information,such as the recommended dose, as well as any warnings,
etc., related to the use of that medicine.
At the utterance level, three functions are theoretically
required for the realization of these discourse functions by the
MCA. Function (a) above would be realized by information-
eliciting utterances. Function (b) would be realized by
information-giving and advice-giving utterances.
For the purpose of this study, the information requisite to
achieving function (a) was defined as WWHAM. Discourse
function (b) would be expected to be realized through
advice-giving and information-giving utterances. Advice
may be general or specific. General advice relates to the type
of medication (if any) that would meet the customer’s needs:
for example, a recommendation to use an expectorant cough
medicine. Specific advice relates to the relative merits of
particular products. Advice may be stated in positive terms:
It’s a good idea to take these with food
or in negative terms (a warning or, less frequently, a
prohibition):
It’s not advisable to drink alcohol when you’re taking these
tablets.
Don’t drink alcohol if you’re taking these tablets.
Information in this function is commonly specific and
related to the particular product in question, for example,
concerning side effects or interactions with other medica-
tion. It is also possible to give specific information about a
range of products in order to enable the customer to make a
choice between them.
2.3. Analysis of the corpus
Utterances in the corpus were tagged, using standard
linguistic analysis processes, for specified functions by four
M. Garner, M.C. Watson / Patient Education and Counseling 65 (2007) 51–5754
research assistants working independently. The tagging
process was iterative: the taggings were compared,
differences discussed, and doubtful utterances re-tagged,
until no further agreement was possible. The few remaining
disputed taggings were ignored, and the utterances discarded
as unclassifiable (n = 22).
The utterances of the MCAs were tagged for the
functions: eliciting information; giving information; and
giving advice. Due to time constraints, it was decided not to
classify the specific nature of the information or advice in
each case. The responses to each of these by the customer
were also tagged according to the following categories:
� u
Ta
Tri
Tri
str
I
II
III
IV
V
VI
nelaborated information or acknowledgment (e.g., yes/
no/uh-huh)
� e
laborated information or acknowledgment (e.g., yes, I’vebeen taking these for several months)
� n
o verbal response (although there may have been aphysical response such as nodding)
� o
therTraining for MCAs, by focusing on the use of the
WWHAM mnemonic, stresses the elicitation of information.
The successful outcome of the interaction, however, cannot
be judged merely by whether the appropriate information
was elicited. The information elicited must also be used by
the MCA as a basis for advising or informing the customer
about the particular medicine in question.
Thus, the analysis focused on triads, or three-part
sequences [25], beginning and ending with an MCA
utterance:
MCAcustomer’s responseMCA2
The MCA2 is the critical utterance, as it reveals the
functional response to the customer’s utterance. Each MCA2
was tagged for the three MCA1 functions, together two
others. ‘Unelaborated acknowledgement’ refers to ‘yes’,
’OK’, ‘uh-uh’ or similar, and ‘no response’ refers to verbal
response. The triad as the unit of analysis focuses on the
interaction from the perspective of the initiating speaker (the
MCA), whose response to the customer’s utterance
completes the communicative unit.
ble 1
adic structure of utterances
adic
ucture
MCA1 Customer Number
of triadsa,
n (%)
No
resp
(%)
Information eliciting Yes/no 254 (50.3) 59
Information eliciting Elaborated response 112 (22.2) 78
Information giving Simple acknowledgement 30 (5.9) 57
Information giving Elaborated response 36 (7.1) 47
Advice giving Simple acknowledgement 47 (9.3) 55
Advice giving Elaborated response 26 (5.1) 35
a Total N = 505.
3. Results
3.1. Use of WWHAM
One of the preliminary aims of the analysis was to
ascertain whether WWHAM was used by MCAs in
interaction with customers. A visual inspection of the
corpus showed that the entire WWHAM mnemonic never
occurred, either in its suggested form or in some other
version. The two questions most frequently asked were ‘who
is the medication for’, and ‘is any other medication currently
being taken’.
3.2. Triads and utterance functions
The linguistic corpus thus defined consisted of 168
recorded and transcribed interactions. The interactions
consisted of between 4 and 19 utterances each, with an
average of 9.33 utterances per exchange. A total of 505
triads were identified 11 (2.1%) of which were initiated by
clauses that were also the third clause (i.e., MCA2) of an
immediately preceding triad. These were found to function
like, and were analysed as, independent triads. Of the 773
utterances, 60.9% were information eliciting, 12.6%
information giving, 13.5% advice giving and 11.3% were
‘‘other’’ or unclassifiable i.e. utterances with apparently
clear functions that were irrelevant to the immediate
interaction (such as utterances addressed to someone other
than the MCA), as well as those which were unintelligible or
whose function was unclear. Six combinations of MCA1 + -
customer’s response were possible in the two utterances that
precede the crucial MCA2, listed as triadic structures I–VI
(Table 1). Since the communicative focus was on the third
utterance (MCA2), Table 1 presents the percentages of each
type of third utterance in the triad (MCA2) in relation to
these six combinations of preliminary utterances.
Of the triads initiated by an information-eliciting clause,
regardless of whether the customer’s response was simple or
elaborated, in almost 100% of cases the MCA made no
further interactional use of the information. (That is not to
say that the information was ignored in every case. It is
possible that in some cases the MCA made a mental check of
the response and was satisfied that no further development of
onse
Simple
acknowledgement
(%)
Information
given (%)
Advice
given
(%)
Other
(%)
Further
information
given (%)
31 8
22
30 3 6 3
53
19 8 17
58 7
M. Garner, M.C. Watson / Patient Education and Counseling 65 (2007) 51–57 55
it was necessary. Nonetheless, it is hard to avoid the
conclusion that MCAs are unclear about what to do once the
customer has responded with the information requested.)
4. Discussion and conclusion
4.1. Discussion
A clear picture emerges of the communicative practices
of MCAs from this study, which has implications for the
future training of MCAs. Triadic structures I and II, which
commenced with the MCA eliciting information from the
customer, comprised the majority of triads. This is
unsurprising, given that the role of the MCA depends on
obtaining a clear picture of the customer’s needs and
relevant health-care history. The use of WWHAM in the
MCAs’ training reinforces the emphasis on obtaining
information. Of concern is that in every case the MCA
did not draw on the information in constructing the third
utterance of the triad: information was simply elicited and
(at best) acknowledged, e.g.:
� M
CA: Have you used these, the [brand name] eye dropsbefore?
� C
ustomer: Yes� M
CA: Uh-uh OK� M
CA: . . . Has [your husband] used [brand name] before?� C
ustomer: Yes, but not for a few years� M
CA: Uh-uhIn future studies, it may be helpful to tag information-
eliciting utterances for the nature of the information being
transacted, which may have an influence on the ensuing
utterances. For the purposes of the present study, however,
this would have made no difference as MCAs consistently
made no communicative use of information elicited in their
follow-up utterance (i.e. MCA2). An elaborated response by
the customer was significantly more likely to evoke a verbal
acknowledgement in MCA2. No effort was made to
incorporate the information provided in any MCA2.
Triadic structures III–VI were inaugurated by advice-
giving or information-giving utterances in MCA1. Advising
and informing are distinct communicative functions, but the
distinction between them is often very unclear in normal
conversation. Because giving advice can appear to imply
that the speaker has the right to control another’s behaviour,
the pragmatics of politeness make it more usual to be
indirect. The speaker will provide the relevant information
and allow the other to infer the appropriate advice or
warning:
The park is a dangerous place to walk at night.
Even more indirectly (and thus politely), the expression
of a personal viewpoint can be intended to function as a
warning:
I wouldn’t walk through the park at night myself.
In such a case, the listener is required to deduce that the
speaker has information on which to base this viewpoint, and
that this information is valid for the listener as well. The
distinction between informing and advising/warning is,
however, of practical communicative significance in certain
types of institutional interactions. In police operational
communication, for example, it can be dangerous to assume
that because an officer has been given information he or she
will make the required response. It is safer to provide both
the information and the advice and to mark them as distinct
in some way [26].
The same principle applies in communication involving
MCAs. Politeness is not an issue here, as it is assumed that
the MCA has specialist knowledge and is expected to make
that available to the customer. The successful outcome of an
interaction is judged by whether the customer obtains an
appropriate medication and is provided with both sufficient
information and appropriate advice or warnings in relation
to its use. The two functions need to be made explicit:
information does not by itself constitute advice. Equally, the
two functions also need to be explicitly connected, in one of
two ways. If information is given first, the implications of
the information should be stated along the lines of:
� M
CA: These tablets can cause drowsiness� C
ustomer: All right� M
CA: So you shouldn’t take them if you’re going to driveConversely, if advice is given first, the information on
which it is based should be provided:
� M
CA: You shouldn’t take these tablets if you’re going todrive
� C
ustomer: All right� M
CA: Because they can cause drowsinessIn the light of these considerations, triadic structures
III–VI reveal significant shortcomings in current practice.
When information was given in MCA1 and was simply
acknowledged by the customer (III), in only 6% of triads did
MCA2 provide follow-up advice, and in 3% further
information was given. Even more striking is the fact that,
if the customer made an elaborated response to the
information given in MCA1 (IV), advice was never given
in MCA2. In these cases, the MCA may unconsciously
assume that an elaborated response is sufficient indication
that the customer needs no further advice:
� M
CA: Right OK, something with all the travel sicknesspills, they may cause drowsiness, that’s just a side effect of
the medication
� C
ustomer: It’s just that I thought I’d take them and see, aslast year I was queasy
� M
CA: That’s OK, these ones are a good one here
M. Garner, M.C. Watson / Patient Education and Counseling 65 (2007) 51–5756
Whilst this assumption may happen to be correct, it is
nevertheless the MCA’s responsibility to ensure that relevant
advice is made explicit. A more appropriate MCA2 in the
example above would have been something like:
� M
CA: So you yourself shouldn’t driveThe lack of awareness of the need to connect information
and advice was even more evident in those triads initiated by
advice giving in utterance MCA1. There were no instances in
the corpus in which the MCA2 of these triads provided the
information on which the advice was based, as in the
following example:
� M
CA: . . . You are best to let it run its course for the first24 h
� C
ustomer: She knows that. I think she’s just thinking—� M
CA: yes ah-hahAs with the triads initiated by information giving (III
and IV), there is evidence of the effect of an elaborated
response by the customer on the function of MCA2. A simple
acknowledgement of an advice-giving MCA1 (V), prompted
further advice in the MCA2 in 17% of the triads. By contrast,
an elaborated response by the customer (VI) elicited no
further advice (or information) in the MCA2.
Whilst the findings of this study are suggestive, they
were based on data from a geographically limited area (the
north-east of Scotland), and there is at least the possibility
that they were influenced by regional communicative
style. Similar studies in other areas are needed before
definitive statements and widely applicable recommenda-
tions can be made for the communication training of
MCAs.
Limitations of resources meant that other potentially
significant features of MCAs’ discourse were not investi-
gated. These include the nature of the information transacted
in the ‘information eliciting’ and ‘information giving’
utterances (for example, information about price and
quantities was not distinguished from health information),
and it relationship, if any, to the structure of the triad. Advice
and warning, too, may be functionally different in their
effects, but these were not distinguished in the study. Finally,
as a piece of analytical research with an intended applied
outcome, this study represents only the first stage of a
process. Even though the results presented here were
remarkably clear, before they can be fully evaluated they
will need to be incorporated into a communication
intervention programme and evaluated. The evaluation
may in turn lead to further analysis of these and/or new data
before all relevant aspects of the discourse of MCAs can be
described. Nonetheless, the findings of this study were
sufficiently clear both to indicate directions for commu-
nication training and, more broadly, to suggest that there is
ample scope for worthwhile applied research in the health
services field.
4.2. Conclusion
Whilst it was evident from the transcripts that MCAs
attempted to interact with customers in a friendly and
helpful manner, they did not appear to have been made
sufficiently aware of the ways in which their exchanges,
having very specific intended outcomes, are crucially
different from natural conversation. MCAs’ communication
is dominated by the function of eliciting information. The
preponderance of this function is likely to be influenced by
the fact that most MCAs are taught the mnemonic WWHAM
(each of which stands for an information-eliciting question)
as part of their training. Moreover, asking questions
(information-eliciting utterances) is an easy and well-
defined communicative act that may seem to the MCAs to
give structure and perhaps authority to their role during
consultations. The findings indicate, however, that MCAs
lack a clear conception of the next communicative step after
information-eliciting utterances.
4.3. Practice implications
In order to help customers decide upon an appropriate
NPM, the MCA has the role of both informing and advising.
Information is the basis on which a well-informed decision
is made; advice provides a guide to the nature of that
decision. Training should focus on informing and advising
as distinct functions, and the potential problems that can be
caused by conflating them. As well as making MCAs aware
of the communication issues raised by this study, it will be
important to train them with the linguistic means to achieve
the desired outcomes. A well-designed training programme,
properly informed by linguistic research, will make a major
contribution to ensuring that they are able to use this tool to
maximum advantage.
Acknowledgements
We thank the medicine counter assistants, pharmacists
and customers who participated in the study from which
these data were derived. We also thank Ms. Jo Hart, PhD
student, who collected the audiotaped consultations. MW
was funded by a Medical Research Council Special Training
Fellowship in Health Services Research.
References
[1] Royal Pharmaceutical Society of Great Britain. Bibliography: Pre-
scription only medicines reclassified to pharmacy only medicines;
October 2004.
[2] Royal Pharmaceutical Society of Great Britain. Med Ethics Pract
2005;29.
[3] Porteous T, Bond C, Hannaford P, Sinclair H. How and why are
non-prescription analgesics used in Scotland? Fam Pract 2005;22:
78–85.
M. Garner, M.C. Watson / Patient Education and Counseling 65 (2007) 51–57 57
[4] Shi C, Gralnek IM, Dulai GS, Towfigh A, Asch S. Consumer usage
patterns of nonprescription histamine2-receptor antagonists. Am J
Gastroenterol 2004;99:606–10.
[5] Ferris D, Nyirjesy P, Sobel J, Soper D, Pavletic A, Litaker M.
Over-the-counter antifungal drug misuse associated with patient-
diagnosed vulvovaginal candidiasis. Obstet Gynaecol 2002;99:
419–24.
[6] Sihvo S, Klaukka T, Martikainen J, Hemminki E. Frequency of daily
over-the-counter drug use and potential clinically significant over-the-
counter-prescription drug interactions in the Finnish adult population.
Eur J Clin Pharmacol 2000;56:495–9.
[7] Sinclair HK, Bond C, Hannaford P. Over-the-counter ibuprofen: how
and why is it used? IJPP 2000;8:121–7.
[8] Bond C, Hannaford P. Issues related to monitoring the safety of over-
the-counter (OTC) medicines. Drug Safety 2003;26:1065–74.
[9] Watson MC, Bond CM, Grimshaw JM, Mollison J, Ludbrook A,
Walker A. Educational strategies to promote evidence-based commu-
nity pharmacy practice: a cluster randomised controlled trial (RCT).
Fam Pract 2002;19:529–36.
[10] Flint J. Training requirements for medicines counter assistants. Pharm
J 1996;256:858–9.
[11] NPA launches training with a W-WHAM. Pharm J 1989;243:40.
[12] Sharpe S, Norris G, Ibbitt M, Staton T, Riley J. Protocols: getting
started. Pharm J 1994;253:804–5.
[13] Watson MC, Bond CM. The evidence based supply of non-prescrip-
tion medicines: barriers and beliefs. IJPP 2004;12:65–72.
[14] Watson MC, Bond CM, Grimshaw JM, Johnston M. Factors predicting
the guideline compliant supply (or non-supply) of nonprescription
medicines in the community pharmacy setting. Qual Saf Health Care
2006;15:53–7.
[15] Parrott R. Emphasizing ‘communication’ in health communication. J
Commun 2004;54:751–87.
[16] Sirangi S. Towards a communication mentality in medical and
healthcare practice. Commun Med 2004;1:1–11.
[17] Widdowson H. On the limitations of linguistics applied. Appl Linguist
2000;21:3–25.
[18] Roberts C. In: Sirangi S, van Leeuwen T, editors. Applied linguistics.
2003. p. 133–49.
[19] Axley S. Managerial and organizational communication in terms of
the conduit metaphor. Acad Manage Rev 1984;428–37.
[20] Johnson E, Garner M. Operational communication: theory and prac-
tice. London: Palgrave Macmillan; 2006.
[21] Garner M, Johnson E. An approach to applied research into police call-
handling. Int J Speech Language Law, in press.
[22] Cameron D. Working with spoken discourse. Sage Publications; 2001.
[23] van Dijk T. Discourse as interaction in society. Sage Publications; 1997.
[24] Drew P, Heritage J. Talk at work: interaction in institutional settings.
Cambridge: Cambridge University Press; 1992.
[25] Renkema J. Introduction to discourse studies. Amsterdam: John
Benjamins Publishing Company; 2004. p. 166.
[26] Garner M, Hick S, Johnson E, Matthews D. Police Commun Language
Channel Tunnel. Cambridge: Policespeak Publications; 1993.
