The disclosure of bad news calls for compassionate yet effective communication skills. Breaking bad news is one of the most challenging experiences for the practicing clinician. The disclosure of cancer diagnosis is especially challenging, because approximately 25–35% of newly diagnosed patients experience high levels of emotional distress.1 Pre-existing psychosocial factors (eg, social support, medical history, life stressors, education, employment status) have previously been shown to influence a patient's ability to cope with cancer.1
There is a growing body of literature dedicated to disclosure of bad news, and more recently an even greater focus on cancer patient communication experiences and preferences has developed.2–6 However, prior literature is largely nonempirical, consisting of letters, opinion pieces, correspondence, and reviews.7–11 Stepwise protocols and guidelines6 have been published, but less than one-fifth of studies involve an intervention trial and greater than 90% of studies fail to measure psychosocial outcomes for patients.12
Identifying modifiable factors in the disclosure process has potential to optimize patient care, enhance the physician–patient relationship, and improve transition of care to specialists. Our primary objective was to characterize the experience of cancer diagnosis disclosure among gynecologic oncology patients. We identified patient, physician, and environmental factors and assessed which were associated with higher patient satisfaction. Our secondary objectives were to measure a patient's anxiety level at diagnosis disclosure and assess its effect on patient satisfaction. We hypothesized that higher patient satisfaction would be associated with lower anxiety levels and longer conversations that take place in a personal setting. Among patient, physician, and environmental factors, we considered that physician communication skills and behavior would have the most influence on higher patient satisfaction.
MATERIALS AND METHODS
We performed a single-institution survey study of 100 women recently diagnosed with a gynecologic cancer (cervical, endometrial, fallopian tube, ovarian, peritoneal, vaginal, or vulvar cancer) within 6 months before study enrollment. Before the initiation of the study, all procedures were reviewed and approved by Washington University's Human Research Protection Office (HRPO #201103279). All participants were at least 18 years of age, able to read and write in English, and provide signed informed consent. We excluded patients with recurrent disease and confirmed all diagnoses by documented histology on pathology reports. Gynecologic oncology patients treated by members of the Division of Gynecologic Oncology at Washington University School of Medicine and Siteman Cancer Center (a National Cancer Institute-designated comprehensive cancer center) were invited to participate between December 2011 and September 2012.
Patients were approached if eligible and were assured that participation was completely voluntary and anonymous. They were allowed to complete the questionnaire at the time of their outpatient appointment in a private room or take the survey home and return it in a self-addressed envelope provided. Those who did not return the survey after 2 weeks were followed up with a telephone call as a courtesy reminder. Our study population was a convenience sample of 100 patients. This was determined primarily based on clinical feasibility rather than statistical power because prior literature was insufficient to allow estimation of strength of associations for power calculations. One hundred twenty-one patients were eligible. Twenty-one did not return their survey and 11 patients were approached, but declined enrollment. Seventeen patients completed the entire 83-question survey (see the Appendix, available online at http://links.lww.com/AOG/A439); however, the overall response rate was high. The average number of questions missing per patient was 6.7 (standard deviation 6.6, median 4, range 1–30).
An 83-item questionnaire was administered based on three previously validated assessment tools as well as a literature search of prior surveys on disclosure of cancer diagnosis. Data collected included patient demographics and factors related to diagnosis disclosure (eg, location, people present, and length of discussion). Questions were grouped into three broad categories: patient, physician, and environmental factors. Each question was scored separately and within each category a summary score was calculated by adding the scores from each question and dividing by the number of items answered.
Fourteen questions from the Communication Assessment Tool13 were used, asking respondents to rate different dimensions of the communication and interpersonal skills of the physician using a 5-point rating scale (1=poor, 2=fair, 3=good, 4=very good, 5=excellent). In addition, the five-item Wake Forest Trust Scale14,15 was incorporated to evaluate patients' trust in their physician, measuring physician professionalism and patient care skills. A 5-point scale was used ranging from 1=strongly disagree to 5=strongly agree, with a higher score indicating a higher level of trust. The mean overall score for each questionnaire was calculated. Lastly, the 20-item State-Trait Anxiety Inventory-Trait scale16 was used to measure trait anxiety, a stable propensity to experience anxiety or view stressful situations as threatening. The State-Trait Anxiety Inventory-Trait anxiety measure is balanced for acquiescence with both directly scored and reverse-scored items. After reverse scoring nine items, a total score was computed with higher scores indicating greater trait anxiety.
Patients were also asked to rate their overall experience at the time of diagnosis disclosure on a scale of 1–100. Associations between patients' satisfaction and each of patient, physician, and environmental factors were analyzed using the Wilcoxon rank-sum, Kruskal-Wallis, and Spearman's rho tests as appropriate. The mean value of patients' reported satisfaction score was used as the cutoff to define a dichotomized satisfaction score. Poisson regression with robust error variance was used to assess factors associated with satisfaction and estimate relative risk (RR) and 95% confidence intervals (CIs). This analytic approach is a reasonable alternative for logistic regression in cross-sectional studies, providing an unbiased estimate of the RR when the outcome is common.17 The significant factors from the univariate analysis were included in the multivariate analysis using a stepwise approach. Analysis was performed using Stata 10 and an association was considered statistically significant if P<.05.
Data were collected on 100 women regarding their experience at the time of diagnosis disclosure of gynecologic cancer. Their mean age was 58 years and the majority were married (52%), white (90%), and diagnosed with uterine cancer (71%). Ovarian, fallopian tube, or peritoneal cancer composed 20% of the study population followed by cervical, vulvar, or vaginal cancer (11%). Two patients had synchronous tumors (Table 1).
Twenty-four percent were told their diagnoses over the telephone, 60% in the physician's office, and 16% in the hospital (Table 2). The majority of disclosures were performed by an obstetrician–gynecologist (58%) compared with 26% by a gynecologic oncologist, 8% by a primary care physician, and 8% by a member of the patient's care team. These included a reproductive endocrine infertility physician, an emergency medicine physician, a physician of unspecified specialty, two physician assistants, a resident, and a women's health nurse practitioner. Fifty-two percent of women were accompanied by a support person, most often a spouse, child, or parent. Forty-eight percent of disclosure discussions lasted between 1 and 10 minutes, 46% between 11 and 30 minutes (46%), and two patients recalled a discussion length less than 1 minute. Treatment options were included 73% of the time. Future fertility was only relevant for 32 patients, and of these, 17 (53%) reported that their physician discussed how their diagnosis of gynecologic cancer would affect their ability to become pregnant in the future.
Patient satisfaction regarding their overall experience with the disclosure process was rated on a scale of 1–100. The mean score was 86.5, standard deviation 24.4, and the median was 100. We found that face-to-face disclosure (P=.02), location of disclosure in a personal setting (P=.004), and length of conversation more than 10 minutes (P<.001) were all associated with higher patient satisfaction ratings (Table 3).
Patient-centered factors were also associated with higher patient satisfaction (P<.001). They were scored per question and represented as a mean cumulative score. Examples of these survey statements included, “The doctor let me set the pace of the conversation,” “I was given the chance to ask questions,” and “The doctor took into account what I already knew about the situation.”
Physician-centered factors were also represented as a cumulative mean score and subdivided into behavior and communication (Table 3). Both were associated with higher patient satisfaction scores (P<.001). Examples included: “The doctor did not seem nervous as I received the news,” “The doctor checked with me several days later to see whether I had any questions or concerns,” and “The doctor used body language and nonverbal cues to let me know that bad news was forthcoming.” Other factors associated with higher patient satisfaction included strong trust in their physician (P<.001) and environmental factors such as a quiet, private, and comfortable setting without interruptions (P<.001). Disease-related factors such as tumor site and stage were not significantly associated with higher patient satisfaction.
Lower anxiety levels were also associated with higher patient satisfaction (P<.001). Among the 86 patients who had complete anxiety data, the mean total State-Trait Anxiety Inventory-Trait score was 38, standard deviation 12.2, and median 36. Using the standard cutoff (45.81) for a female adult according to the manual for the State-Trait Anxiety Inventory-Trait, 25 patients in our study (29%) had at least a moderately high level of stress.
We noted associations between how different types of physicians disclosed a cancer diagnosis and each of several aspects of the disclosure experience (Table 4). Regardless of specialty, there were no significant differences in length of discussion; whether a support person was present; patient, physician, or environmental factors; level of physician trust; or patient anxiety at the time of diagnosis disclosure. Gynecologic oncologists were more likely to disclose a cancer diagnosis in person (P=.016) and in a personal setting (P=.03) compared with their professional counterparts. There was no significant difference in patient satisfaction by physician type (P=.19); however, we acknowledge that our study was underpowered to adequately evaluate this relationship.
Poisson regression analysis was performed using a dichotomized satisfaction defined using the cutoff at the mean score. Significant factors from the univariate analysis were entered into the multivariate Poisson model using a stepwise approach. In this multivariate analysis, patient-centered factors (RR 2.8, 95% CI 1.75–4.49, P=.013) and patient perception of physician communication skills (RR 4.05, 95% CI 1.34–12.26, P<.001) were significantly associated with higher patient satisfaction regarding the disclosure experience. Among 11 patient-centered statements, two were most significantly associated with higher patient satisfaction: 1) “The doctor seemed to think about my needs during the interaction”; and 2) “Rather than sticking to a plan, the doctor allowed my reaction to dictate the flow of the transaction.” With regard to physician communication skills, the following statements were most significantly associated with higher patient satisfaction: 1) “The doctor understood my main health concerns”; 2) “The doctor talked in terms I could understand”; 3) “The doctor showed care and concern”; and 4) “The doctor spent the right amount of time with me.”
Our study characterized the experience of cancer diagnosis disclosure among gynecologic oncology patients. Overall patient satisfaction was high, but equally importantly, our results offer insights into how disclosure can be improved, especially with regard to discussion length of time. Conversations should take place in a private setting, for greater than 10 minutes, and include discussion of treatment options if appropriate. Furthermore, opportunities to ask questions, express emotion, and set the pace of conversation are important patient-centered factors that should be taken into consideration to optimize the disclosure process. Our findings not only strengthen the current body of literature regarding communication of bad news to patients, but also expand the scope of evidence by identifying issues specific to gynecologic oncology patients.
While interpreting our results, consideration should be given to the fact that cancer diagnosis disclosure is often complicated by the need to expedite referrals to oncology specialists. This time-sensitive issue has potential to play a major role in how disclosures are performed and ultimately may have a negative effect on a patient's experience. Establishing with patients upfront (eg, at the time of the initial procedure consent) how they would prefer results disclosed may facilitate more individualized counseling and ultimately enhance their experiences when dealing with bad news. Unfortunately, our study questionnaire did not specifically inquire whether patients were asked how they would like test results disclosed. However, this could be incorporated into a future intervention trial to further evaluate patient preferences and the feasibility of individually tailored diagnosis disclosures.
Other limitations of our survey study stem from the inherent nature a cross-sectional design and asking patients to reflect on a past experience. The former limits our ability to infer causality from associations observed between the outcome, patient satisfaction, and specific factors of interest. Furthermore, given the sensitive nature of this topic and weight of emotions to cloud memories, recall bias has potential to affect our results. To minimize this effect, we restricted study eligibility to patients who were diagnosed within 6 months of study enrollment and were approached no later than their 6-week postoperative appointment. This also minimized exposure to any adjuvant treatment, which also has potential to cause cognitive deficits in recall. Additionally, we excluded patients with recurrent or progressive cancer because disease status may confound results and exaggerate bias.
Patient selection bias is another limitation that may have influenced our results. However, the overwhelming majority of our practice is referral-based and our results confirm that diagnosis disclosure was most often performed by the referring gynecologist. It is possible that members of the research team avoided asking patients who appeared upset or distraught or, alternatively, were more inclined to approach patients who they had established a good relationship. Again, it should be noted that 74% of patients answered their survey questions based on experiences with their referring physician. Unfortunately, there is no way to know if those that declined did so based on their dissatisfaction with the disclosure experience.
Racial disparities and cultural differences also have potential to affect our results. Given that the majority of our study population was white (90%), our results may not be generalizable to all gynecologic oncology patients. For example, Fujimori and Uchitomi found marked cross-cultural differences and wide variation with respect to having relatives present when receiving bad news and desire to receive information regarding their life expectancy.3
Despite this racially homogenous sample, our study highlights an important point that gynecologic oncology patients are willing to share their diagnosis-disclosure experiences. There may be some presumptions that asking patients to recall their diagnosis disclosure will provoke anxiety with potential to negatively affect a patient's well-being. Previously published research has reported that women with gynecologic cancer have significant psychological problems caused by diagnosis and treatment, and that the most persistent of them is fear of death (23%).18 However, this should not discourage future studies related to this topic. Our results provide useful insights into the disclosure process and can be tested in other populations. Moreover, results from this study have potential to improve clinical training, enhance the physician–patient relationship as well as optimize patient care and its effect on family members. It may also facilitate transition of care from primary physician offices to specialists and referral centers.
Disclosing a cancer diagnosis is as much of an art form as the treatment options that come afterward. Discussions should be tailored to the individual and involved family members, approached with an open mind, and adaptive to different patient needs and reactions to receiving news of a cancer diagnosis. One of our study participants wrote, “There is no easy way to tell a person they have cancer. I elected to stay home from work, because I thought I would hear from my obstetrician–gynecologist regarding my biopsy. In a way, I was glad to be home…so I could gather myself together and get ready for the fight of my life.” We hope our findings underscore the importance of effective communication, especially when the diagnosis is gynecologic cancer, because such conversations are a fundamental component of training for the practicing obstetrician–gynecologist.
1. Zabora J, BrintzenhofeSzoc K, Curbow B, Hooker C, Piantadosi S. The prevalence of psychological distress by cancer site. Psychooncology 2001;10:19–28.
2. Butow PN, Kazemi JN, Beeney LJ, Griffin AM, Dunn SM, Tattersall MH. When the diagnosis is cancer: patient communication experiences and preferences. Cancer 1996;77:2630–7.
3. Fujimori M, Uchitomi Y. Preferences of cancer patients regarding communication of bad news: a systematic literature review. Jpn J Clin Oncol 2009;39:201–16.
4. Parker PA, Baile WF, de Moor C, Lenzi R, Kudelka AP, Cohen L. Breaking bad news about cancer: patients' preferences for communication. J Clin Oncol 2001;19:2049–56.
5. Ptacek JT, Ptacek JJ. Patients' perceptions of receiving bad news about cancer. J Clin Oncol 2001;19:4160–4.
6. Figg WD, Smith EK, Price DK, English BC, Thurman PW, Steinberg SM, et al.. Disclosing a diagnosis of cancer: where and how does it occur? J Clin Oncol 2010;28:3630–5.
7. Baile WF, Buckman R, Lenzi R, Glober G, Beale EA, Kudelka AP. SPIKES-A six-step protocol for delivering bad news: application to the patient with cancer. Oncologist 2000;5:302–11.
8. Eggly S, Penner L, Albrecht TL, Cline RJ, Foster T, Naughton M, et al.. Discussing bad news in the outpatient oncology clinic: rethinking current communication guidelines. J Clin Oncol 2006;24:716–9.
9. Girgis A, Sanson-Fisher RW. Breaking bad news: consensus guidelines for medical practitioners. J Clin Oncol 1995;13:2449–56.
10. Schofield PE, Butow PN. Towards better communication in cancer care: a framework for developing evidence-based interventions. Patient Educ Couns 2004;55:32–9.
11. Walsh D, Nelson KA. Communication of a cancer diagnosis: patients' perceptions of when they were first told they had cancer. Am J Hosp Palliat Care 2003;20:52–6.
12. Paul CL, Clinton-McHarg T, Sanson-Fisher RW, Douglas H, Webb G. Are we there yet? The state of the evidence base for guidelines on breaking bad news to cancer patients. Eur J Cancer 2009;45:2960–6.
13. Makoul G, Krupat E, Chang CH. Measuring patient views of physician communication skills: development and testing of the Communication Assessment Tool. Patient Educ Couns 2007;67:333–42.
14. Dugan E, Trachtenberg F, Hall MA. Development of abbreviated measures to assess patient trust in a physician, a health insurer, and the medical profession. BMC Health Serv Res 2005;5:64.
15. Hall MA, Camacho F, Dugan E, Balkrishnan R. Trust in the medical profession: conceptual and measurement issues. Health Serv Res 2002;37:1419–39.
16. Spielberger C. Manual for the State-Trait Anxiety Inventory STAI (Form Y). Palo Alto (CA): Mind Garden; 1983.
17. Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003;3:21.
18. Steginga SK, Dunn J. Women's experiences following treatment for gynecologic cancer. Oncol Nurs Forum 1997;24:1403–8.