Discussing weight loss opportunistically and effectively in family practice: a qualitative study of clinical interactions using conversation analysis in UK family practice.

BACKGROUND: GPs are encouraged to make brief interventions to support weight loss, but they report concern about these conversations, stating that they need more details on what to say. Knowing how engage in these conversations could encourage GPs to deliver brief interventions for weight loss more frequently.
OBJECTIVE: To examine which specific words and phrases were successful in achieving conversational alignment and minimizing misunderstanding, contributing to effective interventions.
METHODS: A conversation analysis of English family practice patients participating in a trial of opportunistic weight-management interventions, which incorporated the offer of referral to community weight-management services (CWMS). Qualitative conversation analysis was applied to 246 consultation recordings to identify communication patterns, which contributed to clear, efficient interventions.
RESULTS: Analysis showed variation in how GPs delivered interventions. Some ways of talking created misunderstandings or misalignment, while others avoided these. There were five components of clear and efficient opportunistic weight-management referrals. These were (i) exemplifying CWMS with a recognizable brand name (ii) saying weight-management 'programme' or 'service', rather than 'group' or 'club' (iii) stating that the referral is 'free' early on (iv) saying the number CWMS visits available on referral (v) stating that the CWMS programme available was 'local'.
CONCLUSIONS: When making a brief opportunistic intervention to support weight loss, clinicians can follow these five steps to create a smooth and efficient intervention. Knowing this may allay clinicians' fears about these consultations being awkward and improve adherence to guidelines.

Doctors are encouraged to offer referral to community weight-management services.

They rarely do so, stating they that do not know what to say.

By analysing recordings, we found out how to deliver clear, efficient referrals.

Early inclusion of key information contributed to clarity and efficiency.

Word choice could contribute to misunderstandings and interactional problems.

Small communication changes could avoid problems and support decision making.

Introduction

International guidelines, including those in the USA, Canada and UK, encourage family practice clinicians to opportunistically intervene on obesity (1–6). Evidence shows interventions that comprise referrals to behaviourally informed community weight-management services (CWMS) could reduce population mean weight and are acceptable to patients and clinicians (7).

These services provide multi-component behavioural interventions. The NHS contracts with programmes that provide dietary advice, physical activity advice and behaviour change components (8). International guidelines recommend family practice clinicians to offer opportunistic referral to such services, to support patient weight loss. However, research has demonstrated that discussions about weight occur rarely in family practice when seeking advice for weight loss is not the primary purpose of a patient’s visit (9). In a survey of 366 people with overweight and obesity, most stated wanting ‘more help with weight management than they are getting from their primary care physicians’ (10).

A recent trial (Brief Interventions for Weight Loss [BWeL]) showed that offering family practice patients CWMS referrals was acceptable to patients and resulted in a mean weight change of 2.43 kg in those who went on to attend (7). However, while we know these interventions can be successful, evidence shows that family practice clinicians, known in the UK as GPs, are reluctant to initiate discussions about weight (11). They state that talking about weight is difficult, and guidelines are too vague to provide support (11). Clinicians also stated concerns about opening a ‘Pandora’s box’ (12) that would take significant consultation time (13). However, should they intervene, GPs prefer offering programmes that encourage lifestyle change (14). Clinicians expressed a need for advice on specific words and phrases to use when delivering these interventions (11,13).

In this article, we address this evidence gap. We analyse GP–patient weight-management conversations during the BWeL trial, where consecutively attending patients with obesity were offered a free CWMS referral. We examine how GPs made an offer of a CWMS referral in the consultation. We explicate the core conversational practices that comprise such discussions, and consider which types of approach were successful in achieving maximum alignment between doctor and patient, and minimizing misunderstanding. We identify how interactional troubles and lengthy discussions could be avoided. We used conversation analysis (CA) to conduct a detailed empirical analysis of these interventions. Conversation analysis allows researchers to identify and build an evidence-base of effective conversational strategies (15). Identifying what to say to encourage patient understanding and avoid interactional troubles could encourage clinicians to engage more frequently in brief interventions for weight loss (11).

Methods

Context—the BWeL trial

In this qualitative study, we use conversation analysis to examine audio-recorded consultations from the brief interventions for weight loss (BWeL) trial. The BWeL trial was a parallel two-arm, randomized controlled trial assessing the effects of GP-delivered brief weight-loss interventions in family practice. Trial researchers asked to weigh, measure, and estimate the body fat of every patient waiting to see one of 137 participating GPs. Researchers aimed to enrol patients with a body mass index ≥ 30 kg/m2 (or ≥25 kg/m2 if Asian), aged ≥18 years. Patients excluded from the study were as follows: pregnant people (or those who were planning pregnancy); people who had experienced or were scheduled for bariatric surgery; people who had completed a weight-management programme 3 months before recruitment; people seeing their GP to discuss weight; and people who did not speak English.

Between 4 June 2013 and 23 December 2014, 1882 eligible patients consented to take part in the study and were randomized to an intervention arm (940 patients) or control arm (942 patients). Full details on randomization processes are available in the BWeL trial protocol and results paper (7,16).

Recording collection

At the end of a typical consultation, GPs in the intervention arm made an opportunistic intervention, which comprised endorsing, offering, and facilitating a referral to a behaviourally informed CWMS (either Slimming World or Rosemary Conley). These two services were chosen as they were commonly available in the NHS and had been shown to be effective (17). Half of patients were randomly selected for audio-recording. GPs audio-recorded using hand-held devices visible to GP and patient. Patients had the option to participate in the trial but decline audio-recording or to request deletion after the intervention had been delivered. Additionally, some GPs did not record; some recordings were unusable for technical reasons; some GPs delivered advice not support; and many recordings were not uploaded by the research team as they did not consider it a priority. This provided 246 recordings for analysis, in British English. Patient allocation and randomization for recording is illustrated in Figure 1. Recordings were from 77 doctors in 37 practices, and ranged in length from 8 to 458 seconds, with an average of 95 seconds. Data were stored on secure departmental servers at the Nuffield Department of Primary Care Health Sciences, and only the trial team had access to these. The BWeL trial was registered with the ISRCTN Registry, ISRCTN26563137 and approval was granted by NHS Research Ethics Service (reference no. 13/SC/0028).

Figure 1.

Patient allocation and randomization for recording.

Data analysis

CVAA, a conversation analyst specializing in advice giving in primary care, led the CA, transcribing available recordings using the standard Jefferson system for conversation analysis (18), which captures information about how turns at talk are delivered, including intonation, and onset of overlap. CVAA, HW and ES mapped the referral sequences systematically, identifying how GPs offered referrals and with what patient response. Webb is a conversation analyst experienced in analysing obesity consultations, and Stokoe is a professor of social interaction with expertise in institutional talk. In CA, the effectiveness or otherwise of each turn is revealed in the response in the next turn (called the ‘next turn proof procedure’ [NTPP]), rather than in the analysts’ subjective interpretation. We used the NTPP to identify what GPs did that patients (mis)understood and responded to well (or not so well). We conducted a detailed analysis of word choice, action format (e.g. how offers, explanations, etc., were designed), sequential positioning, prosody and action, to identify the core features of effective practice. We considered deviant cases and looked at responses in relation to the wider interactional sequence. Data were handled using NVivo11. Reporting follows the ‘Standards for reporting qualitative research’ (19).

Results

Analysis of 246 consultation recordings where GPs opportunistically offered patients a free CWMS referral showed that the following GP-initiated actions comprised these sequences:

Establish the patient has obesity

Assert evidence behind CWMS

State that a referral is available

Provide information about the referral

Ask if a patient would like to attend

We found that step 4, providing information, was especially relevant for securing frictionless uptake from patients. This was because this point in the interaction provided key details about the referral. We identified variation in how information was provided. Some ways resulted in misunderstandings or misalignment between doctor and patient. While others promoted alignment and understanding. We focus here on step 4 identifying conversational features that can encourage patient understanding and avoid interactional troubles.

We identified five components of the brief intervention discussion that, if absent from the ‘provide information’ stage could result in misunderstandings or misalignment. If these occurred, doctors needed to do significant work to clarify and rectify these troubles, extending the intervention length. If present however, these components could avoid conversational troubles, promoting clear and efficient conversations.

These aspects were:

Exemplifying CWMS with a recognizable brand name

Saying weight-management ‘programme’ or ‘service’

Stating that the referral is ‘free’ early in the intervention

Saying the number CWMS visits available

Stating that the CWMS was local

We illustrate these components of clear and efficient referrals with consultation extracts. Transcripts have been simplified from Jeffersonian format (Transcription Key, Table 1).

Table 1.

Simplified Jeffersonian transcription conventions

=	Equal signs indicate latching talk
[	Square brackets indicate the start of overlapping talk
→	Draws the reader’s attention to a particular line of the transcript
(0.3)	Numbers in parentheses indicate pauses in talk, measured in tenths of a second
(.)	A full stop in parentheses indicates a hearable pause of <0.3 seconds
Wo::rld	Colons indicate elongation of the immediately prior sound. The number of colons show the length of elongation
LOOK	Capitals indicate talk is markedly louder than the surrounding talk
.hh	A dot followed by a row of ‘h’s indicates an inbreath. The number of ‘h’s indicates the length of the inbreath
hh	‘H’s without a dot Indicates aspiration, such as outbreaths
-	A hyphen shows that a word or part of a word has been cut off
Y(h)e(h)s	‘H’s in brackets within a word indicates laughter during speech
°management°	Degrees signs either side of a word or TCU shows that the enclosed talk was markedly quiet or soft
↓ or ↑	Up and down arrows mark notable and/or sharp rises or falls in pitch
Yep. Fine.	Boldfaced consonants indicate that consonants were delivered with unusually hardened sounds
£Tha:t’s ri:ght.£	Pound-signs either side of a word or TCU shows that the enclosed talk was delivered with ‘smiley voice’ and indicates the speaker was smiling
It’s Weight Watchers.	A full stop marks falling intonation, giving some sense of completion
lifestyle change,	A comma marks gently rising tone, giving a sense of speaker continuation
How do you feel about that?	A question mark indicates a rising tone
(have)	Words in brackets mark a lack of certainty as to the exact word that was said

Exemplifying CWMS with a recognizable brand name

GPs often spoke about ‘weight-management services’, but evidence from patients’ responses showed that they often displayed that they did not understand this phrase.

For example, in Extract 1 (Table 2), the GP asserts that the best way to lose weight is through a ‘commercial weight-management service’ (lines 1–4). However, he does not say what these are; his statement presupposes that the patient will know. The GP then goes on to ask if the patient would like a referral (line 7). When he does not receive a response, the GP asks the question a second time (line 9), indicating orientation to trouble in the patient’s delayed response about attendance, not problems in understanding about the service itself. However, in his response (lines 10–11), the patient requests clarification on what a commercial weight-management service ‘actually is’ displaying that he has a problem understanding the way the GP had described this service. Patient responses across the data highlighted that the term ‘weight-management service’ could be problematic.

Table 2.

Transcribed extracts from audio-recorded GP-delivered referrals to community weight-management services

Exemplifying services with well-known brand names could avoid these troubles in understanding. In Extract 2 (Table 2), for example, the patient shows they do not understand, saying ‘a whatta?’ (line 4). The GP orients to this as a problem in hearing (not understanding), as they repeat ‘commercial weight-management system,’ rather than changing their terminology. The patient responds minimally to this, and it is only after the doctor exemplifies using a recognizable brand ‘like Slimming World’ (lines 7–8) that the patient displays that they have understood (line 9), with a change-of-state (20).

Saying weight-management ‘programme’ or ‘service’

GPs used a range of terms to describe the CWMS. They most frequently referred to a weight-management programme, group and service. Analysis showed that words ‘group’ and ‘club’ could cause interactional difficulties, as patients oriented to them as informal and social, rather than as a structured service. For example, in Extracts 3 and 4 (Table 2), the GPs mention weight-management ‘groups’.

In Extract 3, after the GP states that free group membership is available, the patient orients to the term ‘group’ as informal rather than structured support, stating he has his own ‘little group’ with some friends (line 5). In Extract 4, the patient also orients to problems with the term ‘group’ stating that he would not mix well as he is not a ‘sociable person’ (line 8). Here the GP subsequently undertakes additional interactional work to emphasize that it is not about mixing but receiving information.

While our data showed a pattern where ‘club’ and ‘group’ could be highlighted as problematic, there were few instances of interactional troubles when GPs spoke about ‘programmes’ or ‘services’.

Stating that the referral is ‘free’ early in the intervention

In these data, referrals were offered for free. We found that the word ‘free’ was important. For example, in Extract 5 (Table 2), the GP does not initially state that the referral is ‘free’. The patient displays initial reluctance to agree to attend the referral. She does not answer the GP’s question, and instead asks one of her own ‘the first thing is how much does it cost’ (lines 7–8). This indicates that she has many questions before responding to the GP’s question about referral and that her primary one refers to cost. The GPs then states that the referral is free, and the patient agrees, without asking the further questions to which she had alluded, indicating they are no longer relevant, and were potentially cost-contingent.

As well as highlighting the importance of the word ‘free’, we also found that where it was mentioned in the unfolding sequence, and in what kind of action it was embedded, was important for patient understanding and subsequent referral uptake. In some cases, GPs delayed stating that the referral was free and intervention length greatly increased. Figure 2 illustrates this, showing an extended discussion where the GP initially provides information about the referral from lines 1–5, omitting stating that the referral is free. The patient responds with a minimal ‘Yeah.’ at line 6, and the GP talks more about weight loss, but again omits saying that the referral is free. Following the referral question ‘would you be interested’ (line 14), the patient responds negatively, stating that she does not ‘have time’. The GP continues to talk about the benefits of attending, but the patient still does not respond positively. However, from lines 24–25, the GP delivers the information omitted earlier—the referral is free. The patient responds with a news-receipt ‘Oh right’, targeting the delayed delivery of ‘free’ as news-for-her (21). Following this, when the GP asks the referral question a second time, at line 39, the patient responds positively, with her assessment ‘Sounds good’. The patient targets this cost-related information as key in her decision making.

Figure 2.

An extended discussion where the GP initially omits stating that the referral is free.

These results illustrate a consistent pattern that showed that saying ‘free’ early in the discussion, before the patient had initially responded, facilitated patient displays of understanding about the offer, contributing to both conversational alignment and a briefer discussion.

Saying the number CWMS visits available

We also found that it was important for GPs to state the length of the referral. Patients often showed they thought only a portion of the referral was free, rather than twelve-weeks. For example, patients in Extracts 6 and 7 (Table 2) oriented to the need for more information, asking if it was ‘one day free’ or ‘just the joining fee’. When GPs stated ‘twelve-weeks’ (Extract 6) or ‘the full programme’ (Extract 7), patients did not request further details, frequently received this information positively, and the referral could progress smoothly to the next stage.

Stating that the CWMS was local

Evidence showed that the location of the referral was important for patients. GPs often did not state the referral location during the information-giving stage, and moved on to ask if a patient would like to attend. This presented interactional problems as patients often delayed answering the GP’s question about attendance, and instead asked questions about location. For example, in Extract 8 (Table 2), the GP has moved onto the next stage of the interaction, asking if the patient would like to attend. The patient does not respond to this question, indicating some interactional trouble (22). The GP orients to this trouble as a lack of understanding about the trial, providing more information about ‘questionnaires’. The patient responds to this quickly with a ‘yeah’ (line 5), indicating that this was not a problem. Instead they highlight location as a problem (line 7), asking if it will be ‘local’. The GP provides this information (line 8), and the patient responds positively, agreeing to attend.

If GPs stated that referrals were ‘local’ early in the discussion, patients did not ask for further clarification on location, and referrals usually progressed smoothly.

Discussion

Summary

GPs use of specific words and phrases and their location in the overall unfolding interaction can engage patients in brief opportunistic interventions for weight loss, smoothly progressing the conversation. Our analysis showed that some words or phrases hindered the clarity of referrals, and could lead to misunderstanding. We highlighted solutions that could facilitate smooth and efficient discussions. We first showed that patients often displayed trouble with what a CWMS referral ‘actually is’. We demonstrated that this could be avoided by exemplifying with recognizable brands, prefaced with ‘like’, to show these were examples of the type of group to which a patient could be referred. We also demonstrated that the words used when describing weight-management programmes were important. Descriptions that sounded ‘professional’ such as weight-management ‘programme’ or ‘service’ best facilitated a smooth move to next actions, while ‘group’ or ‘club’ could cause interactional difficulties. We identified that providing details and specificities about the cost, duration, and location of the referral early on seemed to expedite these interventions.

GPs have stated a need for more specific detail on how to intervene on patient weight and what to say when they do (11,13). We have shown how specific words and phrases, and their sequential placement, contribute to patient displays of understanding, smoothly progressing the consultation. This information is mostly absent from current guidelines (2,3), which only provide general advice. This detailed analysis highlighted one area that is present in current guidelines—talking about the location of referral. National Institute for Health and Care Excellence (NICE) guidelines on offering referral do exhort GPs to ‘discuss sources of long term support, such as from a…local support group or weight-management programme’ (Recommendation 6 (3)). They do not, however, explicitly encourage GPs to say the term ‘local’ or ‘free’, nor do they state where in a conversation this is best placed. Our analyses presented here showed that troubles could occur if the term ‘local’ was absent. Saying ‘local’ during information-giving, facilitated a smooth move to next actions. Our results here align with NICE guidelines, providing specific detail on how they can be best implemented.

Existing literature on talking about weight-management in family practice largely relies on post hoc reports from GPs and patients emphasizing their perceptions and experiences (12,23). These results complement these post hoc accounts, illuminating how weight-management interactions are carried out in practice.

Conversation analysis of weight-management discussions has mostly focussed on weight-loss advice (24). This is the first conversation analysis study to our knowledge to examine how GPs deliver very brief interventions for weight loss, incorporating the offer of CWMS referral.

Subsequent research from our team will build on these results, exploring relationships between in-consultation communication practices used by clinicians, and patient attendance at the weight-management service.

Strengths and limitations

A strength of this study was the conversation analysis of recorded data, meaning analysis was not limited by recall or social desirability biases. A further strength is the use of the next turn proof procedure, meaning analysis were grounded in aspects of the conversation highlighted by patients during the interaction, rather than by a priori assumptions. Data were collected across a number of surgeries and from diverse patient groups. A limitation was that we could not analyse multi-modal communication, as data were audio only. A potential limitation is that we do not know if the patient and GP had a prior relationship. Another limitation is that patients needed to orient to a particular word or phrase as problematic or useful for us to identify it. Other words and phrases may also hinder or facilitate the smooth running of the consultation, but we could not identify these from our data.

Conclusion

Family practice clinicians express concern about how best to intervene on patient weight and what to say to avoid initiating a lengthy discussion. Our analysis shows that including specific words during brief opportunistic weight-loss interventions facilitated a smooth referral. Saying ‘free’, ‘local’, talking about weight-management ‘services’ or ‘programmes’, and exemplifying services with recognizable brands avoided misunderstandings, and time spent rectifying these. By including these specific aspects during brief interventions, clinics can avoid inadvertently lengthening referrals and deliver brief interventions clearly and efficiently.