Annals of Oncology Advance Access published February 6, 2014

original article Annals of Oncology 00: 1–5 , 2014

doi:10.1093/annonc/mdt582

Breaking bad news–what patients want and what they

get: evaluating the SPIKES protocol in Germany
C. Seifart1*, M. Hofmann2, T. Bär1, J. Riera Knorrenschild3, U. Seifart4 & W. Rief2
1
Institutional Review Board; 2Department of Clinical Psychology and Psychotherapy; 3Department of Internal Medicine, Division of Hematology and Oncology, Philipps-
University of Marburg, Marburg; 4Hospital Sonnenblick, Germany

Received 30 October 2013; accepted 16 December 2013

Downloaded from http://annonc.oxfordjournals.org/ at Universitaetsbibliothek Muenchen on July 2, 2014

Background: Evaluation of the SPIKES protocol, a recommended guideline for breaking bad news, is sparse, and infor-
mation about patients’ preferences for bad-news delivery in Germany is lacking. Being the first actual–theoretical com-
parison of a ‘breaking bad news’ guideline, the present study evaluates the recommended steps of the SPIKES protocol.
Moreover, emotional consequences and quality of bad-news delivery are investigated.
Patients and methods: A total of 350 cancer patients answered the MABBAN (Marburg Breaking Bad News Scale), a
questionnaire representing the six SPIKES subscales, asking for the procedure, perception and satisfaction of the first
cancer disclosure and patient’s assign to these items.
Results: Only 46.2% of the asked cancer patients are completely satisfied with how bad news had been broken to
them. The overall quality is significantly related to the emotional state after receiving bad news (r = −0.261, P < 0.001).
Patients’ preferences differ highly significantly from the way bad news were delivered, and the resulting rang list of patients’
preferences indicates that the SPIKES protocol do not fully meet the priorities of cancer patients in Germany.

original article
Conclusions: It could be postulated that the low satisfaction of patients observed in this study reflects the highly signifi-
cant difference between patients’ preferences and bad-news delivery. Therefore, some adjunctions to the SPIKES proto-
col should be considered, including a frequent reassurance of listeners’ understanding, the perpetual possibility to ask
question, respect for prearrangement needs and the conception of bad-news delivery in a two-step procedure.
Key words: breaking bad news, spikes protocol, bad-news delivery, cancer diagnosis

introduction and steps of these protocols. One comprehensive guideline that

had been developed in Australia [5] is based on the consensus of
In the medical context, bad news are ‘any news that drastically patients, nurses and doctors. After implementation, the evalu-
and negatively alters the patient’s view of her or his future’ [1]. ation of the protocol in Australia in melanoma patients showed
Bad news means a kind of information that starts a new life era that several modifications of the protocol should be reasonable
for the patient. Breaking this kind of news is a frequent and following patients’ preferences [8]. Another study evaluated the
difficult task for every physician, independent of her or his spe- consensus between breast cancer patients, oncologists and on-
cialty. It is particularly common in the oncological setting that cology nurses on the guidelines for breaking bad news and
life-threatening and life-influencing diagnoses are frequently showed, using a consensus of 70%, that there is a high level of
given to the patients [2, 3], such as newly diagnosed cancer or agreement between the groups. However, there was only little
undesirable developments of a known cancer. The quality of the consistency in the ranking of the checked items [9], suggesting
delivery of bad news to patients seems to be directly related to significant differences between doctors’ and patients’ preferences
patients’ stress and anxiety, their adjustment to the bad news, on how to break bad news. Additionally, there is rare evidence
coping and satisfaction with care and health outcome [4]. about the adherence to these guidelines and about the impact
There are a few established recommendations for the delivery these guidelines may have on satisfaction among patients.
of bad news in the United States [4, 5], Australia [6] and the UK The most popular guideline, the SPIKES protocol [1], recom-
[7]. These protocols are primarily based more on expert mends a six-step protocol for delivering bad news, with a special
opinion, but less on empirical evidence. Only a few studies have application for cancer patients [2]. It was evaluated for structur-
evaluated patient-based evidence for the recommended items ing the delivery of bad news in the United States and it had
reached guideline status in America and in a number of other
countries [10], including Germany [11]. It is used as a guide for
*Correspondence to: Dr Carola Seifart, Philipps-University of Marburg, Dekanat,
Institutional Review Board, Marburg, Baldingerstraße 1, 35039 Marburg, Germany. this sensitive practice and for communication skills training in
E-mail: carola.seifart@staff.uni-marburg.de this context [11, 12]. The acronym SPIKES refers to six steps

© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: journals.permissions@oup.com.
original article Annals of Oncology

recommended for breaking bad news: (i) Setting up the inter- statistical analysis of the present study. Full statistical analyses of quality cri-
view, (ii) assessing the patient’s Perception, (iii) obtaining the teria of the MABBAN have been reported elsewhere [13].
patient’s Invitation, (iv) giving Knowledge and information to The Hospital Anxiety and Depression Scale (HADS) was used to measure
the patient, (v) addressing the patient’s Emotions with empathic psychological morbidity. It consists of two subscales, each comprising seven
responses and (vi) Strategy and Summary [1]. items. The subscales assess depression and anxiety separately.
Evaluation of the SPIKES protocol is sparse and insufficient,
and further studies are needed. Neither the SPIKES protocol has results
been evaluated in Europe before nor is there sufficient informa-
tion how bad news are broken in Germany and about patients’ patients
preferences for delivering bad news. Therefore, we conducted a Of the 350 returned questionnaires, six were filled in incom-
survey based on the items of the SPIKES protocol and asked pletely. The remaining 344 were able to be used for the analysis.
cancer patients how bad news had been broken to them and Main reason for answer-missing was comprehension difficulties
about their preferences for the related communication process. of participants due to old age. Demographic characteristics of
Being the first actual–theoretical comparison of a ‘breaking bad the study population are listed in supplementary Table S1, avail-
news’ guideline, the recommended steps of the SPIKES protocol able at Annals of Oncology online. The age range of the partici-

Downloaded from http://annonc.oxfordjournals.org/ at Universitaetsbibliothek Muenchen on July 2, 2014

were evaluated, each concerning patient’s preferences and the pants was 20–85 years, with a mean of 58.4 years (SD = 11.8).
perception of how they had been realized in bad-news delivery. The most common types of disease were breast cancer (27.6%),
Moreover, emotional consequences of bad-news delivery are hemic diseases (22.4%) and colon rectum cancer (11%). Only
investigated. 11.3% had been diagnosed with recurrent disease. The mean
The study revealed four main results: (i) only a minority of time since diagnosis was 24.36 months (SD = 40.28 months),
the asked cancer patients are completely satisfied with the com- and 56.6% of the participants got the diagnosis within the past
munication process by which bad news had been broken to year.
them, (ii) patients’ preferences differ significantly from the way
bad news were delivered, (iii) a rang list of patients’ preferences ‘breaking bad news’ characteristics
indicating that the SPIKES protocol do not fully meet the prior- Characteristics of the setting and the bearer of bad news are
ities of patients in Germany for breaking bad news has been shown in supplementary Table S2, available at Annals of
identified and (iv) the overall quality of breaking bad news is Oncology online. Over 50% of the patients did not know the
significantly related to the emotional state after receiving bad bearer before the conversation, 70.1% did not get the following
news. treatment from the bearer and only 42.2% are currently in
contact with him (69.7% male) or her (30.3% female).
patients and methods
ratings of patients’ preferences concerning
The questionnaire was handed out to 350 cancer patients in an inpatient and important aspects of breaking bad news
outpatient setting of the University Hospital Marburg (UKGM) and the re-
Of the 37 questionnaire items to measure patients’ preferences,
habilitation center Klinik Sonnenblick, Marburg, Germany. In addition to
we identified those that the participants rated the most relevant.
written informed consent, inclusion criteria required the oncological diagno-
The mean ratings are listed in the descending order of import-
sis of a malignant neoplasm, sufficient German language skills and a
ance in Table 1. ‘Having clarity about suffering and progress of
minimum age of 18.
the disease’ (SPIKES 4) was rated the most important. Of the 10
items receiving the highest ratings, 5 addressed aspects of the
ethics physicians giving knowledge and information to the patient
The study was approved by the institutional review board of the Medical (SPIKE 4). Three items of the ‘top ten’ were designed to
School, University of Marburg. Before participation, subjects gave written measure SPIKE 6 (strategy and summary, e.g. possibility to ask
informed consent. questions), and two items addressed aspects of the setting
(SPIKES 1).

questionnaire patients’ preferences compared with reality

A total of 37 items of the MABBAN (Marburg Breaking Bad News Scale)
The results of comparing patients’ preferences with the subject-
[13] representing the six SPIKES subscales (Setting, Perception, Invitation,
ively perceived reality are also reported in Table 1. The test sta-
Knowledge, Emotion and Summary and Strategy) were generated [1]. The
questionnaire is composed of two main parts: the first one asks for the pro-
tistics show that there are significant differences in each variable
cedure, perception and satisfaction of the first cancer disclosure according to
between patients’ preferences and the reality. The ratings of ‘en-
the recommended steps of the SPIKES protocol, while the second one con- tirely affirmation’ were significantly higher on preference side
sists of corresponding questions asking for the importance of patient’s assign than on the reality side. The greatest difference between what
to these items. Most of the items were rated on a Likert scale from 1 (‘entire- patients want and what they get was in relation to getting a
ly’) to 4 (‘not at all’). definite explanation of the course of disease (z = −11.600,
Additionally, four items concern patients’ preferences for the doctor P < 0.001, r = −0.633). Although there is obviously a high need
characteristics (e.g. profession, sex, level of familiarity) and four other items for information, 22.6% of the patients wanted not to be involved
had a different format (e.g. dichotomous variable). Together, these eight in the following decisions by the physician, and 42.3% of the
items were classified as ‘informative items’ and were not included in the patients stated that they were not able to make important

 | Seifart et al.
Annals of Oncology original article
Table 1. Highest ratings of patients’ preferences regarding different aspects of breaking bad news compared with experienced reality

Item (SPIKES number) Patients’ preference Reality Test statisticsa

‘Entirely’ (%)b M (SD)c ‘Entirely’ (%)b M (SD) c
z r n

Having clarity about suffering and progress after BBN (4) 96.8 1.03 (0.18) 77.9 1.35 (1.22) −7.111*** −0.388 335
Reinsurance about patients’ comprehension (4/6) 94.7 1.05 (0.22) — — — — —
Having enough time (1) 94.5 1.06 (0.23) 64.4 1.62 (0.96) −9.164*** −0.496 342
Possibility to ask questions (6) 93.8 1.06 (0.24) 60.9 1.67 (0.97) −9.567*** −0.523 335
Having the feeling that planned treatment is the best (4) 93.8 1.07 (0.28) 60.9 1.77 (1.11) −9.539*** −0.526 329
Elaborate and coherent explanation of the disease (4) 93.2 1.07 (0.27) 61.3 1.58 (0.87) −9.297*** −0.506 337
Definite explanation of the diagnosis (4) 89.8 1.11 (0.36) — — — — —
Undisturbed atmosphere (1) 86.9 1.17 (0.51) 63.3 1.87 (1.25) −8.439*** −0.460 337
Involving the patient in further planning (6(4)) 84.2 1.19 (0.47) 48.1 2.14 (1.27) −10.194*** −0.565 325
Definite explanation of the course of disease (6) 82.9 1.21 (0.53) 40.7 2.18 (1.17) −11.600*** −0.633 336

a
Wilcoxon signed-rank test. bValid percentage of response option ‘entirely’. cAll items are rated on a 1–4 scale.

Downloaded from http://annonc.oxfordjournals.org/ at Universitaetsbibliothek Muenchen on July 2, 2014

***P < 0.001.

decisions in the first bad-news discussion. More than half of the discussion
patient group (50.2%) stated that a second talk would be neces-
sary, preferably accompanied by a relative or loved one (76.5% With regard to breaking bad news, the physician–patient com-
of the patients voted for a second talk). munication process becomes particularly a difficult task for
physicians, requiring social and communicative competence,
special management and responsibility. For this reason, several
relationship between demographic and medical
recommendations for the delivery of bad news has been estab-
variables and patients’ preferences
lished, the most popular being the SPIKES protocol [1] and its
There were no differences between ‘type of disease’ and the top- application to cancer patients [2]. This protocol has reached
10 patients’ preferences. Significant correlations were found guideline status in several countries, including Germany, and is
between the ‘age’ of the patients and the need for ‘elaborate and therefore frequently used for communication skills training in
coherent explanation of the disease’ (r = −0.123, P < 0.05) and the context of bad-news communications. However, it is not
‘definite explanation of the course of disease’ (r = −0.186, known whether and how the elementary steps of the SPIKES
P ≤ 0.001), indicating a pronounced need in younger patients. protocol are represented in common bad-news discussions, and
There were only a few significant differences between the ‘sexes’ whether the protocol meets the needs of cancer patients in
and their preferences. Women rated getting an ‘elaborate and Germany.
coherent explanation of the disease’ as more important than did The overall quality of breaking bad news (retrospectively) in
men (P = 0.044). The analyses have shown that there are differ- the present study correlates significantly with the reported emo-
ences in patients’ preferences depending on the ‘educational tional state after receiving bad news (r = −0.261, P < 0.001), em-
status’ of the patients. The main differences were ‘the need for phasizing the importance for a high-quality communication
definite explanation of the course of disease’. This seemed to be process in bad-news delivery. However, compared with previous
of more importance for patients with lower education-levels studies [14, 15], in this study, cancer patients report lower satis-
(9–10 years of school) compared to those with high school faction with the way bad news was communicated, as only
graduation (13 years of school with the graduation ‘Abitur’) 46.1% were satisfied with the disclosure process of their cancer
(P = 0.027) and for patients with Certificate of Basic Secondary diagnosis. In general, satisfaction with physician–patient com-
Education (9–10 years, ‘Hauptschulabschluss’) to those with munication is known to depend on different factors, including
Certificate of Advanced Secondary Education (10 years, ‘Real communication skills and style of the physician, physician–
schulabschluss’) (P = 0.045) and those with high school gradu- patient relationship as well as biomedical and psychosocial
ation (13 years, ‘Abitur’) (P = 0.004). factors [3, 13, 16, 17]. However, it could be postulated that the
low satisfaction in the present study also reflects the significantly
quality of ‘breaking bad news’ discussion and high difference between patient’s preferences and the way bad
emotional state after receiving a bad news news were delivered. In regard to the first 10 preferences,
Less than a moiety (46.2%) of the patients were entirely satisfied patients reported a highly significant difference between request
with the communication process by which bad news had been and realization, pointing to considerable deficiencies in the bad-
broken to them. The overall quality of breaking bad news was news communication process from the patients’ viewpoint.
significantly related to the emotional state ([14-item scale of Although the study is limited by the missing information about
different emotional states, including depression, anxiety and communication skills and styles of the physicians and their
sleeplessness [HADS]) after receiving bad news (r = −0.261, approach, a guided communication process that is closer to
P < 0.001). patient’s preferences is likely to be more helpful.

doi:10.1093/annonc/mdt582 | 
original article Annals of Oncology

Following patient’s preference information (SPIKES 4: knowl- be considered that insufficient realization is due not only to indi-
edge), prearrangement (SPIKES 1: setting) and physicians’ ap- vidual physician attitudes, but also to the changes of clinical
proach are of central importance, supporting the findings from daily life in the last years, with increasing work load and less
previous studies [13, 18 –20]. First, patients wish to gain clarity privacy room capacity. Therefore, hospital business manage-
about the expected distress and progression of the disease. This ment has to be aware of these patient needs and should accom-
is not identical to an exact prognosis or anticipated life expect- modate the facilities of oncology medical care.
ation, as these specific information items range considerably Next to the prearrangement aspects, emotional support is one
lower. Secondly, the difference between patients’ request and of the central concern of the SPIKES protocol. However, similar
reality was greatest in relation to getting a definite explanation to observations in a UK cancer center [14], supportive aspects
of the course of disease (z = −11.600, P < 0.001, r = −0.633). were rated considerably lower compared with information or
Rather, patients’ interest refers to the expected alteration of their prerequisite items in the present study, such as the possibility to
everyday life and course of the disease. be emotional (77.9%; rank 15), empathy (65.6%; rank 16) and
Clarity about the expected distress and progression of the familiarity of the physician (52.4%; rank 26). However, it could
disease, the possibility to ask questions and the reinsurance not be concluded that support aspects might be less important
of understanding by the physician were rated next on the top of for a high quality of bad-news delivery, although they seem to be

Downloaded from http://annonc.oxfordjournals.org/ at Universitaetsbibliothek Muenchen on July 2, 2014

the patients’ preference list. Particularly, the high number of less interest in patients view.
of requests for physician’s reassurance and the possibility to ask Following the results of the present study, we suggest some
questions emphasizes the importance to involve patients actively recommendations for a modification of the SPIKES protocol.
into the communication process. Previously, the approach Patients’ needs for prerequisites should be more respected and
‘ask–tell–ask’ was recommended [21]. Following patients’ pre- realized by the physician and the hospital management. The
ferences in the present study, we suggest to change the approach physician should ask for patient’s information needs and focus
to ‘ask–tell–invite to ask–reassure understanding’. Moreover, on the course of the disease and implications for patient’s daily
the importance of the identification of patient’s preferences life. In every step of the protocol, the physician should routinely
before a discussion of bad news should not be disregarded. ask for understanding of the patient and invite the patient to ask
Although there is obviously a high need for information, questions. Because a single communication might not be suffi-
22.6% of the patients wanted not to be involved in the following cient and satisfactory for the listener and because a high per-
decisions by the physician, with the majority being elderly men. centage of patients feel incapable of making any decisions after
In a study evaluating preferences of elderly patients with meta- receiving bad news, we recommend bad-news communication
static cancer, more than half of the patients (52%) preferred a as a two-step procedure.
passive role in the decision-making process [22]. On the other There are several potential limitations of this study, including a
hand, information needs are high. Therefore, having a look into lack of information regarding physicians’ approach and commu-
the ethical waiver debate, the observation of this high need for nication style, physicians’ training in communication skills and
information with a striking number of patients who prefer a preferences for breaking bad news in a cross-cultural population.
passive role in decision-making makes clear that information The survey was based on the recommendation of the SPIKES
and decision renunciation must be differentiated. Otherwise, the protocol, and therefore, other potential important aspects might
low satisfaction and the reported difference between request and be missed. Additionally, patients were asked to recall their percep-
realization of information items might be the result of the tion of the first cancer disclosure. It is likely that their perception
nature and impact of bad-news discussions and the time man- and emotional state have changed over time. Therefore, a longitu-
agement of delivery. In the present study, only 23.6% of patients dinal study should be carried out that includes data about
received a second talk, reflecting the routine of bad-news deliv- physician communication skills and training.
ery in a one-step discussion. Often, the information given to the Despite these limitations, the study indicates a clear need
patient could not be recalled and understood in the face of up- for the improvement of breaking bad-news communication in
setting news, particularly if the patients are too stunned to regis- Germany. Therefore, some adjunctions to the SPIKES protocol
ter any further information [23]. At the end, patients may stick should be considered, including repeated reassurance of listeners
to the feeling of lacking information, resulting in depression, understanding, the perpetual possibility to ask question, respect
uncertainty and anxiety [24]. Interestingly, 42.3% of the patients for prearrangement needs and the conception of bad-news
stated that it is not possible to make important decisions during delivery in a two-step procedure.
the first bad-news discussion and 50.2% wished to have a
second talk, preferably accompanied by a relative or loved one. acknowledgements
Therefore, bad-news discussions in a two-time procedure
should generally be strongly recommended. The authors thank Prof. Gerd Richter for supporting this study.
Another important scope in ‘breaking bad news’ communica-
tions related to patients’ satisfaction is prearrangement aspects
[13–15]. In the present study, adequate time and undisturbed
disclosure
surrounding were two of the most important requests (94.5% WR received honoraria from Berlin Chemie, Astra Zeneca and
and 86.9%, respectively), which were realized only in ∼60% of Heel for consultation and presentations on placebo effects and
the patients satisfactorily. As already recommended by the medication adherence. US received honoraria Amgen and Lilly
SPIKES protocol [1, 4], these findings emphasize further the dis- for consultation and presentations on lung cancer. All remain-
tinct importance of prearrangement aspects. However, it has to ing authors have declared no conflicts of interest.

 | Seifart et al.
Annals of Oncology original article
references 13. Hofmann M, Seifart U, Rief W et al. Assessing patient’s preferences for Breaking
Bad News: development of the MABBANs. Psychooncology 2013, submitted.
1. Buckman R. How to Break Bad News: A Guide for Health Care Professionals. 14. Brown VA, Parker PA, Furber L et al. Patients preferences for the delivery of bad
Baltimore: Johns Hopkins Press, 1992. news—the experience of a UK Cancer Centre. Eur J Cancer Care 2011; 20:
2. Fellowfield L, Jenkins V. Communicating sad, bad, and difficult news in medicine. 56–61.
Lancet 2004; 363: 312–319. 15. Ptacek JT, Ptacek JJ. Patients’ perceptions of receiving bad news about cancer.
3. Baile WF, Buckman R, Lenzi R et al. A six-step protocol for delivering bad news: J Clin Oncol 2001; 19: 4160–4164.
application to the patient with cancer. The oncologist 2000; 5: 302–311. 16. Schmid Mast M, Kindlimann A, Langewitz W. Recipients’ perspective on breaking
4. Fellowfield LJ, Jenkins V. Effective communication skills are the key to good bad news: how you put it really makes a difference. Pat Educ Couns 2005; 58:
cancer care. Eur J Cancer 1999; 35: 1592–1597. 244–251.
5. Rabow MW, McPhee SJ. Beyond breaking bad news: how to help patients who 17. Zachariae R, Pedersen CG, Jensen AB et al. Association of perceived physician
suffer. West J Med 1999; 171: 260–263. communication style with patient satisfaction, distress, cancer-related self-
6. Girgis A, Sanson-Fisher RW. Breaking bad news: consensus guidelines for efficacy, and perceived control over the disease. Br J Cancer 2003; 88:
medical practitioners. J Clin Oncol 1995; 13: 2449–2456. 658–665.
7. SCOPE. Right from the Start Template: Good Practice in Sharing the News. 18. Fujimori M, Akechi T, Morita T et al. Preferences of cancer patients regarding the
London: Department of Health, 2003; 8. disclosure of bad news. Psychooncology 2007; 16: 573–581.
8. Schofield PE, Beeney LJ, Thompson JF et al. Hearing the bad news of a cancer 19. Parker PA, Baile WF, de Moor C et al. Breaking bad news about cancer: patient’s

Downloaded from http://annonc.oxfordjournals.org/ at Universitaetsbibliothek Muenchen on July 2, 2014

diagnosis: the Australian melanoma patient́s perspective. Ann Oncol 2001; 12: preferences for communication. J Clin Oncol 2001; 19: 2049–2056.
365–371. 20. Sastre MTM, Sorum PC, Mullet E. Breaking bad news: the patient’s viewpoint.
9. Girgis A, Sanson-Fisher RW, Schofield MJ. Is there consensus between breast Health Comm 2011; 26: 649–655.
cancer patients and providers on guidelines on braking bad news? Behav Med 21. Back AL, Arnold RM, Baile WF et al. Efficacy of communication skills training for
2010; 25: 69–77. giving bad news and discussing transitions to palliative care. Arch Intern Med
10. Teike Lüthi F, Cantin B. Breaking bad news: ‘EPICES’, a French style as a learning 2007; 167: 453–460.
method. Rev Med Suisse 2011; 12: 85–87. 22. Elkin EB, Kim SH, Casper ES et al. Desire for information and involvement in
11. Wand S, Schildmann J, Burchardi N et al. Die Aufklärungsgesprächbewertungsskala treatment decisions: elderly cancer patients’ preferences and their physicians’
(AGBS): Ein Instrument zur Bewertung kommunikativer Kompetenzen bei der perceptions. J Clin Oncol 2007; 25: 5275–5280.
Aufklärung von Patienten über Tumorerkrankungen. Z ärztl Fortbild Qual Gesundh. 23. Hogbin B, Fallowfield L. Getting it taped: the ‘bad news’ consultation with cancer
wes 2008; 101: 645–651. patients. Br J Hosp Med 1989; 41: 330–333.
12. Cegala DJ, Lenzmeier Broz S. Physician communication skills training: a 24. Newall DJ, Gadd EM, Priestman TJ. Presenting of information to cancer patients: a
review of theoretical backgrounds, objectives and skills. Med Educ 2002; 36: comparison of two centres in the UK and USA. Br J Med Psychol 1987; 60:
1004–1016. 127–131.

doi:10.1093/annonc/mdt582 | 