Psycho-OncologyPsycho-Oncology 20: 242–251 (2011)Published online 24 March 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/pon.1730

Discussion of emotional and social impact of cancerduring outpatient oncology consultations

Sally Taylor1�, Clare Harley1, Lyndsay J. Campbell1, Laura Bingham1, Emma Joanne Podmore1,Alex C. Newsham1, Peter J. Selby1, Julia M. Brown2 and Galina Velikova1

1Cancer Research UK Psychosocial and Clinical Practice Research Group, St James’s Institute of Oncology, Leeds, UK2Clinical Trials and Research Unit, University of Leeds, Clinical Trials Research House, Leeds, UK

Abstract

Objective: Following publication of national guidelines on detection and management of

psychosocial problems in oncology, this study explores frequency of discussion of emotional

and social issues in outpatient oncology consultations.

Methods: Analysis of baseline data from 212 outpatients participating in a randomized

controlled trial. Baseline data included content analysis of audio recordings of consultations,

Functional Assessment of Cancer Therapy-General (FACT-G) questionnaire subscale scores,

and patient and clinician self-rated preferences and perceptions of communication.

Results: Fifty-nine percent patients and 75% clinicians expressed preferences to discuss

emotional issues during consultations. Analysis of audio recordings showed that they were

discussed in 27% of the consultations, regardless of severity of emotional problems reported by

patients (FACT-G Emotional well-being subscale). Fifty percent of clinicians reported

discussing emotional issues ‘often’ or ‘almost always’, compared with 18% of patients. Forty-

four percent patients and 39% clinicians reported that they would discuss social activities, but

they were actually discussed in 46% of consultations. Patients predominantly initiated

discussion of emotional and social issues (85 and 60% consultations, respectively).

Conclusions: Low prevalence of discussion of psychosocial issues cannot be accounted for by

patient or clinician communication preferences. If clinicians rely on patients to initiate

discussion of psychosocial issues, patients’ problems may go unaddressed.

Copyright r 2010 John Wiley & Sons, Ltd.

Keywords: oncology; cancer; communication; emotional; social

Introduction

Modern medical practice recognizes the impor-tance of a patient-centered model of care, whichseeks a medical diagnosis and patient’s illnessexperience [1,2]. Illness experience includes patientfeelings, their fears of being ill, and the impact ofdisease on function. Patient experience can only beunderstood in the context of their life setting,family, and social networks.Devastating effects of cancer on patients and

their families are well recognized. Diagnosis andtreatment has significant impact on patients’ lives,causing not only serious symptoms (such as pain,fatigue, sickness) [3], but also leading to emotionaldistress and limiting roles and social activities [4].During treatment, health professionals focus oncontrolling symptoms and side effects. Discussionof psychosocial issues is less common, so patients’concerns may be left unaddressed [5,6].The patient-centered approach can be related to

positive health outcomes, such as patient recovery,

emotional health, and physical functioning ingeneral medicine [7]. Training programs encoura-ging physicians to communicate in a patient-centered way resulted in improved patient out-comes in general medicine and oncology [8,9].However, teaching and educating physicians inpatient-centered care remains a challenge. Guide-lines have been developed and introduced forassessment and management of emotional andsocial problems to encourage physicians to changetheir practice. In oncology, the National Institutefor Clinical Excellence (NICE) guidelines in theUnited Kingdom recommend that psychosocialissues are routinely assessed and discussed [10].Guidelines on distress management have beenintroduced in the United States by the NationalComprehensive Cancer Network (NCCN) [11].Their definition of ‘distress’ includes psychological,social, and spiritual aspects [12].

Evaluating the effectiveness of this approach ischallenging. After introduction of NCCN guide-lines, an audit showed eight institutions (53%) had

* Correspondence to: CancerResearch UK Psychosocialand Clinical PracticeResearch Group, St James’sInstitute of Oncology, BeckettStreet, Leeds LS9 7TF, UK.E-mail: s.s.taylor@leeds.ac.uk

Received: 11 August 2009

Revised: 8 December 2009

Accepted: 25 January 2010

Copyright r 2010 John Wiley & Sons, Ltd.

implemented the guidelines and routinely assessedpatients for distress [13]. Despite this result, anumber of studies have highlighted continued highprevalence of unaddressed psychosocial problemsamongst cancer patients [14,15]. This observationalstudy provides a specific example from routineoutpatient oncology consultations, exploring fre-quency of adopting a patient-centered approach(discussion of emotional/social issues), followingpublication of national guidelines on detection andmanagement of psychosocial problems in cancerpatients.The majority of research examining discussion of

psychosocial problems in oncology practice focuseson emotional issues [6,16,17], and suggests they arediscussed in 35–40% of consultations [6,18].Physicians often focused on medical/technicalissues [19,20] while emotional issues were typicallyinitiated by patients [6]. When patients mentionedemotional issues, physicians often terminated thediscussion rather than further exploration [16].Additional guidelines published in the UnitedStates, in 2007, stressed the need for psychosocialcare to be integrated into routine oncology practice[21]. To achieve this, physicians may requireadditional support in addressing psychosocialissues. The use of screening instruments mightfacilitate this [15].It is common for patients to experience a wide

range of social problems as a result of cancerdiagnosis and treatment [22]. Few studies haveexplored how these issues are detected andaddressed by health professionals. The mostcommon social problems experienced by patientsinclude relationship issues and difficulties withdomestic chores and social activities [23]. Thereis limited evidence to indicate how often theseissues are discussed in outpatient consultations.Research conducted before the publicationof NICE guidelines suggests that socialproblems are discussed in 41% of oncologyconsultations [18].Analysis of outpatient oncology consultations

has not been conducted in the United Kingdomsince the publication of NICE guidelines in 2004.There is little up-to-date information to determinewhether clinical practice has changed since theirpublication. The data presented in this paper wascollected after the introduction of the guidelines(2004–2006). The primary aim of this study wasto examine frequency of discussion of psychosocialissues in outpatient oncology consultationsduring treatment. This study expands on thecurrent literature by examining frequency ofdiscussion of emotional/social problems. Thesecondary aim was to examine whether commu-nication preferences of patients and physicians orseverity of patients’ self-reported emotional/socialproblems were related to frequency of theirdiscussion during consultations.

Materials and methods

Study samples and procedures

This paper is a descriptive analysis of baseline datacollected during a randomized controlled trial(RCT) [24]. The RCT was designed to examinewhether regular touch-screen computer completionof a quality of life (QOL) questionnaire (EORTC-QLQC30) in oncology outpatient clinics (withoutproviding QOL scores to oncologists) increasedpatient–physician communication and improvedpatient well-being. Outpatients attending one regio-nal cancer center and two district general hospitalswere eligible to participate in the RCT, if they werecommencing treatment. Patients were entered intothe study when attending outpatients for review fortheir first, second or third cycle of chemotherapy.All patients should have attended the clinic pre-viously. Not all patients were receiving first-linetreatment, so may have more experience of thedepartment. All oncology consultants and specialistregistrars were invited to participate. Physiciansworked in teams of 4–7 doctors, so over timepatients saw different doctors. Patients were rando-mized to intervention or control arm. Both groupshad their consultations audio recorded at baselineand at three subsequent visits. At baseline, patientscompleted the Functional Assessment of CancerTherapy-General questionnaire (FACT-G) and aquestionnaire exploring their preferences and per-ceptions of communication (PPC) during consul-tations (detailed below). Physicians completed aversion of the PPC on entering the study. Patientsocio-demographic and clinical data was collected.This study reports the following baseline data

from the RCT for control and intervention groups:content analysis of audio consultations, FACT-Gsubscale scores, and PPC questionnaire scores.The project was approved by the institutional

Ethical Committees. Written informed consent wasobtained from patients and physicians.

Study measures

Patient/physician characteristics

Patients’ clinical information (diagnosis, stage ofdisease, performance status, treatment details) wererecorded from medical notes. Physicians completed ademographic and professional characteristics survey.

Preferences for discussion of emotional and social

problems

The PPC questionnaire is a 14-item study-specificquestionnaire used to assess patients’ view ofcommunication in clinic [25]. The first seven itemsrelate to preferences for discussion of physicalsymptoms, physical activities, how the patient feelsemotionally, impact of disease or treatment on

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

Emotional and social impact of cancer 243

work, social activities, relationships, and memory.Response options: ‘Yes I would like to discuss thistopic’, ‘Yes provided the doctor mentions it first’,and ‘Rather not’ indicating that patients wouldeither initiate the discussion, wait for the doctor toinitiate it, or would prefer not to discuss it at all.The next seven items relate to the perceivedfrequency of discussion in consultations and coveroverall health, symptoms/side effects, limitation inphysical activity, limitation in work, emotionaldistress, impact on relationships, and social activ-ities. Response options: ‘Never’, ‘Rarely’, ‘Some-times’, ‘Often’, ‘Almost always’. Physicians alsocompleted the PPC but the response option ‘Yesprovided the doctor mentions it first’ was changedto ‘Yes provided the patient mentioned it first’. Wereport patient/physician responses to discussinghow the patient feels emotionally, and the impactof disease and treatment on social activities.

Patient self-reported QOL

All patients completed FACT-G [26] at baseline.The FACT-G is a 27-item instrument used to assessfunctional status of cancer patients in trials. It hasfour scales: Physical well-being, Social/family well-being (SFWB), Emotional well-being (EWB) andFunctional well-being. This study reports onlyEWB and SFWB scales to determine the extentof patient self-reported emotional/social problems.Scores for EWB and SFWB subscales range from0–24, higher scores indicate better functioning.

Patient–physician communication

Outpatient consultations were audio recorded andcoded directly from audio recordings using contentanalysis. The coding scheme recorded discussion ofsymptoms and functioning issues covered in theEORTC-QLQ C30: physical, role, emotional,social, cognitive, financial, sexual, and generalwell-being. Coding included whether the discussionwas instigated by patient or physician. Discussionof functions was only coded once per consultation,regardless of the number of times the topic wasraised. Actions taken by physicians as a result ofdiscussing functional issues were also coded. Anaction could include further exploration, making areferral to specialist services, or prescribing medi-cation. Guidelines for coding were written by theprincipal investigator (consultant oncologist).The consultations were coded by four trained

researchers. Training continued until all research-ers achieved acceptable agreement (470%). Cod-ing was conducted in pairs which were changedfrequently to prevent coding biases. Regularconsensus meetings were held. Twenty consulta-tions were randomly selected and re-analyzed2 months later by the same researchers to checkfor drift over time. Kappa statistics showed good

levels of agreement (0.69 and 1.00 for emotionaland social problems, respectively).

Data analysis

Patient/physician preferences for discussing emotional/

social functioning issues

Descriptive analysis of responses to emotional/social items on the PPC determined the proportionof patients/physicians who reported they wouldinitiate discussion of emotional/social issues duringconsultations. w2 analyses determined whetherpreferences for communication differed betweenpatient age groups (p60, X61) or genders.

Content of patient–physician communication

Descriptive analysis of the consultation codingdetermined the frequency of discussion of emo-tional/social issues during consultations andwhether these discussions were initiated by patientsor physicians.

Severity of patients’ self-reported problems and com-

munication

Patients’ self-reported FACT-G scores were usedto determine the severity of patients’ emotional/social functioning problems. The FACT-G has nopublished cut-points suggesting when a scoreindicates a serious problem. Therefore, FACT-GEWB/SFWB subscale scores of patients were splitinto quartiles and the lowest quartile for each ofthe scales defined significant problems. For EWB,this wasp13 (out of 24) and for SFWB p19 (outof 24). PPC scores and content of consultationswere examined for patients reporting elevatedproblems on the EWB and SFWB subscales, todetermine whether the severity of self-reportedproblems had an effect on discussion of the issue inconsultations.

Association between individual physicians and content

of communication

As the same physicians saw a number of differentpatients in this study, we examined whetherindividual differences between physicians influ-enced the discussion of emotional/social issues.Logistic regression for emotional and social issuesbeing discussed was performed to investigate theeffect of individual doctors while controlling forpatients’ EWB and SFWB scores on FACT-G. Wethen calculated, for each physician, the proportionof consultations in which they discussed emotionalissues and the proportion in which they discussedsocial issues. The data were examined for patternsof consistency for individual physicians. Fivephysicians were then removed from the analysesas they had only conducted one consultation.Physicians were grouped according to their PPC

244 S. Taylor et al.

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

scores for discussion of emotional/social issuesinto those who agreed with the statement ‘YesI would like to discuss’ and those who agreedwith the statement ‘Yes if the patient mentions itfirst’ (grouped separately for their response toemotional or social issues). No physicians reportedthat they would not discuss social or emotionalissues. Physicians were also grouped accordingto whether or not they had received communi-cation skills training. Independent samples t-testsdetermined whether there were any differencesbetween physician groups and frequency ofdiscussion of emotional/social issues in theirconsultations.

Results

Study sample

Five hundred and two patients were invited toparticipate in the RCT. Of these, 356 patientsconsented (71%), but 120 discontinued (49 patientsdied, 33 refused later, 21 had no further appoint-ment, 15 missed for administrative reasons, and 2unknown). A total of 236 patients completed thestudy; only 212 had four complete consultationrecordings owing to administrative errors (failedrecordings/poor quality sound). Only consultationsof patients who had completed the study wereanalyzed owing to resource constraints. Apartfrom extent of disease, there were no differencesin age, gender, diagnosis, or performance statusbetween the 212 patients who completed the studyand the 120 patients who discontinued or the 24without complete recordings. Patient characteris-tics are presented in Table 1. The medianconsultation length was 13.5min (range:2.5–37.5), 27% under 10min, 53% 10–20, and20% for more than 20min.Thirty-six physicians agreed to participate in the

study. Demographic and professional characteris-tics are presented in Table 1. Twenty-four physi-cians (66%) reported they had received formalcommunication skills training.

Emotional functioning

Patient/physician preferences for discussion ofemotional functioning issues (Figure 1((a), (b))The majority of patients and physicians (59 and75%, respectively) said they would initiate emo-tional discussions, 30% patients and 25% physi-cians would wait for the patient/physician toinitiate discussion. Only 13 patients (6%) reportedthat they would rather not discuss emotionalissues. However, patients and physicians percep-tions differed on how frequently emotional issueswere discussed. Eighteen physicians (50%) re-ported they discussed emotional issues ‘often’ or

‘almost always’, whereas only 37 patients (18%)agreed. There were no significant differences inpatients’ reported preferences for discussing emo-tional issues according to gender or age.

Patient–physician communicationEmotional issues were mentioned in 58/212 (27%)consultations and this led to a conversationbetween patient and physician in 24 instances.Emotional issues were generally raised by thepatient (49/58 (85%)). Female patients were morelikely to discuss emotional issues than malepatients (w2 (2, N5 212)5 4.4, p5 0.036), but therewere no differences by age group. Of the 190patients who reported they want to discuss emo-tional issues, 54 consultations (28%) includeddiscussion of emotional issues.

Table 1. Patient and physician characteristics

Patient characteristics Number %

Gender

Female 147 69

Male 65 31

Age in years (mean, SD) 58.9, 13.78

Diagnosis

Ovarian/gynecological 72 34

Breast 52 24

Lung 29 14

Germ cell 19 9

Colorectal 17 8

Bladder/renal 12 6

Other 11 5

Extent of disease

Primary local 71 34

Local recurrence 24 11

Metastatic 117 55

Performance status

0 36 17

1 117 55

2 44 21

3 14 6

Missing 1 1

FACT-G scale scores

Emotional well-being (mean, SD) 17.1, 5.11

Social well-being (mean, SD) 21, 3.65

Physician characteristics Number

Gender

Female 15

Male 21

Age (median, range) 35, 28–53

Qualification level

Consultant 14

Associate specialist 1

Clinical assistant 2

Specialist registrar 17

Senior house officer 1

Staff grade 1

Number of year practicing medicine (median, range) 10, 4–30

Number of years in oncology (median, range) 4, 0–22

Formal training in communication skills

Yes 24

No 12

Emotional and social impact of cancer 245

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

Anxiety and depression were the most commonlydiscussed topics (21 and 31%, respectively). Only 6patients (22%) who discussed feelings associatedwith depression used the word ‘depressed’. Eightpatients (30%) in this group described ‘feelingdown’ or ‘low’. When referring to anxiety, 7 patients(39%) used ‘anxiety/anxious’, others felt ‘workedup’ or ‘worried’. Some emotional issues were relatedto a specific symptom or side effect, such as feelingscared to lie down as breathing was difficult orscared to wash thinning hair. Fifteen patients (17%)were optimistic and felt ‘positive’ or ‘elated’.Physicians acted upon the discussion of emotional

issues in 13/58 (22%) consultations. The mostcommon response (six consultations) was a detaileddiscussion of anxiety/depression, with the aim ofdiagnosis. Other actions included current drugprescription enquiries, drug modifications, or dis-cussing referrals to clinical psychology. Two physi-cians offered referrals, but the patients declined.

Relation between severity of patients’ self-reportedproblems and communication (Figure 2(a))Forty-four (21%) patients scored 13 or less onEWB (Figure 2(a)), indicating elevated problems.Forty-two (96%) of these patients reported that

they would like to discuss emotional issues withtheir doctor. Twenty-eight patients (64%) wouldinitiate the discussion of emotional issues and 14patients (32%) would wait for the doctor to initiatediscussion. Of the 44 patients scoring 13 or less onthe EWB subscale, emotional issues were discussedin 14 consultations (32%), a similar proportion asin the whole group (41 (26%)).

Association between individual physicians and con-tent of communicationEach doctor conducted an average of 5.5 consulta-tions (SD5 3.5, range 1–12; see Table 2). Logisticregression with emotional issues being discussed asthe outcome variable revealed no significant effectof individual doctors (Wald’s w2(34)5 17.11,p5 0.99), including patients’ EWB score in themodel (Wald’s w2(1)5 0.84, p5 0.36). The fre-quency of discussing emotional issues betweendoctors was skewed with the majority of physiciansincluding these in fewer than half of theirconsultations. There were no noticeable differencesin practice style between physicians.There was no significant difference in frequency

of discussion of emotional issues between physi-cians who said they would initiate discussion of

Figure 1. Patient and physician preferences for discussing emotional (a) and social issues (c). Patient and physician perceptions ofhow often they feel they discuss emotional (b) and social issues (d)

246 S. Taylor et al.

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

emotional issues (M5 20.7%, SD5 4.23) andphysicians who would wait for the patient toinitiate discussion (M5 23.5%, SD5 8.87);t(29)5 0.24, p5 0.81). Although physicians whohad received communication skills training men-tioned emotional issues more frequently(M5 39.3%, SD5 19.25) than physicians whohad not (M5 17.3%, SD5 23.17), this differencewas not significant (t(29)5�1.53, p5 0.14).

Social functioning

Patient/physician preferences for discussion of socialfunctioning issues (Figure 1(c), (d ))Ninety-four patients (44%) reported they wouldinitiate discussion of social issues during consulta-tions. Eighty-three patients (39%) would wait forthe physician to initiate discussion. Twenty-fourpatients (11%) would rather not discuss socialissues with their physician. Twenty-one physicians(58%) would wait for the patient to initiatediscussion of social issues, whilst 14 (39%) wouldinitiate the discussion. Patients perceived thatsocial issues were discussed in consultations lessfrequently than physicians did (Figure 1(d)). Therewere no significant differences in patients’ reportedPPC scores by gender or age group.

Patient–physician communicationSocial issues were discussed in 97/212 (46%)consultations, and this led to a conversationbetween patient and physician in 37 of theseinstances. Social issues were often raised by thepatient (68/97 (60%)). Of the 177 patients whoreported wanting to discuss social issues, 88consultations (50%) included discussion of socialissues. Discussions of social issues encompasseda variety of topics, such as recreational activities,family issues, and holidays. There were no

significant differences in frequency of social issuesdiscussions between patient age groups, gender, ordiagnosis.Physicians acted upon discussion of social issues

in 4/97 (4%) consultations. In two cases, they gaveadvice about recreational activities. The other twocases related to support at home. One physicianoffered to contact the district nurse to investigateavailable support. One physician offered to help thepatient communicate with their children aboutcancer.

Severity of patients’ self-reported problems andcommunication (Figure 2(b))Forty-eight patients (23%) scored 19 or less onSFWB, indicating elevated problems. Of thesepatients, 41 (85%) reported they would discusssocial activities, 26 (63%) would initiate discussion,and 15 (31%) would discuss the issue if thephysician initiated discussion. Only 13/26 (50%)patients who said they would initiate discussionactually discussed these issues in their consultation.Nine out of fifteen patients (60%), who would waitfor the doctor to initiate the issue, discussed it intheir consultations. Social issues were discussed in24/48 (50%) of the consultations for patientsreporting serious problems, a very similar propor-tion as in the whole group (69 (46%)).

Association between individual physicians and con-tent of communicationLogistic regression, with social issues being dis-cussed at baseline as the outcome variable, revealedno significant effect of individual doctors (Wald’sw2(34)5 13.69, p5 1.0), including patients’ SFWBscore in the model (Wald’s w2(1)5 0.57, p5 0.45).The frequency of discussing social issues betweendoctors was normally distributed with physicians,including these in about half of their consultationsas the number of consultations per doctor

Table 2. Number of consultations recorded per doctor and the proportion of consultations where emotional and social issueswere discussed

Number of

consultations

per doctor

Number

of doctors

Proportion of consultations in which emotional

issues were discussed (count)

Proportion of consultations in which social

issues were discussed (count)

0 o50 50 450 100 0 o50 50 450 100

% %

1 5 4 1 4 1

2 6 4 2 3 1 2

3 1 1 1

4 4 1 2 1 1 1 2

5 3 3 2 1

6 3 2 1 2 1

7 3 2 1 1 2

8 2 1 1 1 1

9 1 1 1

10 5 5 3 1 1

11 2 2 1 1

12 1 1 1

Emotional and social impact of cancer 247

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

Figure 2. Examination of subgroups of patients with significant emotional distress (a) or social problems (b): Patient preferences fordiscussion, if they would initiate the discussion and the actual discussion during the consultations

248 S. Taylor et al.

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

increased (see Table 2). There were no noticeabledifferences in practice style between physicians.There was no significant difference in frequency ofdiscussion of social issues between physicians whoreported they would initiate discussion (M5 40.4,SD5 35.99) and physicians who reported theywould wait for the patient to initiate discussion(M5 45.3, SD5 32.5); t(33)5 0.43, p5 0.67). Therewas no significant difference in discussion of socialissues between physicians who had received com-munication skills training (M5 49.6%, SD5 29.03)and physicians who had not (M5 41.3%,SD5 32.95); t(16)5�0.72, p5 0.48).

Discussion

The majority of patients and physicians reportedthey would discuss emotional issues during cancertreatment. However, even after the formal pub-lication of national guidelines, physicians did notaddress these issues to the same degree theyindicate as appropriate. Emotional distress wasdiscussed in less than a third of outpatientconsultations.Results show slightly lower frequencies of dis-

cussion of emotional issues than earlier research[6,16,18]. In contrast to other findings [6], there wasno relationship between discussion of emotionalissues and severity of patient self-reported emo-tional difficulties. Patients were more likely toinitiate emotional discussion; only 15% wereinitiated by physicians. When emotional issueswere discussed, the most common response fromphysicians was to explore in more detail. In mostcases, physicians did not offer medication orreferrals to psychological services. Two patientswere offered a referral but both declined. Talkingthrough emotional issues with physicians may besufficient to ease patient concerns, with furtherintervention being unnecessary [16,27].In contrast to earlier research [27], depression

was the most commonly discussed emotional issue.Patients rarely used the term depression, usinginstead more colloquial language. In comparisonto other research, patients discussed feelingsassociated with anxiety more frequently, positiveemotional experiences were discussed, and feelingsof fear were mentioned less frequently [27]. Similarto other findings [6], emotional issues were morelikely to be raised by female patients.There are many possible reasons why emotional

issues are not discussed during consultations:medical issues take precedence during activetreatment, poor patient–physician communication,patients feeling that it is not the physician’s role todeal with emotional difficulties [27], or patientswithholding information so as not to burden theirphysician [16]. Sixty-seven percent of physicianshad communication skills training, but there was

no relationship between training and discussion ofemotional issues. Physicians may not be fully awareof services available to support patients experien-cing emotional difficulties, or access to supportservices may be limited. Physicians may feeluncomfortable initiating discussion of emotionalissues they feel unable to support or resolve.Conversely, physicians may be confident in addres-sing emotional issues but less confident identifyingproblems requiring intervention [16]. The cancercenter participating in this study has an activepsychosocial oncology department, led by seniorclinical psychologist and a liaison psychiatrist withexisting referral pathways.A large proportion of patients reported they

would discuss social issues with their doctor, but30% would wait for the doctor to initiate discus-sion. Most physicians reported they would wait forthe patient to initiate discussion. If a subsample ofphysicians and patients are waiting for the other toinitiate discussion, patients’ concerns may beunaddressed. Few studies have reported frequencyof discussion of social issues in oncology consulta-tions. In an earlier study [18], we reporteddiscussion of social activities in 41% consultations,similar to the frequency identified here.Physicians were more likely to discuss social

issues than emotional issues. This may occurbecause many social activities are dependent onphysical functioning. Since physicians tend to focusdiscussions toward more medical issues, such asmonitoring physical symptoms and side effects [27],the discussion of social activity may be a usefulindicator of functional status, performance, or levelof support or care received. Emotional issues maybe perceived as being less relevant to medicalsymptoms and side effects of treatment, or provi-sion of medical care.The study has a number of limitations. All

patients participated at the beginning of che-motherapy, at a time when physicians are morelikely to focus on medical issues, such as che-motherapy toxicities. Another limitation is that theaudio recordings represent a single outpatient visit.It is possible that emotional/social issues werediscussed at other clinic visits or with other healthprofessionals. The majority of the patients werefemale, reflecting the demographics of patientsattending the clinics. Women were more likely todiscuss psychosocial issues compared with men,suggesting results may overestimate the overallfrequency of discussion of psychosocial issues givenan equal male/female sample. Finally, the studypresents the results of a basic coding technique,reporting the frequency of emotional/social discus-sions during a single interaction. This codingscheme does not reflect the complexities ofpatient–physician interaction which have beenobserved in other studies [16]. The coding schemewas, however, generous as a broad spectrum of

Emotional and social impact of cancer 249

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

issues was coded as ‘emotional’. The record ofemotional discussions presented here could be anoverestimate of how often more complex emotionalissues are discussed in clinical practice. There maybe many more patients whose emotional problemsare left unaddressed.

Conclusion

High-quality, patient-centered care relies on effec-tive patient–physician communication [16]. Lowprevalence of discussion of emotional or socialfunctioning cannot be accounted for by patient/physician attitudes. Patients are more likely toinitiate discussion of emotional issues, but in thecase of social issues, a significant proportion ofpatients/physicians would prefer the other party toinitiate discussion. Despite high up-take of physi-cian communication skills training, additionaltraining may be required to help physicians identifyindirect cues used by patients that signal psycho-social problems [27].The study highlights the need for further devel-

opments in undergraduate and postgraduate profes-sional education to enhance communication skills.Supplementary strategies could be recommended,such as routine use of screening instruments inclinical practice or increasing focus on advisingpatients that it is reasonable for them to discussemotional/social issues during consultations.

Acknowledgements

We thank all the patients and oncology of staff for theircooperation and support. We are grateful to Ada Keding forperforming the logistic regression analysis looking atindividual doctor effects.Funding: The study was supported by grants from CancerResearch UK (GV and PJS; grant number: C7775/A7424)and NHS Research & Development (JMB). The funderswere not involved in study design, data collection, analyses,interpretation of the results, the decision to submit the paperfor publication, or the writing of the paper.Sponsors: The study was sponsored by the University of Leeds.The study was presented in part at IPOS 2008 as a posterand won best poster award.The authors of this paper, Sally Taylor, Clare Harley,Lyndsay J. Campbell, Laura Bingham, Emma JoannePodmore, Alex C. Newsham, Peter J. Selby, Julia M.Brown, and Galina Velikova, have no competing interests.

References

1. Balint M, Hunt J, Joyce D, Markinker M, Woodcock J.Treatment or Diagnosis: A Study of Repeat Prescriptionsin General Practice. J. B. Lippincott: Toronto, 1970.

2. Levenstein JH, McCracken EC, McWhinney IR,Stewart MA, Brown JB. The patient-centred clinicalmethod. 1. A model for the doctor-patient interactionin family medicine. Fam Pract 1986;3(1):24–30.

3. Lau P, Stewart K, Dooley M. The ten most commonadverse drug reactions (ADRs) in oncology patients: do

they matter to you? Support Care Cancer 2004;12:626–633.

4. Mitchell T. The social and emotional toll ofchemotherapy—patients’ perspectives. Eur J CancerCare (Engl ) 2007;16(1):39–47.

5. Ford S, Fallowfield L, Lewis S. Doctor-patientinteractions in oncology. Soc Sci Med 1996;42(11):1511–1519.

6. Detmar SB, Muller MJ, Wever LD, Schornagel JH,Aaronson NK. The patient–physician relationship.Patient-physician communication during outpatientpalliative treatment visits: an observational study. J AmMed Assoc 2001;285(10):1351–1357.

7. Stewart M, Brown JB, Donner A et al. The impact ofpatient-centered care on outcomes. J Fam Pract 2000;49(9):796–804.

8. Roter DL, Hall JA, Kern DE, Barker LR, Cole KA,Roca RP. Improving physicians’ interviewing skillsand reducing patients’ emotional distress. A rando-mized clinical trial. Arch Intern Med 1995;155(17):1877–1884.

9. Fallowfield L, Jenkins V, Farewell V, Saul J,Duffy A, Eves R. Efficacy of a cancer research UKcommunication skills training model for oncologists:a randomised controlled trial. Lancet 2002;359(9307):650–656.

10. NICE. Improving Supportive and Palliative Care forAdults with Cancer: The Manual, National Institutefor Clinical Excellence, Editor. 2004.

11. NCCN practice guidelines for the management ofpsychosocial distress. National comprehensive cancernetwork. Oncology 1999;13(5A):113–147.

12. Holland JC, Bultz BD. The NCCN guidelinefor distress management: a case for making distressthe sixth vital sign. J Natl Compr Canc Netw 2007;5(1):3–7.

13. Jacobsen PB, Ransom S. Implementation of NCCNdistress management guidelines by member institutions.J Natl Compr Canc Netw 2007;5(1):99–103.

14. Strong V, Waters R, Hibberd C et al. Emotionaldistress in cancer patients: the Edinburgh cancer centresymptom study. Br J Cancer 2007;96(6):868–874.

15. Pruyn JFA, Heule-Dieleman HAG, Knegt PP et al.On the enhancement of efficiency in care for cancerpatients in outpatient clinics: an instrument toaccelerate psychosocial screening and referral. PatientEduc Couns 2004;53(2):135–140.

16. Pollak KI, Arnold RM, Jeffreys AS et al. Oncologistcommunication about emotion during visits withpatients with advanced cancer. J Clin Oncol 2007;25(36):5748–5752.

17. Kennifer SL, Alexander SC, Pollak KI et al. Negativeemotions in cancer care: do oncologists’ responsesdepend on severity and type of emotion? Patient EducCouns 2008;76(1):51–56.

18. Velikova G, Booth L, Smith AB et al. Measuringquality of life in routine oncology practice improvescommunication and patient well-being: a randomizedcontrolled trial. J Clin Oncol 2004;22(4):714–724.

19. Dimoska A, Butow PN, Dent E, Arnold B, Brown RF,Tattersall MH. An examination of the initial cancerconsultation of medical and radiation oncologistsusing the Cancode interaction analysis system. Br JCancer 2008;98(9):1508–1514.

20. Fagerlind H, Lindblad AK, Bergstrom I et al.Patient-physician communication during oncologyconsultations. Psycho-Oncology 2008;17(10):975–985.

21. Institute of Medicine. Cancer Care for the WholePatient: Meeting Psychosocial Health Needs. NationalAcademies Press: Washington, DC, 2007.

250 S. Taylor et al.

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon

22. Thomas CM, Soothill K, McIllmurray K, Francis B,Harman J. What are the psychosocial needs ofcancer patients and their main carers? The Institutefor Health Research, Lancaster University 2001;1–260.

23. Wright EP, Kiely MA, Lynch P, Cull A, Selby PJ.Social problems in oncology. Br J Cancer 2002;87(10):1099–1104.

24. Velikova G, Sheppard S, Campbell L et al. Rando-mized trial of quality-of-life measurement in oncologypractice: do oncologists need to know? ASCO AnnualMeeting Proceedings (Post-Meeting Edition). J ClinOncol 2008;26.

25. Detmar SB, Aaronson NK, Wever LD, Muller M,Schornagel JH. How are you feeling? Who wantsto know? Patients’ and oncologists’ preferences fordiscussing health-related quality-of-life issues. J ClinOncol 2000;18(18):3295–3301.

26. Cella DF, Tulsky DS, Gray G et al. The functionalassessment of cancer therapy scale: development andvalidation of the general measure. J Clin Oncol1993;11(3):570–579.

27. Anderson WG, Alexander SC, Rodriguez KL et al.‘What concerns me is...’ Expression of emotion byadvanced cancer patients during outpatient visits.Support Care Cancer 2008;16(7): 803–811.

Emotional and social impact of cancer 251

Copyright r 2010 John Wiley & Sons, Ltd. Psycho-Oncology 20: 242–251 (2011)

DOI: 10.1002/pon