Although studies have shown the usefulness of improving health professionals' communication skills by training, to the authors' knowledge none have demonstrated the efficacy of communication skill training (CST) for health professionals in terms of improving patient outcomes. This study aimed to assess the efficacy of CST for nurses in improving psychologic distress and coping among patients after being informed of a cancer diagnosis.
Nurses who mainly provide patients with psychologic and informational support after being informed of their cancer diagnosis by physicians at a cancer screening center were randomly assigned to either an experimental or a control group; patients were supported by either group of nurses. Patient selection criteria were: age >18 years with gastric, colorectal, or breast cancer that was not in advanced stage. Intervention consisted of 3 1-on-1 nurses' interviews (on the day of, 1 week after, and 1 month after diagnosis). Efficacy was assessed through patients' psychologic distress and coping by administering the Hospital Anxiety and Depression Scale (HADS) and Mental Adjustment to Cancer scale (MAC), at 3 time points (1 week, 1 month, and 3 months after diagnosis).
Eighty-nine patients participated. Repeated measures analysis of variance demonstrated a significant group-by-time decrease in patients' psychologic distress on HADS (P = .03), and significant group-by-time increase in fighting spirit and decrease of fatalism (P = .01 and P = .04, respectively), in addition to significant between-group difference of anxious preoccupation on the MAC (P = .003).
Communicating effectively in delivering bad news is reportedly an especially important skill for nurses in oncology, because they encounter many opportunities to confirm difficult information.1, 2 Nevertheless, the majortity of oncology nurses experience difficulty in communicating, for example, when supporting patients and families given bad news such as a cancer diagnosis.3, 4
Communication skill training (CST) methods have been developed for nurses in Western countries to enable better communication skills (CS) in oncology practice,4–7 and their effectiveness has been evaluated and summarized in several reviews.8–12 Although these reviews found that participants' confidence and attitudes toward communication with patients improved after CST, to our knowledge, no studies have yet demonstrated the efficacy of CST for health professionals in improving patient outcomes such as psychologic distress and adjustment to cancer. Studies and reviews have suggested several reasons for this, including lack of randomized controlled design, evaluation of patient psychosocial outcomes, uniformity and sufficient duration of programs, and homogeneity of the study sample.4–6, 8–12 In light of these suggestions, we based our CST program on programs developed in the West and made it longer for nurses. In addition, our eligibility criteria were aimed at selecting homogenous groups of patients based on age, cancer site, cancer stage, and the type of bad news received (first-time cancer diagnosis). We then investigated the efficacy of CST for homogeneous patients' outcomes.
The purpose of the current study was to investigate in a randomized design whether a CST program to assist nurses would be useful to reduce psychologic distress and improve coping among patients informed of a cancer diagnosis just after a cancer screening test.
MATERIALS AND METHODS
We chose cancer screening institutions in eastern Japan, which have a follow-up system by nurses after physicians have informed patients of their cancer diagnosis. At first, we approached 4 institutions and visited head nurses to ask them to participate. The head nurses who agreed allowed us to contact nurses who worked in the follow-up system. Consent was obtained in writing from each of them.
Patients were selected from among those who had undergone a cancer screening test and had the support of nurses after a physician consultation informing them of a cancer diagnosis. The study protocol was reviewed and approved by the Institutional Review Board and the Ethics Committee of Tokyo Metropolitan University.
Patient eligibility criteria for the study were 1) newly diagnosed and informed of cancer by a physician consultation after undergoing a cancer screening test within the study period (from January through December 2006); 2) nurse in charge present at the physician consultation; 3) age >18 years; 4) disease that was not advanced and was at an operable stage; 5) diagnosis of gastric, colorectal, or breast cancer, the 3 major sites most tested at the center, where nurses usually provide support for all these patients after cancer diagnosis; and 6) written informed consent.
Patients were excluded if they had not been informed of a cancer diagnosis by physicians and had severe psychologic status based on physician assessment.
The efficacy of the CST program for nurses on patient outcomes was assessed by assigning nurses randomly either to the experimental groups or to the wait-listed control groups (Fig. 1). Before the study, nurses in the experimental group attended a 1-day CST program twice in the course of 3 months (October and December of 2005). Nurses assigned to the wait-listed control group were invited to participate in the CST program after the study period.
The study was explained during the study period to patients who met the eligibility criteria, and they were consecutively asked to participate by nurses in charge of each group. It was explained to the patients that: 1) the study was performed to develop a better nursing support style within the usual care method or a new method using CS by nurses; 2) the participants would be randomly assigned to either group using the usual method or a new method, and participants would be blinded to their assignment; and 3) they would have 3 interviews by nurses and 3 assessments. Patient recruitment and first-time assessment were initiated at T1 (1 week after cancer diagnosis), because the manager of the institution participating in this study (Tokyo Metropolitan Cancer Screening Center) did not approve of starting the patient assessment at T0 (on the day of cancer diagnosis). Patients in the experimental group were interviewed 3 times by nurses who attended the CST program, and controls were also interviewed at the same 3 times by nurses in the control group in the usual support system. In the conventional system, nurses in charge are always present at physician consultations when patients are informed of a cancer diagnosis. The physicians then entrust nurses with the support of patients thereafter. In the system, patients were routinely supported at 1 week and 1 month as well as on the day of the diagnosis. In our protocol, therefore, the interviews were scheduled on the day of cancer diagnosis (T0 [baseline]), 1 week after the diagnosis (T1), and 1 month after the diagnosis (T2), through 1-on-1 psychologic and informational support (Fig. 2).
CST Program for Nurses
The program consisted of 2 workshops: 1 at the start and the other at the end of the 3-month period. The program was based on those developed by Baile et al13, 14 and Fujimori et al.15 The workshop was highly structured and brief, lasting 6 hours each time, and the study focused on CS for nurses. A step-by-step approach is used consisting of 6 steps, referred to by the acronym SPIKES: S, setting up the interview; P, assessing the patient's perception of the illness; I, obtaining an invitation by the patient to disclose information; K, giving knowledge and information to the patient; E, addressing the patient's emotions with empathic responses; and S, strategy and summary (Fig. 1).
The program began with a large group meeting for 2 hours, focusing on theoretic information, at which participants were lectured on the impact of communication between patients and health professionals, the principles of CS for breaking bad news, and how to handle distress in cancer patients. In addition, nurses were given original educational materials and a checklist to unify and confirm their CS.
Nurses were then divided into small groups of 2 or 3, in which 1 or 2 facilitators were assigned. One participant volunteered to play a nurse and another to play a patient in various scenarios involving a nurse's follow-up situation. We prepared the following 3 scenarios, which took into account a suitable Japanese medical system and culture: the first is a young female diagnosed with early-stage breast cancer; the second is a man diagnosed with advanced gastric cancer; and the third is an elderly person diagnosed with severe inoperable cancer whose daughter is the main caregiver. The “nurse” conducted a simulated interview with the “patient” according to the SPIKES steps, and other small group members acted as commentators. Facilitators urged the participants to discuss their concerns and uncertainties in these role-play scenarios. All 3 nurse facilitators had >5 years of counseling experience in oncology, and they also had received 2 days of training in facilitating workshops on CS.
Patients in the control group had the same interviews as the experimental group by control-group nurses in the usual support system.
Patients' Sociodemographic Data
Data regarding patient demographic and clinical characteristics (age, sex, employment status, marital status, number of family members living together, number of confidants, satisfaction with social support, cancer site, cancer stage, and whether they had undergone surgery) were obtained by reviewing patient records.
Patients' Psychologic Distress and Coping
Efficacy was assessed through patients' psychologic distress and coping by administering the Hospital Anxiety and Depression Scale (HADS) and Mental Adjustment to Cancer (MAC) scale 3 times: 1 week after diagnosis (T1), 1 month after (T2), and 3 months after (T3).
The HADS is a 14-item self-rating scale,16 with each item rated on a scale of 0 to 3. Higher scores indicate a greater tendency to anxiety and depression. The Japanese version of HADS has been demonstrated to have adequate validity and reliability.17, 18
The MAC scale is a 40-item self-rating scale used to assess cancer patients' coping style. It consists of 5 subscales: fighting spirit, anxious preoccupation, fatalism, helplessness/hopelessness, and avoidance. Each item is rated on a scale of 1 to 4, ranging from “definitely does not apply to me” to “definitely applies to me,” with higher scores indicating a greater tendency to adopt that coping style. Previous studies indicated that the MAC scale has adequate validity and reliability.19, 20 The Japanese version of the MAC scale also showed adequate validity and reliability.21
Nurses, Physicians, and Interview Characteristics
Data regarding physicians' age and sex, nurses' age and sex, physicians' and nurses' experience in oncology practice (in years), time of physician consultation informing the patient of their cancer diagnosis, time of nurse interview after physician consultation, and whether the physician/nurse had some previous CST within the last year were obtained by reviewing the nurses' checklist.
For the experimental group, interviews were all audiotaped and transcribed. The transcripts were then assessed by 2 trained investigators to determine the number of CS used within the 6-step SPIKES. The 2 investigators rated each interview independently. If the results were different, the investigators had a discussion and decided the number of CS used in that interview. For the control group, we asked the nurses about the contents of their interview with the patients.
Demographic and clinical data of the patients, nurses, and physicians as well as psychologic scores of each assessment time (T1-T3) were tested by the Student t test, chi-square test, or Mann-Whitney U test to assess comparability between the groups.
Effects of an intervention on each measure were assessed using repeated measures analysis of covariance (ANCOVA) to test for a difference between the experimental group and the control group over time, after adjusting for variables related with P < .10 between groups in the univariate analysis. Group and group-by-time changes were processed using ANCOVA. The P value was set at .05, and all data analyses were conducted using the SAS statistical software (version 9.1; SAS Institute, Inc, Cary, NC).
Sample of Institution, Nurses, and Physicians
A cancer institution, Tokyo Metropolitan Cancer Screening Center, consented to participate in the study. All of 9 follow-up nurses consented to participate in the study and completed the program. In this screening center, 18 other nurses provide technical support for potential cancer patients undergoing screening tests.
Because a nurse went on a maternity leave just after attending the CST program, 8 nurses were randomly assigned to either group (4 to the experimental and 4 to the control) at the initiation of the study, and gathered patient data consecutively during the period. Twelve physicians provided consultations for registered patients in both groups. The demographic characteristics of nurses and physicians are listed in Table 1. Comparison of these characteristics in the experimental and control group demonstrated no statistically significant differences.
Table 1. Demographic Characteristics of Nurses and Physicians
Mean±SD or No. (%)
SD indicates standard deviation; CST, communication skill training.
Nurses' characteristics (n=8; 100%)
Clinical experience, years in oncology practice
Attendance to CST workshops within the last year
Physicians' characteristics (n=12; 100%)
Clinical experience, years in oncology practice
Attendance to CST workshops within the last year
Of the consecutive 175 eligible patients during the study period in both groups, 4 from both groups (2%) were excluded because of physicians' assessment of severe psychologic distress, and 82 (47%) refused to participate (57 because of the burden of participating in the study and 25 because of social barriers). Therefore, a total of 89 patients (51%; 42 in the experimental group, 47 in the control group) consented to participate after the explanation by the nurse in charge. Among these participants, 41 in the experimental group and 45 in the control group completed questionnaires more than twice (T1 and T2 assessments), including analyses of the study. Eleven patients (12.3%, 4 in the experimental group and 7 in the control group) dropped out after the second assessment, refusing further assessment (Fig. 2).
No statistically significant differences were found between subjects who consented to participate and those who refused to participate or dropped out with regard to age, social support, cancer stage, cancer site, and treatment. No statistically significant differences were found between the experimental and control groups within the patient demographic and clinical variables.
In the experimental group, the mean number of CS used was 4.6 ± 1.5. All 4 trained nurses could conduct the interviews according to the 6-step SPIKES fairly well. In the 41 interviews by these 4 trained nurses for the experimental group patients, 22 (54%) interviews improved communication using all 6 steps. Ten (24%) of the remaining interviews did not sufficiently follow the second step “P: perception,” 12 (29%) did not follow the third “I: invitation,” and 9 (22%) did not follow the fourth “E: emotion.” The other 3 steps were sufficiently followed in >80% of the total 41 interviews. In the interviews, the majority of patients expressed their negative experiences as well as satisfaction with the nurses' supportive guidance of communication. Examples of responses include “Although I could understand the cancer diagnosis given me by my physician, I was not sure what the physician said afterward. However, I could know what to do, after your calm and gradual help” and “Why did I have cancer? I panicked when I heard it. Your composed support was so helpful to me.”
With regard to the interview characteristics between groups, the time of the nurse's interview at baseline (T0) was statistically different (mean time of experimental group: 21.9 ± 7.6 minutes; control group: 18.3 ± 8.0 minutes). Therefore, this variable was controlled in the multiple analysis of each of the outcomes. No other statistical differences were found between the groups with the variables in Table 2.
Table 2. Comparison of Patient Variables Between Groups
Experimental, N=41 (100%), Mean±SD or No. (%)
Control, N=45 (100%), Mean±SD or No. (%)
SD indicates standard deviation; NS, not significant; T0, the day of the cancer diagnosis; T1, 1 week after the cancer diagnosis; T2, 1 month after the cancer diagnosis; CS, communication skills.
Satisfaction with social support was estimated, with possible answers ranging from 1 (extremely dissatisfied) to 7 (extremely satisfied).
Evaluation of communication skills of the nurse through the interview transcription by investigators (not evaluated for control-group nurses' interviews). SPIKES is an acronym for the following steps: S, setting up the interview; P, assessing the patient's perception of the illness; I, obtaining an invitation by the patient to disclose information; K, giving knowledge and information to the patient; E, addressing the patient's emotions with empathic responses; and S, strategy and summary.
Patient demographic characteristics
Working part or full time
No. in family living together (including patients)
Time of physicians' consultation informing of cancer diagnosis, min
Time of nurses' interview after physicians' consultation, min
No. of CS used within 6-step SPIKES through nurses' interview (range, 0-6)†
Conversely, all the control group nurses answered that they conducted interviews by passively reacting to patients' expressions. For example, they provided emotional support focused on what patients expressed, and they mentioned the cancer diagnosis again only when patients wished to confirm it. All 4 control group nurses also stated that they always communicated with patients in their own way, but had no confidence in it, because they never had a chance to learn the strategy.
Effect of Intervention on Patient Psychologic Distress and Coping
The repeated measures ANCOVAs, which controlled the time of the nurse's interview at baseline, revealed significant group-by-time differences in the depression score and the total HADS score (P = .03 and P = .03, respectively). No other difference over the study period was found for any HADS variables (Table 3).
Table 3. Effects of Nurses' Communication Skill on Patients' Psychologic Distress as Measured by the Hospital Anxiety and Depression Scale
On the MAC subscales, a significant between-group difference over the study period in the subscale of anxious preoccupation was revealed (P = .003). ANCOVA also revealed significant group-by-time differences in the fighting spirit and fatalism subscale scores (P = .01 and P = .04, respectively). No statistically significant between-group differences or group-by-time interactions were observed in other subscales (Table 4).
Table 4. Effects of Nurses' Communication Skill on Patients' Coping With Cancer Measured by the Mental Adjustment to Cancer Scale
There were no significant differences among nurses in charge or physicians in both experimental and control groups in any variables measured by HADS and MAC for all 3 study periods (T1, T2, T3).
This randomized trial demonstrated the effects of a CST program for nurses on improving psychologic distress and coping style among patients who were informed of their cancer diagnosis just after a cancer screening test. To our knowledge, this is the first report to demonstrate the effect of CST for health professionals in improving cancer patient outcomes. One possible reason for this result may be the study design, which called for selecting homogeneous subjects.4, 8–12 We attempted to select homogeneous groups of patients and investigated the efficacy of CST for this population receiving the same bad news (first-time cancer diagnosis), and with the same cancer site and stage.
Another potential reason for the result may be the nurses' training schedule and Japanese nurses' basically lower CS in delivering bad news than in the West.22 Several studies indicated that CST programs must be longer and more intensive so that health professionals acquire satisfactory levels of skills.4, 11, 23–25 A previous study reported that health professionals benefit by learning valuable communication strategies and reviewing demonstrations of skills in realistic clinical situations.26 On the basis of these suggestions, we set a 3-month period for a CST program and nurses trained twice. In the current study, nurses may have been able to acquire CS by seeing examples of valuable communication strategies from role playing in training and by demonstrating skills repeatedly in clinical settings during the 3-month period. In Japan, as the control group nurses stated, they communicate with patients in their own way, because they have never had a chance to learn the strategy for delivering bad news. Thus, were they properly trained, Japanese nurses would be better able to communicate with their patients.
Still another reason might be the influence of the cultural attitude of the Japanese nurses on the result. A previous study pointed out that any attempt to apply Western-developed intervention models to Japanese individuals without considering differences in cultural contexts, such as communication style and family structure, would fail to achieve its effect.26 Accordingly, given the needs of the nurses who attended the 2-time CST program, we intensively discussed how to communicate with physicians and other health professionals, how to provide medical information, and how to communicate with and support not only patients but also family members, while giving Japanese culture its due place in the program. Hence, our study may have met these nurses' needs and effectively helped them provide more culturally appropriate support for their patients.
Results may also have been influenced by the cultural attitude of Japanese patients and the basically poor support system for patients with newly diagnosed cancer that exists to date in Japan. Some studies have shown that Japanese patients do not seek professional assistance for psychosocial problems created by cancer experience,27 and that Japanese cancer patients usually repress their emotions about having cancer.28 Because Japanese people are accustomed to these attitudes, nurses' improved communication using 6-step CS may stimulate and increase patients' need for psychologic and emotional support and thus may change their attitude. This interpretation may be confirmed by the results of their improved coping style. Thus, subjects in the experimental group evidenced an intensified adaptive coping style (fighting spirit in the MAC) and weakened maladaptive coping style (anxious preoccupation and fatalism in the MAC). Moreover, patients in the experimental group actually expressed their negative feelings immediately after the cancer diagnosis, but later many of them were satisfied with the support from the nurses. Considering the poor support system for such patients by nurses in Japan, patients' psychologic distress might be alleviated by the nurses' support through their improved CS. The results of the current study suggest the importance of improved nurse CS to reduce patients' long-term psychologic distress,8, 29 although further evidence must be provided by future studies.
The generalizability of the current study findings is uncertain. First, recruitment was difficult, as in previous studies.23–25 We could recruit from only 1 institution, and the number of subjects was small. Second, because we performed transcript analysis of the interviews only in the experimental group, we could not compare the differences in actual communication between groups. Moreover, neither inter-rater reliability nor intrarater reliability were confirmed on the transcript analysis, although nearly all results from the 2 investigators were the same, and if the result was different, they conferred and concurred on a result. Third, we did not assess physician communication precisely. However, nurses in charge were always present at the physician consultations and supported patients afterward in light of what physicians told the patients. In addition, between the experimental and control groups, we confirmed there were no differences in physicians' clinical experience and time of consultation.
Although further studies are needed to resolve these limitations, to our knowledge this is the first study published to date to use a randomized design to assess the impact of a nurses' CS training program on patient outcomes. Our study may thus provide encouragement and a direction for future research in the application of CST for more health professionals in oncology practice to better deal with unrecognized distress among patients diagnosed with cancer.
We thank the Tokyo-Metropolitan Cancer Screening Center and health professionals for their generous cooperation in this study. We are also grateful to Ms. Chiaki Kusano, Ms. Masayo Tokuhiro, Ms. Kazuko Tanaka, and Ms. Asako Tamura for their research assistance.