Patient Education and Counseling 91 (2013) 126–128

Reflective Practice

Reflective research: Supporting researchers engaged in analyzing end-of-lifecommunication§

Lee Ellington a,*, Maija Reblin a, Patricia Berry a, Janine Giese-Davis b, Margaret F. Clayton a

a University of Utah College of Nursing, Salt Lake City, UT, USAb University of Calgary, Department of Oncology, Division of Psychosocial Oncology and Tom Baker Cancer Centre, Psychosocial Resources, Calgary, Alberta, Canada

A R T I C L E I N F O

Article history:

Received 16 July 2010

Received in revised form 9 July 2012

Accepted 22 September 2012

Keywords:

Communication

End-of-life

Research

Mindfulness

A B S T R A C T

Objectives: The objectives of this commentary are to: (1) identify unique aspects of communication at

end-of-life as compared to more commonly studied healthcare encounters; (2) outline challenges and

opportunities research team members may encounter in coding and analyzing end-of-life communica-

tion; and (3) offer recommendations for quantitative communication research teams.

This commentary is based on ours and others’ experiences in studying the communication among

patients with terminal illnesses, their families and their providers.

Provider-patient/caregiver communication at end-of-life has some differences as compared to other

forms of healthcare communication and, thus, has implications for researcher response. Challenges and

opportunities for research team members include gaining new knowledge, developing feelings of

attachment to study participants, burn-out, and emotional and existential impact on personal life.

Practical and psychological considerations for conducting end-of-life communication research are

discussed.

Conclusions: Researchers engaged in the study of end-of-life interactions may experience unique

challenges, potentially compromising the validity of the research results and interpretation of

communication processes. A balance between likely coder emotions and objectivity must be achieved.

Implications: Anticipating and facilitating the research teams’ reactions to end-of-life interactions while

still maintaining scientific standards may be achieved by using a reflective, mindful approach.

� 2012 Elsevier Ireland Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

Patient Education and Counseling

jo ur n al h o mep ag e: w ww .e lsev ier . co m / loc ate /p ated u co u

When promoting improved communication, attention has beendevoted to supporting patients/families at end-of-life and tosupporting healthcare providers to prevent burn-out and reducestress [1–5]. Yet, the potential stress experienced by researcherswho study these interactions has been largely ignored. Instead, thefocus for quantitative communication researchers tends to be onmaintaining objectivity: coders are expected to mechanisticallyreview data to maintain reliability and validity. However, coderswho are immersed in the study of interactions in palliative caremay experience strong reactions, thus potentially compromisingthe validity and reliability of the research results and interpreta-tion of communication processes under study.

§ For more information on the Reflective Practice section please see: Hatem D,

Rider EA. Sharing stories: narrative medicine in an evidence-based world. Patient

Education and Counseling 2004;54:251–253.

* Corresponding author at: College of Nursing, University of Utah, 10 South 2000

East, Salt Lake City, UT 84112-5880, USA. Tel.: +1 801 585 9844;

fax: +1 801 587 9838.

E-mail address: lee.ellington@nurs.utah.edu (L. Ellington).

0738-3991/$ – see front matter � 2012 Elsevier Ireland Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.pec.2012.09.007

1. Objectives

This commentary has three objectives. First, we identifyunique aspects of communication at end-of-life as compared tomore commonly studied healthcare encounters. Second, weoutline challenges and opportunities that research team membersmay encounter. Third, we offer recommendations for quantitativecommunication research teams, particularly for those who codeand analyze end-of-life communication.

2. Nature of end-of-life communication

Provider-patient/caregiver communication at end-of-life maybe different from other forms of healthcare communication inboth process and content, which may influence how researchersrespond to the data. In many cases the patient is surroundedby family members talking to the patient, each other, and tohealthcare providers. Palliative care is often a team effort and thus,multiple care providers may be involved. At a practical level, thesemultiple speakers can be frustrating for coders to discern who istalking and to whom, resulting in additional coder time and energy.

L. Ellington et al. / Patient Education and Counseling 91 (2013) 126–128 127

Although involvement of more individuals increases the complexityof coding, it also provides a more comprehensive picture andrelational context of communication during the dying process.

In addition to the presence of multiple persons in end-of-lifeencounters, the topics covered are wide-ranging and often outsidethe realm of more typical healthcare encounters. Conversationscommonly involve beliefs, relationships, reminiscence, and infor-mation related to the physical care of the dying patient. Althoughcurrent coding systems may broadly include these topics, thebreadth and depth of end-of-life care may be a new challenge forcoders, especially for those who have previously coded shorter,disease/treatment focused healthcare encounters.

End-of-life communication is typically more emotional thanother healthcare conversations. Coders repeatedly listen to deeplypersonal, detailed discussions which can increase coder burden byadding another layer of complexity on top of the task of decidingwhich codes to assign. Coders may vary in their comfort level inwitnessing emotional expressions, but all will likely be affected insome way. Even lack of emotional response in participants mayevoke strong coder reactions.

Our recent work with hospice nurses and family caregiversunderscores the unique nature of end-of-life interactions [6].Coded home-visits were emotionally complex and included thepresence of multiple family members. They tended to be long(average 55 min), thus testing coder stamina. Additionally, therewere multiple recorded visits for the same case over time. For thisproject, we made the decision to code visits sequentially believingit would improve coder understanding with repeated exposure tothe same speakers occurring in sequence. Coding across sequentialhome visits quickly gave the coders a sense of the of the patients’decline. During initial visits patient dialogue was evident, but astime passed, the patient became increasingly ill and unresponsiveand exchanges were largely between nurses and caregivers.Finally, when coming into the home, the nurse navigates withinthe family’s environment (e.g., meeting pets, looking at photos,listening to music), rather than the family navigating within aclinical environment. The home setting conveys a holistic sense ofthe family, which may lead coders to make more contextuallybased decisions, but the intimacy of the home environment mayalso impact coder mood. These experiences led us to take a closerconceptual look at the types of challenges and opportunities thatresearch teams face.

3. Challenges and opportunities

Team members may have a range of existing beliefs andassumptions about dying and death based on prior experiences.Unless coders are experienced end-of-life healthcare professionals,they will gain significant new knowledge about dying, particularlydetails about symptoms, including pain. This new information mayincrease awareness of the heterogeneity of the dying process anddispel myths. One of our coders described this additionalknowledge as a ‘‘blessing and a curse’’ and another stated ‘‘I’mnot sure I want this knowledge,’’ but that she felt better informed.

Given the intimate nature of conversations that occur duringthe dying process, coders may develop feelings of attachment toparticipants. This was particularly true in our project when codersreviewed multiple visits with the same nurses, patients, andcaregivers. These feelings of attachment may lead to a sense of lossas the patient dies. Coders may respond to nurses, patients, andcaregivers with personal feelings of sadness, isolation, helpless-ness, frustration, desire for distance or a sense of numbness.Moreover they may revisit personal experiences with dying anddeath [7]. These coder responses, if left unattended, can lead tobiased coding decisions. For example, frustration with a nurse’sstyle or uncontrolled patient symptoms may lead a coder to

overlook or misinterpret communication. Moreover, as with anylongitudinal coding, coders’ responses to early interactions maycolor coding decisions on subsequent interactions and these strongemotions may have even more impact at end of life.

Coder burn-out and turn-over is a concern for all teamsconducting quantitative coding, because the training and codingare time consuming and resource intensive. This concern isheightened in teams studying end-of-life encounters due to thepotential for compassion fatigue [8]. Repeated exposure to deathand dying conversations may have cumulative effects increasingthe risk of coder distress [9]. Members of our team expressedconcerns that they would become desensitized to emotionallyladen experiences in their personal lives. On the positive side, teammembers actually found their work life-changing. Coders statedthey became more aware and sensitive to the suffering of othersand had re-prioritized close relationships in their personal lives. InGiese-Davis’ work [10] coders report increased confidence inmaking important life decisions because they have been exposedto existential end-of-life discussions. Both positive and negativereactions of the research team are consistent with findings on thecoping practices of palliative care providers [11].

4. Considerations for the future

Both practical and psychological considerations may be impor-tant to take into account when embarking on end-of-life communi-cation research. Practical considerations include special attentionwhen hiring research staff including an honest assessment bypotential team members as to whether they feel they canconsistently work with end-of-life data. Additional interview topicsmight include exploration of attitudes toward death, tolerance forintense emotions and suffering, and previous exposure to health-care, end-of-life care, and symptom management.

Because of the unique aspects of end-of-life communicationcoding research, investigators may want to budget and plan foradditional staff, resources and time in comparison to moretraditional coding projects. Coder meetings may be longer andinclude coding decision discussions, but also allowing time fordebriefing and reflection. This extra time may help coders to reflectupon their personal reactions and management of emotionalresponses. Experienced team members are likely to be helpful infacilitating these meetings to ensure that coders feel supported,engender team-building, and assist them in maintaining appro-priate boundaries, thus, allowing them to move forward with theresearch objectives.

Psychological strategies that help healthcare providers copewith the challenges of end-of-life care and prevent burnout [12]can potentially help researchers. Principles in end-of-life educationfor healthcare providers also apply to preparing end-of-liferesearch staff. For example, our palliative/end-of-life care co-investigator (PB) educated our team about the physical stages ofdying and prepared coders for the range of emotions, potentialsounds and physical manifestations of the dying process. Evencoders with end-of-life experience will need to repeatedly replaysections of recordings in order to make coding decisions, and mayfind some aspects of the interactions unsettling, in particular,when the patient appears to be suffering. It may be critical tonormalize coder reactions and to encourage self-care (e.g., noticingwhen to disengage from coding and when to re-engage). Giese-Davis’ research team holds memorial services to recognizedeceased study participants with time for sharing of emotionalreactions. Coders report that both the weekly debriefings and thememorial services help them manage their personal reactions.

Because of the unique inter/intrapersonal challenges ofcommunication surrounding death, healthcare providers havebeen encouraged to practice mindfulness. Mindfulness is defined

L. Ellington et al. / Patient Education and Counseling 91 (2013) 126–128128

as an awareness of emotions, biases, and other reactions to the taskat hand. Just as there has been a recent push for mindfulnesspractice in healthcare [12], there is need for mindful and reflectiveresearch practices. Qualitative researchers have historically beenmore attentive to how their methods emotionally impactresearchers [9,13–17] than quantitative investigators. As teammembers become increasingly aware of their own experiences andaccepting of their responses to participant dialogue, codingdecisions will more fully represent the participant voice.

Interaction analysis of end-of-life communication demandsthoughtful attention to the processes behind producing validrepresentations of the dialogue. Participants allow us as research-ers to observe a private and vulnerable period in their lives inhopes that their experience will ultimately benefit others. In orderto honor their participation, we must first appreciate how thistouches us. Only then can we effectively utilize the scientificapproach to better understand and improve end-of-life care.

Acknowledgments

This project was supported by Huntsman Cancer Institute pilotfunds from the National Cancer Institute, award numberP30CA042014. Salary and Emotion Coding Lab support forGiese-Davis was funded through ACRI-Recruitment and RetentionGrant #4739 and #24397, and the Enbridge Endowed Chair inPsychosocial Oncology Research. The authors wish to thank AmieeMaxwell and Valerie Thomas for sharing their reflections and fortheir contributions to our coding lab.

References

[1] Armstrong J, Holland J. Surviving the stresses of clinical oncology by improvingcommunication. Oncology 2004;18:363.

[2] Hauser JM, Kramer BJ. Family caregivers in palliative care. Clin Geriatr Med2004;20:671–88.

[3] Kearney MK, Weininger RB, Vachon MLS, Harrison RL, Mount BM. Self-care ofphysicians caring for patients at the end of life. J Amer Med Assoc2009;301:1155–64.

[4] Penson RT, Dignan FL, Canellos GP, Picard CL, Lynch Jr TJ. Burnout: caring forthe caregivers. Oncologist 2000;5:425–34.

[5] Rabow MW, Hauser JM, Adams J. Supporting family caregivers at the end oflife: ‘‘they don’t know what they don’t know’’. J Amer Med Assoc 2004;291:483–91.

[6] Ellington L, Reblin M, Clayton MF, Berry P, Mooney K. Hospice nurse commu-nication with cancer patients and their family caregivers. J Palliative Med2012;15:262–8.

[7] Katz RS. When our personal selves influence our professional work: anintroduction to emotions and countertransference in end-of-life care. In:Katz RS, Johnson TA, editors. When professionals weep: emotional and coun-tertransference responses in end-of-life care. New York, NY: Routledge/Taylor& Francis Group; 2006. p. 3–9.

[8] Sabo BM. Compassion fatigue and nursing work: can we accurately capture theconsequences of caring work? Int J Nurs Pract 2006;12:136–42.

[9] Woodby LL, Williams BR, Wittich AR, Burgio KL. Expanding the notion ofresearcher distress: the cumulative effects of coding. Qual Health Res2011;21:830–8.

[10] Giese-Davis J, Piemme KA, Dillon C, Twirbutt S. Macrovariables in affectiveexpression in women with breast cancer participating in support groups. In:Harrigan JA, Rosenthal R, Scherer KR, editors. The new handbook of methodsin nonverbal behavior research. New York, NY, US: Oxford University Press;2005 . p. 399–445.

[11] Yedidia MJ. Transforming doctor–patient relationships to promote patient-centered care: lessons from palliative care. J Pain Symptom Manage2007;33:40–57.

[12] Krasner MS, Epstein RM, Beckman H, Suchman AL, Chapman B, Mooney CJ,et al. Association of an educational program in mindful communication withburnout, empathy, and attitudes among primary care physicians. J Amer MedAssoc 2009;302:1284–93.

[13] Dickson-Swift V, James EL, Kippen S, Liamputtong P. Risk to researchers inqualitative research on sensitive topics: issues and strategies. Qual Health Res2008;18:133–44.

[14] Dickson-Swift V, James EL, Kippen S, Liamputtong P. Blurring boundaries inqualitative health research on sensitive topics. Qual Health Res 2006;16:853–71.

[15] Johnson B, Clarke JM. Collecting sensitive data: the impact on researchers.Qual Health Res 2003;13:421–34.

[16] Kitson GC, Clark RD, Rushforth NB, Brinich PM, Sudak HS, Zyzanski SJ. Researchon difficult family topics. Family Relat 1996;45:183–8.

[17] Rowling L. Being in, being out, being with: affect and the role of the qualitativeresearcher in loss and grief research. Mortality 1999;4:167–81.