Hospital consultants breaking bad news with simulated patients: An ...

Patient Education and Counseling 83 (2011) 185–194

Contents lists available at ScienceDirect

Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou

Communication study

Hospital consultants breaking bad news with simulated patients: An analysis of communication using the Roter Interaction Analysis System Laura Vail a,*, Harbinder Sandhu b, Joanne Fisher b, Heather Cooke b, Jeremy Dale b, Mandy Barnett b a b

Health Sciences Research Institute, Warwick Medical School, University of Warwick, Coventry, UK Warwick Medical School, University of Warwick, UK

A R T I C L E I N F O

A B S T R A C T

Article history: Received 1 December 2009 Received in revised form 2 April 2010 Accepted 15 May 2010

Objective: To explore how experienced clinicians from wide ranging specialities deliver bad news, and to investigate the relationship between physician characteristics and patient centredness. Methods: Consultations involving 46 hospital consultants from 22 different specialties were coded using the Roter Interaction Analysis System. Results: Consultants mainly focussed upon providing biomedical information and did not discuss lifestyle and psychosocial issues frequently. Doctor gender, age, place of qualification, and speciality were not significantly related to patient centredness. Conclusion: Hospital consultants from wide ranging specialities tend to adopt a disease-centred approach when delivering bad news. Consultant characteristics had little impact upon patient centredness. Further large-scale studies are needed to examine the effect of doctor characteristics on behaviour during breaking bad news consultations. Practice implications: It is possible to observe breaking bad news encounters by video-recording interactions between clinicians and simulated patients. Future training programmes should focus on increasing patient-centred behaviours which include actively involving patients in the consultation, initiating psychosocial discussion, and providing patients with opportunities to ask questions. ß 2010 Elsevier Ireland Ltd. All rights reserved.

Keywords: Breaking bad news Patient centredness Roter Interaction Analysis System Doctor–patient communication

Breaking bad news is one of the most challenging tasks faced by health professionals, for which they often have little training [1]. It requires effective communication skills [2], and the way that bad news is communicated can affect patients’ level of anxiety [3,4], emotional adjustment to their diagnosis [5] and satisfaction with the consultation [6]. Drawing upon Brewin’s model [7], Schmid Mast et al. [8] describe three approaches which physicians can take when delivering bad news; disease, emotion, or a patient-centred approach. The disease-centred approach places little importance on communication within the bad news consultation, as it is thought that patients’ distress cannot be avoided. The physician’s role is to provide information with very little encouragement for patient participation. Alternatively, physicians can adopt a ‘kind and sad’ communication style [7], otherwise referred to as ‘emotion-centred’ [8]. Within this type of approach the news is given ‘gravely and solemnly’ with an emphasis upon the severity of the message. Lastly, the physician may adopt an ‘understanding

* Corresponding author at: Health Sciences Research Institute, Warwick Medical School, University of Warwick, Gibbett Hill Road, Coventry, CV4 7AL, UK. Tel.: +44 02476573957; fax: +44 02476528375. E-mail address: [email protected] (L. Vail). 0738-3991/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.pec.2010.05.016

and positive’ or patient-centred approach. This involves adapting behaviour according to the patient’s response, whilst maintaining an element of positive thinking and reassurance [7]. Schmid Mast et al. [8] found that the patient-centred approach was the most beneficial for patients, in relation to how satisfied they were with the consultation, and their emotional state after the bad news had been delivered. Many authors have presented recommendations for improving the quality of the breaking bad news consultation [9]. Baile et al. [1] produced the SPIKES protocol which recommends a six-stage process: setting up the consultation, assessing the patient’s perception of the illness, obtaining the patient’s permission to deliver the bad news, giving the knowledge, addressing the patient’s emotions with empathetic responses, and producing a strategy for the future and summarising. However, Ptacek and Eberhardt’s [10] review of the literature highlights that much of these recommendations lack an empirical basis, hence bringing to light the need for more rigorous research on how behaviour can be improved. Patients’ perspectives on the breaking bad news consultation have been explored in previous research [11,12]. Ptacek and Ptacek [11] found that cancer patients reported that they were satisfied with their ‘bad news’ consultation and that they reported that their physicians behaved consistently with published

186

L. Vail et al. / Patient Education and Counseling 83 (2011) 185–194

recommendations. Fujimori and Uchitomi’s [12] systematic review has highlighted that patient preferences can vary depending on demographic factors such as age, gender, educational background, and ethnicity. Detailed information and emotional support was preferred by younger patients, female patients, and highly educated patients, whereas less discussion of life expectancy and having a relative present were preferred more by Asian patients than Westerners. Health care providers’ perceptions on their behaviour during bad news delivery have also been explored [13–17]. A study by Ptacek and Ellison [13] has shown that physicians feel that they typically adopt a ‘warm and caring tone’ but do not always explore patient’s emotions or provide information at the patients pace. Other authors have found that physicians are aware of and actively carry out recommendations from the literature during the process of breaking bad news [14,16] but that there is scope for improving the discussion surrounding prognosis and referral to support groups [14]. However, the validity of self-reported data has been questioned by Ptacek and Ellison [13] who argue that physician accounts can provide a misrepresentation of what they actually do, hence highlighting the need for studies which investigate actual behaviours during the breaking bad news consultation. Despite the growing interest surrounding communication within consultations in which bad news is broken, there has been little empirical research looking at physicians’ actual behaviours when conveying bad news. Previous studies have focussed primarily upon observing the behaviour of medical students [18], trainees [19] or one occupational group of senior clinicians [20]. In general, the findings of these studies show that inadequacies in communication exist, hence highlighting the need for improving the interaction between doctor and patient when bad news is being broken. The work of Eggly et al. [19: 397] revealed that trainee internal medicine residents’ behaviour reflected a ‘lack of competence’, with physicians receiving low scores for the majority of skills deemed necessary for the effective delivery of bad news. With regards to oncologists, Ford et al. [20] found that whilst they provided patients with adequate biomedical information, they rarely explored patients’ emotions or instigated psychosocial discussion. A study [18: 1017] which investigated medical students performance found that they most commonly used non-medical (‘‘small talk and other topics not relevant to the health problem’’) and ‘information giving’ talk during the process of delivering bad news, and that they were perceived as lacking in effectiveness, empathy, care, and pleasantness by raters. Other studies provide insights into communication within cancer consultations, but lack a specific emphasis upon how bad news is broken [21,22]. Although oncologists are closely involved in sharing bad news, the initial explanation regarding a cancer diagnosis is often the responsibility of a clinician in the relevant diagnostic specialty [6]. Furthermore, clinicians from nearly all specialities regularly face the task of disclosing unwelcome information to patients [23,24]. Hence, it is important that research looks at how clinicians in general deliver bad news so that effective approaches to training are developed. Towards achieving this aim, we developed a multi-phase study investigating hospital specialists breaking bad news skills and the impact of training interventions. Findings from other phases of the study have been or will be reported elsewhere [24–29]. In this paper we focus on the following objectives: (1) To explore how clinicians from wide ranging specialities deliver bad news within the consultation, using simulated specialty-specific patients. (2) To investigate the relationship between physician characteristics (including specialty) and patient centredness.

We are therefore presenting the findings from the preintervention simulated patient consultations. 1. Methods 1.1. Participants in wider study The overall study population comprised all consultants (n = 285) working in a broad range of clinical specialties (Table 1) in three acute hospital trusts in the UK (two district general hospitals and a cancer centre, all affiliated to a new medical school). Each consultant was sent an invitation to participate, a questionnaire exploring their training, experience and attitudes towards breaking bad news [13], a description of the study, and a reply slip. Non-responders were sent up to two reminders. Consultants who responded were invited to take part in an educational training programme in which was preceded by a video-taped simulated consultation. 1.2. Data collection A bank of detailed breaking bad news scenarios was developed by three members of the research team (MB, JF and HC). Each contained similar elements involving a presentation of cancer, either newly diagnosed or recurrent, but in a context modified to be appropriate to each clinician’s specialty (see Box 1). 1.2.1. Simulator patients Simulators comprised actors recruited from local amateur dramatic groups, and occasionally members of faculty not known to the participating consultants. This provided a range of ages and mixed genders. 1.2.2. Setting up of simulator consultations Consultations took place in a purpose built video suite on the University campus, at a time convenient to the consultant and the simulator. The studio was laid out with chairs and a desk Table 1 Doctor characteristics (n = 46). Gender

Specialty

Age

Qualified inside or outside UK

Female n = 9

Medical n = 19

Qualified inside UK n = 32

Male n = 37

Care of the elderly n = 1 Clinical Oncology n = 1

Mean age = 44.5 SD = 7.016 Range 32–64

Rheumatology n = 1 Nephrology n = 3 Genito urinary medicine n = 1 General medicine n = 1 Stroke medicine n = 1 Haematology n = 1 Cardiology n = 1 Dermatology n = 2 A&E n = 3 Rehabilitation n = 1 Paediatrics n = 2 Surgical n = 12 Orthopaedics n = 4 Oral maxillofacial surgery n = 1 General surgery n = 3 Cardiothoracic surgery n = 1 Obstetrics and gynaecology n = 3 Anaesthetics n = 11 Radiology n = 4

Qualified outside UK n = 14 Sri Lanka (n = 1) Germany (n = 1) Belgium (n = 1) India (n = 6) West Indies (n = 1) Holland (n = 1) Pakistan (n = 1) Caribbean (n = 1) South Africa (n = 1)


187

Box 1. Example scenarios for consultants by speciality. Speciality

Scenario

Radiology

Jeff Adams is a 39-year-old Planning Officer (a keen runner). He is sent up for a plain film of his knee on a Friday afternoon, after presenting to his GP with a hot swollen knee and no history of injury and a short history of shortness of breath. He is refusing to leave the department until he has more information as he is planning a major run on Sunday. The film shows the unmistakable signs of an osteosarcoma in the lower femur. A chest X-ray was also requested and shows cannonball metastases in both lungs. The GPs surgery number is on answering machine. He is married with two young children.

Dermatology

Caroline Harris, a 32-year-old housewife, is referred to you urgently by her GP, having presented with a suspicious mole on her calf. You examine her and find an obvious malignant melanoma and clearly she has extensive metastatic lymph node spread in her groin. She also has a swollen leg. She is anxious to know what you are looking for. She is married and has a young child.

Obstetrics and Gynaecology

Your patient is a 70-year-old retired housing officer (Mrs. Edith Cooper) who had a hysterectomy and radiotherapy for carcinoma of the cervix last year. She attended follow-up last week complaining of lower abdominal pain and groin pain. Ultra-sound has revealed extensive pelvic recurrence. You also suspect the presence of bony secondaries, but a bone scan has yet to be carried out. She is married and her husband is in the early stages of senile dementia.

approximating to the clinical outpatient layout, but participants were allowed to arrange furniture as they wished. Filming was carried out by two cameras and a microphone located remotely so that their position was not visually intrusive. Both simulator and participating consultant were given individualised written summaries of the respective patient/ relative scenario. The scenario was set up so that the doctor and patient had not met before and this was made clear to participants. Participants were allowed several minutes to familiarise themselves with their scenario and ask for any factual clarification. Simulators were informed that they were going to receive a piece of news relating to the written scenario. They were aware that the consultation was being recorded as part of a research exercise and that the doctors were experienced health professionals. They were not given specific cues to respond to but were asked to consider the information they were given about the person they were portraying and then to respond to the doctor as they would in real life. Consultants were told that the interview would be filmed for up to 15 min, and that they could finish at any point but would receive a warning knock at 13 min into the consultation. 1.3. Coding of data We used the Roter Interaction Analysis system (RIAS) to code the consultations. This system, which is based on Bales Process Analysis [30] is a widely used system for the assessment of medical interaction and has been used in over 75 communication studies [31]. Analysis involves the classification of ‘‘utterances’’ to categories that reflect the content and context of the discussion. Utterances are ‘‘the smallest discriminatible speech segment to which a classification may be assigned’’ and coding is carried out directly from the tape in order to allow for the assessment of tonal qualities ([32]: 4). The RIAS contains 40 categories which can be classified as either task focused/instrumental (e.g. giving instructions or providing information about a condition) or social-emotional/affective (e.g. showing reassurance or concern). For analysis in this study these categories were further classified into composite groups which are shown in Table 2 (clinician categories) and Table 3 (patient categories). The composite groups focus upon data gathering (open and closed questions), patient education and counselling (information giving and persuasive talk), affective talk (building a rapport with the patient—social, positive, negative and emotional) and process-orientated (partnership building) talk. The coder for this study was highly experienced in using the RIAS, had been trained by the developers and had worked on a number of studies using the system previously. A random sample of 11% (n = 9) consultations was selected for double coding which revealed a good intra-rater reliability of 0.94 (range 0.84–1.00) using Pearson correlation coefficient.

1.4. Data analysis We used descriptive statistics to explore the content of consultations. In addition, patient-centredness scores were Table 2 Clinician categories. Group

Combination of RIAS categories

Data gathering—biomedical

Asks closed-ended question—medical condition Asks closed-ended question—therapeutic regimen Asks closed-ended question—other Asks open-ended question—medical condition Asks open-ended question—therapeutic regimen Asks open-ended question—other Bid for repetition

Data gathering—lifestyle/ psychosocial

Asks Asks Asks Asks

Patient education and counselling—biomedical

Gives information—medical condition Gives information—therapeutic regimen Gives information—other Counsels—medical condition/ therapeutic regimen

Patient education and counselling—lifestyle/ psychosocial

Gives information—lifestyle Counsels—lifestyle/psychosocial

Facilitation and patient activation

Asks for opinion Asks for permission Asks for reassurance Asks for understanding Back-channel responses Paraphrase/check for understanding

Rapport building—positive

Laughs, tells jokes Shows approval—direct Gives compliment—general Shows agreement, understanding

Rapport building—emotional

Empathy/legitimation Shows concern or worry Reassures encourages or shows optimism Partnership statements Self-disclosure statements

Rapport building—negative

Shows disapproval—direct Shows criticism—general

Rapport building—social

Personal remarks, social conversation.

Procedural

Transition words Gives orientation, instructions Unintelligible utterances

closed-ended question—lifestyle closed-ended question—psychosocial open-ended question—lifestyle open-ended question—psychosocial


188 Table 3 Patient categories. Grouped categories

Combination of RIAS categories

Question asking—biomedical

All questions medical All questions therapeutic regimen All questions other Bid for repetition

Question asking—lifestyle/ psychosocial

All questions lifestyle All questions psychosocial

Information giving—biomedical

Gives information—medical condition Gives information—therapeutic regimen Gives information—other

Information giving— lifestyle/psychosocial

Gives information—lifestyle Gives information—psychosocial

Patient activation and engagement

Asks for service Asks for reassurance Asks for understanding Paraphrase/check for understanding

Rapport building/positive

Laughs, tells jokes. Shows approval—direct Gives compliment—general Shows agreement, understanding

Rapport building/emotional

Empathy/legitimizing statements Shows concern or worry Reassures encourages or shows optimism

Rapport building/negative

Shows disapproval—direct Shows criticism—general

Rapport building/social

Personal remarks, social conversation

Procedural

Transition words Gives orientation, instructions Unintelligible utterances

calculated for each consultation using the following formula [the sum of the patient’s utterances related to biomedical/psychosocial/ lifestyle information giving, emotional statements, lifestyle/ psychosocial question asking, and the physicians lifestyle/ psychosocial questions and counselling statements, partnering and emotional statements] divided by [the sum of the physician’s utterances related to medical/therapeutic regimen questions, medical/therapeutic regimen information giving and counselling statements, procedural statements, and patient utterances related to medical/therapeutic regimen question asking]. Data were skewed and therefore we conducted Mann– Whitney’s U-tests to determine the relationship between patient centredness scores and doctor gender, patient gender, age of doctor, and place of qualification. A Kruskal–Wallis test was performed to investigate the effect of speciality on patient centredness score. Descriptive statistics are reported in Table 6. We investigated whether consultants followed aspects of steps 2, 3, 4 and 5 of the SPIKES protocol [1]. Three proficiency codes were added into the RIAS database to allow for assessment of the extent to which doctors: (1) enquired about patients perceptions of their illness before the bad news was delivered, (2) obtained patients invitation to deliver the bad news, (3) provided a warning that the bad news was coming step 5 of the protocol (addressing the patient’s emotions with empathetic responses) was assessed using the RIAS empathy code. 2. Results 2.1. Response rate and respondent profile Of the 285 invitations to participate that were sent to consultants, 9 were returned as the specialist was not in post (3 had left, 3 retired, 1 was on long term leave, and 2 were suspended for disciplinary reasons). Of the remaining 276, 173 (63%) replied

with 153 (55%) returning questionnaires, and 120 (44%) agreeing to participate in the educational intervention. Of these, 49 consultants (17.8% of possible respondents) were video-taped within the timescale of the study. One participant subsequently asked to have the video destroyed and did not consent to its formal analysis; a second participant asked to retain the video for professional development purposes, so preventing it from being coded; a third the recording equipment malfunctioned. The characteristics of these doctors are summarised in Table 1. Those video-taped represented a broad range of clinical specialties (22/33 specialties originally approached). Their gender distribution was representative of the wider consultant population, the sample population approached comprised 85% male, 15% female; whilst participants were 80% male, 20%. Likewise, 25% of the wider population were international graduates compared to 30% of those that participated. 2.1.1. Simulator characteristics Within the 46 consultations, 28 were carried out using a female simulated patient and 18 with a male simulated patient. Twentyone of the actors were acting as patients aged over 50, and 25 were acting as a patient aged 50 or under. 2.2. Consultation attributes The median length of consultations was 16 min (range 7–32, inter-quartile range 13.5–17) and the median number of utterances per consultation was 319.5 (range 167–671, inter-quartile range 260–354). Of these, the median number of utterances spoken by the doctor was 199 (range 95–397, inter-quartile range 158– 235) and for the patient 116 (range 48–274, inter-quartile range 92–137). The median patient centredness score was 1.0 (range = 0.3–2.5, inter-quartile range 0.8–1.5). 2.2.1. Doctor communication Table 4 reports descriptive statistics for the content of doctors’ communication. As indicated, the majority of doctors’ utterances were concerned with providing biomedical information (43%; 3983/9240 utterances). This included information concerning diagnosis, prognosis, future appointments, and actions to be undertaken by the patient: ‘‘There is something there squashing the bone’’ (Tape 111, RIAS category: gives information—medical). ‘‘There are two things we need to do for this’’ (Tape 248, RIAS category: gives information—therapeutic regimen). ‘‘If you’ve got any questions, even when you get out of this room, write them down’’ (Tape 040, RIAS category: counsels— medical/therapeutic regimen). In contrast, relatively few utterances were focussed upon lifestyle or psychosocial information (5.9%; 548/9240): ‘‘Keep your family and friends around you’’ (Tape 169, RIAS category: counsels—lifestyle/psychosocial). The second most frequent type of utterance was associated with rapport building, with a particular emphasis upon emotions; for example, concern, empathy, reassurance, forming an alliance with patient, and self-disclosure (20.9% 1893/9240 utterances): ‘‘We’re always here, we’re always on the phone’’ (Tape 005, RIAS category: partnership). ‘‘Today’s treatments are very good’’ (Tape 146, RIAS category: reassures, encourages or shows optimism). ‘‘I’m sorry to say the mole was abnormal’’ (Tape 283, RIAS category: concern).


189

Table 4 Content of talk during the breaking bad news consultation (clinician). Communication category

Number of utterances per consultation: median, range, inter-quartile range

Data gathering—biomedical

Range = 0–13 Median = 3.0 Inter-quartile range = 2–5

173 (1.8%)

Asks closed-ended question—medical condition—62 Asks closed-ended question—therapeutic regimen 19 Asks closed-ended question—other—0 Asks open-ended question—medical condition—76 Asks open-ended question—therapeutic regimen—12 Asks open-ended question—other—0 Bid for repetition—4

Data gathering—lifestyle/psychosocial


146 (1.6%)

Asks Asks Asks Asks

Patient education and counselling—biomedical


Patient education and counselling—lifestyle/psychosocial

Range = 0–58 Median = 10.0 Inter-quartile range = 4–17.3

Facilitation and patient activation

Range = 6–73 Median = 24 Inter-quartile range = 13.8–36




Range = 8–86 Median = 41 Inter-quartile range = 31–50.1


Range = 0–5 Median = 0 Inter-quartile range = 0.0–1


Range = 0–6 Median = 3.0 Inter-quartile range = 1.75–4

124 (1.3%)

Personal remarks, social conversation—124

Procedural


203 (2.2%)

Transition words—120 Gives orientation, instructions—83

Patient-centredness score

Range = 0.3–2.5 Median = 1.0 Inter-quartile range = 0.8–1.5

Number of utterances (% total utterances)

3983 (43%)

548 (5.9%)

1214 (13%)

932 (10%)

1893 (20.9%)

24 (0.3%)

Combination of RIAS categories (with total number of utterances)

closed-ended question—lifestyle—67 closed-ended question—psychosocial—20 open-ended question—lifestyle—26 open-ended question—psychosocial—33

Gives information—medical condition—1541 Gives information—therapeutic regimen—2282 Gives information—other—5 Counsels—medical condition/therapeutic regimen—155 Gives information—lifestyle—89 Counsels—lifestyle/psychosocial—332 Gives information—psychosocial—127 Asks for opinion—166 Asks for permission—2 Asks for reassurance—2 Asks for understanding—157 Back-channel responses—508 Paraphrase/check for understanding—379 Laughs, tells jokes—37 Shows approval—direct—43 Gives compliment—general—14 Shows agreement, understanding—838 Empathy—126 Legitimation—75 Shows concern or worry—859 Reassures encourages or shows optimism—700 Partnership statements—91 Self-disclosure statements—42 Shows disapproval—direct—21 Shows criticism—general—3

The behaviours associated with a higher patient-centredness score are highlighted in bold; the behaviours which impact negatively on patient-centredness scores are in italics.

‘‘It’s a bit scary, I understand’’ (Tape 286, RIAS category: empathy). ‘‘I spoke to a friend last week’’ (Tape 111, RIAS category: selfdisclosure). Partnership building statements accounted for 13% (1214/ 9240) of doctor utterances, with the majority of these statements being back-channel responses and checks for understanding: ‘‘Is it okay if we talk to your wife about this?’’ (Tape 306, RIAS category: asks for opinion). ‘‘You came into hospital three days ago?’’ (Tape 040, RIAS category: Paraphrase/checks for understanding).

Patient: She knew she wasn’t well. . .she didn’t know. . . Doctor: Yes. Patient: she was on the brink. . .. (Tape 190, RIAS category: back-channel response). In total, doctors asked 168 closed questions and 147 open questions across the 46 consultations related to the four RIAS categories; lifestyle, psychosocial, medical condition and therapeutic regimen. These tended to be biomedical questions (1.8% utterances) rather than psychosocial (1.6%). ‘‘Tell me what’s been happening to you?’’ (Tape 023, RIAS category: Open Question—Medical).


190

‘‘Do you have a neighbour or a friend that you are close to?’’ (Tape 146, RIAS category: Closed Question—Psychosocial). Social ‘chit chat’ statements were uncommon and when they did occur they appeared most commonly within the opening segment of the consultation. Statements expressing negative emotions such as criticism and disapproval were also rare. 2.2.2. Patient communication The most frequent type of patient utterance was associated with positive rapport building (34.2% 1880/5497 utterances) which was mainly due to the high level of ‘agreement’ statements: Doctor: ‘‘I’m afraid it’s something we have to check’’ Patient: ‘‘Yeah’’ (Tape 183, RIAS Category: agree). Patients frequently gave information about lifestyle or psychosocial issues (18.4% 1011/5497 utterances), and about their emotions concerning their illness (15.1%; 832/5497), but offered less biomedical information (12.2% 670/5497): ‘‘My husband is in Brussels. He gets back on Friday’’ (Tape 262, RIAS Category: gives information—lifestyle, 2 utterances).

‘‘I don’t want to think this is the end of the road. I’m not that sort of person’’ (Tape 111, RIAS Category: Gives information— psychosocial, both utterances). ‘‘I’m still a bit chesty’’ (Tape 164, RIAS Category: Gives information—medical).

Statements representing patient activation and engagement in the consultation were relatively low, accounting for 6.5% (354/ 5497) of utterances. Such statements reflected the patients’ level of interaction with the doctor and involved making requests for services, asking for reassurance, asking the doctor if they have understood the information that they have provided, and reflecting information back to the doctor to clarify what had been said. Although patients did not ask questions frequently, when they did, they asked more from a biomedical (7.7%; 427/5497 utterances) than a lifestyle/psychosocial perspective (0.9% 50/ 5497 utterances) (see Table 5 for the descriptive statistics for patient communication during the consultations). 2.3. Association between doctor characteristics and patient centredness Table 6 provides the descriptive statistics for patient centredness scores by gender, speciality, age and place of qualification.

Table 5 Content of talk during the breaking bad news consultation (patient). Number of utterances (% total utterances)

Combination of RIAS categories (with total number of utterances)

Communication category

Number of utterances per consultation: median, range, inter-quartile range

Question asking—biomedical


427 (7.7%)

Question asking—lifestyle/psychosocial


50 (0.9%)

All questions lifestyle—34 All questions psychosocial—16

Information giving—biomedical

Range = 2–52 Median = 12 Inter-quartile range = 8–21

(670 (12.2%))

Gives information—medical condition—462 Gives information—therapeutic regimen—205 Gives information—other—3

Information giving—lifestyle/psychosocial


Patient activation and engagement


354 (6.5%)



1880 (34.2%)


Range = 1–56 Median = 14.5 Inter-quartile range = 8.8–23.5


Range = 0–7 Median = 0 Inter-quartile range = 0–1.3

59 (1.1%)

Shows disapproval—direct—33 Shows criticism—general—26



51 (0.9%)

Personal remarks, social conversation—51

Procedural


1011 (18.4%)

832 (15.1%)

163 (3%)

All questions medical—199 All questions therapeutic regimen—223 All questions other—0 Bid for repetition—5

Gives information—lifestyle—392 Gives information—psychosocial—619

Asks for service—3 Asks for reassurance—141 Asks for understanding—25 Paraphrase/check for understanding—185 Laughs, tells jokes—53 Shows approval—direct—124 Gives compliment—general—14 Shows agreement, understanding—1689 Empathy—0 Legitimizing statements—4 Shows concern or worry—726 Reassures encourages or shows optimism—102

Transition words—151 Gives orientation, instructions—12

L. Vail et al. / Patient Education and Counseling 83 (2011) 185–194 Table 6 Clinician characteristics and patient centredness. Clinician characteristic

Patient-centredness score

Male

Mean = 1.0 Median = 1.0 Range .34–2.20 Inter-quartile range = 1.0–1.3

Female

Mean = 1.5 Median = 1.4 Range = .79–2.52 Inter-quartile range = .9–2.2

Qualified inside the UK

Mean = 1.2 Median = 1.1 Range = 0.34–2.52 Inter-quartile range = .85–1.5

Qualified outside the UK

Mean = .94 Median = .97 Range = 0.45–1.53 Inter-quartile range = .73–.97

Medical


Surgical

Mean = 0.9 Median = .96 Range = 0.50–1.42 Inter-quartile range = .60–1.1

Anaesthetics

Mean = 1.0 Median = 1.1 Range = .34–1.62 Inter-quartile range = .8–1.4

Radiology


44


Doctor gender, speciality, place of qualification and age were not significantly related to patient centredness. 2.4. Consistency with the SPIKES protocol Within the 46 consultations, patient perceptions of their illness were explored in just over half (24 consultations). A warning shot was used within 25 consultations and an invitation to deliver the bad news was demonstrated in 3 consultations. Across the 46 consultations, a total of 126 empathy statements were used by the clinicians, with 36 containing at least one empathy utterance (median = 2, range = 0–10, inter-quartile range = 1–4). 3. Discussion and conclusion 3.1. Discussion This study is the first that we are aware of to describe how experienced consultants from a broad range of specialties break bad news. Its findings suggest that bad news is delivered generally in a way that focuses upon providing biomedical information, with

191

little consideration to psychological issues. High level of ‘agreements’ from patients were observed, which appeared to reflect patients spending much of the consultation listening to the clinician rather than being more actively involved. This pattern of consultation is similar to that reported for oncologists, whose behaviour during the breaking bad news consultation has been described by Ford et al. [20]. There are some key differences between the findings of our study and those of previous research investigating doctor–patient communication during the breaking bad news consultation. Whereas oncologists were found by Ford et al. [20] to use closed rather than open questions, in our study open and closed questions were quite evenly spread. Our sample also received higher patient-centredness scores than the oncologists within the study by Ford et al. (0.33–0.41), who also used a RIAS based scoring method to assess patient centred behaviour. Our findings also contrast with those of Gilloti et al. [18], who found that non-medical talk was the most common communicative behaviour during bad news delivery. However, this study involved medical students and used different coding tools. In addition, the bad news was a diagnosis of HIV, provoking a discussion of lifestyle and psychosocial events leading up to infection. The extent to which the pattern of communication during the bad news consultation is influenced by the patient’s diagnosis needs further investigation. The breaking bad news consultations observed in our study contained elements of all three approaches described by Schmid Mast et al. [8], although the ‘disease centred’ approach appeared to be the most prominent. Essentially the clinicians were focussed upon providing biomedical information. However, the high level of emotional talk, especially concerns, within our consultations correlates with the emotion-centred approach to breaking bad news, whilst the high frequency of reassuring statements is more in line with the patient-centred approach. Guidelines for breaking bad news tend to focus upon specific actions to be carried out (e.g. finding out what the patient already knows about their condition before disclosure) [1] within consultations and do not refer directly to the proportions or use of different types of communicative behaviours. This raises the question about how much biomedical information is sufficient within a breaking bad news consultation, and to what extent doctors should be discussing psychosocial issues whilst disclosing bad news. Research has shown that patients prefer a doctor to adopt a ‘patient-centred’ approach during cancer consultations [33]. However, there is some evidence that patients place higher importance on the doctor’s information giving role and expertise than on their exploration of psychosocial issues during the bad news consultation [34]. Other studies have demonstrated that using checklists containing both symptom and psychosocial items to prompt patient participation lead both to more patient-centred communication and lower patient anxiety in the context of oncology consultations [35]. This group also found that patient preferences to discuss psychosocial issues varied both with patient characteristics and at points in their disease history, particularly over preferences regarding discussion of prognosis [36]. Concerns about raising prognosis has been found in other patient interview studies [37], but appears more likely to be discussed when both patient and doctor are mutually prompted to do so [38]. Clearly there is a subtle balance to be struck, and further exploration is needed to investigate the extent to which psychological issues should be explored within the breaking bad news consultation across all specialties, and how this is to be incorporated within the discussion between doctor and patient. Although our study failed to identify evidence that consultants’ age, speciality, place of qualification, and gender impacts upon patient-centred behaviour, this needs to be interpreted carefully owing to the small sample size. Other research [39,40] has found

192


that male and female doctors communicate differently in medical consultations. Female doctors have been shown to adopt a more ‘patient-centred’ interviewing style by displaying more warmth, responsiveness, empathy and engaging in more psychological counselling and question asking, in comparison to male doctors. In relation to breaking bad news specifically, self-reported data has shown that females take longer to deliver bad news, and that they carry out more ‘emotionally supportive’ behaviours [14]. When drawing upon this literature, it would be expected that communicative behaviour within breaking bad news consultations would differ significantly according to gender, and that females may receive higher patient-centredness scores, due to a higher frequency of rapport building (positive and emotional), and psychological data gathering and counselling statements. In our study the number of female participants was too small for such a difference to be seen, although it was representative of the proportion of females in the wider consultant body. From the patient perspective, a recent study of real patients found that the majority had no preference for the gender of the doctor conveying bad news to them, assuming otherwise equal skills [41]. Further research is needed to explore the impact of gender specifically on communicative behaviour during breaking bad news consultations, and also how this relates to patient satisfaction and outcomes. Our research did not reveal any differences in patientcentredness scores between younger (44) consultants. It might have been anticipated that younger members of the sample would achieve higher patient centredness scores as communication skills training has only recently been a formal requirement of medical school curriculum [42]. Patient characteristics, such as gender and age, have been shown to impact upon communication in health care consultations [39]. Across our 46 consultations, simulator patient age was relatively evenly spread. The majority of our patients, however, were female, and this may have influenced our findings. Research has shown that female patients receive more medical information/ jargon [43], empathy [44] and are asked about their opinions and feelings more frequently [45]. In terms of patient’s own behaviour, females have been shown to express more emotion [46] and ask more questions [47,48]. Compliance with the SPIKES protocol was inconsistent, with just over half of the consultations including an enquiry about patient perceptions of their illness, and just over half including a warning shot. Very few consultants complied with step 3 of the protocol by obtaining the patients invitation to deliver the bad news (occurring in only 3 consultations). However, the majority did comply with step 5, to some extent, as empathy was displayed in the majority of consultations (36). Our findings suggest that compliance with recommendations from the literature is patchy, and that there is considerable scope to improve consultants’ behaviours. In contrast to our findings, selfreported data has shown that physicians feel that they do carry out many of the communicative behaviours outlined within the literature, which suggests that there can be mismatches between what physicians say they do, and what is observed within empirical investigations using observational methods. Further large-scale research studies, using observational methods, are needed to explore the extent to which physicians comply with recommended practice and whether physicians’ characteristics impact upon compliance. 3.2. Limitations of the study The study targeted a population of nearly 300 consultants, of whom only a relatively small proportion was successfully recruited and video-recorded within the time period of the study. Given the

small sample from each specialty, the failure to identify interspecialty differences needs to be interpreted cautiously, and further work is needed to explore in more depth doctor characteristics in relation to patient-centred behaviour and the breaking bad news consultation. There may be limitations to our study due to the method used to assess patient centredness. There are two features of patient-centred communication that were not specifically coded within this study. Firstly, the use of silence and active listening on the part of the doctor. Training interventions aimed at doctors have focused on this aspect of communication in particular as an area requiring improvement. The fact that this could not be coded within the existing RIAS software, means that patient centredness may have been under-estimated in consultations with a high percentage of ‘active’ silence. Secondly, nonverbal empathic gestures were not recorded within the RIAS. Therefore, empathy conveyed by non-verbal or non-utterances was not picked up, which may have lead to under-estimation of empathy. Although this group appeared to be representative in terms of age, gender and specialty of the wider study population (based on characteristics of questionnaire responders and a GMC register search of non-responders), it was likely to be selfselecting in favour of those interested in the process of communication. Hence, it is possible that doctors who chose not to participate may have been less patient-centred that those observed here. Our participants were limited to a 15-min consultation which was in line with the time available within the standard outpatient setting. Although they were not given a definitive endpoint that they must reach, the majority completed and closed the consultation close to this time limit. Studies among both general practitioners [49] and consultants representing different disciplines [50] have shown that a perception of time limitation influences clinicians’ use of more doctor-centred behaviours; those working within longer consultation times are more likely to encourage patients to open up psychological concerns, and are in turn awarded higher satisfaction scores by patients [50]. Hence, it is possible that the consideration of a time limit influenced the doctors’ communication, although this clearly applies equally in the actual clinic situation. Another potentially confounding variable is the use of patient simulators. Whilst they provided a realistic proxy for patients with a wide variety of specialty-specific conditions against which doctors’ communications skills could be assessed, they did not provide the range of patient characteristics that are encountered across everyday clinical practice. For example, they did not include patients who had difficulty communicating in English. Facilitating behaviours may be more important to patients who are less confident communicating with their doctor, or for whom English was their second language [51]. 4. Conclusion This study is unique in exploring the interaction between hospital consultants from wide ranging specialities and simulated patients during ‘breaking bad news’ consultations. It provides evidence that suggests consultants are mainly focussed upon providing biomedical information and do not actively involve patients within the consultation. Doctor characteristics such as age, gender, place of qualification and speciality did not emerge as being related to patient-centred behaviour during bad news delivery. However, due to the small sample size used within our research, further investigations are needed to clarify the relationship between doctor characteristics, and patient-centred communication within breaking bad news consultations.


4.1. Practice implications This study has shown that it is feasible to video-record breaking bad news consultations involving hospital consultants from widely ranging disciplines by utilising simulated patients in a simulated clinical setting. Our findings indicate a predominantly ‘diseasecentred’ approach to breaking bad news, and that there is considerable scope for increasing patient-centred behaviours. The importance of actively involving patients in the consultation, initiating psychosocial discussion, and providing patients with opportunities to ask questions is widely recognised, but the extent to which this can be changed as a result of training at undergraduate and postgraduate levels needs further investigation. 4.2. Areas for future research As our research was small scale, we are only able to suggest tentatively that physician characteristics are not related to patientcentred behaviour during the breaking bad news consultation. We are, however, able to provide directions for future research, as our study had highlighted a number of gaps in the literature. Most importantly, there is a need for larger scale research looking at the impact of physician characteristics on communicative behaviour during bad news delivery to further elaborate on aspects of the doctor that influence patient centredness within the breaking bad news consultation. Funding The study was funded by the NHS Executive West Midlands: Locally Organised Research Scheme (LORS) Project Grant £70,000 2000–2002. Ethical approval Coventry Research Ethics Committee. Acknowledgements We would like to thank the hospital consultants who agreed to be video-taped for the purposes of this study, and our clinical colleagues among the medical school faculty who helped us to develop specialty-appropriate scenarios. We are also grateful to the actors for volunteering their time. References [1] Baile WF, Buckman R, Lenzi R, Glober G, Beale EA, Kudelka AP. SPIKES—a sixstep protocol for delivering bad news: application to the patient with cancer. Oncologist 2000;5:302–11. [2] Ngo-Metzger Q, August KJ, Srinivasan M, Liao S, Meyskens FL. End-of-life care: guidelines for patient centered communication. Am Fam Physician 2008; 77:167–74. [3] Takayama T, Yamazaki Y, Katsumata N. Relationship between outpatients’ perceptions of physicians’ communication styles and patients’ anxiety levels in a Japanese oncology setting. Soc Sci Med 2001;53:1335–50. [4] Maguire P. Breaking bad news. Eur J Surg Oncol 1998;24:188–99. [5] Sardell AN, Trierweiler SJ. Disclosing the cancer diagnosis. Procedures that influence patient hopefulness. Cancer 1993;72:3355–65. [6] Barnett M, Fisher J, Wild A, Cooke H, Irwin C, Dale J. An audit of documentation of breaking bad news: can we tell who said what to whom? Clin Manage 2002;11:181–4. [7] Brewin TB. Three ways of giving bad news. Lancet 1991;337:1207–9. [8] Schmid Mast M, Kindlimann A, Langewitz W. Recipients’ perspective on breaking bad news: how you put it really makes a difference. Patient Educ Couns 2005;58:244–51. [9] Walker G, Bradburn J, Maher J. Breaking bad news—establishing an auditable procedure for giving the cancer diagnosis. London: King’s Fund; 1996. [10] Ptacek JT, Eberhardt TL. Breaking bad news. A review of the literature. J Amer Med Assoc 1996;276:496–502.

193

[11] Ptacek JT, Ptacek JJ. Patients’ perceptions of receiving bad news about cancer. J Clin Oncol 2001;19:4160–4. [12] Fujimori M, Uchitomi Y. Preferences of cancer patients regarding communication of bad news: a systematic review. Jpn J Clin Oncol 2009;39: 201–16. [13] Ptacek JT, Ellison NM. Health care providers’ perspectives on breaking bad news to patients. Crit Care Nurs 2000;23:51–9. [14] Farber NJ, Urban SY, Vollier VU, Weiner J, Polite RG, Davis EB, et al. The good news about giving bad news to patients. J Gen Intern Med 2002;17: 914–22. [15] Dickson D, Hargie O, Brunger K, Stapleton K. Health professionals’ perceptions of breaking bad news. Int J Health Care Qual Assur 2002;15:324–36. [16] Ptacek JT, Fries EA, Erberhardt TL, Ptacek JJ. Breaking bad news to patients: physicians’ perceptions of the process. Support Care Cancer 1999;7: 113–20. [17] Ptacek JT, Ptacek JJ, Ellison NM. ‘‘I’m, Sorry To Tell You. . .’’ physicians reports of breaking bad news. J Behav Med 2001;24:205–17. [18] Gillotti C, Thompson T, McNeilis K. Communicative competence in the delivery of bad news. Soc Sci Med 2002;54:1011–23. [19] Eggly S, Afonso N, Rojas G, Baker M, Cardozo L, Robertson S. An assessment of residents’ competence in the delivery of bad news to patients. Acad Med 1997;72:397–9. [20] Ford S, Fallowfield L, Lewis S. Doctor–patient interactions in oncology. Soc Sci Med 1996;42:1511–9. [21] Ishikawa H, Takayama T, Yamazaki Y, Seki Y, Katsumata N. Physician–patient communication and patient satisfaction in Japanese cancer consultations. Soc Sci Med 2002;55:301–11. [22] Ong LML, Visser MRM, Kruyver IPM, Bensing JM, Van Den Brink-Muinen A, Stouthard JML, et al. The Roter Interaction Analysis System (RIAS) in oncological consultations: psychometric properties. Psychooncology 1998;7:387– 401. [23] Fallowfield L, Jenkins V. Communicating sad, bad, and difficult news in medicine. Lancet 2004;363:312–9. [24] Barnett M, Fisher J, Cooke H, James P, Dale J. Breaking bad news: consultants’ experience, previous education and views on educational format and timing. Med Educ 2007;41:947–56. [25] Barnett M. Effect of breaking bad news on patients’ perceptions of doctors. J R Soc Med 2002;95:343–7. [26] Barnett M, Fisher J, Cooke H, Dale J. Developing an audit instrument for breaking bad news. Qual Prim Care 2003;11:287–91. [27] Barnett M, Fisher J, Cooke H, Dale J, Irwin C. Effectiveness of training in breaking bad news among hospital consultants: a prospective evaluation of three brief training methods. In: Paper, EAPC congress; 2003. [28] Cooke H, Barnett M, Fisher J, Dale J. Breaking bad news good practice recommendations—the patient’s voice. In: Poster, 1st interdisciplinary conference on communication, medicine and ethics; 2003. [29] Fisher JD, Cooke H, Barnett M, Dale J, Irwin C. What do patients suspect about their cancer diagnosis and why?In: Poster: BPOS annual conference; 2001. Abstract: Psychooncology 11:549–550, 2002. [30] Bales RF. Interaction process analysis. Cambridge: Addison–Wesley; 1950. [31] Roter D, Larson S. The Roter Interaction Analysis System (RIAS): utility and flexibility for analysis of medical interactions. Patient Educ Couns 2002; 46:243–51. [32] Roter D. The roter method of interaction process analysis. RIAS manual. Baltimore; 2008. Available from URL: http://www.rias.org/manual. [33] Dowsett SM, Saul JL, Butow PN, Dunn SM, Boyer MJ, Findlow R, et al. Communication styles in the cancer consultation: preferences for a patient-centred approach. Psychooncology 2000;9:147–56. [34] Parker PA, Baile WF, de Moor C, Lenzi R, Kudelka AP, Cohen L. Breaking bad news about cancer: patients’ preferences for communication. J Clin Oncol 2001;19:2049–56. [35] Velikova G, Booth L, Smith AB, Brown PM, Lynch P, Brown JM, et al. Measuring quality of life in routine oncology practice improves communication and patient well-being: a randomized controlled trial. J Clin Oncol 2004;22: 714–24. [36] Velikova G, Awad N, Coles-Gale R, Wright EP, Brown JM, Selby PJ. The clinical value of quality of life assessment in oncology practice – a qualitative study of patient and physician views. Psychooncology 2008;17:690–8. [37] Barnett MM. Does It Hurt To Know the Worst?—psychological morbidity, information preferences and understanding of prognosis in patients with advanced cancer. Psychooncology 2006;15:44–55. [38] Brown RF, Butow PN, Dunn SM, Tattersall MH. Promoting patient participation and shortening cancer consultations: a randomised trial. Brit J Cancer 2001;85:1273–9. [39] Roter DL, Hall JA. Doctors talking with patients/patients talking with doctors. Westport: Praeger Publishers; 2006. [40] Roter DL, Hall JA, Aoki Y. Physician gender effects in medical communication. J Amer Med Assoc 2002;288:756–64. [41] Cox A, Jenkins V, Catt S, Langridge C, Fallowfield L. Information needs and experiences: an audit of UK cancer patients. Eur J Oncol Nurs 2006;10:263–72. [42] General Medical Counsel. Tomorrow’s doctors. London: GMC; 1993. [43] Hall JA, Irish JT, Roter DL, Ehrlich CM, Miller LH. Gender in medical encounters: an analysis of physician and patient communication in a primary care setting. Health Psychol 1994;13:384–92. [44] Hooper EM, Comstock LM, Goodwin JM, Goodwin JS. Patient characteristics that influence physician behaviour. Med Care 1982;20:630–8.

194


[45] Stewart M. Patient characteristics which are related to the doctor–patient interaction. Fam Pract 1984;1:30–6. [46] Hall JA, Roter DL. Patient gender and communication with physicians: results of a community-based study. Women’s health: research on gender. Behav Policy 1995;1:77–95. [47] Wallen J, Waitzkin H, Stoeckle JD. Physician stereotypes about female health and illness. Women Health 1979;4:171–6. [48] Pendleton DA, Brochner S. The communication of medical information in general practice consultations as a function of patients’ social class. Soc Sci Med 1980;14A:669–73.

[49] Howie JG, Porter AM, Heaney DJ, Hopton JL. Long to short consultation ratio: a proxy measure of quality of care for general practice. Br J Gen Pract 1991;41:48–54. [50] O’Neill J, Williams JR, Kay LJ. Doctor–patient communication in a musculoskeletal unit: relationship between an observer-rated structured scoring system and patient opinion. Rheumatology 2003;42: 1518–22. [51] Zandbelt LC, Smets EMA, Oort FJ, Godfried MH, De Haes HC. Medical specialists’ patient-centered communication and patient-reported outcomes. Med Care 2007;45:330–9.