Abstract
The breaking of bad news is a routine but difficult task for many health professionals. There are numerous anecdotes of insensitive practice but the subject has attracted little systematic research. We therefore interviewed 106 patients with advanced cancer (from an original sample of 195) to assess their perceptions of the doctors involved in their care. Aspects of the ‘breaking bad news’ event were recorded during discussion of the illness history and were subsequently rated. Participants were also asked to nominate doctors under the headings ‘most helpful’ and ‘less helpful’, and completed standardized psychological screening questionnaires.
In 94 of the 106 cases the bad news had been given by a doctor, usually a surgeon. Of the 13 doctors categorized as ‘most helpful’ when breaking bad news, 8 were general practitioners; of the 7 categorized as ‘less helpful’ all were surgeons. 69% of patients were neutral or positive about the bad-news consultation, but 20% were negative and 6% very negative. Doctors in surgical specialties were significantly more likely to be rated poorly than non-surgical specialists or general practitioners.
Surgeons were the group of doctors most likely to break bad news, but non-surgical doctors were rated more positively in performance of the task. This finding has implications for training.
INTRODUCTION
The breaking of bad news is an emotive subject for both health professionals and patients. Anecdotes abound of insensitive practice and patient distress, but there has been little systematic study1. The task is feared by junior doctors2 and struggled with by surgeons3,4. Even experienced oncologists acknowledge difficulty in detecting psychological distress5. In an American study of 55 patients, 74% had been told their diagnosis by a surgeon, only 11% by their primary care doctor. Those informed over the telephone or in the recovery room were most likely to describe the encounter in negative terms6. In the UK, Macmillan Cancer Relief surveyed 2000 patients7. Reports varied widely, but the mean time for the consultation was only 13 minutes, reflecting a generally hurried approach. Most respondents desired more information and 20% described harrowing experiences. In contrast, Seale8 found that bereaved relatives were largely positive about the manner in which they and the deceased patient had been informed.
The overall view is that a positive or negative bad-news experience can affect a patient's subsequent adjustment9, but few have measured the long-term effects. One attempt to link the bad-news ‘event’ with later psychological distress in parents of terminally ill children yielded no correlation, though parents were noted to retain vivid memories of the interview and were sometimes still preoccupied with it many years later10. A recent study of breast cancer survivors did find a relationship between positive perceptions of physician behaviour during the diagnostic consultation and psychological adjustment, but the effect was modest11. In this study, we aimed to examine the long-term psychological adjustment of patients in the terminal phase of their illness and to compare this with patient perceptions of their doctors' attitudes and skills, including their experiences of receiving bad news.
PATIENTS AND METHODS
195 adult (>18 years) patients with advanced cancer (estimated prognosis <1 year) were identified consecutively from various sources—cancer registries, oncology and respiratory clinics, palliative care nurses, general practitioners—and asked to participate in a study considering their experience of medical care. For the present substudy the interview included the Hospital Anxiety and Depression Scale12 and the Rotterdam Symptom Checklist13. A semi-structured subsection on the breaking of bad news was piloted in 6 patients but proved difficult to complete. This was modified to open discussion of the patient's history, details of the breaking-bad-news experience being recorded as they emerged, with specific comments noted verbatim. Where there was more than one episode (e.g. original diagnosis and later recurrence) participants were asked to describe the event that was most important to them.
Contemporaneously recorded interview notes were analysed and coded by the researcher (retrospectively for the first 66 participants, directly after the interview for the remaining 40). Responses were graded positive, neutral, negative, very negative, uncodable. The quality of the information given could not be assessed systematically; thus, grading reflected a global assessment of the patient's account, with emphasis on the personal qualities and overall supportiveness of the bad-news breaker, unless explanation issues were specifically raised. The ‘uncodable’ category applied where text was substantially lacking or uninformative, while ‘neutral’ was employed where the account either indicated no strong feelings or offered only factual information that did not allow for interpretation of the patient's views. While this might tend to skew the results towards neutral, it was expected to offset single-researcher bias by ensuring that the analysis was confined to interviews in which the patient had commented specifically on the way bad news had been given. The coding was then given a numerical rating.
Further on in the interview, patients were asked to consider all doctors with whom they had had contact in the context of their current illness, and to nominate a ‘most helpful’ and a ‘less helpful’ doctor. In this paper we report only on doctors who had given the bad news.
For statistical analysis, breaking-bad-news ratings in relation to doctors' specialties was examined by Fisher's exact test. The Hospital Anxiety Depression scale (HAD) has separate subscales for anxiety and depression; the Rotterdam Symptom Checklist (RSCL) has physical and psychological distress subscales. Individual scores were defined in terms of ‘caseness’14 on each subscale, with standard cut-offs of > 10. For comparisons, psychological scores were grouped together into three ‘psychological distress bands’ as follows: (1) ‘case’ on one or both HAD scales and/or ‘case’ on the RSCL (cut-off > 10); (2) ‘borderline’ on one or both HAD scales (score 8-10) (there is no borderline score described for the RSCL); (3) ‘non-case’ on all three. Psychological ratings, time from diagnosis and breaking-bad-news ratings could then be compared by one-way anova and chi-square.
RESULTS
Of 195 patients who were approached, 126 responded positively and 109 were interviewed at home: 3 were subsequently excluded (not terminal prognosis), leaving 106 for analysis.
In most instances (94/106), the bad news had been broken by a doctor: specialty was clearly identifiable in 85; a further 5 could be identified as hospital doctors, though 4 without specialty or grade and 1 was said to be a private consultant, specialty unknown. 77 doctors were fully identified by name.
73 patients (86%) had been given the news by a hospital specialist, usually (48/85) a surgeon; general practitioners were involved far less (13/85) and oncologists least (4/85). Grade was identifiable in 71/73 specialists and in 55 it was a consultant; where the task was performed by junior doctors (16/73) most were in surgical specialties (10/16). Nearly all patients were told face to face rather than by telephone or letter.
49% of patient accounts were neutral and 20% positive. However, in 20% memory of the event was negative and in 6% very negative (Table 1). Box 1 gives extracts of verbatim comments. Patients especially recalled individuals who were brusque, unsympathetic or impatient. In addition, the need for simple clear information was a persistent theme; one participant was particularly distressed by a junior doctor who was perceived to have given conflicting information on different occasions but subsequently denied doing so.
Table 1.
Breaking-bad-news coding (all patients, n=106)
| Bad news code | No. of patients (%) | No. of identifiable doctors | No. of partly identifiable doctors | No. of unidentified doctors | Unknown or indirect (includes family) source | Total |
|---|---|---|---|---|---|---|
| Not codable | 6 (6%) | 3 | 1 | 0 | 2 | 6 |
| Positive | 21 (20%) | 18 | 2 | 1 | 0 | 21 |
| Neutral | 52 (49%) | 37 | 0 | 6 | 9 | 52 |
| Negative | 21 (20%) | 14 | 4 | 2 | 1 (family) | 21 |
| Very negative | 6 (6%) | 5 | 1 | 0 | 0 | 6 |
| Total | 106 (101%) | 77 | 8 | 9 | 12 | 106 |
Box 1 Extracts from interview notes
Positive—Mrs A: The tumour was found at the hospital, but it was Dr X her own doctor (GP) who told her about it. He did it beautifully, kept her calm
Neutral—Mrs B: Referred to consultant surgeon with breast lump for needle aspiration; he told her the diagnosis—he was quite sympathetic, and she wasn't surprised.
Negative—Mr C: Had a colostomy done for a ‘blockage’. Later surgical team came and told him about operation; they had done all they could; nothing else to be done in hospital; it was now up to his GP. Patient's daughter rang consultant to ask for more information. He then came ‘storming’ onto the ward to see the patient, saying ‘I thought I'd told you’
Very negative—Mrs D: Developed a lump in her breast at 35. When seen after its removal, consultant surgeon told her she had cancer across her chest. She cried when she heard. He told her to ‘stop that, you've a lot more to go through’—and this was seeing him as a private patient (recalled 11 years after event).
A spread of ratings was obtained across all hospital specialists (Table 2). When doctors in surgical specialties were compared with those in non-surgical specialties, the distribution of negative/ very negative and neutral/ positive ratings differed, with surgeons performing worse (P exact=0.018). For the purpose of this analysis, oncologists were excluded. When gynaecologists (with their hybrid role) were excluded from the surgical group the result remained significant (P=0.04). Surgeons scored worse than general practitioners, though not significantly so (P exact=0.08). The grade of doctor had no significant effect, but the number of juniors was small.
Table 2.
Specialty, grade and ratings for all partially or fully identified doctors (n=85) (numbers in parentheses indicate juniors (jun) included in total)
| Spec | BBN+ | BBN neutl | BBN- | BBN v- | Not codable | Total |
|---|---|---|---|---|---|---|
| Oncology | 1 | 2 | 0 | 0 | 1 | 4 |
| Surg (jun) | 9 (3) | 18 (1) | 14 (4) | 5 (1) | 2 (1) | 48 (10) |
| Chest phys | 5 (1) | 4 | 1 | 0 | 0 | 10 (1) |
| Phys (jun) | 1 (1) | 3 | 0 | 0 | 0 | 4 (1) |
| Gynae (jun) | 1 (1) | 3 (2) | 1 (1) | 0 | 0 | 5 (4) |
| Cons unkn | 1 | 0 | 0 | 0 | 0 | 1 |
| GP | 2 | 8 | 1 | 1 | 1 | 13 |
| Total | 20 (6) | 38 (3) | 17 (5) | 6 (1) | 4 (1) | 85 (16) |
BBN+=positive experience of receiving bad news; BBN neutl=neutral experience; BBN-=negative experience; BBN v-=very negative experience; jun=junior; phys=physician; cons unkn=consultant of unknown specialty
In the categorization of ‘most helpful’ and ‘less helpful’ doctors in the whole course of the illness, 20 of the nominations were the doctors who had given the bad news. These categories were consistent with observer ratings of the bad-news experience (P exact=0.027) (Table 3). For acute hospital bad-news breakers there were 10 nominations—3 favourable, 7 (all surgeons) unfavourable. General practitioners, who were the bad-news breakers in 13 cases, received 8 nominations, all ‘most helpful’. Oncologists, bad-news breakers in 4 cases, had 2 nominations, both ‘most helpful’.
Table 3.
Breaking-bad-news coding where doctor breaking bad news was nominated (20 patients)
| Bad-news code | No. of ‘most helpful’ | No. of ‘less helpful’ |
|---|---|---|
| Not codable | 1 (8%) | 0 (0%) |
| Positive | 4 (31%) | 2 (29%) |
| Neutral | 6 (46%) | 0 (0%) |
| Negative | 2 (15%) | 4 (57%) |
| Very negative | 0 (0%) | 1 (14%) |
A total of 23/106 (22%) patients reached ‘caseness’ (i.e. significant psychological distress) on the HAD and/or the RSCL. However, many patients had very low scores (i.e. good psychological adjustment): the median scores for each subscale were HAD anxiety 4, HAD depression 5, RSCL psychological subscale 6.5.
There was no significant association between patients' current psychological morbidity and their perceptions of the bad-news event, in either direction; that is, patients with a positive memory were not protected from depression, and patients with highly negative memories were not more likely to be depressed (P=0.68).
There was a trend (by one-way anova) for shorter adaptation time from diagnosis to be associated with greater psychological distress, particularly in the case of recurrence (Table 4). However, there was no relation between time from diagnosis and breaking-bad-news rating; in other words, patients' recall and description of highly positive or negative events was not attenuated or enhanced by the passage of time (P=0.38).
Table 4.
Average time since diagnosis, according to depression band
|
Psychological distress band
|
|||
|---|---|---|---|
| Case (n=23) | Borderline (n=19) | Normal (n=64) | |
| Average time in wk since diagnosis (SD) | 36.09 (45.43) | 33.42 (31.62) | 37.52 (53.26) |
| Average time in wk since recurrence, or diagnosis if no recurrence (SD) | 15.87 (15.84) | 17.63 (21.07) | 20.52 (29.29) |
SD=Standard deviation
DISCUSSION
To learn that one has a life-threatening disease is a major event, and many patients recalled it with clarity even years later. This has been noted before6,10,15. The event itself did not seem to cause long-term psychological morbidity, although the findings must be interpreted with caution in view of the range of time from diagnosis to interview (unavoidable in a diagnostically heterogeneous sample of terminally ill patients). Another reassuring negative finding was that patients did not ‘shoot the messenger’16. Although doctors broke the bad news on 94 occasions, only 7 of these were categorized as ‘less helpful’. It was the patients at the two extremes—those who recalled the interview as very well or very badly done—who were most likely to recall the exact circumstances and the words used.
Bad news is broken most commonly by hospital specialists, and this is what most patients expect and desire17. In this study the specialist was usually a surgeon, and our diagnostically heterogeneous patient sample was probably representative. Doctors in surgical specialties were significantly more likely to be rated negatively, and to receive ‘less helpful’ nominations. Although these nominations were applicable to any doctor involved in the patient's care, the correlation between the bad-news rating and the nomination status did support the interpretation of causality, especially in the ‘less helpful’ category.
The numbers are small, and the study has several methodological flaws, both in its retrospective design and in the global rating system. These criticisms notwithstanding, there does appear to be a difference in patients' perceptions of doctors between surgical and other specialities in the breaking of bad news.
The reasons could include, first, heightened patient anxiety; many hospital specialists were involved at an earlier and more uncertain stage in the patient's illness. This could lead to a greater recall bias, though it applies equally to surgical and non-surgical diagnosticians. Secondly, patients had less contact with this group (2-6 encounters), so had less opportunity to develop rapport. In addition, treatment may have been limited or unsuccessful, so these doctors might have been perceived as ‘less helpful’ for these reasons over and above the breaking of bad news. Thirdly, the circumstances of consultation—e.g. busy ward or out-patient clinic versus familiar surgery—may be relevant and were not specifically recorded here. However, while these factors may differentiate between specialists and general practitioners, these too apply equally to surgical and non-surgical specialties. Time pressure is another factor cited by doctors as a particular problem. However, results from the Doctor-Patient Relationship Questionnaire in the full study showed that patients did not distinguish between ‘most helpful’ and ‘least helpful’ doctors on this item (i.e. they perceived all doctors as busy, but felt that they had enough time with both groups)18. Fourthly, the diagnosis could have influenced the nature of the discussions.
There is a fifth possible explanation—that doctors in surgical specialties were less effective communicators. While the General Medical Council recommended the inclusion of communication skills training in the undergraduate curriculum in 199319, this only recently became a formal requirement, and a survey in the early 1990s revealed a lack of training emphasis, both in time allocated and in formal assessment20. Among postgraduates, vocational training schemes have incorporated communication skills training for many years, and the Royal College of General Practitioners introduced formal assessment into its Membership examination in 1995. However, other Royal Colleges lagged behind and trainees in hospital specialties still rely largely on the example of seniors and their own experience. In a personal survey of 201 doctors18, only 14% had received any formal training as undergraduates in how to break bad news. 40% reported communication skills training at postgraduate level, but most of these were general practitioners. Among nominated hospital specialists (n=33) only 4 (12%) had received formal postgraduate training in how to break bad news. Our findings were reinforced by an interview survey of consultants regularly involved in breaking bad news of a cancer diagnosis21. While this group acknowledged the importance of this task, few had undertaken any formal training and were sceptical of its value, advocating the ‘apprenticeship’ model. This scepticism is widespread22, despite evidence that training can be useful and is retained23, that bad role models can perpetuate bad practice24,25, and that junior doctors lack competence in delivering bad news26. Among oncologists, for whom continuing contact with patients with progressive cancer is inherent in their daily practice, specially tailored workshops have proved both acceptable and beneficial27. However, this is not the group with primary responsibility for initial breaking of bad news.
Herein lies the nub of the problem. While this is not the first study to demonstrate that surgeons are the group most likely to make an initial cancer diagnosis, there is little or no assessment of performance in communication during surgical training. This can no longer be left to individual choice. In its NHS Plan28, and in the light of a public inquiry29, the Government has made clear that it expects all doctors to undertake communication skills training. We are now conducting a prospective study of the breaking of bad news in an unselected cohort of hospital consultants working in an acute trust that incorporates a cancer centre, and evaluating the effect of a range of educational interventions on consultant attitudes and practice and patient satisfaction.
Acknowledgments
I thank Professor J Dale, Centre for Primary Health Care Studies, University of Warwick, for critical reading; and Bristol Oncology Centre for provision of facilities and administrative support during the study fieldwork.
Ethical approval was granted by Bristol and Weston, Southmead and Frenchay ethics committees.
The study was funded by Macmillan Cancer Relief, Bristol and Weston District Research Committee and Special Trustees of United Bristol Hospitals Trust.
References
- 1.Ptacek JT, Elserhardt T. Breaking bad news: a review of the literature. JAMA 1996;276: 496-502 [PubMed] [Google Scholar]
- 2.Sykes N. Medical students' fears about breaking bad news. Lancet 1989;iii: 564. [DOI] [PubMed] [Google Scholar]
- 3.Burton MV, Parker RW. Psychological aspects of cancer surgery: surgeons' attitudes and opinions. Psycho-Oncology 1997;6: 47-64 [DOI] [PubMed] [Google Scholar]
- 4.Girgis A, Sanson-Fisher RW, McCarthy WH. Communicating with patients: surgeons' perceptions of their skills and need for training. Aust N Z J Surg 1997;67: 775-80 [DOI] [PubMed] [Google Scholar]
- 5.Ford S, Fallowfield L, Lewis S. Can oncologists detect distress in their outpatients and how satisfied are they with their performance during bad news consultations? Br J Cancer 1994;66: 767-70 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lind SE, DelVecchio Good MJ, Seidel S, Csordes T, Good BJ. Telling the diagnosis of cancer. J Clin Oncol 1989;7: 583-9 [DOI] [PubMed] [Google Scholar]
- 7.Cancer Relief Macmillan Fund. The Experience of a Cancer Diagnosis: Research Report. London: CRMF, 1991
- 8.Seale C. Communication and awareness about death: a study of a random sample of dying people. Soc Sci Med 1991;32: 943-52 [DOI] [PubMed] [Google Scholar]
- 9.Maguire P. Breaking bad news. Europ J Surg Oncol 1998;24: 188-91 [DOI] [PubMed] [Google Scholar]
- 10.Woolley H, Stein A, Forrest GC, Baum JD. Imparting the diagnosis of life threatening illness in children. BMJ 1989;298: 1623-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Mager WM, Andrykowski MA. Communication in the cancer “bad news” consultation: patient perceptions and psychological adjustment. Psycho Oncology 2002;11: 35-46 [DOI] [PubMed] [Google Scholar]
- 12.Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67: 361-70 [DOI] [PubMed] [Google Scholar]
- 13.de Haes JCJM, van Knippenberg FCE, Neijt JP. Measuring psychological and physical distress in cancer patients: structure and application of the Rotterdam Symptom Checklist. Br J Cancer 1990;62: 1034-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hopwood T, Howell A, Maguire P. Screening for psychiatric morbidity in patients with advanced breast cancer: validation of two self-report questionnaires. Br J Cancer 1991;64: 353-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fallowfield LJ. Giving sad and bad news. Lancet 1993;341: 476-8 [DOI] [PubMed] [Google Scholar]
- 16.Buckman R. Breaking bad news: why is it still so difficult? BMJ 1984;288: 1597-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Meredith C, Symonds P, Webster L, Lamont D, Pyper E, Gillis CR, Fallowfield L. Information needs of cancer patients in west Scotland: cross sectional survey of patients' views. BMJ 1996;313: 724-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Barnett MM. The Doctor—Patient Relationship in Advanced Cancer. [thesis] Bristol: University of Bristol, 1995
- 19.General Medical Council. Tomorrow's Doctors. Recommendations on Undergraduate Medical Education. London: GMC, 1993
- 20.Frederickson L, Bull P. An appraisal of the current status of communication skills training in British medical schools. Soc Sci Med 1992;34: 515-22 [DOI] [PubMed] [Google Scholar]
- 21.Walker G, Bradburn J, Maher J. Breaking Bad News—Establishing an Auditable Procedure for Giving the Cancer Diagnosis. London: King's Fund, 1996
- 22.Razavi D, Delvaux N. Communication skills and psychological training in oncology. Europ J Cancer 1997;33(suppl 6): S15-21 [DOI] [PubMed] [Google Scholar]
- 23.Maguire GP. Can communication skills be taught? Br J Hosp Med 1997;43: 215-16 [PubMed] [Google Scholar]
- 24.de Groot L. Pliable but not receptive: concerning the marginal influence of a medical psychology course on the socialization process of doctors. Med Ed 1987;21: 419-25 [DOI] [PubMed] [Google Scholar]
- 25.Moorhead R, Winefield H. Teaching counselling skills to fourth year medical students: a dilemma concerning goals. Family Practice 1991;8: 343-6 [DOI] [PubMed] [Google Scholar]
- 26.Eggly S, Afonso N, Rojas G, Baker M, Cardozo L, Robertson RS. An assessment of residents' competence in the delivery of bad news to patients. Acad Med 1997;72: 397-9 [DOI] [PubMed] [Google Scholar]
- 27.Fallowfield L, Lipkin M, Hall A. Teaching senior oncologists communication skills: results from Phase I of a comprehensive longitudinal program in the United Kingdom. J Clin Oncol 1998;16: 1961-8 [DOI] [PubMed] [Google Scholar]
- 28.Department of Health. NHS Plan. London: DoH, 2000
- 29.Kennedy I (chair), Learning from Bristol: the Report of the Public Inquiry into Children's Heart Surgery at the Bristol Royal Infirmary 1984-1995 [www.bristol-inquiry.org.uk ] [DOI] [PubMed]